@ 5802; ; Material Nonlinear Analysis - TEMP(LOAD) - Superelements vs.

Non -Superelements If superelements are used in a nonlinear run with temperature dependent material properties, the superelement properties are calculated using the TEMP(INIT), rather than the TEMP(LOAD). This is a limitation since superelements are linear, their material properties are not changed during the analysis. This can lead to different answers when comparing superelement models to non-superelement models. Avoidance: In order to use the TEMP(LOAD) for the material properties, use the following DMAP alter in Version 68: COMPILE SELG, SOUIN=MSCSOU, NOLIST, NOREF ALTER 11,11 SSG1 SLT,BGPDTS,CSTMS,SILS,EST,MPTS,ETT,EDT,MJJ,CASESX, DIT,,,,/PJX,PTELEM,/LUSETS/1//'NLST' $ @ 5803; 69; Miscellaneous - GINO - SYSTEM FATAL ERROR (bio_c.read) UNIX versions of MSC/NASTRAN Version 68 and MSC/EMAS Version 3 may report a SYSTEM FATAL ERROR (bio_c.read) in the LOG file but will complete the job normally. Avoidance: No avoidance is necessary, the error message can be ignored. @ 5804; 7 "NSINIT ERROR 20 *** SYNTAX ERROR IN CONTROL FILE"or "UFM 17 UNIDENTIFIED NASTRAN CARD KEYWORD" occurs if NASTRAN OLDQ4K is placed in a runtime configuration file or the data deck. Refer to section 3.4 of the MSC/NASTRAN Version 68 Installation and Operation Instructions for information on the location and format of the runtime configuration files. Avoidance: Two avoidances exist. Either OLDQ4K can not be the first system cell on the NASTRAN command line or the equivalent system cell, SYSTEM(173), must be used in place of the keyword. @ 5805; , UIM 4550 ; Nonlinear Transient Analysis - NLTRD, RESTART - SFM 4276 EC 1044 69; Miscellaneous - RC File - NASTRAN OLDQ4K, NSINIT ERROR 20, UFM 1

IF a model containing superelements is executed in solution 106, and then restarted into solution 129 (using PARAM, SLOOPID), and the "METHOD" field on the TSTEPNL entry is set to AUTO, then the job fails and the following fatal message is issued: SYSTEM FATAL MESSAGE 4276 ERROR CODE 1044 (BIOMSG: ERROR 1044 HAS OCCURRED IN ROUTINE OPEN FILE INDEX (FILEX) = 0) Immediately preceeding this fatal message, USER INFORMATION MESSAGE 4550, NEW STIFFNESS MATRIX REQUIRED is issued. This error may happen even though the nonlinear iteration module output indicates that all time-steps have been successfully completed.

Avoidance: Use the default ADAPT method. @ 5807; 70; DMAP - COMPILE - UFM 423 If a user attempts to compile a user dmap where the name of the dmap begins with the letter "E", the job may fail with the message "USER FATAL MESSAGE 423 (ATTDBS) THE FOLLOWING PHYSICAL FILE DOES NOT EXIST /msc/msc68/emas3/sunos/del/ESS.MASTERA". Avoidance: Change the main dmap and SOL name so that it does not begin with "E". This apparent error is the result of an MSC/NASTRAN feature that will run MSC/EMAS if a deck with a 'SOL 300' or 'SOL E*' entry is submitted. If such a SOL entry is found, the script attempts to attach the MSC/EMAS Solution Sequences. This delivery database does not exist in a pure MSC/NASTRAN installation. @ 5808; 68.2; Differential Stiffness and Buckling - EIGRL, EIGB - Wrong Answe rs or UFM 4683 If the a-set and g-set are the same (no SPCs, MPCs, or reduction) in any superelement or the residual structure, then the differential stiffness matix used in solution 105 (buckling) or the buckling option in solution 200 will not be correct. The error may lead to wrong answers or "USER FATAL MESSAGE 4683, DIFF. STIFFNESS MATRIX NEEDED FOR EIGENVALUE ANALYSIS". DMAP avoidance: COMPILE SEKDR, SOUIN=MSCSOU, NOLIST, NOREF $ ALTER 42, 42 $ EQUIVX KDGGX/KDAA/-1 $ @ 5819; 69; Static Analysis - TEMP(LOAD), LSEQ - UFM 9058 If the LOADSET, LSEQ combination and TEMP(LOAD) are specified in MSC/NASTRAN Version 68, then "USER FATAL MESSAGE 9058" may be issued in SOL 101. This error occurs when there are unreferenced LSEQ entries. Avoidance: Remove the unreferenced LSEQ entries or insert the following alter: compile sestatic souin=mscsou noref nolist alter 59 $ PARAML CASEsx//'DTI'/nskip/7//S,N,TEMPLD $ @ 5820; 68.2; Optimization/Design Sensitivity - ANALYSIS=SAERO - Wrong Answers Applied loads, for example to simulate subcase dependent fuel loads, are ignored and wrong answers are produced if a static aeroelastic design task is submitted in solution 200. Avoidance: Insert the following DMAP avoidance: compile phase1dr souin=mscsou nolist noref alter 81,81 alter 89 equivx casesx/casew/-1 $

@ 5823; forces 68.1; Inertia Relief - Multiple Subcases, SUPORT1 entries - Wrong SPC If inertia relief is used in conjunction with multiple subcases each referencing a different set of SUPORT DOFs, then the spcforces will be incorrect. This is a Version 68 error since prior versions do not support multiple inertia relief sets. This feature is activated with the use of the SUPORT1 entries. Avoidance: Insert the following DMAP alter avoidance. compile sestatic souin=mscsou nolist noref alter 100,100 $ CALL SUPER3 CASECC, ,UL, , , , , , , , CRX, , ,,,, PCDB,XYCDB,POSTCDB,FORCE,EMAP,MAPS,EQEXINS, PVTS,CASES,SLT,ETT,GOAT,,LOO,CRJ,LAO,UOO,UOX,GPLS,USET, SILS,PJ,EDT,YS,GM,PSS,KFS,KSS,QR,,,,,, DYNAMICS,BGPDTS,CSTMS,MPTS,DIT,EST,GEOM2S,GEOM3S,ECTS, EPTS,INDTA,KELM,KDICT,GPECT,VELEM,,SILX,EQEXINX,ECTX, BGPDTX,,GEOM4S,PG,,,,EPT,MPT,GEOM1,GEOM2,GEOM3,GEOM4, PTELEM,,, ,,MEDGE,VIEWTB,ESTVU,GEOM1VU,SILVU, CSTMVU,DEQATN,DEQIND,GDNTAB,OINT,PELSETS, KSF,KMG,,,,,,,,,,,,/ UG,QG,CASEDR,XYCDBDR,PUG,GPSF, , OUG1,OPG1,OQG1,OEF1,OES1,OEE1,OGPS1, OUG2,OPG2,OQG2,OEF2,OES2,OEE2,OGPS2, OESM1,OESG1,OEEM1,OEEG1,GPS,ONRGY,OGPFB/ APP/APP/RSONLY/FALSE /FALSE /FALSE/DROPT/FALSE/ S,PFILE/S,CARDNO/0/FALSE/0/PEXIST/ ADPTINDX/ADPTEXIT $ $ @ 5829; 69; Plotting - SEUPPLOT - Missing Plot Output In SOL 103, if the SEUPPLOT command is used to generate assembly plots, plots will be available for only the first half of the modes in the solution. Avoidance: Apply the following DMAP avoidance in Version 68: compile semodes, souin=mscsou, nolist, noref alter 1 $ disable scratch LTU logic - otherwise the datablocks will $ not be there when needed for plotting putsys(0,192) $ compile sedrcvr, souin=mscsou, list, noref alter 1 type db,phidlg $unused datablocks in SOL 103 $ alter 48 $ after sdr2 copy ug/ugcopy $ append ug,ugcopy/ugtwice $ get "double" set of solutions $ SDR2 to get "double plot vectors"

message //'seid='/seid $ SDR2 CASEDR,CSTMS,MPTS,DIT,EQEXINS,,ETT,OL2,BGPDP, PJ1,QG,UGtwice,EST ,XYCDBDR,OINT,PELSETS,VIEWTB/ ,,OUGVxxx,,,phidlg/APP1/S,N,NOSORT2xx/ NOCOMPS/////ACOUT/PREFDB/-1./-1./ ADPTINDX/ADPTEXIT/FALSE $ $ compile super3, souin=mscsou, list, noref alter 227,227 $ replace PLTVEC type db,phidlg dbdict datablk=(phidlg,phidrg,pug), select(name,size,seid,peid,zname,version) $ DBVIEW PhidlgF=Phidlg (WHERE SEID=* AND WILDCARD=TRUE) $ PLTVEC EMAP,DRLIST,PCDB,EQEXINX,SILX,SILSF,GPLSF,phidlgf/ PUGX,PCDBX/SEID/'SEID' $ @ 5831; 68.2; DMAP - MATMOD, FBS, TRNSP - Wrong Answers, SFM There is a special form of an identity matrix called a form 8. Conventional identity matrices are form 6. This type of matrix can be generated with MATMOD option 2, or may be generated internally in modules. The new sparse options may have trouble with this form of matrix. Troubles observed to date include: Wrong answers from FBS when more than one pass is required. The symptoms for the FBS problem found to date include a solution which has about 50% more non-zero terms than it should. A SYSTEM FATAL MESSAGE in the TRNSP module. Avoidance: The error in FBS may be avoided by switching back to the non sparse methods by setting SYSTEM(126)=1. It is recommended this be done only locally around the offending module with PUTSYS statements. The error in TRNSP may only be avoided by producing a form 6 identity matrix. MATMOD option 1 outputs a conventional identity matrix. @ 5835; 68.0.2; Miscellaneous - RC File - Directory Does Not Exist If a scratch directory is specified in an RC File that does not exist, the run fails even though a directory that does exist is specified in a subsequent RC File that is meant to take precedence over the first specification. Avoidance: Change the directory specification to a directory that exists. @ 5838; 69; Loads - TEMP(LOAD) - UWM 4012, UFM 4017 When doing a static analysis with temperature loads in Version 68, it is necessary to specify the temperature for all grid points in the model. If this is not done, the following messages will be issued, "USER WARNING MESSAGE 4012, THERE IS NO ELEMENT, GRID POINT, OR DEFAULT TEMPERATURE DATA FOR TEMPERATURE SET 102, WITH RESPECT TO ELEMENT ID = 0","USER FATAL MESSAGE 4017. THERE IS NO TEMPERATURE DATA FOR ELEMENT 15 IN SET 2". Unfortunately, the element ID and set numbers are incorrect. Avoidance: Always include a TEMPD bulk data entry in any static analysis with temperature loads.

@ 5839; 68.2; Heat Transfer - heatconv, SPC - Cannot Convert Large Field Forma t, Wrong Answers The heatconv program does not correctly handle all large field SPC entries. In particular, given the following large format SPC entry: SPC* 2.00000000E+01 1 1 1

heatconv will incorrectly copy this to a small format entry as follows: SPC 1 1 1 2.000000

Whenever the temperature fields, i.e., fields 5 & 8, exceed eight characters in length, incorrect results will occur.

@ 5841; 2001.0.1; Composites - PCOMP, MAT2 - UFM 4291 Certain MAT2 entries that are referenced from a PCOMP are not checked for correctness. In particular, if 13 and/or 23 terms are non-zero and the 33 term is zero, the matrix is not positive definite. In this case MSC/NASTRAN will create potentially incorrect PSHELL/MAT2 combinations from the PCOMP/MAT2 inputs. Avoidance: Enter a positive definite MAT2 entry. @ 5843; 68.2; P-Elements - RFORCE - Wrong Answers or SFM 4276 Although it is not documented, RFORCEs are not implemented for p-elements. If RFORCES are applied to p-elements either wrong answers or a SYSTEM FATAL MESSAGE 4276 results. Avoidance: Do not apply RFORCEs to p-elements. @ 5844; 69; P-Elements - HEX20, PENTA15, TETRA10 - Missing Documentation The CHEXA element with 20 grids, the CPENTA element with 15 grids, and the CTETRA element with 10 grids may not be used as p-elements by referencing the PVAL entry. If attempted, the following message is issued: *** USER FATAL MESSAGE 6390, (GP0A), ELEMENT ID = 1 TYPE HEXA20 OR PENTA15 AND/OR TETRA10 ARE NOT SUPPORTED IN P-ELEMENT ANALYSIS. USER ACTION: CHECK ELEMENT CONNECTION ENTRY OR SET ENTRY. @ 5848; 68.2; Optimization/Design Sensitivity - DSCREEN - Poor Convergence The DSCREEN entry permits the user to override default screening parameters for design optimization. This capability does not work for responses TDISP,TVELO,TACCL,TSPC,TSTRE,TFORC,TRIM,STABDER and FLUTTER. Avoidance: The only avoidance is to do a TABEDIT on the DSCREN datablock. The following words correspond to the data locations for the responses

TDISP TVELO TACCL TSPC TSTRE TFORC TRIM STABDER FLUTTER

TRS 179 182 185 188 191 194 245 248 251

NSTR 180 183 186 189 192 195 246 249 252

A candidate DMAP for performing the TABEDIT is: compile desopt souin=mscsou nolist noref alter 41 $ call dbstore dti,dtindx,,,//1/2/'dball'/0 $ compile predom souin=mscsou nolist noref alter 12 $ call dbfetch /dti,dtindx,,,/1/2/0/0/0 $ dtiin dti,dtindx/control,,,,,,,,, $ tabedit dscren,control,,,/ndscren/-1 $ equivx ndscren/dscren/always $ $ $ the following dti modifies word 195 as required $ for a TFORC request and nstr $ dti,control,0 dti,control,1,cr,1,dw,195,195,1 ,1 dti,control,2,er,endrec $ @ 5849; 68.1; Matrix Methods - FBS - UWM 4398, Wrong Answers MSC/NASTRAN Version 67.7 and 68 on the CRAY C90 computer may produce wrong answers for models where FBS method 2 is selected. The error may be evidenced by "USER WARNING MESSAGE 4398" which indicates "DEGREES OF FREEDOM DO NOT PASS SYMMETRY CHECK". The error only occurs when there are strings in the factor that are longer than 64 terms. Note that except in superelement models, that FBS method 2 is rarely selected. Avoidance: FBS Method 2 should be deselected by setting "FBSOPT=-1" in the system RC file on Version 67.7 and 68 CRAY C90 computers. @ 5850; Buffer 68.1; Miscellaneous - IFP, XRCARD - Infinite CPU Loop, Overflow Input Some invalid bulk data input may result in an infinite cpu loop in the routine XRCARD. Although the exact type of input that results in this error is not known, the specifics of the input line that resulted in this error is that the line was over 80 characters, contained many leading blanks and included an embedded negative, "-", sign. The infinite cpu loop has only been reproduced on the CRAY computer but may occur on other systems. Avoidance: There is no practical avoidance. @

5852; x

; Documentation - DMAP Module Dictionary, MODTRL - Bad Input Matri If the MODTRL module is used to increase the number of columns in a matrix, the resulting matrix is unusable in most modules, including matprn and ADD. The cause of the error is that the matrix does not have column headers for the added null columns. Avoidance: The MATGEN and ADD module should be used to increase the number of columns in a matrix. An example follows: matgen ,/nullab/7/rn/cn $ null matrix with rn rows and cn cols. user must set rn and cn above. add nullab,oldmat/newmat $ trailer will match first input equivx newmat/oldmat $ oldmat is now 'bigger'

@ 5858; 68.2; Optimization/Design Sensitivity - SPCFORCE and MPCFORCE - Wrong Answers If scr=yes is used in SOL 200 with dynamic optimization or sensitivity, then MPCFORCE and SPCFORCE output does not include mass and damping effects after the first design cycle. (Version 68 only) Avoidance: Specify scr=no. An extensive DMAP alter is also available if scr=yes is desired. @ 5860; 69; Aeroelastic Analysis - GI, CAERO1, CAERO2 - Wrong Answers MSC/NASTRAN Version 68 on Sun computers may produce wrong answers for aeroelastic flutter problems when aerodynamic bodies, i.e. CAERO1 or CAERO2 elements, are present. Avoidance: There is no avoidance. @ 5861; 69; Dynamics - PARAM,MODACC - Poor Displacement Results The mode acceleration method is intended to improve element stress data recovery in dynamic solutions using modal formulations. However an error may occur in the total displacements when free-body motions are present. This error will NOT affect strains, stresses or constraint forces. The major problem is that the angular displacements or pointing directions will be affected. The problem occurs because the "improved" relative displacements are not orthogonal to the free-body modes and create small additional displacements and rotations. The error is proportional to the local frexibility of the SUPORT degrees of freedom. This error is considered fixed in Version 69 by the addition of a new mode acceleration option which is selected by setting PARAM,MODACC,1 instead of PARAM,MODACC,0. The avoidance is to use conventional dynamic recovery methods in a modal formulation when free-body modes are present. The "Residual Vector" method is an accurate alternative. @ 5862; ; Database - DBUNLOAD - UFM 3007, illegal input to subroutine PVT If the dbunload is done on a long word machine such as Cray and dbload is

attemted on a short word machine, then UFM 3007, illegal input to subroutine PVT may be issued. This error occurs only if the initial run which created the databases had parameter specifications in the Case Control. This occurs since the diliminator for the parameter in the Case Control is machine dependent. Avoidance: There are two avoidances: 1. Remove the parameter specifications from the initial run which creates the databases. or. 2. Use the following alter in the dbload run. ACQUIRE NDDL ASSIGN DBLOAD='neutrala.cray',FORM=FORMATTED DBLOAD FORMAT=NEUTRAL $ $ avoidance is to add the following alter $ $ SOL DBTRANS TIME 5 compile dbtrans souin=mscsou list noref alter 114,114 CEND BEGIN BULK ENDDATA @ 5863; 68.2; P-Elements - GP4 - UFM 4276, CSTM DOES NOT EXIST (GP4) If a model contains p-elements and there is insufficient memory in the GP4 module the message "USER FATAL MESSAGE 4276, CSTM DOES NOT EXIST (GP4) may be output. Avoidence: Request more memory and rerun the job. @ 5864; atives 68.2; Aeroelastic Analysis - DIAG 50 - Unsplined Rigid Stability Deriv For Version 68, static aeroelastic analyses (SOLs 144 and/or 200) produce Unsplined Rigid Stability Derivatives. These data are presented only to provide a check of the splining and are not used in the solution. For antisymmetric models and models with no symmetry (SYMXZ = -1 or 0), the numbers may be in error. For sideforces the error is likely to be simply one of sign. For yawing and rolling moments, the error is less obvious. @ 5866; 68.2; Heat Transfer - RADMTX - Duplicate Continuation Entry with Multi ple Cavities For the multiple cavity radiation view factor calculation, each cavity will have a RADLST and RADMTX entry associated with it. The punch file which contains the MSC/NASTRAN calculated RADMTX entries for each cavity contains duplicate RADMTX continuation identifiers. Thus, subsequent restarts using the punch file fail.

Avoidance: Since continuation identifiers are not necessary if the input is in sequential order, blank out all the continuation identifiers in field 10 and field 1 of the RADMTX entries. @ 5867; s The Quick Reference Guide indicates that the DTOR value input on the SPLINE2 entry must be greater than 0.0. However, this requirement is never checked. If the user leaves DTOR field blank the run could give wrong answers or fail due to a singular spline matrix. Avoidance: Specify a value greater than 0.0 for DTOR. @ 5868; d List 68.1; Installation - Authorization - UFM 3060, BUCKLING not in Approve An authorization code for the workstation option is suppose to include statics, normal modes and buckling. An error in the authorization code generator resulted in workstation options not including buckling. Avoidance: The error in the authorization code generator has been corrected. Contact MSC for a new code. @ 5870; 68.2; Multipoint Constraint - SPOINT - UFM 2101A DOF, Wrong Answers. If the user inadvertently specifies a degree of freedom greater than 1 for an SPOINT in an MPC equation, then wrong answers may result or the message "USER FATAL MESSAGE 2101A, GRID POINT xxx COMPONENT yyy ILLEGALLY DEFINED IN SETS zzz" may be output referring to the wrong grid point/degree of freedom. Avoidance: Other than correcting the SPOINT degree of freedom, there is no avoidance. @ 5871; 68.2; Restarts - RESTART PROJECT - Inefficient Restart In Version 68, the PROJECT keyword is ignored on the RESTART statement which will result in an inefficient restart. For example, 1. The initial run creates project A: PROJECT='A' ASSIGN MASTER='A.MASTER', DELETE 2. The subsequent run restarts from project A into project B: PROJ='B' ASSIGN MASTER='A.MASTER' RESTART PROJECT='A',KEEP Avoidance: For SOLs 101 through 200 insert the following DMAP alter: compile ifpl souin=mscsou noref nolist alter 80,80 type parm,,i,,projid=1,versid=1 $ projid=1 $ set the projid to the integer equivalent of the project you are restarting from. versid=1 $ set the versid to the version you are restarting from $ dbdict datablk=geom1q $ 68.2; Mulitpoint Constraints - SPLINE2 - Singular Matrix, Wrong Answer

alter 118,131 DBVIEW GEOM1R=GEOM1Q(WHERE version=versid and PROJECT=PROJID) $ DBVIEW GEOM2R=GEOM2(WHERE version=versid and PROJECT=PROJID) $ DBVIEW GEOM3R=GEOM3(WHERE version=versid and PROJECT=PROJID) $ DBVIEW GEOM4R=GEOM4(WHERE version=versid and PROJECT=PROJID) $ DBVIEW EPTR=EPT(WHERE version=versid and PROJECT=PROJID) $ DBVIEW MPTR=MPT(WHERE version=versid and PROJECT=PROJID) $ DBVIEW EDTR=EDT(WHERE version=versid and PROJECT=PROJID) $ DBVIEW DITR=DIT(WHERE version=versid and PROJECT=PROJID) $ DBVIEW DYNAMICR=DYNAMIC(WHERE version=versid and PROJECT=PROJID) $ DBVIEW AXICR=AXIC(WHERE version=versid and PROJECT=PROJID) $ DBVIEW PVTR=PVT(WHERE version=versid and PROJECT=PROJID) $ DBVIEW MATPOOLR=MATPOOL (WHERE version=versid and PROJECT=PROJID) $ DBVIEW caseccrr=caseccr (WHERE version=versid and PROJECT=PROJID) $ DBVIEW EDOMR=EDOM WHERE (version=versid and PROJECT=PROJID) $ compile phase0 souin=mscsou noref nolist alter 281,297 type parm,,i,,projid=1,versid=1 $ projid=1 $ set the projid to the integer equivalent of the project you are restarting from. versid=1 $ set the versid to the version you are restarting from $ DBVIEW CASESR=CASES (WHERE version=versid AND PROJECT=PROJID) $ DBVIEW GEOM1SR=GEOM1S (WHERE version=versid AND PROJECT=PROJID) $ DBVIEW GEOM2SR=GEOM2S (WHERE version=versid AND PROJECT=PROJID) $ DBVIEW GEOM3SR=GEOM3S (WHERE version=versid AND PROJECT=PROJID) $ DBVIEW USET0R=USET0 (WHERE version=versid AND PROJECT=PROJID) $ DBVIEW RMGR=RMG (WHERE version=versid AND PROJECT=PROJID) $ DBVIEW YGR=YG (WHERE version=versid AND PROJECT=PROJID) $ DBVIEW DYNAMCSR=DYNAMICS (WHERE version=versid AND PROJECT=PROJID) $ DBVIEW EPTSR=EPTS (WHERE version=versid AND PROJECT=PROJID) $ DBVIEW MPTSR=MPTS (WHERE version=versid AND PROJECT=PROJID) $ DBVIEW MAPSR=MAPS (WHERE version=versid AND PROJECT=PROJID) $ DBVIEW PVTSR=PVTS (WHERE version=versid AND PROJECT=PROJID) $ DBVIEW BACKR=BACK (WHERE version=versid AND PROJECT=PROJID) $ DBVIEW HARMR=HARM (WHERE version=versid AND PROJECT=PROJID) $ DBVIEW KVALR=KVAL (WHERE version=versid AND PROJECT=PROJID) $ DBVIEW FORER=FORE (WHERE version=versid AND PROJECT=PROJID) $ alter 352,352 DBVIEW EQEXINSR=EQEXINS (WHERE version=versid AND PROJECT=PROJID) $ @ 5876; , SCR300 69; File Management Section - EXPAND, RESTART - SFM 4276 XIO, EXPAND

An attempt to expand a database fails when restart is attempted in the same run. A sample File Management Section demonstrating the error follows: assign dbup1a='scr7a.dbup1a' expand dbup1, logical=(dbup1a(100)) restart version=1 keep The sample input leads to the following fatal message due to the scr300 file being improperly defined: "SYSTEM FATAL MESSAGE 1030 (XIO) BLOCK NUMBER MISMATCH IN READ SYSTEM FATAL ERROR 4276.XIO ERROR CODE 4276"

Avoidance: 1) run an "expand" job only (i.e. expand and endjob) 2) run the restart job @ 5877; 67.5.0.1; Miscellaneous - SCO/UNIX - UFM 3060 The SCO/UNIX Version 67.5 of MSC/NASTRAN may issue "USER FATAL MESSAGE 3060" indicating authorization code violation during long running job with many module changes (e.g. nonlinear). Avoidance: There is no avoidance. This problem will be fixed on a rerelease version (MSC/NASTRAN v67.5.0.1 for SCO/UNIX). @ 5878; 69; Sequencer - SEQP - SFM 3008, Insufficient Memory A small p-element model ( 1 Hexap and 2 Pentap ) fails in the sequencer, module SEQP, with the following error message from subroutine seqos: SYSTEM FATAL MESSAGE 3008 MODULE = SEQP 0INSUFFICIENT MEMORY AVAILABLE FOR SUBROUTINE SEQOS Avoidance: Avoid the sequencer by setting PARAM,NEWSEQ,-1. @ 5880; 68.2; P-Elements - TEMP - UFM 4017 All p-version models which contain thermal loads via TEMP entries and more than one type of solid element (HEXA, PENTA, TETRA), fail with the following fatal message, "USER FATAL MESSAGE 4017. THERE IS NO TEMPERATURE DATA FOR ELEMENT y IN SET x". Avoidance: If the thermal loading is a uniform temperature increase throughout the entire structure, use a TEMPD Entry. Otherwise, the user must use an equivalent TEMPF Entry. Also, it may be possible to use only one element type in the model. @ 5881; 69; Documentation - PARAM,BAILOUT - Missing Documentation MSC/NASTRAN Version 68 supports multiple boundary conditions in the structured solution sequences. However, PARAM BAILOUT cannot be changed between subcases. @ 5889; 68.2; Optimization/Design Sensitivity - DRESP2 - Wrong Answers If a DRESP2 entry refers to two or more DRESP1 entries and there are DVPREL2 entries in the input file, then the sensitivities printed in the columns of the DSCM2 matrix associated with the DRESP2 entry are in error. Avoidance: There is no avoidance other than to recast the problem so that it does not require the offending DRESP2,DRESP1,DVPREL2 combination. @ 5891; 70.5.1; Elements - PENTA - Possible Wrong Answers, UFM 4654 If a pyramid-shaped degenerate PENTA element is input to MSC/NASTRAN, USER

FATAL MESSAGE 4654 is issued warning of two corner points with the same coordinates but the run will continue. Pyramid-shaped PENTA elements are not supported and should not be used. @ 5893; 002 69; Sequencer - SEQP, PARAM,NEWSEQ,6 PARAM,OLDSEQ,6 - SFM 4276, EC 3 It is possible that jobs that request param,newseq,6 prior to Version 69 or param,oldseq,6 in Version 69 and after may encounter SYSTEM FATAL MESSAGE 4276, ERROR CODE 3002 in the SEQP module. By default sequencing in the SEQP module was turned off in Version 69 and beyond. PARAM,NEWSEQ was changed to PARAM,OLDSEQ to force the user to go out of his way to request sequencing. Thus, this error is considered corrected in Version 69 by the fact that sequencing is no longer recommended. @ 5894; 2001; Transient Response Analysis - TIC - Initial Conditions Not Refle cted in Results An initial velocity, V0, TIC (transient initial condition) is not properly reflected in the Velocity (first time derivative) results. The discrepancy between the value specified and the value output occurs because the program uses a central difference approach. The displacements are correct. @ 5896; 68.2; Miscellaneous - NASTRAN SYSTEM(i) - Bad Processing Order MSC/NASTRAN processes systems cells specified by the format "NASTRAN SYSTEM(i)=value" before those specified by "SYSTEM(i)=value". Thus system cells specified in the latter format will take precedence regardless of where they are specified. Avoidance: Always use the format "NASTRAN SYSTEM(i)=value" in the system, architecture, user, and local RC files. @ 5900; GL; Material Nonlinear Analysis - RESTART, NLPARM - DIM 9005, UFM 20 44, SFM 3007 If thermal loads are applied in a SOL 106 cold start, and, if in the restart, PARAM,LOOPID is not the last loop id for the subcase prior to PARAM, SUBID then the program will issue various misleading fatal messages. In one example the message "USER FATAL MESSAGE 2044 UNDEFINED TEMPERATURE SET, SYSTEM FATAL MESSAGE 3007 MODULE=NLCOMB ILLEGAL INPUT TO SUBROUTINE NLCPTM" was encountered. Avoidance: Other than specifying the correct LOOPID, there is no avoidance. @ 5901; 68.2; Installation - Multiple Auth Codes - Only One Code in Auth File If authorization codes for MSC/EMAS and MSC/NASTRAN are entered at different times during the same installation, only the last code entered is placed in the authorize.dat file. Avoidance: Enter all codes for all products at the same time in the installation. The installation script allows you to enter a series of codes. If one of the products is using an existing authorize.dat file, modify the appropriate <install-dir>/conf/"rc file" to have the AUTH

keyword point to the correct file. @ 5911; 68.2; Complex Eigenvalue Analysis - SPCForces - Wrong answers In MSC/NASTRAN Version 68, the SPCForce results are incorrect in SOLs 107, 110, and 145 only. The SPCForces do not include the effects of the imaginary part of the eigenvalue which results in "undamped" SPCforces. Avoidance: Insert the following DMAP alter. COMPILE CFORCE SOUIN=MSCSOU noLIST NOREF $ ALTER 47,47 $ replace DSAP PARAML OL//'TRAILER'/1/S,N,NFREQS $ NUMBER OF Eigenvalues TYPE PARM,,I,N,II=0,iip3,iip4 $ LOOP COUNTER TYPE PARM,,CS,N,IOMEGAX,MOMEGX $ FILE qx1=APPEND $ DO WHILE (II<NFREQS) $ II = II + 1 iip3=6*(ii-1)+3 $ clama word 3 iip4=6*(ii-1)+4 $ clama word 4 PARAML OL//'DTI'/2/IIP3/S,N,omega $ real eigenvalue PARAML OL//'DTI'/2/IIP4/S,N,omegai $ imag eigenvalue IOMEGAX = CMPLX(omega,OMEGAi) $ MOMEGX = IOMEGAX**2 $ MATMOD KUx,,,,,/KI,/1/II $ COLUMN EXTRACTION MATMOD BUx,,,,,/BI,/1/II $ MATMOD MUx,,,,,/MI,/1/II $ ADD5 KI,BI,MI,,/FI/1.0/IOMEGAX/MOMEGX $ APPEND FI,/qx1/2 $ ENDDO $ NFREQS @ 5914; 68.2; Nonlinear Transient Analysis - PARAM,G, PARAM,W3 - Incorrect St ructural Damping In MSC/NASTRAN Version 68, solution 129, the structural damping for linear elements specified by parameters G and W3 is doubled. The structural damping specified by GE and parameter W4 are correct. (V68 only) Avoidance: Apply the following DMAP alter: compile nltran, souin=mscsou, nolist, noref $ alter 224, 224 $ ADD5 BDD0,KDDNL,K4DDNL,,/BRDD//GW3/RW4 $ endalter $ @ 5915; 69; Case Control - GPSDCON, ELSDCON - Missing Output Element and grid point stress discontinuity output (ELSDCON and GPSDCON Case Control commands) cannot be requested for a single subcase or a set of subcases. This output is only generated if the request is made for all subcases in Case Control. Avoidance: The user must place these commands above the subcase level or in each subcase. @ 5918; 69; Complex Eigenvalue Analysis - CEAD - Poor Message, UFM 4215

When running a complex eigenvalue analysis, the message "USER FATAL ERROR 4215. COLUMN 1 OF DATA BLOCK SCRATCH IS NULL BUT ROW 1 IS NOT" may be issued. However, this null column does not affect the solution. Avoidance: There is no avoidance. @ 5920; 68.2; Transient Heat Transfer - MATT4,TABLEM4 - SFM 2090 Contrary to the documentation , the temperature dependent specific heat, T(CP), on the MATT4 entry, cannot be specified with the TABLEM4 entry (coefficient of a power series). If this definition is attempted, the message "SYSTEM FATAL MESSAGE 2090, TABLE DICTIONARY ENTRY 201 MISSING" is output. Avoidance: Use the TABLEM1, TABLEM2 or TABLEM3 entry to specify temperature dependent specific heat. @ 5921; 70; Nonlinear Analysis - TEMPP1, TEMPP3 - Wrong Answers If the user specifies a temperature gradient using the TEMPP1 or TEMPP3 entry in a nonlinear static thermal stress analysis, then the TEMP(INIT) case control command treats the thermal gradient as zero despite the value specified by the user for the thermal gradient. Avoidance: The user may avoid this problem if the thermal expansion coefficient is not temperature dependent. In this case, the thermal load from the thermal gradients depends only on the difference of the thermal gradients. The user should, on the TEMPP1 entry for TEMP(INIT) place a 0.0 or blank in the T' field, and on the TEMPP1 entry for TEMP(LOAD) place T'(load) - T'(init) in the T' field. @ 5922; 69; Miscellaneous - ALPHA OSF/1 Only - UFM 1012 DBSET DBALL IS FULL On the ALPHA OSF/1 release of MSC/NASTRAN when the system runs out of allocated disk space the job may hang. When the process is killed, the system sends the message that the job is finished and the output files look as they should with no indication of a problem. This error has also been observed in connection with SFM 4393 followed by SFM 4276 Error Code 3002. Avoidance: Other than killing the job, there is no avoidance. @ 5923; ; Miscellaneous - DIAG 49, F04 Analyze - Missing Output, SWM 1007, LOGIC ERROR 4 Large models that request DIAG 49 (f04 performance summary) may fail and issue the following message in the f04 (Execution Summary) file. Models without DIAG 49 have also less frequently encountered this error. SYSTEM WARNING MESSAGE 1007 (F04SUM) LOGIC ERROR 4 WITH FOLLOWING RECORD(S): or SYSTEM FATAL MESSAGE - bio_c.sysrd.read A BIO operation on a GINO file has failed. File descriptor: 7. Result: 49152.

The error occurs whenever the file size is smaller than the buffsize. Avoidance: Other than running Version 67.5 there is no avoidance. @ 5924; 68.2; Nonlinear Transient Heat Transfer - TEMPBC - SFM 4276 EC 3021 If a TEMPBC entry of TYPE="STAT" and TYPE="TRAN" is referenced in the same nonlinear transient run, then the job fails and the message "SYSTEM FATAL ERROR 4276.WRITE ERROR CODE 3021", is issued. Avoidance: Use SPC's to fix the constant temperature boundary condition. Use TEMPBC TYPE="TRAN" to designate the time-varying boundary condition. @ 5927; 68.2; Matrix Methods - MCE1, RBEi, MPC - UWM 6137, INPUT MATRIX IS RAN K DEFICIENT Although it is not illegal to specify a DOF as independent on one rigid element and dependent on another rigid element, if the definitions form a "closed loop" such that DOF A is dependent on DOF B and DOF B is either directly or indirectly dependent on DOF A, and sparse decomposition is selected the message, "USER WARNING MESSAGE 6137, INPUT MATRIX IS RANK DEFICIENT" is output. Avoidance: This error is corrected in Version 67 and later systems for standard decomposition. Thus, if the user forces standard decomposition by setting the SPARSE keyword to 1, then USER FATAL MESSAGE 6202 is produced which identifies the rigid elements in error. @ 5930; 68.2; Transient Heat Transfer - MATT4, RADMT - UWM 2013 In a transient thermal analysis (SOL 159), if QVECT load is used in conjunction with temperature dependent absorptivity. This input leads to the message "USER WARNING MESSAGE 2013, NO STRUCTURAL ELEMENTS EXIST. FATAL ERROR" being issued. The QVECT load is directional load such as solar load. Avoidance: Defines a constant absorptivity value instead of temperature dependent absorptivity value. @ 5931; ; DMAP - COMPILE - UFM 151 (LINKAR) A DMAP compile error may be encountered when logical parameters are in a call statement. An avoidance is to replace the logical in the call statement with another type, such as an integer or to delete the parameter from the subdmap call altogether. @ 5933; GL; DMAP - DCMP, DECOMP - Wrong (Zero) Results The DECOMP and DCMP modules have an optional set of parameters which should give the scaled determinant (DET) and the scale factor (POWER) of the input matrix. For the sparse methods, they are set to zero. Avoidance: An avoidance is to select the nonsparse decomposition methods

when a determinant is needed. That is set system cell 126 to 1 (a.k.a. NASTRAN SPARSE keyword) for symmetric matrices and system cell 209 equal to 0 (a.k.a. NASTRAN USPARSE keyword) for unsymmetric matrices. @ 5934; 68.1; Documentation - GMBC - UFM 315 The following message is issued when the LID field of the GMBC entry is 0 and it contains a continuation entry, i.e., ENTITY = FEFACE and METHOD = LINEAR. USER FATAL MESSAGE 315 (IFPDRV) FORMAT ERROR ON BULK DATA ENTRY GMBC SORTED CARD COUNT = ** 0

Remark 4 under the Version 68 Quick Reference Guide's description of the GMBC entry should be updated to include this condition and it should be stated that not only is the METHOD field not required but will result in a USER FATAL MESSAGE if included. This error is essentially the same as error 5506, it has been corrected in the 2nd printing of the Quick Reference Guide. @ 5935; 69; Miscellaneous - ADD5 - SFM 4276, EC 70020 An error in the bufferpool logic may result in "SYSTEM FATAL ERROR 4276.BPREAD ERROR CODE 70020" to be issued from the ADD5 module. This error is very much problem and buffersize dependent. Avoidance: Changing the buffersize may avoid the error but the sure avoidance is to turn off bufferpooling prior to the ADD5 module by setting SYSTEM(119) as follows: ALTER xxx $ turn off buffpool just before ADD5 putsys (2,119) ALTER yyy $ turn buffpool back on just after ADD5 putsys (3,119) @ 5941; 69; Optimization/Design Sensitivity - DOPTPRM - Missing Output If the P1 parameter on the DOPTPRM bulk data entry is either set to the default value of 0 or simply not specified and allowed to become the default value of 0, and the P2 field is specifed, then requested data on the P2 field is not printed for the last design cycle. Avoidance: Set parameter P1 to 1 to obtain the output for the last design cycle or apply the following dmap alter: $ $ This alter will print the design optimization output $ for the last design cycle regardless of what the value $ of the 'p1' parameter in the doptprm entry is set to. $ ksh: cray.err: permission denied $ **** THIS ALTER IS FOR VERSION 68.2 ONLY **** $ compile desopt nolist noref souin=mscsou alter 'CALL PREDOM'

$ dummy matgen calls to satisfy dmap complier file cntabrgp=save,ovrwrt/cvalrgp=save,ovrwrt/cvalox=save,ovrwrt/ propip=save,ovrwrt/propox=save,ovrwrt/r1tabrgp=save,ovrwrt/ r1valrgp=save,ovrwrt/r1valox=save,ovrwrt/rsp2rgp=save,ovrwrt/ r2valrgp=save,ovrwrt/r2valox=save,ovrwrt/optprmgp=save,ovrwrt/ objtbgp=save,ovrwrt/drstblgp=save,ovrwrt/frqrprgp=save,ovrwrt $ if ( desiter=-1 ) then matgen ,/cntabrgp/1/ $ matgen ,/cvalrgp/1/ $ matgen ,/cvalox/1/ $ matgen ,/propip/1/ $ matgen ,/propox/1/ $ matgen ,/r1tabrgp/1/ $ matgen ,/r1valrgp/1/ $ matgen ,/r1valox/1/ $ matgen ,/rsp2rgp/1/ $ matgen ,/r2valrgp/1/ $ matgen ,/r2valox/1/ $ matgen ,/optprmgp/1/ $ matgen ,/objtbgp/1/ $ matgen ,/drstblgp/1/ $ matgen ,/frqrprgp/1/ $ endif $ $ alter 'IF ( CNVFLG = 2 ) THEN' $ message //'dom10 call after hard convergence - descycle='/descycle $ $ type parm,,i,n,desp $ desp = descycle - 1 $ DOM10, DESTAB,xprev,xinit,CNTABRGp, CVALRGp, CVALOx, DVPTAB,PROPIp,PROPOx,R1TABRGp,R1VALRGp,R1VALOx, RSP2RGp,R2VALRGp,R2VALOx,OPTPRMGp,OBJTbgp,DRSTBLgp, tol1,fol1,frqrprgp// desp/desp/objpv/OBJOUT/ $ $ alter 'DOM9,' $ $ save dom10 inputs equivx cntabrg/cntabrgp/always $ equivx cvalrg/cvalrgp/always $ equivx cvalo/cvalox/always $ equivx propi/propip/always $ equivx propo/propox/always $ equivx r1tabrg/r1tabrgp/always $ equivx r1valrg/r1valrgp/always $ equivx r1valo/r1valox/always $ equivx rsp2rg/rsp2rgp/always $ equivx r2valrg/r2valrgp/always $ equivx r2valo/r2valox/always $ equivx optprmg/optprmgp/always $ equivx objtbg/objtbgp/always $ equivx drstblg/drstblgp/always $ equivx frqrprg/frqpprg/always $ $ alter 'CNVFLG=1' $ DOM10, DESTAB,XINIT,XO,CNTABRG, CVALRG, CVALO,

DVPTAB,PROPI,PROPO,R1TABRG,R1VALRG,R1VALO, RSP2RG,R2VALRG,R2VALO,OPTPRMG,OBJTbg,DRSTBLg, tol1,fol1,frqrprg// DESCYCLE/DESMAX/OBJIN/OBJOUT/ $ $ endalter @ 5946; 68.2; Optimization/Design Sensitivity - DRESP1 - Uses Von Mises Only Remark 1 for the DRESP1 entry indicates that, for stress/strain item codes that have dual meanings, the selection is made based on the option specified in case control. This pertains to Von Mises and Max Shear selection. Due to a programming error, the remark is incorrect and Von Mises results will be selected, regardless of the case control request, as it was in Version 67. Avoidance: Users desiring to use Max Shear in a design task will have to synthesize it from basic stress results. @ 5948; 70.5; Documentation - QRG, USET, PARTN - SFM 3007 If a USETi bulk data entry lists a standard MSC/NASTRAN set, such as S or M, the program may fail in the PARTN module with the message "SYSTEM FATAL MESSAGE 3007, ILLEGAL INPUT TO SUBROUTINE". Avoidance: This entry should only reference new sets defined on DEFUSET bulk data entries. @ 5953; 68.2; Optimization/Design Sensitivity - DCONADD,DCONSTR - UFM 316 If the DCONADD entry references only a single DCONSTR entry on any continuation entry, then the job fails and the following message is issued: USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA CARD DCONADD CARD COUNT = yy xx SORTED

An example of a DCONADD entry that causes this error is as follows: DCONADD, 1, 10, 11, 12, 13, 14, 15, 16, +DCON1 +DCON1, 17, 18, 19, 20, 21, 22, 23, 24, +DCON2 +DCON2, 25 $ Note that the second continuation entry contains a single DCONSTR entry (25 in this case). Avoidance: Add a dummy DCONSTR to the DCONADD continuation entry. The limits should be set to a large value to ensure that the constraint screening will remove this constraint prior to performing optimization. A suggested lower bound (LALLOW) and upper bound (UALLOW) values would be -1.0E20 and 1.0E20, respectively. @ 5954; 69; Grid Point Force - DEFORM, GPFORCE, GPFDR - SFM 6223 (GPFDR2), S FM 4276, EC 3012 When the DEFORM command is used and ELAS elements exist in the deck, the

following fatal message is issued from the GPFDR module. *** SYSTEM FATAL MESSAGE 6223 (GPFDR2), RECORD POINTED AT BY INDEX IN SCRATCH FILE IS WRONG. USER ACTION: CONTACT MSC CLIENT SUPPORT. CASECC NUMBER LOAD SET NO ELEMENT TYPE EXT ELEMENT ID *** SYSTEM FATAL = 4 = 4 = 1 = 101 ERROR 4276.OPEN

ERROR CODE 3012

Avoidance: Do not request GPFORCE output for the subcase with DEFORM loads. @ 5956; 68.2; Old Design Sensitivity - PARAM,POST,-2 - Missing Datablocks In MSC/NASTRAN Version 68, if sensitivity output (SENSITY command) is requested using SOL 103 in conjunction with param,post,-2, not all datablocks are written to the output2 file. Note this error applies only to the old design sensitivity method. It does not apply to the new sensitivity calculation (SOL 200). Avoidance: Insert the following DMAP alters. compile super3 souin=mscsou nolist noref $ alter 90,90 $ gp1 GEOM1,GEOM2,,,/GPL,EQEXIN,GPDT,CSTM,BGPDT,SIL,/0/0/0 $ alter 208,208 $ alter 211,211 $ @ 5960; 4276 69; Buckling Analysis - EIGRL, SYSTEM(146) - SFM 5407, SFM 5401, SFM If SYSTEM(146) is set to save scratch space and reduce I/O in sparse decomposition during a Lanczos eigenvalue buckling run, then the job fails. Different fatal messages have been encountered including "SYSTEM FATAL MESSAGE 5407 INERTIA (STURM SEQUENCE) COUNT DISAGREES WITH THE NUMBER OF MODES ACTUALLY COMPUTED IN A (SUB-) INTERVAL", "SYSTEM FATAL MESSAGE 5401 (REIGL) LANCZOS METHOD IS UNABLE TO FIND ALL THE EIGENVALUES IN RANGE" AND "SYSTEM FATAL MESSAGE RECOVERY ERROR CODE 7777" in sparse decomposition. Avoidance: Do not use SYSTEM(146) during a Lanczos buckling run. @ 5961; 68.2; Heat transfer - SET1, RADCAV - Wrong Answers If the SET identification pairs for the calculation of global view factors exceeds 10 pairs, then an erroneous result is output for the global view factors. The documentation states that only 10 pairs can be shown for one run. This limit is expanded to 30 in Version 68.2. @ 5962; 70.5; Acoustic Analysis - REIGL, SUPORT, MERGE - UFM 4276, EC 3007 SUPORT entries do not work in acoustic analysis. USER FATAL MESSAGE 4276, ERROR CODE 3007 from subroutine merge called from the REIGL module

is issued. Other eigenvalue extraction methods result in other fatal messages. Avoidance: Remove SUPORT entries. @ 5963; 69; Case Control - STRAIN(CORNER) - Missing Output If a STRAIN(CORNER) request is made without a STRESS(CORNER) request, only centroidal strains are output. Avoidance: Request STRESS(CORNER,PLOT) in the Case Control Section. @ 5967; 70.5; DMAP - MPYAD - Wrong Answers The MPYAD module performs the following matrix operation: [A][B]+[C]=[D] with the following specification: MPYAD A,B,C/D $ If [A] does not exist then the MPYAD module assumes [A] is null and is supposed to make [D] equivalent to [C]. However, if [C] is complex then [D] contains only the real part of [C]. Avoidance: Test for the presence of [A] before executing MPYAD: PARAML A//'pres'////s,n,noamat $ if ( noamat>-1) then $ MPYAD A,B,C/D $ else $ EQUIVX C/D/-1 $ endif $ @ 5974; 68.2; Miscellaneous - Accounting Files do not Include the Time Zone The accounting files generated on some versions of AIX will not include the time zone when LANG="". Since the various scripts must set this to function properly, the only recourse is for the execution and accounting summary scripts to account for this potential discrepancy. Avoidance: The running summary written by the execution script, i.e., nast68 or emas3, will be incorrect. The accounting summary script can be corrected by manually editing the accounting file to insert the time zone parameter into each line as appropriate. The accounting summary script will then report correct summary information. @ 5977; 69; Buckling Analysis - SPCD - Wrong Answers If SPCD entries are used in SOL 105 and the SUBCASE used for the buckling analysis does not contain a call for a LOAD set using the SPCDs, the SPCD entries are ignored in the solution. Avoidance: Either use different SPC sets for each SUBCASE, or include a LOAD command in the buckling SUBCASE pointing to the SPCD set. @ 5980; 68.2; Executive - XEQUIV - SFM 4276 EC 7777 If during the equivalence of a local scratch datablock the PVA table runs out of space, the expansion of the PVA table may corrupt open core. Jobs

that encounter this error may fail in many unpredictable ways. The most likely failure will be "SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777". The job that encountered this error is an optimization analysis with image superelements. Avoidance: Increase or decrease the size of RAM with the INIT FMS statement. @ 5981; ; Optimization/Design Sensitivity - DSAD, FRRD1 - Insufficient Tim e, SFM 4276 If an optimization run encounters an "INSUFFICIENT TIME" type message during the finite element part of the solution (the example encountered was insufficient time to complete the frequency response solution, FRRD1) it is possible the run will continue into the optimization process and fail in the optimizer (the example fails in DSAD with SFM 4276, ERROR CODE 3001). If an unexplained error occurs during optimization, check the output file for other messages. Avoidance: Specify sufficient time on the TIME statement in the executive control section. @ 5982; 69; Eigenvalue Analysis - AGIV, AHOU - SDCOMP FATAL ERROR, UIM 5458 If an eigenvalue analysis specifies either the AGIV or the AHOU method, then standard GIV and standard HOU are attempted first. If the mass matrix is singular, an unnumbered fatal message "SDCOMP FATAL" is issued even though the problem continues and the program switches to the modified Givens or Householder method. Avoidance: If the problem runs, ignore the FATAL message. @ 5983; 69; Optimization/Design Sensitivity - DOPTPRM, DOM10 - SFM 3001 An optimization run with P1=1 and P2=15 on the DOPTPRM entry fails and issues "SYSTEM FATAL MESSAGE 3001" from the DOM10 module even though the run works fine without these parameters. These are output control parameters and should not effect the solution. Avoidance: If P1 and P2 have been used on the DOPTPRM and the run fails, try the run again without specifying P1 and P2. @ 5984; 68.2; Nonlinear Analysis - PARAM,NMLOOP , SE's - Missing Output When using the nonlinear normal modes option in Sol 106, no modal analysis is performed when there are superelements and PARAM,NMLOOP is specified in the Case Control. Avoidance: Place the PARAM,NMLOOP statement in the bulk data. or Use SUPER=ALL in the case control. or DMAP alter:

compile nlstatic souin=mscsou nolist noref $ alter 480 $ seid= 0 $ pvt pvts,cases/pvtxx/ $ endalter $ @ 5986; 68.2; Superelements - PARAM, EXTRCV - Wrong Answers When using PARAM, EXTRCV with the RESTART FMS statement, wrong answers result if the residual structure subcase qualifiers (LOAD, SPC, MPC, etc.) differ from those specified in the residual structure subcase of the run which generates the external superelement. Avoidances: 1. Maintain the same residual structure subcase qualifiers in all runs. 2. Insert the following alter in the PARAM, EXTRCV run: COMPILE IFPL souin=mscsou, noref nolist alter 152,152 3. Use the multi-MASTER approach in the PARAM, EXTRCV run. Assuming that the external superelement data base is called extse1.master, replace: RESTART and ASSIGN MASTER, if any, with the following: nastran system(143)=1 assi extse1=extse1.master dbloc logi=extse1 ASSIGN MASTER, if any, with the following: nastran system(143)=1 assi extse1=extse1.master dbloc logi=extse1 Insert PARAM, SERST, SEDR into the Case Control or Bulk Data. @ 5987; 68.2; Postprocessing - PARAM,POST,-2 - IDEAS Dataloader Abort In MSC/NASTRAN Version 68, the structured solution sequences with PARAM.CHECKOUT,YES and PARAM,POST,-2 do not write an END-OF-DATA on the FORTRAN file. Avoidance: COMPILE PHASE0 SOUIN=MSCSOU NOREF NOLIST ALTER 527 $ Before CALL PRTSUM OUTPUT2 //-9/OUNIT1//OMAXR $ write END-OF-DATA CEND @ 5992; 68.2; Grid Point Stress - GPSTRESS, SURFACE, QUAD4 - Missing Output Certain setids (120 thru 138, 148 thru 159, 174 thru 181) may cause incomplete output for GPSTRESS when used in combination with the SURFACE

case control command. An example case control which results in the error follows: SET 21=120 STRFIELD=21 GPSTRESS=21 OUTPUT(POST) SET 95=19,20,21,59,60,61 SURFACE 120 SET 95 NORMAL Z SYSTEM ELEMENT BEGIN BULK @ 5993; 69; Documentation - PBEAM - Continuation Entries The documentation for the PBEAM entry in the MSC/NASTRAN Version 68 Quick Reference Guide states that the fields "SO" and "X/XB" on the second continuation entry are required. A footnote should be included (as was in the Version 66 User's Manual) that states that these fields are only required when the second continuation entry is desired to request intermediate stress output. @ 5995; ; Elements - QUAD4, TEMP(LOAD) - Wrong answers Warped QUAD4 elements with thermal loads give wrong answers. This error is similar to the thermal expansion error in elements with rigid offsets, there is no expansion for the offset. If offsets and orthotropic materials are present, the errors are magnified. @ 5998; ut ; Documentation - Linear Static Analysis User's Guide - Wrong outp Incorrect output result files are included in Figures 11-2 and 11-3 of the MSC/NASTRAN Linear Static Analysis User's Guide. Due to a missing carriage control character, a pload1 entry for element # 2 is missing for the inertia2.dat file that is delivered with the delivery tape. Furthermore, due to error 5522, PLOAD1 yields incorrect results for the BAR, BEAM, or BEND elements when used in conjunction with inertia relief. Thus, this example (using PLOAD1) should not be emulated until error 5522 is corrected. Avoidance: Replace the PLOAD1 with equivalent grid point loads and regenerate the output file. @ 6000; 2001; Transient Analysis - RESTART, PARAM,NOAP - Poor Answers If TIC entries are used to set initial velocity values, the computed initial velocity will be slightly different from the specified initial velocity. On a case where v0 was set to 1.0, the output velocity was computed as 0.9969. This model has a dt of 0.1 Avoidance: An avoidance is to run the model once to determine the c omputed initial velocity, and increase the input initial velocity enough so that the computed intital velocity is the value desired on a second run. @ 6004; 68.2; SSSALTER Library - ZFREQ - Wrong Answers

The ZFREQ alter delivered in the Version 68 SSSALTER library allows frequency dependent springs and dampers to be defined. However, the nominal spring stiffness is not updated in the data recovery phase. Thus, element force output for these elements is incorrect. This is a known limitation of the alter that should have been documented in the ZFREQ.RDM file. The information contained in this error report is to be included in the Version 68.2 ZFREQ.RDM file and the error is considered corrected. Avoidance: An avoidance is to place a stiff spring in series with the mount, the forces in the stiff spring represent the correct element force for the mount spring. The error introduced by this technique is proportional to the ratio of the stiff spring rate to the mount spring rate. For example, if the ratio is 1000:1, the overall mount spring rate across both devices is 0.1% lower than the actual mount rate. The user may modify his mount spring rate data to account for this effect. In most instances the mount spring rates are not known to this level of accuracy, so that the effect may be ignored. To avoid numerical difficulties the stiff spring should be of the same order of stiffness as the structural elements to which it attaches. In most instances the structure to which the mount attaches is much stiffer than the mount. @ 6007; ut 69; Documentation - PARAM, USETPRT PARAM, USETEL - K-set is not Outp The MSC/NASTRAN Quick Reference Guide incorrectly states that the parameters usetprt and usetsel can be used to print k-set degrees of freedom. That is, the k-set decimal equivalent is listed as an option in table 1 of the usetprt parameter description. @ 6010; 68.2; Nonlinear Slide Line Contact - BCONP, BLSEG - Wrong Answers An overhang slave node is not considered as "overhang" if the overhang value is less than 5% of the master segment length. For models with large master segments, this may lead to wrong answers. Typical symptoms include the printout of a wrong slip code (e.g., SLIDE instead of OVERHANG) and incorrect surface coordinates. Avoidance: Use smaller master segments when modeling overhang situations. @ 6011; 70; Security - NAO - UFM 3060, Segmentation Fault, Signal Number 11 in Log File Large jobs that use Network Authorization (NAO) may fail in various moudules on the DEC/ALPHA OSF with the message, "Receiving signal number 11: Segmentation fault", in the MSC/NASTRAN log file. Similarly, if using node lock authorization, and running a looping solution sequence like nonlinear and a large problem, then User Fatal Message 1019 "Fortran Unit Open Error" pointing to the authorization file may be issued. This message is followed by USER FATAL MESSAGE 3060. The problem occurs because a socket (file) is left open each time MSC/NASTRAN checks the authorization code and when the system reaches the open sockets limit, the job fails as indicated above. Avoidance: Edit the system configuration table (sysconfigtab) and increase

the value of the OPEN_MAX parameter. (This is a system administrator job that may require assistance from DEC.) Then reboot the machine. To determine the curren OPEN_MAX value, execute the command: getconf OPEN_MAX @ 6012; 69; Dynamic Response Analysis - EPOINT - Single Precision Accuracy If there are EPOINT entries in the bulk data file, the eigenvalue matrix PHIA is expanded to the d-set matrix PHIDH in the GKAM module. On short-word machines PHIA is double precision. At present PHIDH is single precision. As this matrix is used to reduce load matrices and expand modal solutions to physical DOFs this may reduce the precision of the solution. Avoidance: An avoidance is to discard the PHIDH matrix output by the GKAM module and build an alternate matrix with DMAP. The equations defining how PHIDH is formed from PHIA are given in the MSC/NASTRAN REFERENCE MANUAL. @ 6013; 69; DMAP - DMIG - Wrong Answers If a DMIG entry references a degree of freedom which does not exist, the corresponding coefficients are assigned to another degree of freedom, without any diagnostic messages. In one example, an open set was used to define a series of scalar points. A DMIG term referenced a point that was not defined on an SPOINT entry which occurred between two points which were listed on SPOINT entries. The coefficient was placed on the last DOF. @ 6015; 68.2; Documentation - Aeroelastic Users' Guide - UFM 2268 (FA1) Example HA145I in the Version 68 Aeroelastic User's Guide alter that is in error. The alter inserts a MATPRN after should be after line 45. If submitted as documented, the issues the message "USER FATAL MESSAGE 2268 (FA1) FMETHOD FOUND". contains an line 50. It run fails and SET 0 NOT

The data deck delivered in the tpl has been made current. The documentation will be updated to reflect the latest version when it is updated. @ 6016; 50) 68.2; P-Elements - HEXA, PENTA - Excessive Pivot Ratios at p=2 (UFM 90 If a model consists of HEXAs or PENTAs at a p-level=2 or higher connected end-to-end, the following message may occur: DMAP FATAL MESSAGE 9050 (SEKRRS) - RUN TERMINATED DUE TO EXCESSIVE PIVOT RATIOS IN MATRIX KLL. USER PARAMETER BAILOUT MAY BE USED TO CONTINUE THE RUN. This fatal message is caused by the reduced integration used at a p-level=2 and higher. Reduced integration is the current default. Avoidance: SYSTEM(212)=1 may be set to work around the problem. @ 6017; 68.2; Nonlinear Slide Line Contact - BCONP, BLSEG - Extraneous Output The following parameters are printed in the "LINE CONTACT REGION" output

when using the slide line contact capability. CONT BEHA TYPE FO CO

These values should be ignored as they are not implemented in the program. Avoidance: These values have no effect on the rest of the output and they can be safely ignored. @ 6018; 69; Acoustic Analysis - SORTi - SORT2 Output in Place Of SORT1 In solution 108, a request for SORT1 output in an acoustic analysis results in SORT2 output. Avoidance: Add PARAM,CURVPLOT,1 to the bulk data. @ 6024; 68.2; Optimization/Design Sensitivity - DLINK - Wrong Answers If the user inputs a nonzero value for C0 on the DLINK entry, it is incorrectly processed by the optimization algorithm. Values used by the optimizer for the dependent design variables will be incorrect. This will directly affect constraint calculations from the upper and lower bounds input on the associated DESVAR entry. These constraints may not be critical and and the optimization results may be valid. The printed summary values for the dependent design variables will be incorrect. Avoidance: There is no avoidance. @ 6025; 68.2; Plotting - XYPLOT(TEMP) - Infinite Loop If the SET1 entry references non-existing grid IDs, then the XYPLOT command which reference this SET1 entry cannot identify the missing grid point(s), which causes the program to go into an infinite loop. Avoidance: Check all the grid points on the SET1 entry and make sure they exist. @ 6030; 68.2; Optimization/Design Sensitivity - MPCFORCE, PARTN, DSAP - UFM 21 66, UFM 5901 In Version 68, if MPC forces are requested in an optimization run under static analysis, the program fails at line 121 of subDMAP SEDISP with UFM 2166 stating that the number of columns in the partition vector is incorrect. Under frequency response analysis, UFM 5901 will be issued from the DSAP module. Avoidance: Either remove the MPCFORCE output request or apply the following DMAP alter to avoid the error: $for statics approach: compile sedisp, souin=mscsou, nolist, noref alter 109 noqmg=ltoi(delta) $ $ For Frequency approach: compile sedisp, souin=mscsou, nolist, noref alter 238 if (not delta ) then $

alter 239 endif $ @ 6031; 68.2; Documentation - gentim2.dat - Data File Not Found The MSC/NASTRAN V68 Installation and Operation Instructions implies that the gentim2.dat file used to update timing constants is in the same directory with the SSS.MASTERA and USS.MASTERA (ie /msc68/nast68/aix/del) but it is not. The file is located in the /msc68/nast68/del directory. The install guide should be clarified to tell users where the data file is located. @ 6033; 68.2; Elements - TETRA(10), TEMP(LOAD) - Wrong Answers for TETRA(10) S tresses If the temperature at the grid points of a TETRA(10) element vary (i.e. they are not constant), then the stresses for the element are incorrect. Avoidance: The error is proportional to the variance of temperature within the element; however, refining the mesh so that the temperature within each element is near a constant is not considered a practical avoidance. Stresses requested at the gauss points are correct. @ 6034; 3.2; Documentation - Installation and Operation Instructions - Timing Constants Section 3.6 of the MSC/NASTRAN Version 68 Installation and Operation Instructions explains how to install the timing constants. Step 2 of the documented procedure states: "2. The timing data are generated or changed by running gentim2.dat (delivered in the "del" library)..." Later in this step you are instructed to change to the directory install-dir/msc68/{prog}/{mach}/del where the files *SS.MASTERA, etc are stored. In other words, there are two "del" directories 1) install-dir/msc68/{prog}/{mach}/del and 2) install-dir/msc68/{prog}/del The first directory contains the *SS.* files and the second contains the gentim* files. The installation procedure should be more clear. Avoidance: Copy the file install-dir/msc68/{prog}/del/gentim2.dat to install-dir/msc68/{prog}/{mach}/del @ 6035; 69; Nonlinear Analysis - NLITER, NLPARM, INTOUT=NO - UFM 4551 Nonlinear jobs have been found that converge successfully when INTOUT=YES but fail to converge and issue "USER FATAL MESSAGE 4551 STOPPED PROBLEM DUE TO FAILED CONVERGE" when INTOUT=NO. The exact cause of the problem is under investigation. @

6036;

68.2; P-Elements - Punch Output not Supported When stresses and strains are output in the punch file, they are output in the print file (F06) format, with titles and labels, rather than a format suitable for the punch file.

@ 6038; 69; Elements - CTRIA3, TA1 - No Message, Trap Encountered If a CTRIA3 element references a nonexistent material coordinate system (field 7), the job should issue USER FATAL MESSAGE 2025 UNDEFINED COORDINATE SYSTEM. The IBM RS/6000 versions of MSC/NASTRAN do not issue this message. The job stops without a message being issued. The f04, Execution Summary, indicates that the last module executed is either TA1 or PHASE1DR and the log file contains a "Trap Encountered" message. @ 6042; GL; Elements - PENTA(6), TEMP(LOAD) - Wrong Answers for Stresses If the temperature at the grid points of a PENTA(6) element vary (i.e. they are not the same at all points), then the stress output is wrong. Both corner point (default) and Gauss location (selected on PSOLID) options give wrong answers. The stress error is proportional to the amount of temperature change within the element. Avoidance: This is a limitation of the element. The only avoidance is to refine the mesh so that the variation of temperature within the element is minimal. @ 6044; 68.2; Optimization/Design Sensitivity - Incorrect Sensitivities for Su perelements Incorrect design sensitivities are computed for superelement models in which a superelement has design variables but no design responses. This error occurs only in statics. Avoidance: The avoidance is to specify WEIGHT=ALL (for all superelements) as a design response. @ 6045; 2001; Installation - Timing Constants - UWM 6080 The text of "USER WARNING MESSAGE 6080 (SETKRN), THE TIMING CONSTANTS DATA BLOCK TIMEBLK NOT FOUND ON THE DELIVERY DATABASE" refers the user to "AM 7.6". This reference should changed to the a more generic term, such as machine dependent documentation. @ 6046; n Setting PARAM,DESPCH to a negative value will not prevent DESVAR and GRID Entries from being written to the PUNCH file. Avoidance: The following DMAP alter may be used to avoid the problem: compile exitopt nolist noref souin=mscsou alter 24 $ just before executing module DOM11 68.2; Optimization/Design Sensitivity - PARAM,DESPCH - Incorrect Actio

type parm,,i,n,desitert if (despch < 0) then desitert = desiter desiter = 0 endif alter 25 $ right after execution of module DOM11 if (despch < 0) then desiter = desitert endif $$$ end of alter @ 6047; 68.2; SSSALTER Library - RESFLEXR - Wrong Answers The alter to add residual vectors into a dynamic solution delivered in the sssalter directory with Version 68 does not handle SUPORT entries properly for free-free structures. If a SUPORT entry is present, residual vectors are not generated for the residual structure. Avoidance: Call MSC for an updated version of the alter. @ 6050; 68.2; P-Elements - GMLOAD in convective coordinate systems If a GMLOAD is applied in a convective coordinate system defined on a GMCURV or GMSURF where the CIDIN is other than basic coordinate system, then the following message is output from SSG1: USER FATAL MESSAGE 2025, UNDEFINED COORDINATE SYSTEM FATAL ERROR 2

Avoidance: Possible avoidances include defining the GMCURV or GMSURF in the basic coordinate system, and applying the load in a non-convective system, or using PLOAD4. @ 6051; 69; Documentation - MAT1 - Programs Action not Clear Remark 3 under the documentation for the MAT1 entry in the Quick Reference Guide states that the relationship between E, G and NU "must be satisfied". It should be clarified that satisfying the relationship is merely a recommendation, rather than a requirement. The values furnished are used as indicated in Table 1. and if the relationship is not satisfied the program proceeds but issues USER WARNING MESSAGE 2251A and 2251B. Furthermore, the relationship indicates ">" when it should indicate "<". @ 6052; 69; Acoustic Analysis - ACMG - Job Aborted, No Message, Receiving Si gnal Number 10 An acoustic job that specifies DIFF on the ACMODL entry aborted in the ACMG module with no message issued in the f06 file. A message may appear in the log file indicating the job has aborted but the message will be different depending on the computer. One such message is "Receiving signal number 10: Bus error" Avoidance: The cause of the error is under investigation. Selecting the IDENT method on the ACMODL entry will probably allow the problem to execute to complete but unless the fluid and structural mesh are compatible will not yield the correct results. @ 6058; 68.2; P-Elements - Convective output coordinate system

If a convective coordinate system is specified for the output coordinate system of a GMSURF, GMCURV, FEFACE, FEEDGE, or GRID, the following message is output from GP1: *** USER FATAL MESSAGE 2104, UNDEFINED COORDINATE SYSTEM Avoidance: There is no avoidance. @ 6059; ; Nonlinear Static Analysis - Restarts - Missing Displacements In the nonlinear static solution sequences, displacement output is missing from the restart run, when the following conditions are met: 1) A new subcase is added which requests intermediate output (INTOUT=YES on NLPARM). 2) The restart run regresses to a subcase prior to the last executed subcase (using PARAM,LOOPID and PARAM,SUBID). In other words, if the cold start finished with subcase 4, the new restart subcase 5 would start with the last LOOPID of subcase 3 or prior and a SUBID of 5 (effectively skipping subcase 4). When these conditions are met, the .F06 and the punch file contains ONLY the first intermediate displacements for the new subcase, and no other displacement output for that subcase. Other forms of output are unaffected. Avoidance: Avoid one of the above conditions. @ 6065; MAT2 70.7; Composites - PCOMP, MID4, TEMP(LOAD) - Incorrect Term Output for If unsymetric thermal properties and thermal loads are present on a composite plate that is using the PSHELL/MAT2 instead of the PCOMP/MATi method of composite property description, then the PSHELL's MID4 MAT2 must have a ID greater than 400,000,000 and the thermal properties in that MAT2 must be written in a special nonstandard format (see Remark 13 on the MAT2 description of the MSC/NASTRAN Quick Reference Guide). This is because the MSC/NASTRAN composite formulation is different than the noncomposite formulation for unsymmetric thermal loading. The MID4 MAT2 ID is the flag that tells MSC/NASTRAN whether the PSHELL/MAT2s are composite or not composite. PSHELL/MAT2s that are internally generated from PCOMPs and which are written to the .f06 file or to the .pch file have the correct material IDs on them, and if used without change, will produce correct results. However, if the MID4 MAT2 ID is changed to a number less than 400,000,000 and unsymmetric thermal material properties and loads are present, or if user supplied MAT2s are used and do not follow the above rules, incorrect results will occur. Note that a double field format is required on the MAT2 in order to have an ID this large. The MAT2s echoed in the .pch file are already in a double field format and thus may be used without change. The MAT2s echoed in the .f06 file however, are in a nonstandard print format and must be manually changed to a standard MSC/NASTRAN double field format. Avoidance: Use a MID4 MAT2 ID > 400,000,000 and the proper thermal property format 13

for composites. @ 6067; 68.2; P-Elements - Punch File Output not Supported for Error Tables When the error table is output in the punch file, it is output in the print file (F06) format, with titles and labels, rather than a format suitable for the punch file. @ 6068; 68.2; P-Elements - GMBC - UIM 6353 If a zero displacement value is entered on the GMBC entry, the following message is issued for each grid point associated with the GMEDGE or GMFACE entry: USER INFORMATION MESSAGE 6353 (EGENDD) ENFORCED DISPLACEMENT VECTOR IS NULL FOR COMPONENT=** OF EDGE ID= *** ON GMBC LOAD ID= 0 AND SPC ID= ** Avoidance: Although there is no way to avoid the information message from being output, there is no other adverse affect on the run and the message may be ignored. @ 6070; 68.2; Aeroelastic Analysis - PARAM,LFREQ, OPPHIPA, HFREQ - UFM 5423 A flutter or dynamic aeroelastic run that uses PARAM LFREQ or HFREQ to eliminate modes and PARAM OPPHIPA to print the normal mode shapes including aerodynamic grid points, fails with the message "USER FATAL MESSAGE 5423 ATTEMPT TO MERGE INCOMPATIBLE MATRICES". Avoidance: There is no avoidance other than do not use PARAM LFREQ or HFREQ in combination with OPPHIPA. @ 6072; 69; Optimization/Design Sensitivity - DEQATN - UFM 224 (PRERMM), If the PI(1) function on the DEQATN entry is used with an integer, it results in a mixed mode arithmetic operation error and the following message is issued "USER FATAL MESSAGE 224 (PRERMM) MIXED MODE USAGE BY THE ARITHMETIC OPERATOR * IS INVALID DURING DEQATN PROCESSING". Avoidance: Use a real number instead of an integer. For example, PI(1.0) is acceptable and produces the desired result. The user could also define pi on a DTABLE entry as a constant and not rely on the DEQATN function. @ 6079; output 69; Dynamics - Data recovery with PARAM,DDRMM,0 (default) - Missing By default (PARAM,DDRMM,0) in dynamic data recovery for modal frequency or modal transient analysis (SOLs 11, 12, 111, 112, 146, and 200) any FORCE, STRESS, SPCF (if PARAM,DYNSPCF,OLD is specified), or STRAIN request in a subcase subsequent to the first subcase is not honored. For example, if the subcase structure shown below is specified then the forces for element 20, 40 and 60 are not output. set 1=100,200 set 2=20,40,60 subcase 10 FORCE = 1 Subcase 20 FORCE = 2

Avoidance: Including PARAM,DDRMM,-1 avoids the problem but this forces the output to be in SORT1 format. Another avoidance is to add a new subcase above the existing subcase structure for the FORCE(PLOT)=ALL and STRESS(PLOT)=ALL as shown below. With this avoidance the output is in SORT2 format. set 2=20,40,60 Subcase 1 FORCE(PLOT)=ALL subcase 10 FORCE = 1 Subcase 20 FORCE = 2 @ 6080; 70.5; Eigenvalue Analysis - REIGL - SFM 5407, SFM 5401 Large Lanczos buckling runs may fail in the REIGL module on the CRAY C-90 release of MSC/NASTRAN with the following messages: SYSTEM FATAL MESSAGE 5407. INERTIA (STURM SEQUENCE) COUNT DISAGREES WITH THE NUMBER OF MODES ACTUALLY COMPUTED IN AN (SUB-)INTERVAL SYSTEM FATAL MESSAGE 5401 (REIGL). USER FATAL MESSAGE 4405, NO EIGENVECTORS COMPUTED FOR COMPONENT MODE SYNTHESIS OR SYSTEM SOLUTION. Avoidance 1: An avoidance is to set SHFSCL on the EIGRL entry near the absolute value of the lowest expected root. Avoidance 2 (preferred over Avoidance 1): include sss alter br.v69 for Nastran version 69.0.2 or br.v691 for Nastran version 69.1. Leave the SHFSCL entry on the EIGRL card blank. @ 6086; 68.2; Superelements - PARAM, ERROR, 0 - Misleading Messages, Incorrect Action PARAM, ERROR, 0 does not work in SOLs 101 through 200 in Version 68. If an error occurs in a superelement, or the residual structure, the program will incorrectly continue processing the same superelement rather than skip to the next superelement, or if the residual structure is being process, terminate the run. Avoidance: Insert the following DMAP alter: compile phase1dr souin=mscsou noref nolist alter 100 if ( not skipse ) then $ alter 101 endif $ alter 103 if ( not skipse ) then $

alter 104 endif $ @ 6087; 68.2; Heat Transfer Analysis - heatconv, CHBDY large field format - Wr ong Conversion The heat transfer converter program, heatconv, cannot process CHBDY large field format with ambient points correctly. Specifically, the GA1, GA2, GA3, GA4 points on the CHBDY continuation entry are not copied to the new Version 68 bulk data CONV entry. Avoidance: Use small field CHBDY format. @ 6093; 68.2; Optimization/Design Sensitivity - BNDGRID, RBEi, RBAR - UFM 2101 A, UFM 3001 When a shape optimization job is run which uses BNDGRID entries to define shape basis vectors, if the same component of a grid referenced on a BNDGRID entry is also referenced by a rigid element, then the message "USER FATAL MESSAGE 2101A" is correctly issued stating that a grid component can not be defined in both S-set and M-set. This is a user input error. For versions prior to 68.2, if a shape optimization job uses Analytic Boundary Shape and also includes MPC or rigid element entries in the primary model, error message "USER FATAL MESSAGE 3001 MODULE = UPARTN THE INPUT DATA BLOCK NAMED USET0 AT POSITION 1 DOES NOT EXIST." may be output. Avoidance: Replace the offending RBE's with extremely stiff elements. @ 6095; 2001; Material Nonlinear Analysis - CQUAD4, CTRIA3, PSHELL, MID3 - No Warning Message Material nonlinear analysis does not support the transverse shear stiffness in a CQUAD4 or CTRIA3. This is documented in the MSC/NASTRAN nonlinear seminar notes. However, if the MID3 references a MATS1 entry, no warning message is issued stating that material nonlinearities are not supported for this stiffness. Avoidance: There is no avoidance other than realizing the fact that the transverse shear stiffness remains linear for CQUAD4 and CTRIA3. @ 6100; ; Aeroelastic Analysis - PK Flutter Analysis - Extra Eigenvectors If a user requests flutter eigenvectors in a pk flutter analysis and there are real roots, an eigenvector is printed for each real root. This is confusing to our users since only half of the real roots are provided in the flutter summary. Avoidance: Ignore the first half of the flutter eigenvectors that have real eigenvalues. @ 6101; ; Aeroelastic Analysis - ADR, AERO - Incorrect Frequency Labels. If the AEROF request is used in a flutter analysis, the frequency used to annotate the forces is incorrect. The frequency is computed based on the

reduced frequency of the flutter analysis. In dimensionalizing this reduced frequency, the velocity from the AERO bulk data entry (VELOCITY) is used rather than the velocity from the flutter analysis (VFLUT). Avoidance: The correct frequency can be computed based on: f correct @ 6111; 68.2; Dynamic Response Analysis - PARAM,DYNSPCF,OLD - Wrong Answers If PARAM,DYNSPCF,OLD is used in a modal solution and the matrix method of data recovery is used (the default), incorrect SPC Force output occurs. Avoidance: Either specify PARAM,DDRMM,-1 or include the following DMAP alter: compile sedrcvr, souin=mscsou, nolist, noref alter 163 TYPE PARM,,CHAR8,Y,DYNSPCF='NEW' $ if (dynspcf='old') then $ DDRMM CASEDR,UH1,OL1,OUGV2,oqg2,OES2,OEF2,XYCDBDR/ OUGV2Y,oqg2y,OES2Y,OEF2Y, $ equivx oqg2y/oqg2/always $ else $ alter 164 endif $ @ 6112; 2001; Documentation - MATT8 - MSC/NASTRAN V68 Reference Manual The documentation in Chapter 5 of the Version 68 MSC/NASTRAN Reference Manual referencing the MATT8 entry is incorrect. The MATT8 entry does not exist and it is not possible to define temperature dependent orthotropic material properties. @ 6116; 68.2; P-Elements - RBE, MPC - UFM 5289 The external id field in the message "USER FATAL MESSAGE 5289, COMPONENT WITH EXTERNAL ID = ***, DOF = **, APPEARS MORE THAN ONCE ON THE SAME MPC OR RIGID ELEMENT" is limited to eight characters. Thus, if the message attempts to identify an internally generated degree of freedom from a p-element "********" are output. Avoidance: There is no avoidance. @ 6118; 68.2; Heat transfer - CHBDYE, TA1 - SFM 4276 EC 201 The message "SYSTEM FATAL ERROR 4276.WRITE ERROR CODE 201" occurs when a CHBDYE element references a non-existing conduction element ID. Avoidance: Make sure all the CHBDYE have a corresponding conduction element ID = f * (VFLUT / VELOCITY) printed

6123; 68.2; Optimization/Design Sensitivity - DOPTPRM, P2=8 - Missing Respon se Output The design optimization print produced when P2 is set to 8 on the DOPTPRM Bulk Data entry does not include transient responses. Avoidance: Insert the following DMAP alter in V68.2: compile desopt noref nolist alter 'dom10','' dbview xxxx=oli where (solapp='mtran') $ DOM10, DESTAB,XINIT,XO,CNTABRG, CVALRG, CVALO, DVPTAB,PROPI,PROPO,R1TABRG,R1VALRG,R1VALO, RSP2RG,R2VALRG,R2VALO,OPTPRMG,OBJTbg,DRSTBLg, xxxx,fol1,frqrprg// DESCYCLE/DESMAX/OBJIN/OBJOUT/ $ @ 6124; 68.2; Elements - PARAM,COUPMASS,1 CBAR, EMG - SFM 4276, EC 7777 A run that contains BAR elements that have zero area specified on the PBAR entry but non zero values for K1 and K2 the area factors for shear, may fail and issue the message "SYSTEM FATAL ERROR 4276, RECOVR ERROR CODE 7777" in the EMG module. Avoidance: Either remove the area factor for shears or input a non-zero value for the area of the PBAR. @ 6126; 69.1; Superelements - PLOADX1 - Wrong Answers The PLOADX1 bulk data entry is not supported when using superelements. No load is generated for a PLOADX1 entry that references elements that belong to an upstream superelement, Avoidance: Place any elements loaded by PLOADX1 entries in the residual structure. @ 6127; 69; Eigenvalue Analysis - EIGR, GIV, MGIV, HOU, MHOU - Wrong Answers If any transformation method (i.e. GIV, MGIV, HOU or MHOU) is used to extract eigenvalues from a two degree of freedom problem the eigenvectors may be reversed. Avoidance: Select the lanczos method of eigenvalue extraction. @ 6129; 69; Elements - CTETRA - Poor Answers, Distorted Element Not Detected Severely distorted TETRA elements may not always be identified, this may lead to poor answers. Avoidance: Other than avoiding severely distorted elements, there is no avoidance. @ 6137; 69; Restart/Superelement - SOL 103/129 - Inefficient Restart

When performing a restart from SOL 103 into SOl 129, if the model has multilevel superelements, the restart logic performs dynamic reduction on the last level of superelements before the residual structure. Avoidance: The following DMAP alter for Version 68 bypasses the automatic restart logic which is causing this error. The user must beware that bypassing the automatic restart logic may result in wrong answers if the model has been changed. compile phase0 souin=mscsou noref nolist alter 336,336 $ v68 alter 267,267 $ v675 endalter

@ 6138; 68; OUTPUT2 - RCOUT2 - BUS ERROR (CORE DUMP) In Version 67.5, if the RCOUT2 program is used to read a formatted output2 file on the HP machine, the job may fail with following error messages: BUS ERROR (CORE DUMP) This error only occurs on the HP Version 67.5 system when reading an output2 that is written in a neutral format. Avoidance: Use the Version 68 RCOUT2 program to read the Version 67.5 output2 file. @ 6142; 68.2; P-Elements - EMG - UFM 4659 For some complicated geometric surfaces, such as a sphere which has singularities at the poles, element edges which are mapped directly to the surface may be distorted enough to cause the the fatal message: USER FATAL MESSAGE 4659 (EMGPMD) ID = *** BAD GEOMETRY HAS BEEN DETECTED. DURING INTEGRATION FOR ELEMENT WITH

Avoidance: There is no current workaround. @ 6147; 68.1 (w/ ZA; Matrix Methods - FBS-Method 2 and Unsymmetric DECOMP (Non-s parse) Errors have been found in the Vector Math Kernels for Hitachi VOS3. The errors will only occur when sparse is deselected (i.e., when NASTRAN SPARSE=1,USPARSE=0 is specified). There are two separate errors -- one in FBS Method 2 and one in Unsymmetric Decomp. Both errors can be corrected using the following ZAP: //... JOB ... //* //* CORRECT ERRORS IN NON-SPARSE VECTOR KERNELS //* //ZAPKVCT EXEC PGM=AMASPZAP //SYSPRINT DD SYSOUT=* //SYSLIB DD DISP=SHR,DSN=...NAST68.NASTRAN ..NASTRAN LIB..

//SYSIN NAME VER REP NAME VER VER REP REP NAME VER VER VER VER VER VER VER VER VER VER VER VER REP REP REP REP REP REP REP REP REP REP REP REP /* //

DD * NASTKVCT $PMDINF 0000 C8C9,E3C1 0018 0000,0002,0000,0016,0000,005F NASTKVCT VOS3VCT2 20CA 47B0,F094 2844 0840,4242,0842,4442 20CA 47D0 2844 0840,4442,0842,4642 NASTKVCT VOS3VC1A 3450 0C40,4442 36E2 2362 373C 2A8A 3742 2B20 3832 2A8A 3838 2B20 3BE2 2362 3C12 2A03 3E72 2362 3EA0 2A03 3FFC 2C44,4C0E 4064 084C,4442,084E,4642 3450 0C40,4242 36E2 2862 373C 2B8A 3742 2A20 3832 2B8A 3838 2A20 3BE2 2862 3C12 2B03 3E72 2862 3EA0 2B03 3FFC 2C44,4C18 4064 084C,4042,084E,4242

There is one important thing to note about this ZAP -- one of the routines being modified is the $PMDINF routine (to change the Vector Kernels module date to indicate that the ZAP has been applied). Note that this routine name starts with a dollar sign ('$'). This is actually character code X'5B' and may be represented as a Yen symbol in some VOS3 systems. It is very important that this character be checked. @ 6148; 68.2; Optimization/Design Sensitivity - DMAP FM 9015 (FEA) USET TABLE NOT PRESENT Shape optimization jobs that contain an auxilary model to define the boundary shapes but do not include a static subcase may fail with the message: DMAP INFORMATION MESSAGE 9052 (FEA) - STATIC ANALYSIS COMPLETED. DESIGN CYCLE NUMBER= 1 DMAP FATAL MESSAGE 9015 (FEA) THE USET TABLE IS NOT PRESENT FOR: ANALYSIS = FREQ MPC = -1 SPC = -1 AUXMID = 1

Avoidance: Apply the following DMAP alter: compile sedisp souin=mscsou list noref alter 206 diagon( 20 ) $ alter 207 diagoff( 20 ) $ endalter compile descon souin=mscsou list noref alter 108 diagon( 20 ) $ alter 109 diagoff( 20 ) $ endalter compile desopt souin=mscsou list noref alter 166 statcc = true $ modecc = false $ buckcc = false $ dfrqcc = false $ mfrqcc = false $ mtrncc = false $ saercc = false $ dvrgcc = false $ flutcc = false $ alter 172,172 CALL BNDSHP CASEXX,EDOM,EQEXIN,GPDT,BGPDT,CSTM,SIL, USET,CASEVEC,AMLIST,UG,GEOM4/ CASENEW,CVEC,DVIDS,USET0,YG,rmg/ AMBS/GMBS/LUSET $ alter 174 MDCASE CASEXXu/,xcasem,xcaseb,xcasef,xcasemt, xcasea,xcasedv,xcasef,,,,xcasead,,// S,N,MODECC/S,N,BUCKCC/S,N,DFRQCC/S,N,MFRQCC/ S,N,MTRNCC/S,N,SAERCC/S,N,DVRGCC/S,N,FLUTCC $ alter 193 bndshp = true $ endalter compile bndshp souin=mscsou nolist noref alter 1,1 SUBDMAP BNDSHP CASESTAT,EDOM,EQEXIN,GPDT,BGPDT,CSTM,SIL, USET,CASEVEC,AMLIST,UG,GEOM4/ CASENEW,CVEC,DVIDS,USET0,YG,rmg/ AMBS/GMBS/LUSET $ alter 35,35 type parm,nddl,i,n, mpc,suport $ mpc = 0 $ suport = 0 $ GP4 CASENEW,GEOM4,EQEXIN,SIL,GPDT,BGPDT,CSTM,,,/ RMG,,USET0,/ LUSET/S,N,NOMSET/0/S,N,NOSSET/0/S,N,NORSET/ NSKIPX/0/0 $ endalter compile fea souin=mscsou nolist noref alter 120 type parm,nddl,logical,n, bndshp $ bndshp = false $ $ message //' bndshp in fea = '/bndshp $ alter 204,204

if ( ((dfrqcc or mfrqcc) and (feaopt='analysis' or feaopt='datarecovery')) or ((itol(-dfrqdsa) or itol(-mfrqdsa)) and feaopt='sensitivity' ) ) then $ alter 363,363 if ( (mtrncc and (feaopt='analysis' or feaopt='datarecovery')) or (itol(-mtrndsa) and feaopt='sensitivity') ) then $ alter 453,453 if ( (saercc and (feaopt='analysis' or feaopt='datarecovery')) or ((itol(-saerodsa) or itol(-statdsa)) and feaopt='sensitivity') ) then $ endalter @ 6152; 68.2; P-Elements - GMCORD, GMCURV, FEEDGE - Wrong Answers When a GMCORD coordinate system is specified on a GMCURV, the 2 and 3-directions are incorrect. The 2-direction is opposite the normal to the associated GMSURF, and the 3-direction is the cross product of the tangent to the edge and the normal to the face. The 3-direction should be the normal, and the 2-direction should be the cross product of the normal to the surface and tangent to the curve. @ 6155; 69; Miscellaneous - XCSA - Poor Performance The documentation states that an alphabetic character must be used to start the ID entry if one is used. If a non-alphabetic character is used, the program will issue UFM 300. On some client machines, it appears that excessive time is taken before this message is issued. Avoidance: Do not start any words on the ID entry with a non-alphabetic character. @ 6156; 68.2; Elements - CTRIAi - Poor Answers, No Skew Check There is no skew check performed on triangular plate elements in MSC/NASTRAN. Excessively skewed elements may result in poor answers. Avoidance: Many pre-processors have skew checks for triangular plate elements. A skew check is in MSC/NASTRAN Version 68.2. The test in Version 68.2 is based upon the three vertex angles. A warning message is issued if any vertex angle is less than "T3SKEW", where "T3SKEW" is a user controllable NASTRAN keyword with a default value of ten degrees. @ 6157; ; Performance - DPD - Excessive Time in DPD Module On large-size models, the main function of the DPD module is to generate the usetd table from the uset table. It has been observed that the time to perform this operation on large models can exceed the time to generate

the uset table and process all rigid elements in the GP4 module. Avoidance: An avoidance on sol 103 models which do not have EPOINTs is to write a DMAP to branch over the DPD module, and equiv uset to usetd, dynamics to eed, and set noue to -1. @ 6159; 69; Documentation - MPCFORCE - Misleading Documentation The MSC/NASTRAN Version 68 Quick Reference Guide description of the MPCFORCES command, remark 1, states that the MPCFORCE requests are honored in the Structured Solution Sequences only. This remark implies that MPCFORCE output is available in all of the SSS. It is not available in the nonlinear solutions SOLs 106, 129, 153, or 159. This fact is documented in Table 3 of the Case Control section of the Quick Reference Guide. @ 6166; 69.1; Eigenvalue Analysis - GIV - Wrong Answers, UIM 4158 Using the GIVENS method to calculate modes of a structure which has a non positive definite mass matrix, should result in the following error message: "USER FATAL MESSAGE 4646---- THE MASS MATRIX IS NOT POSITIVE DEFINITE,USING THE GIVENS METHOD. USE MGIV INSTEAD" However, the run may proceed and incorrect answers may be produced. Since MSC/NASTRAN automatically omits massless degrees of freedom, this error only occurs with models involving nearly massless mechanism. The user is warned of the poor answers by a very large value for the factor diagonal ratio in the following message: USER INFORMATION MESSAGE 4158 ---STATISTICS FOR CHOLESKY DECOMPOSITION OF DATA BLOCK MXX FOLLOW MAXIMUM RATION OF MATRIX DIAGONAL TO FACTOR DIAGONAL = The user may also be warned of the poor results by the issuance of a machine dependent message indicating a floating point overflow. Avoidance: Use MGIV or LANCZOS method of extracting eigenvalues for the structures of these nature. @ 6167; ; Material Nonlinear Analysis - NLITER - SFM 3007 NLITER (NLQNVC) A material nonlinear problem using the Quasi-Newton (QN) method for a nearly elastic-perfectly plastic material (i.e., strain/work hardening slope, H<<1) may fail with following fatal message: *** SYSTEM FATAL MESSAGE 3007 MODULE = NLITER 0ILLEGAL INPUT TO SUBROUTINE NLQNVC The error may occur since the same data point (involved in the QN vector generation) may be used in two successive iterations. Avoidance: Change the number of load steps (i.e. number of increments NINC on NLPARM entry). @ 6168; 68.2; DBC, TRANS, ACCESS - Error in makefile. The access and util makefiles, <install-dir>/msc68/access/makefile and <install-dir>/msc68/util/makefile have two errors:

1) The implicit rules for ".ff.f" are wrong The error is corrected by changing the following rules (in both files): .ff.f: $(CPP) -P $< $@ to: .ff.f: $(CPP) -P $(CPFLAGS) $< $@ 2) The install target does not work (the architecture was already included in the EXE macro). The error is corrected by removing the macro "$(ARCH)/" from the pathname of the install target and all dependent targets. @ 6171; ; Plotting - CGAP - UWM 979 XYPLOTs of gap elements work incorrectly. For instance the command "XYPLOT stress /899(2)" for gap element 899 causes the following error message: 0*** USER WARNING MESSAGE 979, AN XY-OUTPUT REQUEST FOR POINT OR ELEMENT ID 899 -STRE- CURVE IS BEING PASSED OVER. THE ID COULD NOT BE FOUND IN DATA BLOCK 105 Three plots are created with this command. The first plot is the correct plot, the second and third are empty plots. @ 6181; 68.2; Optimization/Design Sensitivity - VECPLOT - UFM 5423 If acoustic analysis is requested in SOL 200 with ANALYSIS=MFREQ, DFREQ, or MTRAN and also ANALYSIS=STATICS, then the run fails in the VECPLOT module during static analysis data recovery. Avoidance: Insert the following DMAP avoidance: compile sedrcvr,souin=mscsou,noref,nolist $ alter 28,28 IF ( NOT(PEXIST) AND NOT(NLHEAT) AND (NSILS*6)<>LUSETS ) THEN $ @ 6184; 70.5.2; Nonlinear Static Analysis - NLPARM - Possible Divergence If the stiffness update method, KMETHOD field, is specified incorrectly on the NLPARM entry, a fatal message is not issued and the default AUTO method is used. Note that the double quotes surrounding the legal choices in the MSC/NASTRAN Version 68 Quick Reference Guide should not be included when specifying the KMETHOD field. If the double quote is used, it will be considered as part of the string specified. Avoidance: Specify a correct method in the KMETHOD field. @ 6185; 68.2; Installation - MSC/NASTRAN will not run on CRAY UNICOS V9 The nast68 script does not properly identify the model when running under CRAY Unicos V9, thus MSC/NASTRAN will not run. The script may be corrected by making the following changes:

Around line 325, change: osraw=`uname -m awk '{print $1}'` test "$osraw" = "CRAY" && { to: uname -m grep '^CRAY' > /dev/null 2>&1 && {

and a few lines later, change: sysmod=`uname -m to: sysmod="Y-MP" /bin/crayc90 && sysmod="C90" This changed script will continue to work on UNICOS V7 and V8. @ 6186; 68.2; Optimization/Design Sensitivity - DSAH - SFM 4276 GREAD EC If a design optimization task contains multiple subcases that include design constraints on dynamic responses (either frequency or transient) and there are no active constraints for the first dynamic subcase, the above fatal error occurs. Avoidance: A possible avoidance is to force retention of a response using a DSCREEN bulk data entry with a threshold set to a large negative number. @ 6192; 69.0.2; Eigenvalue Analysis - EIGRL, REIGL - Poor Message If ND, the number of desired roots, is specified greater than the total number of eigenvalues and Summary "TERMINATION MESSAGE" may indicate "NOT RANGE" even though all possible eigenvalues have on the EIGRL entry to be , the Eigenvalue Analysis ALL EIGENVALUES FOUND IN been found. awk '{print $2}'`

Avoidance: Either specify a frequency range or specify an appropriate value for ND. @ 6196; 68.2; Optimization/Design Sensitivity - DOM9 - SFM 3002 EOR RSP2RG In a design optimization task, if the objective is a synthetic (DRESP2) response and there are no constraints on synthetic responses, the above error may occur. Avoidance: An avoidance is to formulate and constrain a DRESP2 response. For example the DRESP2 = 1.0 * DRESP1. In order to force retention of the constraint, a DSCREEN entry with RTYPE = EQUA and TRS = -1000. should be used. @ 6201; 68.2; Postprocessing - Heat Transfer, PARAM POST,-2 - Missing Output In heat transfer analysis solutions 153 and 159, the temperature and flux results are not properly written to the OUTPUT2 file under when using PARAM, POST, -2.

Avoidance: Insert the following DMAP alter: compile sedrcvr souin=mscsou noref nolist alter 66,66 alter 68,68 IF ( POST=-2 AND GETSYS(NH,56)<>0 ) THEN $ MATMOD ougv1,,,,,/tougv1,/13 $ alter 112,112 IF ( POST=-2 AND GETSYS(NH,56)<>0 ) THEN $ @ 6202; 68.2; Installation - MSC/N Will not Install on HP-UX 10.00 A utility program on HP-UX 10.00 returns unexpected information. As a result, the installation and execution scripts do not work. Avoidance: The only work-around is to replace the HP-UX 10.00 version's expr program, found in /bin/expr (a.k.a. /usr/bin/expr) with the HP-UX 9.0 version (found in /bin/expr). @ 6206; 69; Miscellaneous - IBM Hardware/Software Error The following documentation was received from IBM: Today we are informing you of an Engineering Change (EC) package that provides updates for selected systems which you have received. The package contains two changes that apply to a limited number of the subject machines. The EC package will eliminate the remote possibility that these models could store data to memory incorrectly. While this condition could potentially cause a data error without generating a system error message, it can occur only under an extremely rare set of circumstances. The EC package also includes an enhanced cooling system for the Model 590. This features a new fan and baffle to improve airflow and cooling in the processor complex. The EC package with both changes is available to you at no charge. Our records show that the package applies to at least one of the machines shipped to you. You can quickly determine whether any machine can benefit from the changes by making a simple inquiry from a system-attached console, as shown on the Attachment. If you decide that you wish to have the EC package installed, contact your IBM Service Center at 1-800-IBM-SERV (1-800-426-7378) and request ECA #190.

Attachment to IBM Letter of January 9, 1995 To Determine if a particular machine within the RS/6000 Models 58H, 590, 990, 99E or 99F could benefit from the Engineering Change (EC) package, your Systems Administrator can issue an AIX command at a system-attached console. The procedure should take about 30 seconds to complete and will not affect other processes that are running on your system.

From the AIX Prompt, issue the command: lscfg -l sysplanar0 -v pg

On the display, read the line which says: Device Specific.(Z1)........nnYYnn If YY is: 41 or 42, then the machine, as manufactured, already includes the EC package. or Anything else, then the machine is a candidate for the EC package. If you call the IBM Service number to request the EC package, your Service Representative will perform this procedure to verify the status of your machines before any repair activity begins, and will exchange parts only if required. @ 6221; 69; Miscellaneous - CORD2i, GP1 - Missing Error Messages On the DEC/OSF1 (and possibly other machines), when a large number (e.g., 1.E+35) is used to define the vectors of a secondary coordinate system (e.g., CORD2S), the job may fail without issuing an error message in the F06 file. A message such as "abnormal time termination" may be observed in the LOG file. On other machines, SFM 2005, UFM 316, or SFE 4276 EC 7777 may be issued in the F06 file. Avoidance: Use smaller numbers to define the vectors. @ 6224; 69; Transient Response - DMAP, TRD1 - SFM 3008 The transient response solutions in MSC/NASTRAN do not support complex arithmetic. If a complex matrix is introduced by DMAP, the program attempts to process the problem, but may fail in the module TRD1 with a message similar to the following: "INSUFFICIENT CORE FOR SPARSE FBS IN SPFBS USER ACTION: INCREASE CORE BY xxx WORDS. *** SYSTEM FATAL MESSAGE 3008 MODULE = TRD1 INSUFFICIENT MEMORY AVAILABLE FOR SUBROUTINE." Avoidance: Do not use complex matrices in a transient solution. @ 6226; 68.2; Optimization/Design Sensitivity - SDSC - SFM 4276, Attempt to Re ad Past EOR In a dynamic response sensitivity analysis, the above message can occur when the are multiple superelements and one of the superelements has no active responses. Avoidance: A possible avoidance is to explicitly specify each superelement subcase in the case control and not place a design condition for the superelements that are known not to be active in the sensitivity analysis. A less satisfactory fix is to delete the SDSC and DSTAP2 modules in the EXITOPT subdmap. The sensitivity matrix would then be printed without the IDENTIFICATION OF COLUMNS IN THE DESIGN SENSITIVITY MATRX table.

@ 6228; cient 69; Optimization/Design Sensitivity - DVPRELi - DSCM Matrix Row Defi The number of rows in the DSCM matrix should be equal to the number of independent design variables plus the number of DVPREL2 entries. If there are no DVPREL2 entries and the trailing independent design variables are not referenced by any DVPREL1 entries, the number of rows in DSCM matrix will be deficient by the number of unreferenced trailing design variables. Avoidance: An avoidance is to rearrange the ID's of independent design variables so that the unreferenced ID's are not the highest numbers. @ 6229; 68.2; Heat Transfer - CHBDY, Forced Convection Heat Flow - Inconsisten t Calculations The forced convection heat flow is calculated by upwinding shape functions in the solver and by the average fluid temperature in the data recovery. This inconsistency results in an unbalanced energy between the forced convection heat flow output and the mass flow convection (mdot*Cp*Delta T). This error may lead to poor answers for the "FORCED-CONVECTION" and "HEAT FLOW INTO HBDY ELEMENTS" results. Avoidance: There is no avoidance. @ 6230; 68.2; Heat Transfer - ECNVMD, NCNVMD - Wrong Answers If the Prandtl number convection exponents for heating (EXPPI) and cooling (EXPPO) are not the same, then MSC/NASTRAN incorrectly uses EXPPI for cooling and EXPPO for heating calculations. In addition, the shape functions of forced convection boundary conditions (CONVM) are incorrectly assigned Peclet numbers less than 0.01 which have negative mass flow rates. Avoidance: Exchange the values of EXPPI and EXPPO and do not use negative mass flow rates when Peclet numbers are less than 0.01. @ 6233; 69; Aerodynamic Flutter - SOL 145 - Poor Performance The solution approach used in SOL 145 is different in Version 68 as compared to Version 67.5. Invariant matrices are created in Version 68 for aerodynamic optimization. These matrices are larger than the matrices used in the Version 67.5 approach. Therefore, the performance (e.g., in the FBS module) may be slower in Version 68 as compared to Version 67.5. Due to these larger matrices, the disk space requirement may also be larger. Avoidance: There is no avoidance. @ 6234; 70; Optimization/Design Sensitivity - DEQATN Size Increased to 32000 Characters MSC/NASTRAN Version 70 has increased the allowable number of characters for the DEQATN entry from 12,500 to 32,000.

@ 6235; 68.2; Static - PLOAD1, CBEAM, CBEND - Infinite Loop If a problem contains PLOAD1, CBEAM, and CBEND entries, then MSC/NASTRAN may become trapped in an infinite loop. The error only occurs when no PLOADs are applied to any CBEND elements. Avoidance: Apply a zero dummy load on a CBEND element. @ 6240; M 5225 68.2; Static Analysis - QSETi, SSG3 - SFM 3007, UWM 4698, UWM 5221, UF In static analysis, if QSETi entries are defined for SPOINTs and ASETi entries are defined for physical degrees of freedom then the run will fail in SOL 61 and 101 with SFM 3007 in module SSG3 and subroutine SSG3A. In SOL 24, the run will issue USER WARNING MESSAGEs 4698 and 5221 and then fail with USER FATAL MESSAGE 5225. Avoidance: Insert the following avoidance for SOL 101 or remove the QSETi entries: compile selrrs souin=mscsou noref nolist alter 5,5 PARAML USET//'USET'//////'q'/S,N,NOqSET/ 't'/S,N,NOtSET $ IF ( FIXEDB <> -2 ) THEN $ A-SET PARTITIONING ( Q + T ) EQUIVX PA/PT/NOQSET $ IF ( NOQSET>=0 AND NOTSET>=0 ) UPARTN USET,PA/PT,,,/'A'/'T'/'Q'/1 $ ENDIF $ EQUIVX PT/PL/NORSET $ alter 8,8 UPARTN USET,PT/PL,PR,,/'T'/'L'/'R'/1 $

@ 6241; 69; Random Response Analysis - STRESS, BEAM - Wrong Answers If stresses are requested for BEAM elements in a random response analysis, and the STRESS output is requested as real/imaginary values, then the PSD output for the RMS stresses at point F of end A are incorrect. Avoidance: Request the stresses in magnitude and phase by using the STRESS(PHASE) Case Control command. @ 6242; 69; Optimization/Design Sensitivity - DSAD - Datablock Does not Exis t, Poor Message The message "INPUT DATA BLOCK NAMED OUGVS1 AT POSITION 12 DOES NOT EXIST. DSAD" is produced when the user requests responses at a set of grid points that do not exist. If some of the grid points exist and some of them don't, a more meaningful message is produced. A similar nonexistent datablock error is produced for other response types; e.g., the OES1 datablock is missing if none of the requested elements exist. Avoidance: Do not request responses for grids that do not exist. @ 6244; 68.2; Post Processing - PARAM,POST,-1; PATVER,2.5 - Extra Data Block L

AMA If PARAM,POST,-1 and PARAM,PATVER,2.5 are used in SOL 103, MSC/NASTRAN Version 68 incorrectly writes the LAMA (OLB) datablock to the output2 file for PATRAN. PATNAS cannot recognize this datablock and issues an error message. Avoidance: Use the following DMAP alter: COMPILE SEDRCVR SOUIN=MSCSOU NOLIST NOREF ALTER 92 $ IF (POST=-1 AND PATVER=2.5) THEN $ OUTPUT2 OUGV1//OTAPE2/OUNIT2//OMAXR $ ELSE $ ALTER 93 $ ENDIF $ @ 6249; 69; Miscellaneous - PARAM,NMLOOP - UWM 4500, DMAP FATAL MESSAGE 9015 If "PARAM,NMLOOP" is included in the Bulk Data section for a Structured Solution Sequence input file other than SOL 106 or 153, then the job may fail with one or more of the following error messages: USER WARNING MESSAGE 4500, STIFFNESS MATRIX PURGED. GRID POINT SINGULARITY NOT TEST DMAP FATAL MESSAGE 9015 (SEBUCKL) - THE USET TABLE IS NOT PRESENT Note that this parameter is not used except in SOL 106 and 153. Avoidance: Remove the offending parameter as this parameter is only used for SOL 106 and 153. @ 6252; 69; Documentation - SPCFORCE - PLOT Option is Available The SPCFORCE output request supports the plot option even though it is not listed in the MSC/NASTRAN Quick Reference Guide. @ 6253; s The complex lanczos method of eigenvalue extraction may produce non-orthogonal eivenvectors for models with multiple roots at the same eigenvalue. This error is under investigation, currently it has been experienced on short-word machines only. The CRAY produced correct results. Avoidance: Use the Hessenberg method. @ 6255; 69; Matrix Methods - SOL 307 - SFM 3001 In a run using the default sparse solver if there is insufficient memory, then the program should revert to the active column method of decomposition and continue. In this condition "USER WARNING MESSAGE 6136" is output. However, for SOL 307, after outputting this message, the program issues the following fatal message, "SYSTEM FATAL MESSAGE 3001 MODULE = CEAD THE EXECUTIVE FILE 0 DOES NOT EXIST. PROGRAMMER INFORMATION: SUBROUTINE GOPEN". 68.2; Complex Eigenvalue Analysis - CEAD, CLAN - Incorrect Eigenvector

Avoidance: Increase the memory using the 'mem' command line keyword or turn off the sparse solver with the by setting the SPARSE keyword to 1. @ 6257; 69; Miscellaneous - File Not Found, UFM 423 Prior to Version 69, filenames were limited to 80 characters. Filenames that were longer were trunctated by MSC/NASTRAN, this generally resulted in file not found errors. Version 69 increases the allowable filename length limit to 256 characters. @ 6258; ; Documentation - DMAP Module Dictionary - MATMOD Option 16 The first line of the example DMAP for MATMOD option 16 in the MSC/NASTRAN DMAP Module Dictionary (page 196) is incorrect. Avoidance: In the DMAP example, replace the following line MATMOD EQEXIN/INTEXT/9/0/LUSET $ with MATGEN EQEXIN/INTEXT/9/0/LUSET $ @ 6260; ; Nonlinear Heat Transfer - NOLINi, ADAPT - Wrong Answers The ADAPT strategy and the PW convergence criterion leads to an erroneous solution if NOLIN cards are enforcing square wave type of power input loading that is based on specified temperatures at certain grid locations of the model. Particularly, the iteration strategy fails when the temperature of the curve is nearly constant. This phenomenon occurs for the ablation problem for which NOLIN entries are used to model the heat transfer through out the medium. Avoidance: Set the ADJUST of the TSTEPNL entry to zero, so that the automatic time step adjustment is disabled. Also set the convergence criterion to U, and set the EPSU to 1.0-5 instead of the 1.0-3 default value. This avoidance yields correct answers but the actual CPU time to reach convergence will be much longer than using SOL 89 for Version 67.5. ****New Avoidance**** Set the ADJUST equal to zero , and set the convergence criterion to P only. @ 6261; 70; Miscellaneous - ADAPT,OUTPUT,OUTRCV - UFM 300 If free field format is being used with the ADAPT, OUTPUT, or OUTRCV bulk data entries, and the first comma does not occur in columns 1-8, the fatal message, "USER FATAL MESSAGE 300, DATA ERROR IN FIELD UNDERLINED", occurs. Avoidance: Eliminate spaces between the entry name and the first comma. @ 6264; ; Thermal Stress Analysis - TEMPBC - Wrong Answers The TEMPBC bulk data entry is used to specify time-dependent temperatures (TYPE=TRAN) at grid points for transient thermal analysis (SOL 159). Internally in MSC/NASTRAN, this time dependent load is accomplished using

SLOAD and the CELAS2 bulk data entries. For thermal analysis, these generated CELAS2's produce correct temperatures. However, if a TEMPBC entry with the TRAN option is included in a thermal stress analysis (SOL 101, SOL 106), then the thermal stress calculations are wrong because extra forces are introduced in the problem. Avoidance: Remove TEMPBC entries with TYPE=TRAN. @ 6266; 68.2; Miscellaneous - nastran/emas script - Unexpected Result for post Command In the nast68/emas3 script, the post and app=yes operations may interfere with one another. If a client attempts to compress the output files from MSC/NASTRAN or MSC/EMAS and uses app=yes, the obvious command is: nast68 myjob app=yes post='gzip $out.out' Unfortunately, the append commands are run after the post command, so the file myjob.out does not yet exist. If the user were to change the command to nast68 myjob app=yes post='gzip $out.f06 $out.f04 $out.log' the append operation will fail because the three output files have been compressed and renamed. Avoidance: A work-around that provides the effect of the initial command is: nast68 myjob app=no post='cat $out.f06 $out.f04 $out.log $out.out.gz' NOTE: The versioning operation performed by old=yes will _not_ save files with a ".gz" suffix. @ 6267; 68.2; Static Analysis - SUPORT - UFM 3055 In MSC/NASTRAN Version 68, if multiple SUPORT sets are used in SOL 101, it is possible to get "USER FATAL MESSAGE 3055 ATTEMPT TO MULTIPLY INCOMPATIBLE MATRICES". This error occurs when there are a different number of loading conditions for the various SUPORT points. Avoidance: If multiple SUPORT sets are used, insert the following DMAP alter for SOL 101 compile sestatic, souin=mscsou, nolist, noref alter 69 type parm,,i,n,storit=0 $ call dbstore crx,,,,//suport/0/'DBALL'/s,storit $ $ compile sedisp, souin=mscsou, nolist, noref alter 14 type parm,,i,n,gotit=0 purgex /crx,,,, $ call dbfetch /crx,,,,/suport/0/0/0/s,gotit $ gzip -c >

$ @ 6269; le 68.2; SSSALTERS - DDAMF.FOR - Missing Subroutines, DDAM Will not Compi The ddamf.for FORTRAN program delivered with the SSSALTER library for the ddam related alters is missing two subroutines. This is a Version 67.5 plus error. Avoidance: The updated ddamf.for will be delivered with Version 68.2. If you need a copy of this updated program, you can contact your local technical support representatives. @ 6270; 70.5; Buckling Analysis - EIGRL - SFM 5299, SFM 5401, UFM 4405 In a buckling analysis, if the buckling load is reversible (e.g., a torsional buckling load applied to a cylinder), then the job may fail with the following error messages: *** SYSTEM FATAL MESSAGE 5299.UNRECOVERABLE TERMINATION FROM LANCZOS ITERATION *** SYSTEM FATAL MESSAGE 5401 (REIGL). LANCZOS METHOD IS UNABLE TO FIND ALL EIGENVALUES IN RANGE. *** USER FATAL MESSAGE 4405, NO EIGENVECTORS COMPUTED FOR COMPONENT MODE SYNTHESIS OR SYSTEM SOLUTION. In this case, the Lanczos shift logic finds many negative legitimate roots. The number of these negative roots will increase as the model size increases. Avoidance 1: If you are interested in only a few modes, which is the case for most buckling analysis, use the SINV method. If you wish to use the Lanczos method for buckling and you encounter the above problem, specify a shift scale value on the EIGRL entry. Avoidance 2: (preferred over Avoidance 1 for Nastran versions 69 and 69.1): Include alter bra.v691 from the SSS alter library. This alter computes an estimate for the lowest buckling mode then passes this estimate to REIGL where it is used as a shifting scale. With this alter REIGL solves buckling problems much better. @ 6271; 69; Aeroelastic Analysis - AESTATRS - Wrong Answers In a static aeroelastic analysis, if two dynamic pressures differ by less than 5.0e-4, the analysis will be repeated with the results for the first dynamic pressure duplicated rather than new results being generated for the second. Avoidance: The following DMAP avoidance gives differing results whenever the dynamic pressures differ by more than 5.0e-7: compile aestatrs souin=mscsou nolist noref alter 30,30 IQ=NINT(1000000.*Q) $ @

6272;

69; Documentation - DMIG, M2GG, K2GG, B2GG - No Check for Symmetry User-provided matrices, such as M2GG, K2GG, and B2GG, are not checked by the program for symmetry. If non-symmetric matrices are entered for any matrix which should be symmetric, the program may fail in a later operation, or give incorrect answers. Avoidance: Use the symmetric option on the DMIG entry.

@ 6274; es 68.2; Heat Transfer - RADSET - Wrong Answers for Multiple Large Caviti For multiple radiation enclosures, the contents of the radiation matrix for the first cavity is overwritten if the element size of the second cavity is greater than 200. This error also occurs for large sizes of third, fourth and subsequent cavities. Avoidance: Use a single cavity or keep the sizes of the second, third, and subsequent cavities small. @ 6282; 69; Buckling - Thermal Loads - Wrong Answers If multiple buckling analyses are executed in a single solution 105 run, and thermal loads are present, then wrong answers (eigenvalues, eigenvectors etc.) are generated for the second and greater buckling subcases. The error occurs because MSC/NASTRAN uses incorrect temperature information when generating the differential stiffness matrix for the second and greater buckling subcases. The solutions in the static subcases are correct. Avoidance: Specify all loads in both the static and buckling subcases, and include the following dmap alter: compile sekdr souin=mscsou nolist noref $ alter 25, 25 $ if (seid = 0) then $ case cases,/case1/'tran'/colno $ else $ equivx cases/case1/-1 $ endif $ call setq case1//seid/peid/s,mtemp/s,k2gg/s,m2gg/ s,b2gg/s,mpc/s,spc/s,load/s,deform/s,templd/ s,p2g/s,dyrd/s,meth/s,mfluid $ message //'load '/load/'templd '/templd $ endalter $ Example case control to avoid error: SPC = 1 SUBCASE 1 LOAD = 21 TEMP(LOAD)=11 SUBCASE 2 LOAD=22 TEMP(LOAD)=12 SUBCASE 3

STATSUB=1 LOAD = 21 $ AVOIDANCE ERROR 5977 TEMP(LOAD)=11 $ AVOIDANCE ERROR 6282 METHOD = 1 SUBCASE 4 STATSUB=2 LOAD = 22 $ AVOIDANCE ERROR 5977 TEMP(LOAD)=12 $ AVOIDANCE ERROR 6282 METHOD = 1 BEGIN BULK ETC... @ 6285; ; Nonlinear Transient Analysis - MATi, TRD1 - SFM 4276 EC 105 In a nonlinear transient problem (SOL 129), the job fails in TRD1 and the misleading message "SYSTEM FATAL ERROR 4276.PACK ERROR CODE 105" is output whenever all the mass properties are omitted. Therefore, the message occurs whenever the user forgets to specify the mass density on their MATi bulk data entries and doesn't have any other concentrated masses (CONMi, CMASSi) specified. Avoidance: Specify the mass density on MATi bulk data entries or include a concentrated mass. @ 6289; 69; Static Analysis - GPSDCON, ELSDCON - SFM 4276 EC 3002 When attempting to calculate Grid Point Stress Dicontinuities (GPSDCON) or Element Stress Dicontinuities (ELSDCON) on a model containing HEXA, PENTA and TETRA solid elements it is possible to get the following fatal message "SYSTEM FATAL ERROR 4276.FREAD ERROR CODE 3002" if the calculations are done over a volume containing both types of elements. Avoidance: Perform two runs and define the Volume set in the OUTPUT(POST) section to be PENTA or HEXA only. Also, grouping the element types by element id number allows the STDCON module to run. @ 6291; 69; Miscellaneous - SORT2, PUNCH - Missing SUBCASE ID If punched output is requested in SORT2 format (e.g., STRESS(SORT2,PUNCH)=all), then the punch output does not contain the appropriate subcase labels. This error is considered fixed in Version 69 since the subcase id may be obtained by setting system(210)=1. Avoidance: Either request SORT1 output or specify system(210)=1. @ 6292; Loop 69; Nonlinear Transient Heat Transfer - 20 Node HEXA - Infinite CPU If a 20 node HEXA element is used in a transient thermal analysis (SOL 159), the solution will not converge because the weighting of the capacitance matrix is not correct. The job may go into an infinite CPU loop and will have to be manually killed by the user. Avoidance: Use 8 noded HEXA instead. @

6294;

; Nonlinear Transient Analysis - TSTEPNL - SFM 4276 EC 1044 Solution 129 runs may fail with the message, "SYSTEM FATAL ERROR 4276.OPEN ERROR CODE 1044" in the NLTRD2 module. The error is still under investigation but the error only seems to appear when the NO field is not set to 1 or the timestep field (NDT) is not 1.E**. Avoidance: Specify the NO field to 1 or NDT to 1.E** (i.e., .001, .01, .1, 1. ...).

@ 6296; 69.0.2; Nonlinear Analysis - DBC, PARAM,POST,0 - DBCMSG ERROR 23 FROM GR IDBK/ELMBLK If in a Version 68.2 nonlinear transient analysis with PARAM,POST,0 the following warning messages are encountered the graphic database, xdb, will be corrupt. 0*** SYSTEM WARNING MESSAGE 6065 (DBC) *** DBCMSG ERROR MESSAGE 23 FROM SUBROUTINE GRIDBK OUTPUT LIST SYNCHRONIZATION PROBLEM FOR TOL VALUE = 1.2879E+00, POSITION = 52 0*** SYSTEM WARNING MESSAGE 6065 (DBC) *** DBCMSG ERROR MESSAGE 23 FROM SUBROUTINE ELMBLK OUTPUT LIST SYNCHRONIZATION PROBLEM FOR TOL VALUE = 1.2879E+00, POSITION = 52 Avoidance: The following DMAP avoidance may avoid the error in Version 68.2. The error has been corrected in Version 69.0.2; however most Unix workstations were released as Version 69.0.0 and Version 69.0.2 must be specially requested. COMPILE NLTRAN SOUIN=MSCSOU,LIST,NOREF ALTER 392,392 ALTER 406 IF (POST=0) DBC OESNLXD ,,,,,,,,,,,,,,,,,,,// 'OESNL'//////////////////// -1/DBCPATH/S,N,CP/'NLST'//GEOMU/LOADU/POSTU/ DBCDIAG/DBCCONV/DBCOVWRT $ @ 6307; 68.2; Heat transfer - SSG1, QVECT, RADBND - SFM 3007 If the QVECT (direction heat flux) is used in conjunction with RADBND (wave-length dependent radiation) for more than one element , then you SYSTEM FATAL MESSAGE 3001 will be encountered in the SSG1 module. The exact message is: SYSTEM FATAL MESSAGE 3007 MODULE = SSG1 ILLEGAL INPUT TO SUBROUTINE HMAT FATAL ERROR Avoidance: Instead of using the QVECT entry to specify heat flux, use the QHBDY entry to define the heat load. @ 6308; racy ; Grid Point Stress Discontinuities - GPSTRESS, GPSTR1 - Poor Accu Even though the stress discontinuities are exactly zero in the normal-x, normal-y and shear-xy directions the major and minor directions may

indicate small, non-zero stress discontinuities. @ 6310; 68.2; Buckling - ESE, GPFDR - Missing Output, Infinite CPU Loop, SWM ? ???, UWM 2354 Element strain energy and grid point force does not work in Version 68, SOL 105 buckling runs. A request for ESE or GPFORCE may produce an infinite cpu loop or SYSTEM FATAL MESSAGE 4276 with a traceback in the log file. The job may eventually fail due to lack of disk space because the following warning message is issued repeatedly: USER WARNING MESSAGE 4506--- TOTAL STRAIN ENERGY IS ZERO. THE PERCENT OF TOTAL STRAIN ENERGY IS ALSO SET TO ZERO. USER INFORMATION MESSAGE 4507--- IN SUBCASE **** THERE ARE *** QUAD4 ELEMENTS HAVING STRAIN ENERGY WHICH IS LESS THAN .0010 PERCENT OF THE TOTAL STRAIN ENERGY OF ALL ELEMENTS. The following message is another symptom of this error. SYSTEM WARNING MESSAGE ???? ***, GPFDR FINDS A SYMMETRY SEQUENCE LENGTH 19 WHILE ATTEMPTING TO COMPUTE GRID POINT FORCES AND ENERGIES. THERE IS AN INSUFFICIENT NUMBER OF VECTORS AVAILABLE 0 USER WARNING MESSAGE 2354. GPFDR MODULE IS UNABLE TO CONTINUE AND HAS BEEN TERMINATED DUE TO ERROR MESSAGE PRINTED ABOVE OR BELOW THIS MESSAGE. THIS ERROR OCCURRED IN GPFDR CODE WHERE THE VARIABLE -NERROR- WAS SET = 9 Avoidance: Insert the following DMAP alter: compile sedrcvr souin=mscsou noref nolist alter 264 IF ( APP1='bkl1' ) APP2='reig' $

@ 6311; 70; Optimization/Design Sensitivity - SOL 200, GDR, CMS, QSET - Dyna mic Reduction A request for dynamic reduction ( CMS or GDR ) on upstream superelements is ignored in SOL 200. The error only occurs when a statics subcase is also included along with the modes subcase. Avoidance: Specify the METHOD and/or DYNRED Case Control commands in the static subcases. @ 6312; 68.2; Heat Transfer - CHBDYE - UWM 2154 A CHBDYE which is defined as side 4 of a PENTA element does not work. If attempted the message "USER WARNING MESSAGE 2154, A ZERO AREA OR ILLEGAL CONNECTION FOR CHBDYE ELEMENT NUMBER xxxxx" is generated. Avoidance: If possible reorder the grids so the CHBDYE can be defined for a side other than side 4. @ 6313; 69; DBC/TRANS/ACCESS - CHBDYP, PHBDY, Heat Transfer - SFM T10 (XMIT)

TRANS will not work on heat transfer decks that contain CHBDYP elements with PHBDY property cards. The following message is written out and no neutral file is created: *** SYSTEM FATAL MESSAGE T10 (XMIT) A DATABLOCK WAS ENCOUNTERED THAT HAD MORE FIELDS THAN WERE DEFINED IN THE DATA DEFINITION LANGUAGE. DATABLOCK NAME = PHBDY DATABLOCK LENGTH = 7 DATA DEFINTION LENGTH = 4 Avoidance: The only avoidance is to run MSC/NASTRAN on the machine that you intend to perform post processing on (or other compatible machine) so that TRANS is not required. Note that machines that use the same UNIX IEEE structure machines are compatible and TRANS is not required. The list of compatible machines is documented in Chapter 6 of the Installation and Operation Instructions. IBM RS/6000, SGI, CONVEX, HP 9000, SUN SPARC SUNOS, and SUN SPARC Solaris are compatible machines. @ 6316; 70; Frequency Response/Random Analysis - RANDOM - BIOMSG: ERROR 3001 IN FREAD If the RANDPS entry references subcases that do not exist the following uninformitive message is issued: BIOMSG: ERROR 3001 HAS OCCURRED IN ROUTINE FREAD Avoidance: Correct the RANDPS entry that is in error. @ 6317; 68.2; Optimization/Design Sensitivity - DRESP1 - UFM 6293 "USER FATAL MESSAGE 6293 (DSAD16) GRID NO. *** REFERENCED ON A DRESP1 ENTRY COULD NOT BE FOUND" may be issued in error in a frequency response optimization or design sensitivity analysis if there are multiple frequencies and multiple grid points appearing on DRESP1 entries, but not every response is designed for every frequency. Avoidance: The first check should be to make sure the message is in error. If this appears to be the case, it is necessary to add a constraint for every response at every frequency. Very generous limits could be set on these added constraints so that they play no role in the design process. @ 6319; ; DMAP - INPUTT2 - Poor Message If a DMAP alter using INPUTT2 is used, and a parameter is used to specify the INPUTT2, the parameter must be assigned a value before execution of the INPUTT2 module. If no value is assigned, the module will fail with a FORTRAN read error. The message may be different for each computer type, and appears only in the log file. The .f06 and .f04 files end abruptly and no error message is printed. Avoidance: Be sure the unit number is defined before INPUTT2 is used. @ 6322; GL; Superelements - SEMA - SFM 4252, Inefficient restart On a restart run, if the user changes PARAM,AUTOSPC for the residual

structure, then SYSTEM FATAL MESSAGE 4252 may be issued or all superelements will be recomputed. This error was found on a sol 103 restart to sol 129. In a sol 129 run, the PARAM,AUTOSPC default is "NO" in the residual and it should not be changed. Therefore, single point constraints for singularities that were eliminated by PARAM,AUTOSPC,YES in the sol 103 run must be explicitly defined in sol 129 for the residual. Avoidances: 1) Specify PARAM, SPCGEN, 1 in the cold start run to punch SPC entries based on the the AUTOSPC operation. Include the puched SPC entries in the restart Bulk Data section. 2) Use the following alter compile ifpl souin=mscsou,nolist,noref $alter 89,89 $ v67.5 removes autospc restart check alter 150,150 $ v68, autospc compile phase0 souin=mscsou,nolist,noref $alter 231,231 $ v675 autospc alter 304,304 $ v68 autospc @ 6326; 68.2; Optimization/Design Sensitivity - PARAM,AUTOSPC; MERGE, PARTN Various SFM In SOL 200 with PARAM, AUTOSPC, YES, if the model is a fluid/structure model then the run may fail with various matrix partition or merge errors indicating incompatible matrices. The error occurs after the first design cycle if the AUTOSPC operation produces a different set of constraints then in the previous cycle. (Version 68 only) Avoidance: Insert the following DMAP alter: compile dpd1 souin=mscsou noref nolist alter 29 delete /vafs,,,, $ @ 6328; 69; Authorization - UFM 3060 If an MSC/NASTRAN Version 67.5 based authorization file is merged with a Version 68 authorization file, it is possible that Version 67.5 runs will fail with USER FATAL MESSAGE 3060 indicating that OPTION NOT IN APPROVED LIST even though the file contains a valid Version 67.5 authorization code. This problem only occurs when Version 68 authorization codes precede the Version 67.5 authorization code. Avoidance: Version 68 authorization codes should be appended to the end of the authorization file. @ 6329; ; Optimization/Design Sensitivity - PCOMP - Wrong Answers If the ply orientation angle is selected as as a design quantity and the initial ply orientation angle specified by the DESVAR/DVPREL1 combination is 0.0, incorrect sensitivities can result. The error occurs because the finite difference perturbation used to produce the sensitivity is too small when the initial ply angle is zero. Avoidance: Provide an initial design with a non-zero ply orientation angle.

@ 6332; 68.2; Eigenvalue Analysis - EIGRL, REIGL, FBSMEM, SYSTEM(146) - SFM 54 01, UFM 4405 If the NASTRAN keyword FBSMEM (a.k.a. SYSTEM(146)) is set to too large a value such that there is insufficient memory to honor the request, then LANCZOS stops in the REIGL module and issues the messages: SYSTEM FATAL MESSAGE 5401 (REIGL). LANCZOS METHOD IS UNABLE TO FIND ALL EIGENVALUES IN RANGE. ACCEPTED EIGENVALUES AND ADDITIONAL ERROR MESSAGES MAY BE LISTED ABOVE. USER ACTION: RERUN WITH ANOTHER METHOD OR ANOTHER SETTING ON EIGRL ENTRY. USER FATAL MESSAGE 4405 NO EIGENVECTORS COMPUTED FOR COMPONENT MODE SYNTHESIS OR SYSTEM SOLUTION Avoidance: Large values of FBSMEM are not necessary in Lanczos eigenvalue analysis. A value of 4 is the largest recommended value. Larger values are only a significant benefit to frequency response and transient solution sequences. Either increase the mem request or lower the FBSMEM request. @ 6338; 68.2; Heat Transfer - TRIAX6, QVOL - Wrong Answers TRIAX6 elements with QVOL heat loading produce wrong answers on long word machines such as CRAY Y-MP and CRAY C-90. Avoidance: There is no avoidance. @ 6339; 68.2; Optimization/Design Sensitivity - DESGLB - UFM 6545 (DOPR3) The following fatal error message is issued for superelement optimization when a global response is selected from case control using the DESGLB entry. This error only occurs when a weight or volume response is selected in a non-statics case. USER FATAL MESSAGE 6545 (DOPR3) A DESGLB = 10 COMMAND APPEARS AT THE SUBCASE LEVEL. USER ACTION: EITHER REPLACE THE DESGLB COMMAND WITH A DESSUB COMMAND OR PLACE THE DESGLB REQUEST ABOVE THE FIRST SUBCASE. Avoidance: The only avoidance is to specify a dummy statics subcase and to select the global (weight or volume) response with a DESSUB command from the statics subcase. @ 6348; 69; Optimization/Design Sensitivity - DVPREL1, PBEND - UFM 7052, Mis leading Message If an invalid property type is referenced on a DVPREL1 entry, the message "USER FATAL MESSAGE 7052, DOMPTC. ILLEGAL PROPERTY ENTRY NAME *****" is printed. In most cases this message would be caused by a typographical error on the DVPREL1 entry and would be easy to spot. However, the message is misleading for the legal PBEND property entry. As is documented in remark 3 under the DVPREL1 entry, the PBEND entry is not supported on the DVPREL1 entry. The message should read "USER FATAL MESSAGE 7052, DOMPTC. ILLEGAL PROPERTY NAME ***** REFERENCED ON DVPREL1

ENTRY". Note that UFM 7052 is not documented in the Reference Manual and needs to be added. @ 6353; 68.2; Optimization/Design Sensitivity - DRESP1 - UFM 5502 If a frequency sensitivity or optimization is performed at multiple frequencies and there are multiple element types and only a subset of the output frequencies are invoked on DRESP1 entries, then the following message is issued: USER FATAL MESSAGE 5502 (DSAD17) ELEM NO. **** REFERENCED ON A DRESP1 ENTRY COULD NOT BE FOUND. USER ACTION : USE OF UNREFERENCED ELEMENTS IS NOT RECOMMENDED AND MAY INDICATE A USER INPUT ERROR, RESULTS FOR OTHER RESPONSES ARE NOT AFFECTED. One message is output for each element in the model. Avoidance: Only perform the analysis at frequencies for which design results are required. @ 6355; 2001; Miscellaneous - IFPL, MAT1, MAT8 - SIGNAL NUMBER 11 If the same ID is used a MAT1 and MAT8 entry, the job should fail with the following message: USER FATAL MESSAGE 2017, MATERIAL PROPERTY IDENTIFICATION DUPLICATED. x WAS

On the Intel SCO/UNIX, the job terminates without any error message in the f06 file. A symptom of this error is the appearance of the following message in the log file: SIGNAL NUMBER 11 Avoidance: Use unique ID's for each MATi entry. @ 6359; 68.2; Nonlinear Analysis - Superelements, OLOAD, CLOAD - Incorrect OLO AD Output In nonlinear superelement analysis the OLOAD resultant for the residual includes loads from the upstream superelements. This error has been corrected in SOL 106 in MSC/NASTRAN Version 68.2. Avoidance: compile dmap=NLSTATIC souin=mscsou nolist noref alter 290,290 $ VECPLOT PGRES,BGPDTH,EQEXINS,CSTMS,CASESX2,,,/PGB/ GRDPNT/0/+1/'OLOAD' $ @ 6360; ; Miscellaneous - NFS Mount - Job Hangs Submitting a job from a machine that is nfs mounted may result in the job hanging. The exact cause of this specific problem is under investigation but is often the result of incompatibilities in the operating systems. Avoidance: Prior to MSC/NASTRAN V69, set system cell SYSTEM(207)=1;

after that, set the keyword "sysfield=lock=no". This feature is designed to block file locking for those situations where, despite the correct programming of the file locking calls, the system hangs or aborts. @ 6361; put 68.2; Postprocessing - PARAM,POST,-1/PARAM,OUGCORD,BASIC - Missing out If the neutral format option is requested when using "param,post,-1" and "param,ougcord,basic", then the ougv1pat datablock will not be written to the output2 file. This is due to the fact that the datablock ougv1pat is not defined in the NDDL. Avoidance: Add the following DMAP alter avoidance: compile sedrcvr souin=mscsou nolist noref alter 74,74 $ OUTPUT2 OUGV1PAT//OTAPE2/OUNIT2//OMAXR/'oug' $ V68 @ 6364; 68.2; Single and Multipoint Constraints - MPCFORCE - Wrong Answers for Superelements In static analysis in the structured solutions the MPCFORCE output is wrong for upstream superelements when rigid elements are on the boundary between the residual and the superelement or completely interior to the superelement. Avoidance: Insert the following DMAP alter: compile sedisp souin=mscsou noref nolist alter 121,121 $ PARTN PG,COL,ROW/,,,PMI/1 $ upartn uset,pg/pmx,,,/'G'/'M'/'S'/1 $ if ( seid=0 ) then $ PARTN Pmx,COL,/,,,PMI/1 $ else $ equivx Pmx/pmi/-1 $ endif $ @ 6365; ; Material Nonlinear Analysis - QUADX, TRIAX - Poor Answers The radial degree of freedom should be constrained for all grid points on the axis of symmetry for hyperelastic axisymmetric problems else poor answers may result. @ 6366; 2001; Elements - CROD - Wrong Answers When a CROD or CONROD that has torsional properties only, nonzero axial stresses are printed if the attached grids move. Avoidance: Use a torsion only CBAR or ignore the nonzero stresses. @ 6367; 68.2; Optimization/Design Sensitivity - OFREQ, ANALYSIS=DFREQ - SFE 42 76, EC 3001 If a OFREQ Case Control command is present in a direct frequency response optimization/design sensitivity analysis the following message is issued:

SYSTEM FATAL ERROR 4276.GREAD

ERROR CODE 3001 (DSAD)

Avoidance: Insert the following DMAP alter: $ compile fea souin=mscsou list noref alter 297,298 $ CALL SUPER3, CASECC, ,ULF,FOL1,PPFX,PSF1, , , ,DLT, , , ,,,, PCDB,XYCDB,POSTCDB,,EMAP,MAPS,EQEXINS,PVTS,CASES, SLT,ETT,GOAT,GOAQ,,,,UFOO,,GPLS,USET,SILS, PJ,EDT,YS,GM,,KFS,KSS,QR,,,,MAR,MEA, DYNAMICS,BGPDTS,CSTMS,MPTS,DIT,EST,GEOM2S,GEOM3S,ECTS, EPTS,INDTA,,,GPECT,,,SILX,EQEXINX, ECTX,BGPDTX,,GEOM4S,PG,,,,EPT,MPT, GEOM1,GEOM2,GEOM3,GEOM4,,,,CASERS,,,,,,GPSNTS,PVT,,,,,, KSF,KMG,MMG,MSF,BMG,BSF,K4MG,K4SF, K2MG,M2MG,B2MG,K2SF,M2SF,B2SF/ UPF,QPF,CASEDR,XYCDBDR,PUG,,oli, OUPF1,OPPF1,OQPF1,OFEF1,OFES1,OFEE1,, OUPF2,OPPF2,OQPF2,OFEF2,OFES2,OFEE2,, ,,,,,,/ $ /APP1 /CYCLIC/NLHEAT/AERO APP/APP /RSONLY/FALSE /FALSE /FALSE/DROPT/FS/ S,PFILE/S,CARDNO/DESCYCL1/FALSE/DMRESD/FALSE/0/FALSE $ $ IF ( FEAOPT='ANALYSIS') EQUIVX OLI/FOL1/ALWAYS $ $ ENDALTER $ @ 6368; ; p-Elements - FEEDGE - Undetermined Displacement Coordinate Syste ms on Edges If two faces with different displacement coordinate systems intersect at an edge that has no explicitly defined displacement coordinate system, then the coordinate system used for the edge cannot be easily determined. Avoidance: In this situation, specify the displacement coordinate system for the edge explicitly. @ 6374; 69; Grid Point Force - TEMP(LOAD), SUBCOM, GPFORCE - SFM 6220 If a SUBCOM involves a temperature load, and gpforce is requested as an output, then the job will fail with the following error messages: SYSTEM FATAL MESSAGE 6220 (GPFDR2), ERROR SEARCHING INDEX IN SCRT9 FILE Avoidance: Replace the SUBCOM with an appropriate SUBCASE. @ 6377; 68.2; Documentation - Quick Reference Guide - Superelements, SESET The Gi description for the SESET entry is incorrect. The GRID ID limit of less than 1,000,000 is a pre Version 68 limit. The new limit is less than 100,000,000. @ 6378; ; Element - GENEL - UFM 316

If the [S] matrix is left out (to be generated automatically by MSC/NASTRAN) when using the GENEL element, then {ud} must contain six and only six degrees of freedom. This fact is documented in the MSC/NASTRAN Quick Reference Guide under the GENEL entry description. However, if less than six degrees of freedom is defined when using the Z option, then the following unclear error messages is issued: USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA CARD GENEL CARD COUNT = y x SORTED

On the other hand, if the K option is used, no error message is issued. The job most likely fails downstream with other error messages. Avoidance: If the [S] matrix is to be generated automatically by MSC/NASTRAN, define six and only six degrees of freedom for {ud}. @ 6380; ; Nonlinear Analysis - RESTART, LOOPID, SUBID - Missing Output If PARAM,LOOPID and PARAM,SUBID are used to restart a nonlinear analysis and SUBID references a SUBCASE that is not the next sequential subcase following the subcase associated with LOOPID, then missing output occurs. Avoidance: If the user wants to restart an analysis and continue at the beginning of a new subcase, then the subcase(s) should be added sequentially following the subcase associated with LOOPID. @ 6383; ; P-Elements - MODES - UFM 6500 (ADAPT) The modes command does not appear to work with p-elements and yields an incorrect fatal message. The following Case Control: ADAPT = 150 VUGRID(PLOT)=ALL ADACT = ALL $ METHOD=151 SPC=999 DISP=ALL SUBCASE 1 MODES=1 SUBCASE 2 STRESS=ALL resulted in the following fatal message: USER FATAL MESSAGE 6500 (ADAPT) MULTIPLE SUBCASES WITH DIFFERENT METHOD COMMANDS AND/OR SPC/MPC/SUPORT COMMANDS HAVE BEEN SPECIFIED WITH P-ELEMENTS IN NORMAL MODES ANALYSIS. USER INFORMATION: ONLY ONE BOUNDARY CONDITION AND ONE METHOD COMMAND PER RUN IS ALLOWED. Note that the method command and constraint requests are above the subcase level. Avoidance: There is no avoidance.

@ 6386; 69; Documentation - DMAP Module Dictionary - UFM 2143 If the MPL listing for the module show that the parameters are of the BCD type, then they must be of length 8 as shown by the following example. type parm,,char8,n,major,setj,comp='comp' $ major='g' $ setj='a' $ vec uset/vgacomp1/major/setj/comp $ If other lengths are used, e.g., char4 instead of char8, then the job may fail with the following error message: USER FATAL MESSAGE 2143, MODULE VEC UNABLE TO IDENTIFY SET OR SUBSET DESCRIPTOR G A This fact is not documented in the DMAP Module Dictionary. Avoidance: Define the BCD parameter type as char8. @ 6387; 68.2; Nonlinear Analysis - NLTRD2 PARAM,TSTATIC - SFM 4276, EC 7777, I nefficient An error in the bisection logic may cause nonlinear analysis to overshoot the load factor and the subcase may be repeated. If this error is encountered in a transient nonlinear analysis using PARAM,TSTATIC then (depending on machine) the job may issue SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777 in the NLTRD2 module or may exit the NLTRD2 module without any message.

@ 6388; ; p-Elements - GMSPC, GMBC, CORDiR, CORDiC - Wrong Answers When constraints, either GMSPCs or GMBCs, are defined in a cylindrical or spherical displacement coordinate system on an edge or face, the constraints can only be approximately satisfied. It has been observed that in some cases this leads to non-physical singularities. Avoidance: If possible, the basic coordinate system should be used. @ 6390; .2 68.2; Installation - Installation Script Does Not Work on AIX Post 4.1 The installation script for MSC/NASTRAN, EMAS and MN4Pro Version 68 based systems will not work on AIX Version 4.1.2. (The current version is 4.1.2, and this error was reported on a pre-release version that has not yet been named). The uname command returns different information than in previous releases. As a result, the scripts do not recognize the system as AIX. The correction requires changing the architecture selection code that examines the uname output. Currently the code uses the string "AIX" in case statements. The case selector must be changed to "AIX*". This construct is used in the following files:

bin/nast68 or bin/emas3, bin/util68, msc68/access/access, msc68/util/util, bin/mscdoc1 A typical correction would be to change the line: AIX ) to AIX* ) @ 6391; 69; DMAP - Matrix K4AA Has an Improper Trailer in Solution 129 In a SOL 129 run, the K4AA matrix for the superelement is FORM=1 (square) instead of FORM=6 (symmetric). This may cause DMAP fatal messages in user DMAPs that expect the matrix to be symmetric. No adverse affect in the delivered solution sequences is known. Avoidance: In Version 68, the following alter changes the trailer to be FORM=6: compile semrb, souin=mscsou, nolist, noref alter 13 modtrl xaa////6 $ @ 6396; 69; Documentation - SPLINE1, SPLINE2 - Missing Documentation The description of the SETG field of the SPLINE1 entry in the MSC/NASTRAN Version 68 Quick Reference Guide refers to remark 4 which does not exist. Even though the SETG field on the SPLINE2 entry serves the same function, the description does not refer to any remark at all. Avoidance: The remark's original intent was to inform the user that SETi entries are bulk data entries and as such they are described in the same section. @ 6397; 1 In Version 68.1, if the Lagrange Multiplier Technique is used on the SGI, the job may fail with the following (or possibly other) error messages: SYSTEM FATAL MESSAGE 5407. INERTIA (STURM SEQUENCE) COUNT DISAGREES WITH THE NUMBER OF MODES ACTUALLY COMPUTED IN AN (SUB-)INTERVAL SYSTEM FATAL MESSAGE 5401 (REIGL). LANCZOS METHOD IS UNABLE TO FIND ALL EIGENVALUES IN RANGE. This error is most likely to occur when using one of the Lagrange Multiplier alters from the sssalter library (e.g., lmt11.v68). Avoidance: Set SYSTEM(198) and SYSTEM(205) to 1 at the top of the input file with the following NASTRAN statements: 68.2; Lagrange Multipliers - SYSTEM(198), SYSTEM(205) - SFM 5407, 540

NASTRAN SYSTEM(198)=1 NASTRAN SYSTEM(205)=1 @ 6402; 69; Data Recovery - ESE - SFM 6212 If ESE=n references a SET n then the run fails with SFM 6212. Avoidance: Use ESE=ALL. @ 6404; ; Restart - SCR=YES, DBVIEW - Datablocks Deleted If a restart run uses the DBVIEW statement in conjunction with SCR=YES, it is possible that the DBVIEWed datablocks may be deleted from the database. This is in spite of a RESTART, KEEP FMS statement. Avoidance: On restart runs, do not use SCR=YES. @ 6405; 69; Postprocessing - RCOUT2 - SFM 1208 (LDSTR) on RCOUT2 Using RCOUT2 to convert a neutral binary output file that was ftp'd from an IBM MVS/XA system results in SYSTEM FATAL MESSAGE 1208 (LDSTR). This error occurs because a hex 71 is mistakenly written from MSC/NASTRAN, and it is not an acceptable character in EBCDIC. The error will be fixed in future versions. The client can fix MSC/NASTRAN at their site using the following superzap: NAME NASTRAN UNSTR VER 031C BD71 REP 031C BD5F IDRDATA FXUNSTR @ 6406; 68.2; Nonlinear Analysis - PARAM,TESTNEG,2 - Differential Stiffness no t Ignored In nonlinear static analysis PARAM,TESTNEG,2 option is suppose to allow the user the option to ignore the affects of differential stiffness. This parameter does not work. Avoidance: A DMAP avoidance will be supplied shortly.

@ 6413; 68.2; Miscellaneous - script, SMEM - SMEM Incorrectly Set The runtime script incorrectly sets smem=100 as a default value. Unfortunately, this overrides the system supplied default value. The script should not enter a value for smem unless the user wants to change the default value. The 68.0 and 68.1 scripts may be corrected as follows: Change the line: smem="100" to:

smem="" Change the line: echo "SMEM=${smem}" to: test -n "$smem" && echo "SMEM=${smem}" @ 6416; n For any structured solution sequence which computes static loads, if UFM 2025 (UNDEFINED COORDINATE SYSTEMS) is issued more than once by the SSG1 module then the job will continue through DCMP and attempt a static solution instead of terminating the run. This error is caused by load entries such as GRAV, FORCE, MOMENT that reference a coordinate system that does not exist. Avoidance: Insert following DMAP alter: COMPILE SELG SOUIN=MSCSOU NOREF NOLIST ALTER 12,12 $ v68 IF ( NOGO<0 ) THEN $ @ 6418; (DSAD17) 69; Optimization/Design Sensitivity - SFM 5502 (DSAC02) vs UFM 5502 68.2; Static Analysis - SOL 101, SSG1 - UFM 2025 Does Not Terminate Ru

The message number 5502 is used in different routines for different reasons. The Version 68 Reference Guide documents only one of these messages. Avoidance: No avoidance is necessary, unique numbers will be assigned in Version 69 and the documentation updated to reflect the new numbers. @ 6419; 68.2; Miscellaneous - DBUNLOAD, SMEM=0 - FORTRAN BIO Fatal Message If a DBUNLOAD is requested and a FORTRAN unit number is not explicitly specified, then the default unit number is 51. This number causes a conflict when SMEM=0 on systems with FORTRAN BIO. For Version 67 this error only occurs on Hitachi OSF machines when ESS is used. Avoidance: Specify a FORTRAN unit number on the DBUNLOAD entry that is less than 50. This was fixed in v68.2 for both Hitachi and NEC @ 6420; 9 In a nonlinear transient analysis, if the AUTO method is selected and the job is submitted with scr=yes, then the job may fail with the following error message: (HP only) 0*** SYSTEM FATAL ERROR 4276.SKPFIL Avoidance: There are two avoidances: ERROR CODE 1039 ; Nonlinear Transient Analysis - SCR=YES, NLTRD - SFE 4276, EC 103

1. Submit the job with scr=no or insert the following dmap alter avoidance 2. Compile nltran souin=mscsou nolist noref alter 1 $ putsys(0,192) $ @ 6421; 68.2; Elements - QUAD4, STRESS(CORNER) - Poor Default In MSC/NASTRAN Version 68 there are three different options (CUBIC, BILIN and SGAGE) for the calculation of QUAD4 corner stresses. Any of these options may be explicitly requested on the STRESS entry. The default option obtained by simply selecting STRESS(CORNER) is the CUBIC option. It has been determined that the BILIN option produces better results for a wider range of problem, thus beginning with Version 68.2 the default option is changed to BILIN. Avoidance: In versions prior to Version 68.2 the user should explicitly specify STRESS(BILIN) to obtain the BILIN option. @ 6422; 69; Optimization/Design Sensitivity - DOM6 - SFM 3001 If there are no retained type 1 responses, a request for design sensitivity output (PARAM OPTEXIT = 4 or 7) produces the following fatal message: SYSTEM FATAL MESSAGE 3001 MODULE = DOM6 THE INPUT DATA BLOCK NAMED DSCMG AT POSITION 6 DOES NOT EXIST.

Avoidance: Force the retention on at least one type one response. This can be done by creating a dummy invariant response that is constrained at a level that will require its retention but that will not affect the design task when OPTEXIT=7. @ 6423; 68.2; Miscellaneous - DBMGR, DBSTORE, and DBFETCH, DIAG 20 - Excessive Output DIAG 20 activates a DIAG which produces excessive output if placed in the Executive Control. DIAG 20 is also used to print diagnostic messages in DBMGR, DBSTORE, and DBFETCH. A separate DIAG (DIAG 47) has been added in Version 68.2 so that the diagnostic messages from DBMGR, DBSTORE and DBFETCH may be obtained with out excessive output of DIAG 20. @ 6424; tation ; Nonlinear Analysis - CGAP - UFM 2025, Poor Error Message/Documen If a gap element connects coincident grid it is a documented requirement that a coordinate system must be specified else USER FATAL MESSAGE 2025 will be issued. Although this is a user input error, the fatal message should be improved to include more descriptive text such as: "USER FATAL MESSAGE ----, COINCIDENT GRIDS WITH NO CID FOR THE GAP ELEMENT ----." @ 6426; 69; DMAP - MATMOD - Poor Message, SFM 4276 EC 105 If the first datablock for the MATMOD module does not exist, then SYSTEM FATAL ERROR 4276.PACK ERROR CODE 105 may result. This has not been

observed in the solutions provided with MSC/NASTRAN. It has only been observed when a user DMAP is used. Avoidance: Make sure that the first datablock exists. @ 6430; 69; Optimization/Design Sensitivity - CORDi, DOM11 - SFM 3002 "SYSTEM FATAL MESSAGE 3002 MODULE = DOM11" may occur in an optimization run that has many coordinate systems. Avoidance: The error may be avoided by providing a set of dummy GRID entries in the deck. Let NGRID be the total number of GRID entries including dummy ones and NCS be the total number of coordinate systems defined in the deck. Then, NGRID should be determined as follows: NGRID > NCS * 13 / 8 @ 6431; 69; Optimization/Design Sensitivity - GPSTRESS - UFM 1126 In solution 200, if a GPSTRESS request is made in two subcases with different analysis types, for example analysis=statics and analysis=modes, then the following fatal message is encountered: USER FATAL MESSAGE 1126 (GNFIST) DMAP MODULE SUPER3 IS ATTEMPTING TO OUTPUT DATABLOCK 206 NAME = GPSF , WHICH ALREADY EXISTS. USER ACTION: DELETE THE DATABLOCK, OR USE FILE GPSF =OVRWRT Avoidance: Remove the GPSTRESS request from one of the subcases. @ 6433; 69; Grid Point Stress - PARAM CURV, PARAM,S1M - SFM XXXX In a normal mode analysis (e.g., SOL 63, 103, etc.), if the stress output is requested in a material coordinate system, then the following error messages may be issued: SYSTEM FATAL MESSAGE XXXX. SUBROUTINE CURV2 HAS RETURNED WITH ERROR CONDITION -2 LOC CODE = 600 IN SUBROUTINE CURV2 FILE NUM = 301 In addition to printing the regular stress output, the run will print the transformed stresses for the first mode only. Avoidance: Add modes=x in the Case Control Section. The value "x" is an integer greater than or equal to the number of modes in the model. @ 6436; 68.2; Optimization/Design Sensitivity - DOPTPRM, APRCOD=2 or 3 - SFM 4 276, EC 7777 "SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777" indicating a divide by zero is possible in the DOM9 module when either the default approximation method, Mixed Method, or type 3, Convex Linearization, is used. The approximation method is set by specifying APRCOD on the DOPTPRM entry. Avoidance: If this error is encountered set APRCOD equal to 1, Direct Linearization. Note that Direct Linearization is recommended for shape optimization problems.

@ 6447; 68.2; Complex Eigenvalue Analysis - CEAD, HESS - Poor Answers The epsilon to determine convergence in the Hessenberg Method of complex eigenvalue extraction may be too large for some problems and result in poor answers. Avoidance: There is no avoidance. The epsilon is squared in Version 68.2 to correct this error, that is where in prior versions the epsilon was of the order of 1.0E-7 in Version 68.2 the epsilon will be of the order of 1.0E-14. @ 6454; ; Dynamic Response Analysis - SDAMP, FRRD1 - UFM 2089 The SDAMP case control command selects modal damping and thus should be ignored in direct dynamic solutions. When it is present in a direct solution, it results in the following misleading diagnostic messages: USER FATAL MESSAGE 2089 MODULE = FRRD1 TABLE 4 IS UNDEFINED. USER FATAL MESSAGE 2089 MODULE = FRRD1 TABLE 5 IS UNDEFINED. Avoidance: Remove the SDAMP case control command. @ 6457; 6 When a nonstandard buffsize is used, it is possible that a run will fail and issue a SYSTEM FATAL MESSAGE 4276. To date this error has only been experienced on HP systems in the DSVG4 and APD modules. The DSVG4 module is no longer executed in the Version 68.2 solution sequences. Avoidance: To avoid the error in general specify a standard buffsize. To avoid the error in the DSVG4 module apply the following DMAP avoidance: compile respsen souin=mscsou nolist noref alter 171,171 PARAML PARAML matgen diagonal mpyad @ 6458; 69; Optimization/Design Sensitivity - DSVG1 - UFM 5165 The message: "USER FATAL MESSAGE 5165 (DSVG1D-1) UNDEFINED INPUT FILE = 108, NAME = (NONE) , ERROR CODE = -2, MODULE DSVG1 IS TERMINATED" has been encountered in a static - normal modes design optimization run with PARAM,OPTEXIT,7. (Version 68 and 68.2 only) UGy//'TRAILER'/1/S,N,NCOL $ no. of cols reqd. in SPAW1 UGx//'TRAILER'/1/S,N,nrow $ no. of rows reqd. in SPAW1 ,/spaw1/4/ncol/nrow/0/1/nrow/1/1/nrow $ spaw1/spawn/'whole'/0. $ ugx,spawn,ugy/ugsn $ 69; Optimization/Design Sensitivity - DSVG4, APD, BUFFSIZE - SFM 427

Avoidance: Insert the following DMAP alter: compile desopt souin=mscsou noref nolist alter 257,257 $ version 68 alter 269,269 $ version 68.2 delete /fol1,ppf1,,, $ alter 291,291 $ version 68 alter 305,305 $ version 68.2 delete /fol1,ppf1,,, $ @ 6460; 70.5; Superelements - PARAML - UFM 1127 (GNFIST) DMAP MODULE PARAML If the split database approach is used in the SSS in Version 68 or later releases, if any upstream superelement databases are "OFFLINE" (not available), it is possible that MSC/NASTRAN may fail in module PARAML with USER FATAL MESSAGE 1127 stating that the USET is not present. This only occurs if the program is allowed to attempt to process the superelements whose databases are OFFLINE. Avoidance: As long as there are no changes to the previously processed superelements, the SEALL command may be used pointing to only the superelements to be processed in the current run. If the previously processed superelements are changed, their databases must be ONLINE in order for MSC/NASTRAN to include the effects of the changes. @ 6461; 69; Heat Transfer Analysis - CQUAD8 - Wrong Answers The capacitance matrix for the higher order CQUAD8 element uses a lumped capacitance matrix instead of a consistent capacitance formulation. Therefore, the weighting on the QUAD8 capacitance is not correct. Avoidance: Use the CQUAD4 element instead. @ 6463; 69; Data Recovery - CURV3 - SFM XXXX SUBROUTINE CURV3 The following error message may occurs when PARAM,CURV,1 and PARAM,S1M,1 are used. SYSTEM FATAL MESSAGE XXXX. SUBROUTINE CURV3 HAS RETURNED WITH ERROR CONDITION LOC CODE = 150 IN SUBROUTINE CURV3 FILE NUM = 302 -3

Although the fatal message is printed, the run continues and gives the correct results. The reason for this error is not known at this time. Avoidance: Ignore the error message. @ 6464; 70; Miscellaneous - SPCF=set id, DISP=ALL = Missing Output If a deck contains a request for DISP=ALL and a SPCFORCE output request that refers to a set id, then the displacement output may also be restricted to the set. This error does not occur all the time. The exact cause is under investigation. Avoidance: Request SPCFORCE=ALL instead of a set id. @

6466;

; Miscellaneous - UFM 316 "USER FATAL MESSAGE 316" which indicates a format error may be issued when the large-field input format is used when the format of the entry appears correct. The error message may be caused because the number of large-field entries is odd. Section 3.7.1 of the Reference Manual states that the number of large-field format entries must be even. The program should check and issue a specific fatal message if this rule is violated, or better yet, accept an even or odd number of entries. Avoidance: Add a blank wide field continuation line if you have an odd number of large-field lines.

@ 6468; 69; Eigenvalue Analysis - REIGL - Poor Diagnostic Message, UFM 5401 If the user sets f2 on the eigrl entry less than the first frequency of the component, the design intent is that the run is to continue, and set neigv=-1. The dmap writer then can decide to continue or write his own fatal error. Lanczos (a.k.a. the REIGL module) does all this, but also puts out the following misleading message: *** SYSTEM FATAL MESSAGE 5401 (REIGL). LANCZOS METHOD IS UNABLE TO FIND ALL EIGENVALUES IN RANGE. ACCEPTED EIGENVALUES AND ADDITIONAL ERROR MESSAGES MAY BE LISTED ABOVE. USER ACTION: RERUN WITH ANOTHER METHOD OR ANOTHER SETTING ON EIGRL ENTRY.

@ 6469; 007 69; Nonlinear Transient Analysis - NLTRD2, AUTOSPC - UFM 6137, SFM 3 When using component modes synthesis on superelements in a nonlinear transient analysis with adaptive time stepping (METHOD=ADAPT on the TSTEPNL entry), you must select the exact number of generalized coordinates to match the number of modes; otherwise, your job may fail with one of the following error messages: *** USER FATAL MESSAGE 6137 (DFMNRS/D) INPUT MATRIX IS RANK DEFICIENT, RANK = x

SDCOMP INFORMATION MESSAGE--COLUMN YYY OF DATA BLOCK SCRATCH IS NULL. SYSTEM FATAL MESSAGE 3007 ILLEGAL INPUT TO SUBROUTINE NLTRD2 Avoidance: 1. Specify NASTRAN SYSTEM(69)=16. 2. First determine the number of modes in the range of interest for each component by specifying a small frequency range for SINV on the EIGR entry. Then specify the number of modes desired not the frequency range of interest. 3. Select another METHOD on the TSTEPNL entry.

@ 6476; 2001; Plotting - XYPLOT - Stress Plots Default to Max Shear The MSC/NASTRAN Quick Reference Guide states that von Mises stresses are the default, but if von Mises stresses (item code 9 for QUAD4) are requested through a xyplot request, but there is no corresponding case control request, then max. shear is calculated instead of von Mises. This happens because the stress request in case control is defaulted to von Mises in IFP, but the corresponding default is not set in SDR2 or DDRMM for implied stress recovery using the xyplot section. Avoidance: The user can request STRESS(PLOT)= n in the case control to force von Mises. @ 6477; 69; p-Elements - Missing Mass Matrix for p-elements If the first p-element processed (sorted by element type, then ID) has zero density, then the mass of all elements are ignored. Avoidance: If a zero mass element is desired, then it must be given a higher ID. Another alternative is to assign a tiny density, like 1.e-30. @ 6478; 68.1; Executive - XCSA - Unrecoverable Library Error, Traceback If a CEND is missing in the input deck the CRAY Version 68.0 of MSC/NASTRAN hangs in module XCSA with the following messages: lib-1212 /msc/msc68/exe/unicos/analysis.Y-MP: UNRECOVERABLE library error An internal WRITE tried to write beyond the end of an internal file. apparent state: unit 6 named out//infinite_loop.f06 last format: (A4) lately writing sequential formatted internal SIGNAL - Signal SIGABRT(6) caught, job aborting. See signal(2), CRI manual SR-2012 for further information. TB001 - BEGINNING OF TRACEBACK - $TRBK WAS CALLED BY RECOVR AT 525143a (LINE NUMBER 33) - RECOVR WAS CALLED BY sysrcv AT 525363a (LINE NUMBER 112) - sysrcv WAS CALLED BY __handlr AT 1053431a etc. @ 6481; ; p-Elements - EMGPMD - UFM 4659 P-element models with cylindrical or spherical coordinate systems generated using the Aries evaluator may encounter the message "USER FATAL MESSAGE 4659 (EMGPMD) DURING INTEGRATION FOR ELEMENT WITH ID =". This error is demonstrated by tpl deck pfms03.dat. Avoidance: Other than removing cylindrical and spherical coordinate systems which is probably not practical, there is no guaranteed avoidance. Changing the meshing or geometry of the model slightly may avoid the error. @ 6482; 69; p-Elements - ADAPT, DATAREC, OUTRCV, GP1 - UFM 2104 If the p-Element Case Control commands ADAPT, DATAREC, or OUTRCV reference nonexistent bulk data entries, then the job may fail in the GP1 module with the message "FATAL MESSAGE 2104, UNDEFINED COORDINATE SYSTEM"

Avoidance: Add the appropriate bulk data entries. @ 6483; 69; Postprocessing - OUTPUT2, OGS1, OGPS - UWM 6286 If a param,post,-1 is used in sol106 then following warning message is issued while writing the OGS1 datablock. USER WARNING MESSAGE 6286 (ZREAD) THE NDDL DESCRIPTION OF DATA BLOCK=OGS1 , RECORD=IDENT ,DOES NOT HAVE AN EITHER/OR TO MATCH OPT= 10. ITEMS ARE ASSUMED TO BE UNDEFINED. USER INFORMATION: PROBABLE CAUSE IS WRONG NDDL DESCRIPTION CODED, OR , WRONG GENERIC DATABLOCK USED (E.G. OES INSTEAD OF OEF). @ 6484; 70.5.1; SSSALTER - PARAM,DRMH,1 - SFM 4276, EC 3021 Single precision machines, i.e., Cray, some SGI, and NEC, fail when trying to run the files "CBAR1G.DAT" and "cbar9g.dat" in the sssalter directory. This data deck requests data recovery using client developed modules, drmh1 and drmh3, which have not been ported to single precision (long word) machines. SUN systems may experience problems running the DRMH3 module only. The jobs may end in the DRMH3 module with the message "time: command terminated abnormally" issued to the log file. Note that in versions past v68, the problem test decks have been renamed to alter1g.dat, 2g, 3g, and 9g respectively. @ 6485; 70; DMAP - TRNSP - Poor Message, CALLING PEXIT FROM DUMMY The following error message may be output from the TRNSP module when the DMAP forces single precision to be used by setting system cell(55) in the DMAP (i.e., PARAM //'SYSTEM'/S,N,IPREC/55/1 ): CALLING PEXIT FROM DUMMY Avoidance: A user should avoid forcing the precision. MSC recommends letting the precision default to machine precision. The avoidance is to remove this PARAM module from the dmap. @ 6487; 69; Optimization/Design Sensitivity - DSAL, MODES - SFM 4276 GETSTR E.C. 1409 If the MODES case control command is used in SOL 200, module DSAL will issue the message "SYSTEM FATAL MESSAGE 4276 ERROR CODE 1409". Avoidance: Do not use the MODES command in SOL 200. @ 6488; 69; Executive - COMPILER,LIST,REF - Missing Output, Fails to trap sy ntax error When COMPILE,LIST is input instead of COMPILER,LIST, the job runs failing to trap the syntax error and the listing is not produced.

@ 6490; entifier 69; Optimization/Design Sensitivity - DESOBJ - Misleading Subcase Id

The subcase identifier for the design objective produced when the P2 DOPTPRM parameter is set to 1 lists the first subcase when the objective function is weight or volume. Avoidance: Weight and volume responses are not a function of the subcase and the misleading subcase identifier should be ignored. Version 69 corrects this error by indicating a subcase id of zero. @ 6491; ; DMAP - COMPILE, DIAG 4,14, COMPILER - Extra Output Starting in Version 68, if the COMPILE NDDL command is used to punch out the NDDL and both the SOL and DIAG 4,14 (or COMPILER LIST,REF) statements are specified in the Executive Control Section, then NDDL will be written to the punch file twice. Avoidance: There are two avoidances: 1. Remove the SOL statement. 2. Remove either 4 or 14 from the DIAG (or REF or LIST from the COMPILER) statement. @ 6492; ed EOR 69; Design Sensitivity - DSVG1, MATi - UFM 5171 (DSVG1D-7) Unexpect The following message occurs when mass sensitivities are being calculated and some of the materials associated with the designed elements have a density of zero and others do not. UFM 5171 (DSVG1D-7) Unexpected EOR in File 101. SYSTEM FATAL MESSAGE 3003 MODULE = DSVG1 ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK MDICT (FILE 106) IN SUBROUTINE DSVG1D Avoidance: Assign a small, non-zero density in place of the zero values. @ 6494; 69; Eigenvalue Analysis - SDRX - Infinite Loop In SOLs 39, 63, 103, and 200 with ANALYSIS=MODES, if PLOAD1 entries are present and they are applied to CBAR or CBEAM elements then the run will go into an infinite loop in the SDRX module. (Version 68.2 only) Avoidance: Remove PLOAD1 Bulk Data entries or remove the STRESS and FORCE Case Control command requests. @ 6504; 2001; Rigid Elements - GP4, RBE3 - Poor Message If the components specified on the "UM" set reference the same grid point as the REFGRID, and the "UM" components is an incomplete subset of the REFC components, then the job may fail with the following error message in versions prior to Version 68. 0*** USER FATAL MESSAGE 2332, DEPENDENT MPC COMPONENT HAS BEEN SPECIFIED TWICE. EXTERNAL ID = xx DOF = yy

In Version 68 and above, depending on the solution sequence, the job may terminate with no messages in GP4 or it may fatal with the following unclear error message: FATAL ERROR A sample of an RBE3 that can cause this error is as follows: RBE3, 26, ,33, ,UM, 10,, 4.7, 16, 12, 14 16, 17, 1234, 31 1., 123, 11,

Avoidance: Remove the offending "UM" grid points from the continuation entry. @ 6508; 69; Optimization/Design Sensitivity - ANALYSIS=BUCK - UFM 4683 Shape optimization with the analytic boundary shapes approach and buckling analysis (ANALYSIS=BUCK) in solution 200 produces the following error message: USER FATAL MESSAGE 4683, KDIF MATRIX NEEDED FOR EIGENVALUE ANALYSIS Avoidance: The error is avoidable if the user generates basis vectors using either the direct input of shapes method or the DVGRID (i.e., manual) approach. Using geometric boundary shapes will also produce the same error. @ 6509; 68.2.3; p-Elements - PLOAD4 - Wrong Answers If a PLOAD4 is used to apply a pressure load to a p-element that has the connectivity inverted, i.e. that the right hand rule applied to the numbering of the first face produces a vector pointing out of the element, then the pressure is applied in the wrong direction and wrong answers result. Avoidance: There is no simple avoidance. @ 6514; 69; Optimization/Design Sensitivity - RESPSEN - Poor Performance The RESPSEN subDMAP uses the UG datablock rather than UG1. The result is that much of the efficiency that can be gained from load case and/or frequency deletion is lost. Further, there is no current requirement for static spcforces in this subDMAP. Avoidance: A new RESPSEN subDMAP is available on request. @ 6515; 1B) 69; Optimization/Design Sensitivity - DOPR1B, DLINK - UWM 7001 (DOPR The message "USER WARNING MESSAGE 7001 (DOPR1B) DLINK GIVES WRONG XINIT VALUE" is produced in error whenever NMSG = INT (( ndv - 4 )/ 7 ) - ndvi +1 is a positive number. For example, if there are 30 design variables, but only one independent design variable, 3 incorrect messages are

produced. The user input value for the design variable value is incorrect. The value computed based on the DLINK entry is correct. Thus, the error does not produce wrong answers. Avoidance: No avoidance is necessary since the error does not produce wrong answers but only an incorrect warning message. @ 6517; 69; p-Elements - GMCORD, DBUNLOAD, OUTPUT2 - Bad CSTM Trailer Data decks that contain only convective coordinate systems (GMCORD) and no CORDij type data entries have bad CSTM trailers. This does not cause a problem on homogeneous platforms, however, difficulty may be encountered during the neutral OUTPUT2 or DBUNLOAD operations. Avoidance: Include a dummy CORDij entry in the data deck. @ 6521; ; Nonlinear Analysis - TABLES1 - UFM 2076, Poor Message In nonlinear analysis, USER FATAL MESSAGE 2076 is issued if the MATS1 entry is type PLASTIC and the slopes of the first two segments of the stress-strain curve on the referenced TABLES1 entry are the same. The message: USER FATAL MESSAGE 2076, ELEMENT 0 REFERENCES TABLES1 4 WHICH IS THE INCORRECT FORM FOR NONLINEAR ELASTIC ANALYSIS does not clearly indicate the problem and incorrectly references element 0. @ 6525; 69; Nonlinear Analysis - CGAP - Incorrect SPCF Output When the GAP element has an offset moments may be created by the shear force times the axial offset and the axial force times the lateral offset). These moments are balanced by forces of single point constraint, however, the SPC Force output is incorrect in both magnitude and sign. Avoidance: Remove the offset. @ 6527; 69; Documentation - Quick Ref. Guide, Ref. Manual - PARAM,ACOUT,RMS PARAM,ACOUT,RMS only applies to element output. The current documentation in the MSC/NASTRAN Version 68 Quick Reference Guide erroneously implies that all the pressure output will be RMS. Avoidance: In the MSC/NASTRAN Quick Reference Guide, in the definition for parameter ACOUT, the following: ACOUT specifies the type of sound pressure level output to be used with the DISPLACEMENT or PRESSURE Case Control command in coupled fluid-structural analysis (see the MSC/NASTRAN Reference Manual, Chapter 15). The default requests the peak pressure level. ACOUT = RMS requests RMS pressure level and is only available in Solution Sequences 108 and 111. To obtain sound pressure level in units of dB and dBA, a peak reference pressure must be specified with PARAM, PREFDB. The dB level is defined as:

Should be replaced with: ACOUT specifies whether peak or root-mean-square (RMS) output is to be used with the FORCE Case Control command in coupled fluid-structural analysis (see the MSC/NASTRAN Reference Manual, Chapter 15). The default requests peak. ACOUT = RMS requests RMS and is only available in Solution Sequences 108 and 111. In order for decibel information (dB and dBA) from the DISPLACEMENT, PRESSURE, or FORCE Case Control commands to make sense, a peak reference pressure must be specified with PARAM, PREFDB. The dB level is defined as: In the MSC/NASTRAN Reference Manual, Chapter 15.4.2 Coupled Fluid-Structural Analysis, Output and Plotting, the following: The DISPLACEMENT or PRESSURE Case Control command is used to request: 1. Pressure and peak sound pressure level at the fluid points. 2. RMS sound pressure level is requested by the PARAM,ACOUT,RMS Bulk Data entry. This output is available in Modal and Direct Frequency Response analysis only (SOLs 111, 108, and 200). 3. Sound pressure level in dB and dBA - a peak reference pressure must be specified on the PARAM, PREFDB Bulk Data entry. The dB level is defined as dB = 20 log(P/PREFDB) (for air PREFDB = 20 x 10-6 Pascal). should be replaced with: The DISPLACEMENT or PRESSURE Case Control command is used to request: 1. Peak and root-mean-square (RMS) sound pressure level at the fluid points. 2. Sound pressure level in dB and dBA calculated from the RMS sound pressure. Note, a RMS reference pressure must be specified on the PARAM, PREFDB Bulk Data entry. The dB level is defined as dB = 20 log(P/PREFDB). For example in air, a common RMS reference pressure is PREFDB = 20 x 10-6 Pascal. The FORCE Case Control command is used to request complex fluid accelerations, velocities and pressure levels at the fluid elements. This information is peak by default but can be changed to RMS form if PARAM,ACOUT,RMS is used. This output is available in Modal and Direct Frequency Response analysis only (SOLs 111, 108, and 200). @ 6528; 68.2; Composites - SDR2, SDRCOMP, STRESS(CORNER) - Wrong Answers, SFM

4276 If a STRESS(CORNER) request is made for a deck with composite elements wrong answers result. The run may also issue SYSTEM FATAL MESSAGE 4276 EC 7777 or stop with a "Terminated Abnormally" message in the log file. Avoidance: The only avoidance is to either not request STRESS(CORNER) at all, execute STRESS(CORNER) for noncomposite elements in a cold start and then recover the stresses in the layered composite elements in a restart run. In Version 68.2 this error has been addressed by treating a STRESS(CORNER)=ALL request as STRESS(CORNER) for noncomposite elements only and outputting center stress for the composite elements. @ 6529; 68.2.3; Elements - TETRA, TEMPi, SSG1 - Wrong Answers Incorrect stresses result for the TETRA(10) element with curved edges and thermal loads. The error is due to incorrect thermal loads, and thus even the displacements are poor. This is a result of under-integration, and the magnitude of the error depends upon the amount of element geometry distortion. Avoidance: There is no known avoidance short of modelling with straight edges. @ 6534; 69; Acoustic Analysis - RESTART, GEOM2 - SWM 2072 The message "SYSTEM WARNING MESSAGE 2072, CARD TYPE 7408 NOT FOUND ON DATA BLOCK" may be issued in an acoustic analysis restart. The warning message does not affect the run and may be ignored. Avoidance: No avoidance is necessary. @ 6535; 69; Acoustic Analysis - RESTART - Inefficient Restart Changing the PANEL definition in the restart of an acoustic analysis causes the modes to be unnecessarily re-calculated. Avoidance: Apply the following DMAP alter to MSC/NASTRAN Version 68.1 COMPILE IFPL SOUIN=MSCSOU NOLIST NOREF ALTER 146,146 @ 6537; 68.2.1; Dynamics - STRAIN - Missing output STRAIN output is missing in SOLs 110, 111, 112, 145, 146, and 200, with ANALYSIS=MFREQ, MTRAN, and FLUTTER if PARAM, DDRMM is 0, which is the default. This error occurs in MSC/NASTRAN Version 68.2.0 only. This version of MSC/NASTRAN is included with MSC/Aries 7.0.0 on the IBM/RS6000, HP 9000/700, SUN SUNOS and SUN SOLARIS systems. Version 68.2.0 has not been delivered to non MSC/Aries users. Version 68.2.1 corrects this error and is delivered to non MSC/Aries clients. (Version 68.2. only) Avoidance: Insert PARAM, DDRMM, -1 or the following DMAP alter: compile sedrcvr noref nolist alter 174 DDRMM CASEDR,UH1,OL1,,,OSTR2,,XYCDBDR/,,OSTR2Y,,//3 $

@ 6538; 69; Documentation - Illegal data on EPOINT entry The MSC/NASTRAN Version 68 Quick Reference guide does not document a limit for the ID of an EPOINT. The limit is 1000000. @ 6539; 70; P-Elements - SDRP - SFM 4276 EC 7777 SDRP module The job will stop in the SDRP module and the message "SYSTEM FATAL ERROR 4276.RECOVR ERROR CODE 7777" will be output when multiple output requests are used on the output bulk data entry. The error only occurs on the Convex Version 68 systems. The error only occurs when the output requested is changed between sets.(i.e. when output requested for set 3 is different than output requested for set 2. For example: OUTPUT 1 ELSET=3,DISP=PRINT,STRESS=NONE,STRAIN=NONE,ERROR=NONE, PVAL=NONE,FIRST=NO,BY=0 ELSET=2,DISP=NONE,STRESS=PRINT,STRAIN=NONE,ERROR=NONE, PVAL=NONE,FIRST=NO,BY=0 ELSET=999999,DISP=NONE,STRESS=NONE,STRAIN=NONE,ERROR=PRINT, PVAL=PRINT,FIRST=YES,BY=1

@ 6540; E, CALCV 69; Nonlinear Transient Analysis - RESTART, EPOINT - SFM 3007, UMERG

A solution 129 restart from solution 106 fails with the following message whenever EPOINT bulk data entries are used: SYSTEM FATAL MESSAGE 3007 MODULE = UMERGE ILLEGAL INPUT TO SUBROUTINE CALCV Avoidance: Apply the following DMAP Alter: COMPILE NLTRAN SOUIN=MSCSOU NOLIST NOREF ALTER 112,112 UMERGE USETD,PLX,/IFD/'D'/'A'/'E' $ ENDALTER @ 6544; 69; Optimization/Design Sensitivity - SDSC, DSCREEN - SFM 3001 The message "SYSTEM FATAL MESSAGE 3001 MODULE = SDSC THE INPUT DATA BLOCK NAMED DSCMCOL DOES NOT EXIST" is produced in a sensitivity analysis when there are no active responses; i.e., all the responses have been screened out becuase they do not exceed the threshold required for sensitivity computations. Avoidance: Use the following DMAP avoidance (for V68.2.3): compile exitopt list noref souin=mscsou alter 'putsys' dbstatus R1TABRG,RSP2RG//S,N,NORSP1/S,N,NORSP2 $ if (NORSP1>0 OR NORSP2>0) THEN $

alter 'optexit'(7) endif $ NORSP1>0 OR NORSP2>0 @ 6547; 70; Heat Transfer Analysis - CHBDYG, RMG2 - SFM 4276 EC 211 If a connection order error is made on the CHBDYG, AREA4 element, the message SYSTEM FATAL ERROR 4276.BLDPK ERROR CODE 211

may be output which does not give the user any clude to the input error. Avoidance: There is no simple avoidance other than manually checking your CHBDYG,AREA4 elements for the correct connectivity. @ 6551; 69.1; Elements - HEXA, TEMP(LOAD) - Wrong answers The stress output for the HEXA element has an error. This error occurs only for grid locations (not at the center location). The error occurs only when the following conditions are all true: 1. An eight node HEXA element with bubble functions is used. 2. The temperatures are not the same at all grid points. 3. The stress output coordinate system axes do not align with the element coordinate system axes. Avoidance: The error may be averted by avoiding any one of the aforementioned conditions. @ 6552; 69; Heat Transfer Analysis - VIEW - Missing Output, UFM 6660, Not Up ward Compatible Using the old VIEW module produces a punch file in which the RADMiX ambient element (CHBDYP,POINT) is missing. If the punch file is included in a subsequent job, MSC/NASTRAN fails and the following message is issued: USER FATAL MESSAGE 6660 (VIEW) THE RADIATION MATRIX FOR RADIATION CAVITY ID= 11 DEFINED ON RADLST OR RADMTX BULK DATA ENTRIES HAS WRONG FORMAT OR DATA. USER INFORMATION: THE RADIATION MATRIX MUST BE SPECIFIED COMPLETELY. Avoidance: Use the VIEW3D entry to avoid the problem. @ 6554; 69; Heat Transfer Analysis - PARAM,NDAMP - Phase Change Analysis, Po or Convergence PARAM,NDAMP only allows a maximum value of .1. This is inadequate for the highly nonlinear phase change problem. The value of .1 is equivalent to a beta of .55 using the Crank-Nicholson method. The phase change problem requires a beta value of 1.0. Convergence is extremely difficult for the phase change problem with the current limitation on PARAM,NDAMP. Avoidance: There is no avoidance.

@ 6556; 69; Postprocessing - RFxxD66, POST=-1 - Missing Geometry Datablocks Geometry datablocks will not be written to the OUTPUT2 file with RFxxD66 when PARAM,POST,-1 and PARAM,OGEOM,YES is specified. This is contrary to the instructions for RFxxD66 listed in the alter. Avoidance: For RF63D66 insert the following below the line: IF ( POST=-1 ) OTAPE2=-1 $ IF (POST=-1 AND OGEOM='YES') THEN $ OUNIT1=OUNIT2 OUTPUT2 CSTM,GPL,GPDT,EPT,MPT//OTAPE2/OUNIT1//OMAXR $ OTAPE2=0 OUTPUT2 GEOM2,GEOM3,GEOM4,,//OTAPE2/OUNIT1//OMAXR $ ENDIF Similar changes are required for the other RFxxD66 rfalters. @ 6557; 68.2.3; P-Elements - TETRA, TEMP(LOAD) - Wrong Answers with TEMP Loads Poor accuracy occurs for TETRA p-elements with thermal loads and curved edges. Since the poor accuracy results from an error in the integration of the thermal loads, all results are poor. The error is much less if the edge p-value is three or greater (for a quadratic curved edge, p will be at least two). @ 6558; 69; Matrix Methods - SMPYAD, DBVIEW - Wrong Answers If SMPYAD is asked to do a matrix product of the form A(transpose)*B*C, and matrix C is a DBVIEW version of the datablock with the same name as matrix A, then SMPYAD substitutes matrix A for matrix C and performs A(transpose)*B*A instead, which will lead to wrong answers. For example: type db phg,mgg type parm,nddl,i,n,meth meth=6 dbview phg2=phg(where meth=3) $ Points to phg, but where method=3 not 6! smpyad phg,mgg,phg2,,,/phimphi1/3////1 $ Does phg*mgg*phg, not phg*mgg*phg2!!!!!! Avoidance: The simplest avoidance is to set nastran system cell 129 to 1 prior to the smpyad (e.g. insert PUTSYS(129,1) into the DMAP. The system cell may be reset to zero immediately after the smpyad by inserting PUTSYS(129,0). @ 6570; 69; Nonlinear Transient Heat Transfer - RESTART - UFM 2166 The enforced temperature datablock, YSNL, is not stored in the database. This causes restarts in solution 159 to fail with the message: USER FATAL MESSAGE 2166, MATRIX TO BE PARTITIONED IS OF SIZE xx ROWS BY yy COLUMNS. ROW PARTITION SIZE IS xx COLUMN PARTITION SIZE IS zz (INCOMPATIBLE). Avoidance: Insert the following alter in both the cold start and all

subsequent restarts: compile nltcsh souin=mscsou noref nolist alter 583 $ v68 $alter 585 $ v68.2 call dbfetch /ysnl,,,,/0/0/0/0/0 $ alter 618 $ v68 $alter 620 $ v68.2 call dbstore ysnl,,,,//0/0/' '/0 $ @ 6571; 68.2; Miscellaneous - SSG1 - SFM 4276 EC 7777 In MSC/NASTRAN version 68.0 or 68.1 it is possible to get the message "SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777", in the module SSG1 because of an error in SSG1 which causes overlapping memory regions. This error is therefore data dependent and is most likely to occur on the Hewlett Packard 9000, SGI R8000 or the DEC/Alpha OSF/1 systems. This error has been fixed in the V68.2 system. Avoidance: It may be possible to avoid this error if the model were changed slightly. @ 6572; ; p-Elements - GMCURVs with MSC/Aries Geometry Evaluator For certain cases, when the end GRID or the secant point is mapped back to the GMCURV, the MSC/Aries geometric evaluator (i.e., group MSCGRP2 on the GMCURV or GMSURF entry) returns the following message: SYSTEM FATAL MESSAGE 6498 (Aries B-spline Evaluator, Get Curve Parameter) GEOMETRY API MESSAGE FOLLOWS. USER INFORMATION: Numerical Error in Evaluator. Curve: 150. Point: ( 1.000000e+01 1.288252e+02 1.402565e+02). Avoidance: The problem is under investigation and it is unclear specifically when this error occurs or how to avoid the problem. @ 6573; 69; Documentation - Reference Manual, PLOTPS - Obsolete, Superseded Documentation Section 13.4 of the MSC/NASTRAN Reference Manual is obsolete. It has been superseded by the Utilities section of the Installation Guide for the each computer type. @ 6574; 69; Heat Transfer Analysis - CHBDYE - Wrong Answers If a CHBDYE (Boundary element) is used where side 5 represents an edge of a shell element such as a CQUAD4, CQUAD8, TRIA3, or TRIA6 element, the surface area calculated will be incorrect and wrong answers are produced. Avoidance: Use the CHBDYP,LINE option to represent the convection or radiation off the edge of the shell element. @ 6575; 2001; Nonlinear Analysis - QUAD4, ZOFF - Wrong Answers In Nonlinear Statics Analysis the output element forces from the QUAD4

elements are incorrect when the uniform offset (ZOFF field in the CQUAD4 bulk data) is non-zero. The force values are actually calculated in the non-offset element coordinate system while the stresses are calculated correctly in the offset reference system. All other solutions calculate both outputs in the reference system. Avoidance: Use the stress output and ignore the element force output. @ 6577; 079 2001; Eigenvalue Analysis - DPD, EIGR, EIGB, EIGC - UFM 2075 and UFM 2 When the POINT method of normalizing eigenvectors is selected on EIGR, EIGB or EIGC entries, and the specified grid/dof does not exist, the following mis-leading error messages is issued: 1) If the grid does not exist: SYSTEM OR USER DMAP FATAL MESSAGE 2075, IMPROPER VALUE xx FOR FIRST PARAMETER IN DMAP INSTRUCTION SDR2 2) If the component is invalid (in the case of SPOINT for example): USER FATAL MESSAGE 2079, GRID OR SCALAR POINT COORDINATE REFERENCED ON A NOLINI CARD. 0 HAS AN UNDEFINED

Note that these messages are issued even if the EIGR, EIGB or EIGC at fault is not selected by a METHOD entry in the Case Control section. Avoidance: Remove or change the offending entry. @ 6578; rs ; Heat Transfer - VIEW3D, CHBDYG,AREA8, CHBDYG,AREA6 - Wrong Answe The boundary surface elements CHBDYG type AREA8 and AREA6 were implemented in Version 68 to complement higher-order conduction elements such as the CQUAD8, HEX20, and TRIA6. Using these surface elements in View factor calculations results in an incorrect view factor if the surfaces are not in-plane or on a curve surface. Therefore in order to model surfaces on curvilinear surfaces such as cylinder it is best to use AREA4 and AREA3 elements. Avoidance: Use CHBDYG,AREA4 instead of AREA8 to model curvilinear surface for view factor calculation. Similarly, use CHBDYG,AREA3 instead of AREA6 to model curvilinear surface for view factor calculation. @ 6579; 69; Dynamic Response - SDR2, OFREQUENCY, SET=ALL - SFM 4276 EC 7777 In a frequency response run if the OFREQUENCY references a set where "all" frequencies are specified, then the job may fail with the message "SYSTEM FATAL ERROR 4276.RECOVR ERROR CODE 7777". Avoidance: Remove the ofreq=xxx case control command. @ 6580; 69; Nonlinear Transient Heat Transfer - VIEW - SFM 4276 EC 7777 An integer overflow, SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777, has been

encountered on the VAX VMS system while calculating the matrix density of RECM in the VIEW module. Avoidance: The only way to avoid the error is to reduce total number of HBDYi elements for each radiation cavity. @ 6582; 69.1; Postprocessing - DBC - XDBs are Too Large XDB files generated on a NEC SuperUX system are excessively large (by a factor of 8). These files are still readable from RECEIVE, the corresponding NTRL files are the correct size. Avoidance: The corresponding NTRL files are the correct size, so storing the XDB files in this form resolves the problem. @ 6590; 69; DMAP - ALTER, Missing Blank - No Warning Message When the following DMAP statement: alter 449,449 $ was mistyped as: alter449,449 $ the alter was ignored and no warning message was issued.

@ 6596; ion 70; Documentation - PARAM,USETPRT PARAM,USETSEL - Missing Documentat In the Version 68 MSC/NASTRAN Quick Reference Guide, the documentation for PARAM,USETPRT and USETSEL should include the following set names and their decimal equivalents: U1 U2 U3 U4 U5 U6 @ 6598; GL; Miscellaneous - PARAM - UFM 300, Improper Output MSC/NASTRAN has a restriction on the magnitude of an integer parameter of nine (9) digits. If an integer value with more than 9 digits is used for a parameter value (e.g. PARAM*,USETSEL,2147483648) then the parameter does not function properly. If the value of the parameter is hard coded directly in the DMAP (e.g. TABPRT USET,EQEXIN//'USET'/0/2147483648) the message " USER FATAL MESSAGE 300 FROM CRCARD ABOVE" is issued. Avoidance: There is no avoidance. @ 6599; 69; Frequency Response Analysis - FRRD1, DECOMP - UFM 6137 When singular matrices are encountered in a frequency response analysis, 2147483648 1073741824 536870912 268435456 134217728 67108864

MSC/NASTRAN Version 68 would fail in sparse decomposition with the message: USER FATAL MESSAGE 6137 (DFMNRS/D) INPUT MATRIX IS RANK DEFICIENT, RANK = 2 USER ACTION: CHECK MODEL The most common cause of the singular systems is a free body excited at zero frequency. Non-sparse decomposition, which was the default prior to Version 68, behaves differently. Non-Sparse decomposition would output messages that warn the user that a small term was being added to the diagonal so that decomposition could continue. An example of the messages are: USER WARNING MESSAGE---ROW 3 OF LOW. TRI. FACTOR HAS DIAGONAL TERM = 0( OR .LT. 0 IF CHOLESKY ) USER INFORMATION MESSAGE--- 1.0E-10 SUBSTITUTED FOR DIAG. TERM OF LOW. TRI. FACTOR AT ROW 3 USER INFORMATION MESSAGE 4158---STATISTICS FOR SYMMETRIC DECOMPOSITION OF DATA BLOCK SCRATCH FOLLOW MAXIMUM RATIO OF MATRIX DIAGONAL TO FACTOR DIAGONAL = 1.0E+10 AT ROW NUMBER 3 The solution for this frequency would then have a constant displacement on the order of 1.E+10 and a zero acceleration. Sparse decomposition has been enhanced in Version 69 to behave similarly to non-sparse decomposition. Thus, frequency response analysis at zero frequency with rigid bodies should no longer fail with USER FATAL MESSAGE 6137. Note that DMAP has also been added to issue a warning message and discard the solution for singular systems in frequency response. However, with the aforementioned change to add a small term to the diagonal we do not expect this DMAP to be activated under normal circumstances. Future versions may rely entirely on the DMAP action of discarding the solution rather than then adding a small term to the diagonal in decomposition. @ 6601; 69; Buckling - TEMP(INIT) - Wrong Answers In a linear buckling analysis, if TEMP(INIT) references a temperature that is non-zero, then the eigenvalue calculation will be incorrect, which in turn will lead to a wrong buckling load. The results of the static subcase are correct. Avoidance: Use one of the following avoidances: 1. Set TEMP(INIT) to zero and adjust the applied temperature to yield the same delta temperature or 2. Set the initial temperature using the TREF field on the MATi entry instead of using the TEMP(INIT) Case Control Command. @ 6603; 70; DMAP - SELA, SEMA, SEDR, PLTVEC, SEP4 - SFM 4252, UIM 4570 Starting in Version 67.7, the superelement modules, SELA, SEMA, SEDR, PLTVEC, and SEP4, only recognize a qualifier name of SEID for the qualifier which defines the data block family in the DBVIEW WHERE clauses. For example, in subDMAP SELA, if we modify the DBVIEW PAUP statement for

a family of user- defined data blocks, MYDB, and modify the 2nd parameter in the SELA module, DBVIEW PAUP = MYDB (WHERE MYQUAL=*) $ DBVIEW MAPUPL = MAPS (WHERE SEDWN=TEMP AND SEID=* AND PEID=*) $ SELA PJC,SLIST,EMAP,EQEXINS,PAUP,MAPUPL,GDNTAB/ PG/ SEID/'MYQUAL'/0/S,N,NOPGS $ then the SELA module will not find any members in the family PAUP. Avoidance: Modify the NDDL so that the PATH statement for the data block MYDB contains SEID as a qualifier. Also, modify the DBVIEW statement WHERE clause to include SEID=*. For example, NDDL statements: PATH MYPATH SEID ... DATABLK MYDB PATH=MYPATH ... DMAP statements: DBVIEW PAUP = MYDB (WHERE SEID=*) $ DBVIEW MAPUPL = MAPS (WHERE SEDWN=TEMP AND SEID=* AND PEID=*) $ SELA PJC,SLIST,EMAP,EQEXINS,PAUP,MAPUPL,GDNTAB/ PG/ SEID/'SEID'/0/S,N,NOPGS $

@ 6610; 69; Heat Transfer Analysis - ROD, BAR, BEAM - Inconsistent Capacitan ce Matrix The rod element was changed in Version 68 to have a consistent heat capacitance matrix. However, all the other one dimensional heat conduction elements such as the beam and bar elements have a lumped representation. This difference between the one dimensional element formulations may result in slight differences between a model comparing RODs to BARs and/or BEAMs. For a typical transient thermal analysis comparing the result between ROD and BAR elements, the difference in temperature output is generally less than 0.2%. Avoidance: Use bar or beam element. @ 6613; 69; Heat Transfer Analysis - CHBDYE, SSG1 - Incorrect Loading, Wron g Answers If a CHBDYE surface type element is used to model an edge with mid-side nodes and a heat flux loading is applied to this edge, then the answers will be incorrect. This error occurs because the SSG1 module does not know there is loading on the midside nodes. Avoidance: Do not reference elements with mid=side nodes on the CHBDYE entry. For example, instead of using the higher-order CQUAD8 elements that have mid-side nodes, use the lower order CQUAD4 elements. @ 6614; 69; Acoustic Analysis - ACSRCE - Wrong Answers The ACSRCE which defines the power versus frequency curve for an acoustic

source results in an incorrect acoustic fluid power input. The power is off by a factor of i*omega, the radian frequency. Avoidance: Use the RLOAD data entry instead with Qdot, (volume acceleration). Qdot is calculated as the square-root of 8 times pi times the speed of sound in the fluid times power divided by the density of the fluid: Qdot = sqrt(8*pi*c*power/rho) @ 6618; 69; Documentation - String-based alter - UFM 300 When using the string-based alter feature in Version 68.2, the dmap strings should be enclosed by the ' single-quotation string on both ends (e.g., 'string1'). In the Version 68.2 Release Notes, the dmap strings are shown to be enclosed with the ` single-quotation string on the left end and the inverted single-quotation string on the right end (e.g., `string1') throughout chapter 5. An example of this documentation error is shown on page 40 of the Release Notes as ALTER `CALL.*PHASE1B' $ The correct format for the above string-based alter is ALTER 'CALL.*PHASE1B' $ Note that the V68.2 Release notes can not be corrected. Since this documentation error does not exist anywhere else it is considered fixed in Version 69. @ 6620; 69; Installation - No Authorization Code File - Poor Message During the installation of MSC/NASTRAN V68.2 the user is asked to run a test problem after installation. If the user requests to run a test problem and there is no authorization code file, then the job aborts with: 'The test appears to have failed for an unknown reason.' If a bad authorization code is installed, then the installation test problem determines the correct error. But if no authorization code has been installed, then the noted message is issued. Avoidance: Install a correct authorization code. @ 6626; 69; Optimization/Design Sensitivity - DOPTPRM, GSCAL - Poor Default The GSCAL parameter on the DOPTPRM entry is used to avoid a divide by zero when calculating constraints. It has been found that the default value of 0.001 is too large when constraints are being calculated for property values that can be many orders of magnitude smaller than this. Some models with metric units have displayed this problem. The problem is characterized by the design task finishing prematurely. Avoidance: Set the value of GSCAL on the DOPTPRM entry to a value that is several orders of magnitude smaller than the expected property values. Along with this, it is always a good practice to avoid constraints that have a lower or upper bound of

0.0 @ 6628; 69; Heat Transfer Analysis - QHBDY, IFP - UFM 316 A negative heat flux can not be applied using the QHBDY entry. For example the entry: QHBDY 1 LINE -500.0 6.28 6 7

results in the message: USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA CARD QHBDY CARD COUNT = 82 1 SORTED

Although the use of a negative value is rare in heat transfer, it should be allowed. Avoidance: Use the QBDY entry to apply the negative heat flux value. @ 6631; 69.0.2; Optimization - Superelements, DRESP2 - UFM 5609 If an optimization model contains superelements, UFM 5609 may be incorrectly issued if different response types are combined using a DRESP2. For example, this error occurs if a superelement is defined in a frequency response optimization model, and both FRDISP and FRFORC responses from DRESP1 entries are combined on a DRESP2 entry. Avoidance: If different response items need to be combined, use a residual-structure only model or use external superelements. @ 6632; 69; Documentation - SYSTEM(198), SYSTEM(205) - Missing Documentation Version 68.1 of MSC/NASTRAN introduced a very significant performance improvement to sparse decomposition. This enhancement is called the high-rank update. The capability is controlled by two system cells, SYSTEM(205) and SYSTEM(198). The optimum value of these system cells is machine dependent. In Version 68.1 the optimum value of these two system cells are set in the delivered system rc file. In Version 68.2 the optimum values are set in the code. In either version the values may be overwitten by the user but this is generally not recommended. SYSTEM(205) sets the value for the rank of update. The value used for any run may be determined from user information message 4157 output in the f04 execution summary. SYSTEM(198) sets a lower bound on the number of rows to be eliminated in a front, hence the alias name MINFRONT. The meaning of system cell 198 has not changed from 68.0 to 68.2. In version 68.0 this feature was only helpful on a few very large problems on CRAY computers. In Version 68.1 and 68.2 it becomes more critical and can help performance on many other machines. Normally, it is best to set system cells 198 and 205 to the same value, this is done be default.

@ 6643; 69; Optimization/Design Sensitivity - PARAM,OPTEXIT, DSAL - SFM 4276 , EC 3021 The message "SYSTEM FATAL MESSAGE 4276, ERROR CODE 3021" occurs in the

DSAL module when a converged optimum solution is achieved in a modal frequency response design task and OPTEXIT=7 is used to obtain sensitivities of the final design. Avoidance: Remove the OPTEXIT=7 request. A new run that uses the final design and has OPTEXIT=4 will provide the desired information. @ 6656; 69; Optimization/Design Sensitivity - DEQATN - Segmentation Fault, T rap Encountered A DEQATN bulk data entry may not end with a semicolon(;). If it does, the run fails while processing the input in a machine-dependent manner. This error has appeared as a segmentation fault on SGI machines and a "Trap Encountered" on RS/6000 machines. Avoidance: Do not use a semicolon as the last character on a DEQATN entry. @ 6668; 68.2; DMAP - OUTPUT2 - FORTRAN I/O ERROR 910 On the HP, if a user dmap contains an output2 request and there is no corresponding assign statement, then the job may terminate without any error messages in the f06 file. The following error messages may appear in the log file. *** FORTRAN I/O ERROR 910: ACCESS PAST END OF RECORD ATTEMPTED FILE: ftn11, UNIT: 11 Avoidance: Explicitly define the output2 file using the assign statement in the FMS section. @ 6676; ; Plotting - SPCFORCE, MPCFORCE, XYPLOT - Missing Output If XYPLOTs of SPCFORCEs or MPCFORCEs are requested, the items must be requested in the case control before the OUTPUT(XYPLOT) section. For example, if SPCFORCE plots are to be requested for points 1 and 5, the following output request (or a similar one) would be required in the case control: SET 99 = 1,5 SPCFORCE = 99 @ 6678; ; Nonlinear Analysis - TEMP(LOAD) - Wrong Nonlinear Strains with T emperature Loads In nonlinear solution sequences, the output entitled "Nonlinear Stresses" prints mechanical strains which do not include the thermal strains. The true strains are the sum of the mechanical and thermal strains. The only exception to this limitation are the hyperelastic elements which output the true strain. Avoidance: There is no avoidance. @ 6682; 69; DMAP - MATMOD - MATMOD with option P1=2 and P5<0.0. Option P1=2 of the MATMOD module is intended to filter the matrix and set all terms in the matrix with magnitude smaller than P5 to 0.0. However, if P5 is set to a negative value, no filtering will take place. Avoidance: Use a positive value for P5. @

6684;

69; Optimization/Design Sensitivity - DOPTPRM - UFM 316 (IFPDRV) On the DOPTPRM data entry, a P1 value higher than 4 or 5, depending on the computer, causes the FATAL message: USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA CARD DOPTPRM Avoidance: Select a value for P1 that is equal to or less than 4.

@ 6701; 69; Transient Heat Transfer - TETRA - Wrong Answers The 10-noded tetra for the transient thermal analysis (SOL 159) will obtain wrong answer. Avoidance: Use lower order tetra elements. @ 6705; 69; Optimization/Design Sensitivity - MPCFORCE - Missing Output, Wro ng Answers Running SOL 200 with SCR=NO may result in wrong or missing MPCFORCE output. ANALYSIS=STATICS results in missing MPCFORCE output for all but the first cycle. Other approaches result in wrong but plausible answers for all but the first design cycle. Avoidance: Submit job with SCR=YES. @ 6711; 69; Miscellaneous - INCLUDE - UFM 307 (IFPDRV) If an include statement immediately follows a line which ends in a blank comma field, as in the example below, MSC/NASTRAN may fail with the message "USER FATAL MESSAGE 307 (IFPDRV) ILLEGAL NAME FOR BULK DATA ENTRY INCLUDE, USER FATAL MESSAGE 9002 (IFP) - ERROR ENCOUNTERED IN MODULE IFP".with the two lines below: EIGRL, 1, -0.001, include 'struct.inc' , 2, 0, 2, , $ $

Avoidance: The error can be avoided by separating the lines or removing the trailing commas. @ 6718; ored ; P-Elements - GRMAX, GNSTRT, GNMAX, VGSTRT - Keyword Settings Ign The NASTRAN keywords GRMAX, GNSTRT, GNMAX, VGSTRT,VGMAX,VESTRT,VEMAX, that control p-element parameters may be ignored. Avoidance: Use the actual system cells 175, 178, 179, 180, 181, 182, and 183. @ 6721; 69; Miscellaneous - Network - Error Writing File, Poor Message The following message may occur when MSC/NASTRAN cannot access the disk where the .f06 file is to be written in a reasonable amount of time

because of network problems. The sequential WRITE request cannot be completed because an AIX errno value of 28 was received while writing the sequential file ./csr5504s.f06. The program will continue if ERR= or IOSTAT= was coded in the WRITE statement. 1525-099 Program is stopping because errors have occurred in an I/O request and ERR= or IOSTAT= was not coded in the I/O statement. Avoidance: Either re-run the job when network traffic is reduced or direct the output files to a local disk, not through a network. @ 6722; wers 69; Transient Heat Transfer - TRIA6 element using CHBDYE - Wrong Ans The heat capacitance for the edge of the TRIA6 element is not correct. This results in wrong answers whenever the CHBDYE (boundary element) is used with SIDE other than side 1. Avodiance: Use TRIA3 element with CHBDYE or CHBDYP,LINE for the edge convection. @ 6724; 69; DMAP - Linking with SOL or LINK statement - UFM 149 and SFE 4276 in LINKAR If UFM 149 is issued then, in some cases, SFE 4276 will be issued and the run does not terminate gracefully. The exact symptoms are under investigation. If this occurs during the creation of a delivery data base, then the data base will be corrupted. Avoidance: Other than correcting the DMAP error indicated by UFM 149, there is no avoidance. @ 6725; 68.2; Matrix Methods - Sparse Decomposition, DECOMP - UFM A large model that has run successfully on other machines experiences negative terms on the factor when decomposing KOO on the MSC/NASTRAN Version 68.1 on the SGI R8000 computer. Avoidance: If decomposition errors are encountered that do not appear to be due to modeling problems, then it may be due to the High Rank Update feature introduced in Version 68.1. Disable this feature by setting NASTRAN SYSTEM(198)=1 SYSTEM(205)=1 or switch to Version 68.2 where the error has been corrected. @ 6726; 68.2; Miscellaneous - MPYAD - SFM 4276 Large client model failed in the MPYAD module on the SGI R8000 computer. Avoidance: The problem was avoided by disabling the sparse option for MPYAD. This was done by NASTRAN SPARSE=24, however, a significant penalty in CPU resulted. @ 6727; 69; Cyclic Symmetry - CYSUP - UFM 5423 In MSC/NASTRAN Version 68.2 it is possible to get the message "USER FATAL MESSAGE 5423 --- ATTEMPT TO ADD INCOMPATIBLE MATRICES, (SADD5)" in a cyclic symmetry (SOL 114) run using the CYSUP entry. Version 67.5 produces incorrect answers under the same circumstances.

Avoidance: For V67.5 COMPILE STATCYC SOUIN=MSCSOU NOREF NOLIST ALTER 8,8 NOA=LTOI(NOGSET=NOASET) $ CEND Avoidance: For V68 compile statcyc noref nolist alter 'noa=ltoi','' noa=ltoi(nogset=noaset) $ @ 6728; 69; heat transfer - apply boundary condition on both side on CHBDYE or CHBDYG aqa_02 heat transfer - apply boundary condition on both side on CHBDYE or CHBDYG SEE CSR 7462 for the fix that have corrected this error in version 69 Wrong summary for the H E A T T S F L O W I N T O H B D Y E L E M E N

The user has the option using same boundary element to specify front side and the backside convection and radiation. The temperature vectors calculated from using this option is correct. However, the summary table only show the energy on the front side of the element H E A T F L O W I N T O H B D Y E L E M E N T S (CHBDY)

ELEMENT-ID APPLIED-LOAD FREE-CONVECTION FORCED-CONVECTION RADIATION TOTAL 13 4.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 4.000000E+00 Avodiance: use different CHBDYE or CHBDYG ID for specifying the front side and the backside. @ 6731; 69; Cyclic Symmetry - SPCFORCE, CYSUP - Missing Output The forces of single point constraints in the static cyclic symmetry solution sequence SOL 81 do not include the reactions at the degrees of freedom specified on the CYSUP Bulk Data entry. Avoidance: Use SOL 114 or insert the following DMAP alter: compile sol81 souin=mscsou noref nolist alter 818,818 @ 6733; ; P-Elements - TEMP(LOAD), PVAL,ADAPT - Wrong Answers If thermal loads are applied to p-elements and the PVAL entry (referenced by the ADAPT entry) defines the polynomial order to be 1, then the solution may converge to wrong results. Wrong results definitely occur for models with only one element.

Avoidance: Set POLYi field on the PVAL entry to 3 or more. @ 6738; 69; Superelement Analysis - SEMA - Poor Performance If the residual structure of a large superelement model contains a relatively large percentage of the total grid points, then the total run time may be significantly larger then the same model without superelements. For example, in one client's superelement model with over 100,000 grid points in the residual structure with 16 superelements (each with about 1500 grid points) the total run time increased by a factor of 7. The increased time required was due to the SEMA (superelement assembly) module. SEMA is intended to work on a collector superelement (or residual structure) that is much smaller than the total model size. Avoidance: Try to keep the residual structure and collector superelement size to less than 10 percent of the model. The smaller these components are, the more efficient the assembly process is. A DMAP alter avoidance is also available from your MSC Technical Support representative. @ 6739; 69; Dynamic Analysis - ASETi, OMITi - Performance Degradation If a static reduction is requested in a dynamic solution and there is no mass on the omitted degrees of freedom, the program will still perform the operations to reduce the mass matrix. This operation is not necessary and may be skipped by a DMAP alter. Avoidance: In the SSS, include the following alter for V68: compile semr2, souin=mscsou, list, noref alter 19 $ $ check for null moo and moa $ matmod moo,,,,,/tempaa,/12/s,n,nulmoo/1 $ matmod moa,,,,,/tempbb,/12/s,n,nulmoa/1 $ paraml moo//'trailer'/1/s,n,nomoo $ paraml moa//'trailer'/1/s,n,nomoa $ if (nulmoo=nomoo and nulmoa=nomoa)then $ equivx maa1/maa/always $ else $ alter 23 endif $ @ 6744; e 5423 70.5; Optimization/Design Sensitivity - MTRXIN,GMA - User Fatal Messag The message "USER FATAL MESSAGE 5423" may occur in the MTRXIN module when an attempt is made to read M2PP, B2PP and K2PP matrices a second time. This error occurs in optimization analyses with K2PP, B2PP and TF inputs. This error was fixed in Version 69.0.2 and 69.1 but occurred in Version 70. The error is fixed in Version 70.5. Avoidance: Use the following DMAP to skip MTRXIN for subsequent times:

COMPILE GMA SOUIN=MSCSOU NOREF NOLIST $ ALTER 2 $ TYPE DB,ZUZR01,ZUZR02,ZUZR03 $ TYPE PARM,,I,N,NOM2PP,NOB2PP,NOK2PP $ ALTER 9,9 $ PARAML ZUZR02//'PRESENCE'////S,N,NOZUZ2 $ PARAML ZUZR03//'PRESENCE'////S,N,NOZUZ3 $ IF ( NOZUZ2<0 AND NOZUZ3<0 ) THEN $ MTRXIN CASES,MATPOOL,EQDYN,,TFPOOL/K2PP,M2PP,B2PP/LUSETD/ NOK2PP/NOM2PP/NOB2PP $ COPY M2PP/ZUZR01 $ COPY B2PP/ZUZR02 $ COPY K2PP/ZUZR03 $ ELSE $ COPY ZUZR01/M2PP $ COPY ZUZR02/B2PP $ COPY ZUZR03/K2PP $ NOM2PP = 0 $ NOB2PP = 0 $ NOK2PP = 0 $ ENDIF $ @ 6745; 69; P-Elements - GMBC, SPCADD - Wrong Answers GMBC entries referenced by SPCADD bulk data entries are ignored. Avoidance: Other than not using the SPCADD entry, there is no avoidance. @ 6753; GL; Nonlinear Static - RESTART, PARAM, NMLOOP - Wrong Answers If PARAM,NMLOOP is used to restart from SOL 106 (nonlinear statics) into a linear solution, the data recovery output is performed using linear element properties. Ths design causes displacements and element results to be incorrect. However, the solution is performed using the nonlinear equilibrium position (allowing EIGENVALUES to be correct). Only data recovery is in error, hence displacements and element results are wrong. This design does not affect a restart from a nonlinear run to a nonlinear run. Avoidance: There is no reason to perform a RESTART from solution 106 to 101, perform the desired data recovery in solution 106. @ 6755; ers 69; Documentation - Inertia Relief, SUPORT, PARAM,INREL - Wrong Answ In addition to a SUPORT entry, PARAM,INREL,-1 is required to perform inertia relief in solution 101. This requirement is not documented under the description of the SUPORT entry. It is documented under the SUPORT entry description the SUPORT entry "does not perform the same function as a constraint". In fact if PARAM,INREL,-1 is not included the SUPORT entry will act like a constraint. @ 6759; 69; SSSALTER Library - RFLAG,LMT - Fatal Error The SSSALTER Library files RFLAG and LMT will cause fatal errors when used with SOLS 107, 108, and 109 if the stiffness matrix is singular.

There are two possible avoidances: 1. Modify the dmap alter with the following change: Change the DMAP Alter from: GETLLL=(STATICS OR NORSET>-1 OR NO3SET>-1 OR (MODACC>=0 AND (SOLTYP='MODAL' OR APP='REIG'))) $ to: GETLLL=(STATICS OR NORSET>-1 OR (MODACC>=0 AND (SOLTYP='MODAL' OR APP='REIG'))) $ Removing the NO3SET. This matrix does not need to be decomposed for direct analysis. The alter causes a branch over this unneeded operation. 2. Modify the model by adding enough SUPORT entries to stabilize KLL. These SUPORTs will not modify the answers. @ 6761; ts 69; P-Elements - GMLOAD - GMLOAD Entry Does Not Check if FEFACE Exis The GMLOAD entry allows the following options in the entity field: FEEDGE, GMCURVE, FEFACE, and GMSURF If there are no FEFACE, GMCURV or FEFACE entries in the model, no error message is issued, and the user is not given a clear idea of what is wrong. @ 6768; 69; Documentation - Reference Manual, CURV, RF24D40 - Curve Projecti on Limits The description for the CURV module processing in the Reference Manual under RF24D40 (page 15.8-19) Remark 4.a. states: - If n.i >= 0.4 the projections of the y-axis on the surface of the element is taken as the reference axis. This statement should read: - If n.i > 0.4 the projections of the y axis.... @ 6772; 69; P-Elements - TEMP(LOAD), CHEXA,PSOLID,PVAL - Wrong Answers A skewed p-hexa element with p=1 (on the PVAL entry) produces incorrect results with thermal loading. If the p-hexa is not skewed or if p>1, the answers are correct. Incorrect results may also occur if the p-element is next to an h-element, since the p-element will be reduced to p=1 on that face to be compatible with the h-element. Avoidance: Turn off the bubble functions using the PSOLID entry (5th

field) or set p>1 on the PVAL entry. @ 6785; 70.5; Buckling Analysis - EIGRL - SFM 5401 In a buckling analysis, if only the V1 and ND fields are specified on the EIGRL entry, then the job may fail with the following error messages: SYSTEM FATAL MESSAGE 5401, LANCZOS METHOD IS UNABLE TO FIND ALL EIGENVALUES IN RANGE. A symptom of this error is the continuing shifting of Lanczos when ND roots have been found. Avoidance 1: Specify the V2 field in addition to the V1 and ND fields. Setting SHFSCL on the EIGRL entry near the absolute value of the lowest expected root may also avoid the error and improve efficiency. Another alternative is to use the SINV method if only a few roots are desired (which is true for most buckling analysis). Avoidance 2 (preferred over Avoidance 1): A new alter has been added to the sssalter library that can be used with Version 69.0.2. (the alter can not be used with prior releases) that solves this problem. The alter is named br.v69 for Version 69.0.2 or br.v691 for Version 69.1. Leave the SHFSCL entry on the EIGRL card blank. The alter will be made available on the web and on the 69.1 delivery. @ 6786; 69; Optimization/Design Sensitivity - FBS, QSETi - SFM 2198 and 2199 If QSETi entries are specified in solution 200 for the purposes of dynamic reduction and the a-set contains only the q-set degrees of freedom and there is also a static subcase, then the run will fail with SYSTEM FATAL MESSAGE 2198 and 2199. (Version 68.2 only) Avoidance: There are three possible avoidances: 1. Remove the q-set and the dynamic reduction requests and, if not already done so, then replace with a Lanczos eigenvalue analysis (see EIGRL entry). 2. Specify ASETi entries for one or more physical degrees of freedom. 3. Insert the following DMAP alter: compile fea noref nolist alter 'FBS'(,-1),'FBS' $ DBSTATUS UL,USET,LLL//S,N,NOUL/S,N,NOUSET/S,N,NOLLL $ IF ( NOUL=-1 AND NOLLL>0 ) FBS LLL,,PL/UL $ @ 6788; 2001; Transient Response Analysis - XYPLOT - UWM 982, Missing Output If an xyplot is requested for the velocity of a grid point which does not exist and there is no VELO= request in the CASE control for printed data, then no displacement, velocity, or acceleration data will be printed or plotted and the following message will be issued: *** USER WARNING MESSAGE 982. FORMAT OF OUGV1 DESIGN. CODE = 165 INCOMPATIBLE WITH SDR3

Avoidance: Review the output requests versus the grid and element ids in the model, insuring that all requests reference proper ids. Another avoidance is to add a VELO(PLOT)=ALL command to the case control. This command will produce very large files on large size models, however. @ 6789; Damping ; Nonlinear Transient Analysis - PARAM,LGDISP, PARAM,G - Incorrect

There is an error in the way element damping is implemented in nonlinear transient analysis (SOL 99 and 129). This error can result in answers with apparently very wrong damping when large displacement effects are requested with PARAM LGDISP. These solution sequences create a "B" damping matrix, from which damping forces are computed. This matrix is updated whenever the stiffness is updated. The problem is that many structures have stiffness ratios (and hence damping ratios) in perpendicular directions that are orders of magnitude different. A tiny misalignment of the damping can be intolerable. Avoidance: There is no avoidance. @ 6790; Message 69; Nonlinear Statics - MATGEN, SCE1 - SFM 3007, UFM MATGEN-4, Poor

In nonlinear static analysis, SOLs 66 and 106, if the model has no l-set, a poor message is issued. In SOL 66, SFM 3007 is issued for module SCE1 at line 858 in Version 68. In SOL 106, " USER FATAL MESSAGE MATGEN-4, ILLEGAL VALUE FOR PARAMETER 2 = -1" is issued for module MATGEN at line 143 in Version 68. Avoidance: Insert the following DMAP alter in SOL 106 to obtain the correct message, UFM 4402. compile nlstatic souin=mscsou noref nolist $alter 117 $ V68 alter 119 $ V68.2 IF ( NOLSET=-1 ) CALL ERRPH2 //SUBDMAP/4402 $ @ 6794; ks 69; Postprocessing - Heat Transfer, PARAM,POST,-1 - Missing Databloc Executing solutions 153, heat transfer, or 159, nonlinear transient heat transfer, with PARAM,POST,-1 does not produce the geometry data blocks GEOM1, GEOM2 etc. in the OUTPUT2 file. This error only occurs in MSC/NASTRAN Version 68.2. Avoidance: Apply the following alter: compile super3 noref nolist alter 217,217 alter 220,220 @ 6795; 69; Miscellaneous - SGI R8000 Only - Command Terminated Abnormally Jobs submitted with scr=yes on MSC/NASTRAN Version 68.2 on the SGI R8000 computer may in rare cases not write the "END OF JOB" to the f06 file but instead issue a "Command Terminated Abnormally" message in the log file. In the single case experienced the job otherwise completed processing and

the error was easily avoided by running with scr=no. @ 6797; ; DBC/TRANS/ACCESS - V6 & Later for IBM does not support READONLY Version 6 and later releases for MSC/ACCESS for IBM Mainframes running MVS/XA does not allow for current access of XDB databases via the READONLY request. As per the information from the NASTRAN Product Manager, there will be no more versions of MSC/NASTRAN for the IBM MVS platforms so this error will not be corrected. @ 6799; 69; Optimization/Design Sensitivity - DOM2 not available in V68 Prior to Version 68, design variable values could be passed into the DOM2 module and an EPT datablock based on the values and the property to design variable relations would be produced. This capability was removed from DOM2. The functionality has been re-instated in the DOM9 module in Version 69. @ 6800; 2001; Transient Analysis - DPD - Poor Message, UFM 2071 In solution 109, if an initial condition, ic, is specified without the presence of a DLOAD command in the case control, the following fatal message is produced: USER FATAL MESSAGE 2071 MODULE = DPD DYNAMIC LOAD SET 1 REFERENCES UNDEFINED TABLE 1

This message occurred despite the fact that no DLOAD was selected, and the TLOADi,1 entry in the bulk data was a TLOAD2, which can not reference a table. Avoidance: Input a DLOAD command in the case control which references a TLOADi entry which in turn references a DAREA entry with a zero coefficient. In general, all dynamic load input must be complete, even when it can not be selected. For example, if there is incomplete frequency response load data in a transient analysis input file, they will cause fatal errors in the DPD module, even though they can not be selected in that context. @ 6801; 70; Optimization/Design Sensitivity - DRESP1, SUPER3 - UFM 9058 In solution 200, if a DRESP1 entry is provided for an analysis type that is not present in a case control ANALYSIS command (e.g. RTYPE = DISP, but there is no STATICS of SAERO subcase) then "USER FATAL MESSAGE 9058 (SUPER3) THE SOLUTION FOR THE RESIDUAL STRUCTURE DOES NOT EXIST" may occur. Avoidance: Do not use unreferenced DRESP1 entries. @ 6802; 69; Optimization/Sensitivity - DOM9 - Quits No Message If an optimization problem has a near mechanism and param,bailout is set to force it through static decomposition without a fatal message, then the job may fail on DEC ALPHA systems in the DOM9 module with no message. Other machines will produce meaningless answers. Avoidance: There is no avoidance.

@ 6808; Equations ; Documentation - CBEND, MSC/NASTRAN Reference Manual - Incorrect

There are typographical errors in equation 7 of section 5.2.3 of the Reference Manual. R-sub-R should be R-sub-B and in the integral for Z, "R" in the expression R/R-sub-b should be "r". Avoidance: The correct equations may be found in the old MSC/NASTRAN User's Manual @ 6809; 70; Miscellaneous - XSORT - UFM 1126 (GNFIST) If the input data contains two "BEGIN BULK" entries, then the following misleading message may be produced: USER FATAL MESSAGE 1126 (GNFIST) DMAP MODULE XSORT IS ATTEMPTING TO OUTPUT DATABLOCK 201 NAME = IBULK, WHICH ALREADY EXISTS. USER ACTION: DELETE THE DATABLOCK, OR USE FILE IBULK =OVRWRT @ 6810; 69; Optimization/Design Sensitivity - DESGLB - UFM 6545 (DOPR3) The message "USER FATAL MESSAGE 6545 A DESGLB= XXX COMMAND APPEARS AT THE SUBCASE LEVEL" is issued in error when the DESGLB command appears above the subcase level and there is more than one "super" discipline in case control. There are three super disciplines in MSC/NASTRAN design sensitivity and optimization: Frequency (ANALYSIS=MFREQ or DFREQ), Transient (ANALYSIS=MTRAN) and everything else (ANALYSIS=STATICS,MODES,BUCK,FLUTTER or SAERO). Avoidance: Remove the DESGLB command and invoke the constraint with a DESSUB command from the first subcase. This will force the evaluation of the constraint at the subcase level even though there are no subcase associated responses involved in the constraint evaluation. @ 6812; 69; Eigenvalue Analysis - READ, EIGR - UWM 5402 In normal modes or buckling analyses which are defined by the EIGR Bulk Data entry, if the reduced stiffness matrix is less than 0.0001% dense, and the stiffness is not null, then UWM 5402 is issued and no eigenvalue analysis is performed. Avoidance: Use the EIGRL Bulk Data entry or insert the following DMAP alter: compile moders souin=mscsou noref nolist alter 100 $ Version 68 and 68.2 modtrl kxx///////1 $ set density to .0001 @ 6813; 70.5; Restarts - SOL 144, TRIM, Aero User's Guide - UFM 5423 The RESTART example ha144a_rst in Chapter 4 of the Aeroelastic Analysis User's Guide deletes two of the three subcases from ha144a.dat, the cold start run. This causes the RESTART job to fail when it tries to use the UOOX matrix from the cold start with the new GOAT matrix.

Avoidance: Add a dummy FORCE Bulk Data entry to the Bulk Data section and it need not be referenced by the LOAD command. In general lines should not be deleted from the Case Control, subcases and output requests should be added. @ 6814; ; p-Elements - GMCURV, GMSURF - Poor Answers When p-elements are connected to h-elements the polynomial order at the interface should be p=1. If GMCURV or GMSURF data is used along the interface surface the code ignores the different connections and allows the polynomial order to grow. Severe stress concentrations will be calculated at the connected points. Avoidance: Since the interfaces must be straight there is no need for geometry, remove the gmcurv or gmsurf from the interface. @ 6815; 68.2; Optimization/Design Sensitivity - PARAM,POST - Missing Results f or PATRAN If a SOL 200 analysis is run with PARAM,POST,-1, only the last design cycle results are available for postprocessing in MSC/PATRAN using the output2 file, and there is no any model information. Avoidance: The error has been fixed in MSC/NASTRAN Version 68.2, the user should switch to this version. @ 6818; 69; P-Elements - FEEDGE - UFM 2104 When point entries are used with the FEEDGE entry, the order is not supposed to matter; however, some runs may fail and issue the message "USER FATAL MESSAGE 2104, UNDEFINED COORDINATE SYSTEM". Avoidance: There is no simple avoidance. @ 6821; 69; Acoustic Analysis - ACMODL, ACMG - Trap Encountered, Traceback In acoustic/fluid analysis, if DIFF is used on the ACMODL entry and some fluid grids on the fluid/structure interface are not captured by the FSTOL tolarance, then the job may fail without any message being written to the f06 file. The log file may indicate a traceback as follows: Trap encountered, Traceback: Offset Offset Offset Offset 17dc 1f0 140 0 in in in in procedure procedure procedure procedure acmg_ xsem05_ mainal start

Avoidance: Set a value for FSTOL that captures all of the fluid grids at the fluid/structure interface. @ 6823; 2001; Optimization/Design Sensitivity - PCOMP, IFP - SFM 3003 The auxiliary model technique of shape optimization (solution 200) does not work when PCOMP entries are present. The following message is received:

SYSTEM FATAL MESSAGE 3003 MODULE = IFP ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK IEPTX (FILE 202) IN SUBROUTINE IFP6 Avoidance: The avoidance is to use one of the other three shape optimization methods. @ 6826; 69; Inertia Relief - PARAM,INREL/ SUPORT - UFM 6137, UFM 3005 If the inertia relief feature is used in SOL 101, and the model does not contain a mass matrix, then the job will fail with the following unclear messages: *** USER FATAL MESSAGE 6137 (UDSSDD) INPUT MATRIX IS RANK DEFICIENT, RANK = USER ACTION: CHECK MODEL *** USER FATAL MESSAGE 3005 MODULE = SOLVE 0

In SOL 24, the job will terminate without any messages. A symptom of this error is printing of the parameters after echo of the input file. Avoidance: Include the appropriate mass required for inertia relief. @ 6829; 69; Miscellaneous - Network Job Submittal - Job Not Submitted If the current working directory is auto-mounted, and there is a delay in starting execution of an MSC/NASTRAN or MSC/EMAS job (e.g., submitting job with aft= or que=), then there is a chance that the job will not be submitted. Depending on your network setup, you may receive the following message: /tmp_mnt/net/<node>/<somedir> : not found from the daemon. This error is encountered when it attempts to change the working directory to the current working directory as determined by the nast682 or emas32 script. The error can occur on any system where the "pwd" command does not return the canonical working directory. For example, % cd /net/<node>/<somedir> % pwd /tmp_mnt/net/<node>/<somedir> If sufficient time passes, the system will release the mount of '/tmp_mnt/net/<node>/home/<somedir>' and the subsequent reference to this directory will fail. When this happens, neither the F04, F06, nor the LOG file will be created since the job is never submitted. Note this error only occurs if you are submitting a job over the network and the current directory is auto-mounted. Avoidance: There are two possible avoidances. 1. Do not delay the submittal time for the job. 2. Leave the current directory window open without changing directory. Sadly, this may not work on SunOS.

@ 6831; 69; Installation - Remote Installation - Poor Message If the "Remote Installation" option of the script is chosen, AND the user's shell on the remote system is csh AND the remote system is not properly configured so that test(1) is in the PATH during an rsh command, the installation will fail indicating the CD or media could not be found on the remote system. Avoidance: The installation cannot use the specified remote system until the PATH is corrected. The script has been modified to explicitly test for this user error so that a meaningful message is output. @ 6833; ; Documentation - DMAP Module Dictionary - Examples Result in UFM The example in Section 2.1 of the DMAP Module Dictionary has several errors in it. The following lines of the input file cause fatal errors if used as is: QVOL GRID MAT1 *MAT1 CORD2R, +CORD 1.E+2 1 0 36O.O 30.0E3 16.99235344 3, 0, 0.,0., 1. 0. 1. 123 5 1 9 0.0 0 3456 1. 6.5-6 5.37+2 +MAT1 11.814744801 1.E04 0., 0., 0., 1., +CORD1 THRU 240.0

Avoidance: Change the above input to: QVOL GRID MAT1 +MAT1 *MAT1 16.99235344 11.814744801 1.E04 CORD2R, 3, 0, 0.,0., 0., 0., 0., 1., +CORD1 +CORD1 1. 0. 1. @ 6836; 68.2.3; Miscellaneous - OUTPUT2 - UFM 4130 (OUTPN2) The CRAY Y-MP Version of MSC/NASTRAN may fail on a CRAY J-90 in the OUTPUT2 module with the message "USER FATAL MESSAGE 4130 (OUTPN2) MODULE OUTPUT2 - ILLEGAL TAPE CODE HEADER = NASTRAN FORT TAPE ID CODE -". Avoidance: MSC has released a new version of MSC/NASTRAN specifically for the CRAY J-90 computer that not only corrects this error but runs approximately 30% faster than the Y-MP version. All clients with J-90s should receive this new version. @ 6839; 69; Dynamic Analysis - FREQ, TSTEP, SDAMPING - Missing Documentation , Messages By design intent, solutions 107 through 112 do not support changes of freq=, tstep=, method=, cmethod=, sdamping commands in subcases. For 123 5 1 1.E+2 0 360.0 30.0E3 1 240.0 THRU 0.0 1. 9 0 6.5-6 3456 5.37+2

example, if in sol 111 there are two subcases, and each has a different freq= command, only that in the first subcase will be honored. All similar commands in the second and subsequent subcases will be ignored. The output for all subcases will be for the frequency set selected in the first subcase. Note that solutions 108 and 111 do correctly support changes in the DLOAD command between subcases. Avoidance: There is no avoidance. @ 6849; 69; Aeroelastic Response - PARAM,DDRMM,-1 - Wrong Answers When a transient analysis is performed in SOL 146 with PARAM,DDRMM,-1 wrong answers are produced. Avoidance: Use PARAM,DDRMM,0 (i.e., the default), use SOL 76 or apply the following DMAP avoidance for Version 68.2: compile seaero nolist noref alter 'call mfreqrs' app1=app $ @ 6850; rwritten 69; Optimization/Design Sensitivity - PARAM,OPTEXIT,-4 - Output Ove

If "param,optexit,-4" is used in a design optimization run (SOL 200), then the OUTPUT2 results for table DSCMCOL will be overwritten by the matrix DSCM2. This error is due to assigning both the output2 and output4 files to the same unit. Avoidance: Insert the following dmap avoidance. compile exitopt souin=mscsou nolist noref $ alter 20,21 $ v68.02 output4 dscm2,,,,//-1/12 $ output4 ,,,,,//-2/12 $ @ 6853; 70.5; Frequency Response/Random Analysis - RANDOM - Infinite CPU Loop If a zero hertz excitation frequency is specified in a run which has a free-free model and is using modal frequency response to perform a random analysis, the run may encounter an infinite loop. This problem appears to be due to the zero frequency solution being a numeric infinity. If the same model is used in a direct frequency response, the run will fail during the decomposition due to the 0.0 hertz static solution being singular. In the modal solution, a matrix decomposition is not performed (in the uncoupled solution), therefore, a solution is obtained for the 0.0 hz excitation. Avoidance: Do not request a 0.0 excitation frequency in a modal frequency response analysis. @ 6854; MR ; Sequencer - PARAM,SUPER, SEQP - SFM 4276, EC 7777, Traceback SEQ If a superelement model uses PARAM,SUPER,2 without PARAM,MPCX,sid (where sid is the MPC set to be used), the run may fail with no messages in the sequencer.

Avoidance: Include PARAM,MPCX in the run if PARAM,SUPER,2 is used. @ 6855; 69; SSSALTERS - SEP1, lmt11.v682 - SFM 6357 In version 68.2, if the lagrange multiplier solution dependent sssalters lmt*.v682 are used, then the job may fail with the following error messages: SYSTEM FATAL MESSAGE 6357 - NO GRID POINTS DEFINED FOR SEP1 Avoidance: Use the solution-independent lagrange multiplier sssalter rflag.v682 instead of the lmt*.v682. The rflag.v682 alter combines the function of all the lmt*.v682 alters. @ 6859; 69; DMAP - MODTRL, PARTN - Infinte CPU Loop If an element of a matrix is given an index for a row number larger than the number of rows in the matrix, as defined by the matrix trailer, and partn operates on this matrix, partn will go into an endless loop. The MTRXIN and MODTRL modules, and perhaps inputt4, can generate such matrices when used injudiciously. Avoidance: One way to avoid this error is to first put the matrix through the ADD module. The module will throw away all terms beyond the boundaries of the matrix as defined by the matrix trailer. @ 6860; 69; Postprocessing - PARAM,POST,-2 PARAM,OUNIT2 - Missing EOF, Datal oader Abort In MSC/NASTRAN Version 68.2.3 if PARAM,CHECKOUT,YES and PARAM,POST,-2 are used the end of file is written to unit 11. If PARAM,OUNIT2 is used, the eof will go to the OUNIT2 file, but the data is written to unit 11. Avoidance: Please see error 5987. @ 6866; 69; Miscellaneous - Cray J90, T90 - UFM 4311 If you are running MSC/NASTRAN V68.2 or MSC/EMAS 3.2 on a Cray T90 or J90 which has more than 16 CPUs, then your job may fail with the following error messages: USER FATAL MESSAGE 4311 (XSEM) TIME EXPIRED PRIOR TO LINK NS01, CALLING MODULE - 19 IFPL Note that this error may occur even if you are not using more than one CPU as long as the T90 or J90 has more than 16 CPUs. The error is due to one of the timing kernels on the T90 and J90 returning an erroneous number. Avoidance: A fix for this kernel error is available from Cray Research. Even though this is not MSC's error, the error is marked as fixed in Version 69 to remove it from the list. @ 6867; 69; Plotting - PLOTPS/NEUTRL - No .ps file from .neu According to the documentation, the plotps utility should be able to generate a postscript formatted plot from the MSC/NASTRAN plot file (.plt) or a neutral (.neu) format plot file. (The neutral file format is

necessary for transferring data between different machine types). However, when the neutrl utility translates a plot (.plt) file into a neutral (.neu) format, the plotps utility cannot translate the neutral file and fails with the following message "USER WARNING MESSAGE (PLTFIL) No frames found in file." The plotps utility works fine for translating standard plot (.plt) files. Correction: The error exists in the neutrl utility, which is provided in source form. In subroutine RDBIN, change line 240 from: ibyte = 0 to ibyte = 1 Note: A neutral (.neu) plot file is unnecessary unless the user wishes to use one of the unsupported plot utilities given with a prior version of MSC/NASTRAN. The PostScript formatted file (.ps) is already a neutral format and is transferable across machine types. @ 6868; 69; Data Recovery - CBAR, SDR2 - SFM 4276, EC 7777, Command Terminat ed Abnormally Extremely small stresses for bar elements may lead to extremely large margin of safety output. These large numbers may cause MSC/NASTRAN to abort in the SDR2 module. This error has been observed on the DEC ALPHA OSF/1 and the HP/9000 systems. The DEC OSF/1 issues a "command terminated abnormally" message in the log file, while the HP/9000 issued a "SYSTEM FATAL MESSAGE 4276, ERROR CODE 7777" followed by a traceback in the log file. The traceback in the log file indicates that the job was in the "numcvt" routine at the time of failure. Avoidance: The error may be avoided by not setting PARAM,POST,0 for MSC/XL data recovery. The error may also be avoided by removing the stress limits on the MATi entries which results in the margins of safety being calculated. @ 6875; 69; Utilities - Incorrect Ownership Statement The MSC ownership statement in the utilities source (<ins-dir>/msc682/util/*) is incorrect. The correct ownership statement is: THIS PROGRAM IS CONFIDENTIAL AND A TRADE SECRET OF THE MACNEALSCHWENDLER CORPORATION. THE RECEIPT OR POSSESSION OF THIS PROGRAM DOES NOT CONVEY ANY RIGHTS TO DISCLOSE ITS CONTENTS, SELL, LEASE, OR OTHERWISE TRANSFER IT TO ANY THIRD PARTY, IN WHOLE OR IN PART, WITHOUT THE SPECIFIC WRITTEN CONSENT OF MSC. This statement permits the user to modify these routines for use with MSC/NASTRAN and MSC/EMAS. @ 6876; 69; Matrix Methods - ADD5 - Inefficient Processing NO EXTERNAL ERROR REPORT NECESSARY If the scaling factor (for example, alpha) for any matrix in the ADD5 statement is zero and there are additional matrices input to the module,

the module will scale each term of the associated matrix by 0.0 and add it to the additional matrices. Avoidance: No avoidance is necessary as this error is simply an efficiency issue and wrong answers are not produced. DMAP may be used to check if any scaling factors are 0.0 and the associated matrices may be excluded from the addition. @ 6877; 69; Composites - Stress and Failure Indices - Mislabeled and/or Extr aneous Output In SOL 106, composite stress and failure index output is not labeled with the load step as is found in other nonlinear output. Thus, PATRAN is not able to distinguish composite output at different load steps. In addition, in any solution that supports composite output, even if no failure index output is requested, the header labels for failure index output are always printed. Avoidance: There is no avoidance. @ 6879; 70; Acoustic Analysis - DISP, ACCE - Extra, Incorrect Output. If a fluid structure analysis is performed and accelerations are requested for structural points and displacements are requested for the fluid points, additional (incorrect) output will occur beyond the requested fluid point output. Various machines may behave slightly differently. Some machines may fail with the message "SYSTEM FATAL MESSAGE 4276, ERROR CODE 7777" indicating a numeric overflow, other machines may output "NaNQ" or "**********" indicating a numeric overflow. Avoidance: Do not request structural accelerations in the same run requesting fluid displacements. Perform a restart and keep the requests separate. @ 6880; 7 Nonlinear analysis with QUAD4 elements and PARAM,LGDISP or buckling analysis may fail in MSC/NASTRAN Version 68 on long word machines such as CRAY and NEC. The run may issue SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777 indicating a numeric overflow. Other messages may be written to the log file. Avoidance: The error can be avoided by setting SYSTEM(170)=1 which select the old quad4 differential stiffness formulation. @ 6881; 70.5; Miscellaneous - TA1 - UFM 2010, Missing Properties If a model has a large number of property entries and insufficient memory is allocated, the program may issue USER FATAL MESSAGE 2010 stating that properties which are in the input file do not exist. This occurs in DMAP module TA1. Avoidance: Increase the amount of memory for the run or increase the BUFFSIZE on the NASTRAN statement. @ 6882; 69; Miscellaneous - SFM 4276, EC 923, Insufficient Disk Space, Exces 69; Differential Stiffness - QUAD4, PARAM,LGDISP - SFM 4276, EC 777

sive Output NO EXTERNAL ERROR REPORT UNTIL I GET A DECK AND A BETTER DESCRIPTION AS TO WHEN THIS HAPPENS. If the files used by an MSC/NASTRAN run exceed the available space on a disk, the program will provide SFM 4276 EC 923, indicating that system resources were insufficient. However, the message is followed by a "core-dump" which will be several thousand lines long. This additional output has no meaning to the user. @ 6883; 69; Elements - HEXA(8), TEMP(LOAD), PSOLID - Wrong Answers If an eight node HEXA uses full 27 point integration (IN=3 on the PSOLID entry), incorrect thermal loads are computed resulting in wrong answers. Avoidance: The avoidance is to use the default integration method. @ 6887; 69; Frequency Response Analysis - PARALLEL - Wrong Answers MSC/NASTRAN Version 68 may produce wrong answers for frequency response analysis whenever PARALLEL=n is selected, where n is greater than 0. Avoidance: The error only affects parallel sparse complex decomposition. Parallel may be selected for other operations while being deselected for this operation by: instead instead instead instead @ 6888; lts 69; SSSALTER - mfreqa.v68 and mfreqa.v682 - Error in Scaling of Resu In the sssalter mfreqea.v68 and mfreqea.v682 the scaled results of the modal contributions are incorrectly scaled. Only as many modal contributions as there are excitation frequencies are scaled and the others are not printed. Resolution/Avoidance: Change the following line : mpyad onec,totalt,/scalit $ to be mpyad oner,totalt,/scalit $ In the alter for V68.2, also change the following line: mpyad kinetc,onec,/totalk/1 $ to be: mpyad kinetc,oner,/totalk/1 $ of of of of parallel=1, parallel=2, parallel=3, parallel=4, the the the the workaround workaround workaround workaround is is is is parallel=262145 parallel=262146 parallel=262147 parallel=262148

@ 6893; Traceback ; Complex Eigenvalues, Matrix Methods - CEAD, UDCOMP - No Message,

If sparse unsymmetric decomp spills the run may go into an infinite cpu loop or fail with the following traceback being issued to the log file but no other message: Traceback: Offset Offset Offset Offset Offset Offset Offset Offset Offset Offset Offset 0x00000010 0x00000010 0x00000164 0x000031dc 0x00000440 0x000000e8 0x00000120 0x00000e88 0x00000c24 0x0000006c 0x00000994 in in in in in in in in in in in procedure procedure procedure procedure procedure procedure procedure procedure procedure procedure procedure txuycd_ udspcd_ udsncd_ udsfa_ udcomp_ factru_ classd_ clasdd_ cead_ xsem11_ mainal

Avoidance: Increase memory to avoid spill. In Version 70.7 the pivot threshold has been changed from 1.e-6 to 1.e-10 to make this error less likely. Furthermore, it is now possilble to override the pivot threshold with system cell 91. The threshold is set to 10^(sys91) @ 6894; 70; Miscellaneous - No Output - Intel Windows NT 3.1 During testing of MSC/NASTRAN Version 68.2 some jobs produced no output on the Windows NT 3.51 system. The error could not be debugged because when the command prompt window was closed and re-opened the error could not be reproduced. Thus, this report is being written to assist users who possibly encounter the same problem. @ 6902; 69; SSSALTER - Possible error using alter2g.v682 If the DMAP alters for coupled analysis are used with the OUTPUT2 (MATPOOL) option and a component has no damping matrix, the component matrices may not be stored properly on the database. Avoidance: In the alter, change NOM2GG to NOB2GG in the following DMAP statement mtrxin ,,baap,EQEXINS,SILS,/baaext,,/LUSETS/ S,N,NOM2GG///1 $ @ 6903; 69; p-Elements - FEEDGE, GMLOAD, CORDi - Wrong Answers If a GMLOAD is applied to an FEEDGE or GMCURV which have grids with displacement coordinate systems other than the Basic coordinate system, the loads are not properly applied. Avoidance: Set the grids' displacement coordinate system to the Basic coordinate system. @ 6904; 69; Nonlinear Analysis - MATS1, QUAD4, TRIA3 - Wrong Answers

There are two errors in element force output for nonlinear QUAD4 and TRIA3 elements. The first error results in over-estimated force components. This error exist in Version 68 systems only. This error occurs when geometric nonlinear (LGDISP=1) is turned on and the element is in the nonlinear material portion of the stress-strain curve (MATS1). Forces may be over-estimated by more than 40%. The second error results in under-estimated force components, on the order of 4%, when param,lgdisp is the default, -1. This day one error is directly linked to the default value of param,nlayers,4. Avoidance: There is no avoidance for the over-estimated element forces when param,lgdisp is 1. The under-estimated element force error may be reduced to .7% by setting param,nlayers to 12. @ 6905; 69; Miscellaneous - NOTIFY - Command Garbled A syntax error in the "handlenotify" shell function in the MSC/NASTRAN Version 68.2.3 and EMAS Version 3.2.3 script produces a sed error: sed: command garbled: Avoidance: The problem is in the line that reads: ncmd=`echo "$ncmd" sed "s^{}^\'$msg\'"`

Replace this line with: ncmd=`echo "$ncmd" @ 6906; 69; Documentation - HP-UX Installation Guide - Illegal Option The documented example in the "Installation and Operation Guide" for the ncmd command line option is incorrect for HP-UX computers. The example should read: ncmd=echo {} @ 6910; 70.5; Documentation - OUTPUT4, INPUTT4 - Illegal Unit Number for INPUT T4 -11 If the sparse option is used to write a matrix using OUTPUT4 (IUNIT<0), then the program will fail when reading the matrix using INPUTT4 if the sparse input option is used (IUNIT<0). Avoidance: Do not use the sparse option when reading or outputting a matrix with inputt4 or output4. Use a positive value for IUNIT. . @ 6913; 70.5.2; Superelements - Can Not Perform Dynamic Data Recovery for Extern al S.E. Data recovery is not supported for external superelements in dynamic solution sequences. Avoidance: There is no avoidance. Version 70.5.2 has been enhanced to mailx -s {} `whoami` sed "s^{}^\'$msg\'^g"`

provide this support, please see PARAM,EXTDR. @ 6914; 68.2.3; Installation - LANG - awk: UNIX command not found MSC/NASTRAN Version 68.2 may get the following message, "awk: UNIX command not found", or some other UNIX level fatal message on SGI machines outside the US. Non US language AIX machines are unable to understand the 'mem' command line parameter. Other machines may exhibit different problems. The problems are all caused by the fact that the LANG environmental variable is set to null in the Version 68.2 NASTRAN script and not explicitly to 'C'. Avoidance: All non US Language machines should change the MSC/NASTRAN script to specify C explicitly: # Set the "C" locale LANG="C"; export LANG This error is corrected in the Version 68.2.3 R2 CD. @ 6916; 69; Acoustic Analysis - HEXAPR - Incorrect Phase Output The "COMPLEX ACCELERATION VELOCITY AND PRESSURE LEVEL" output is incorrect for models that include damping. "REAL/IMAGINARY" output will always indicate a zero value for the imaginary part of the acceleration and pressure output and the real part of the velocity output. The nonzero part indicates the magnitude. Likewise "MAGNITUDE/PHASE" output gives a correct magnitude but an incorrect phase angle. An example of the incorrect output is shown below. C O M P L E X A C C E L E R A T I O N S P R E S S U R E L E V E L S V E L O C I T I E S (REAL/IMAGINARY) EL-TYPE X-ACCELERATION Y-ACCELERATION Z-ACCELERATION Y-VELOCITY Z-VELOCITY PRES(DB) HEXPR .0 -1.215000E-09 3.504862E+03 .0 3.E+01 .0 .0 -2.789081E+02 1.004350E-10 Avoidance: There is no avoidance. @ 6921; ce 69; Matrix Methods - MPYAD - Negative Time Estimates, Poor Performan It is possible to get a negative time estimate for MPYAD. This error may lead to poor method selection and thus poor performance. Avoidance: There is no avoidance. @ 6928; 69; Optimization/Design Sensitivity - DESOBJ - Wrong Answers In Versions 68, weight and volume cannot both be selected as design responses in the same solution 200 -1.262103E-09 .0 X-VELOCITY .0 9.668665E-11 A N D

job. The problem will run to completion but one of these responses will have zero value (i.e., it will be ignored in the design process). This leads to incorrect answers. The avoidance is to either: 1.) Use only weight or only volume in the design model 2.) Include the following DMAP alter (version 68, 68.2): compile desinit souin=mscsou nolist noref alter 66 wgtvol = 3 $ endalter @ 6932; M 3002 2001; Optimization/Design Sensitivity - PACABS, ELTPRT - UFM 5165, SF If the B field of the PACABS (acoustic absorber element) data entry is specified as a design variable in an optimization analysis, then the execution will result in the following fatal message: USER FATAL MESSAGE 5165 (DSVG1D-1) UNDEFINED INPUT FILE = 106, NAME = (NONE) , MODULE DSVG1 IS TERMINATED. Avoidance: Use the following alter for 68.2.3: compile pslgdv list noref souin=mscsou alter 239, 239 IF ( DODFREQ>0 OR DOMFREQ>0 OR DOMTRAN>0 ) EMG, EST,CSTMS,MPTS,DIT,,,,,,,GPSNTS,EQEXINS,SILS/ kelmx,kdictx,,,BELM,BDICT//0/s,n,nobgg $ In addition to the above error, optimization of the PACABS properties with the objective to minimize weight fails at module ELTPRT in RESPSEN with the following error message: SYSTEM FATAL MESSAGE 3002 MODULE = ELTPRT 0EOF ENCOUNTERED WHILE READING DATA BLOCK ESTDV2F (FILE 105) IN SUBROUTINE SKPREC Avoidance: Minimize a DRESP2 which references the PACABS properties instead of minimizing weight. @ 6933; 69; Optimization/Design Sensitivity - PMIN - Wrong Answers If the user specifies a value of PMIN that is less than -0.99e35, the value could potentially be incorrectly reset to 1.0e-20. This will happen for I12 on a PBAR or PBEAM and for orientation on the PCOMP entry. Avoidance: Specify PMIN values that are greater than -.99e35 @ 6934; 69; Optimization/Design Sensitivity - DVPREL2 - Incorrect UFM 7618 When USER FATAL MESSAGE 7168 is encountered, the DVPREL2 ID referenced is in error. USER FATAL MESSAGE 7168 (DOPR1Z) EQID XXX REFERENCED ON DVPREL2 YYY IS NOT PRESENT ON A DEQATN ENTRY. ERROR CODE = -6,

Avoidance: A nonexistent EQID has been referenced on a DVPREL2 entry. The burden is on the user to determine which DVPREL2 makes the incorrect reference. @ 6936; 69; Documentation - umsec6.doc - File Does Not Exist The "readme" file in the nast68/doc directory states that the directory contains the file UMSEC6.DOC which includes a list of MSC/NASTRAN diagnostic messages. This file is no longer included with the delivery. The MSC/NASTRAN online encyclopedia, delivered with all Version 68 systems, contains this information (and much more) in an easily accessable format. Thus the online encyclopedia is a far superior method for accessing information. The readme file will be updated for Version 69. @ 6937; 69; DMAP - COMPILE - UFM Multiple COMPILE statements for the same subdmap may result in a user fatal message. For example the following DMAPs should be equivalent but the first may lead to a UFM: COMPILE A ALTER 1 COMPILE B ALTER 2 COMPILE A ALTER 3 COMPILE A ALTER 1 ALTER 3 COMPILE B ALTER 2

Avoidance: Combine the alter so that a single COMPILE statement may be used for the subdmap. Re-arranging the COMPILE statements such that the COMPILE statements for each subdmap follow each other directly should also avoid the problem. @ 6939; 69; Nonlinear Analysis - PARAM,NMLOOP, DBSTATUS - SFM 4276 EC 7777 Nonlinear analysis jobs with PARAM,NMLOOP may fail on the DEC ALPHA OSF/1 Version of MSC/NASTRAN. Avoidance: Insert following DMAP alter: compile nlstatic noref nolist $alter '/s,n,fact','' $ V68.2 alter 486,486 PARAML ESTNL//'DTI'/0/6/S,N,FACT $ LOAD FACTOR @ 6941; 69; SSSALTER - MATMOD, ZFREQ - SFM BLDPACK ERROR The Zfreq alters in the sssalter library require the presence of DMI matrices for both frequency-dependent stiffness and damping if the frequency-dependence part of the alter is to take effect. If a zero term is input for any of these effects a BLDPACK error occurs in MATMOD option 28, when attempting to make a diagonal matrix from a vector. Avoidance: @ 6943; 69; Optimization/Design Sensitivity - SHPCAS, TEMP(INIT) - UIM 9052, SFM 3001 Put in a very small number instead of a 0.0 value.

Using TEMPERATURE(INITIAL) in a solution 200 shape optimization run with an auxiliary model results in the messages: USER INFORMATION MESSAGE 9052 (FEA) - STATIC ANALYSIS COMPLETED. DESIGN CYCLE NUMBER= 1 SYSTEM FATAL MESSAGE 3001 MODULE = SHPCAS THE INPUT DATA BLOCK NAMED YGBNDR AT POSITION 2 DOES NOT EXIST. Avoidance: alter: Specify TREF on the MAT1 entry or apply the following DMAP

compile bndshp nolist noref souin=mscsou alter 'auxmid'(2) dbview ugx=ug (where mtemp=*) $ alter 'partn','partn' PARTN UGx,,AXMVEC/UGA,,,/1 $ alter 'equivx','equivx' EQUIVX UGx/UGA/ALWAYS $

@ 6947; 69; Miscellaneous - INPUTT4 - SFM 4276 EC 7777 When reading a double-precision datablock on an SGI R8000 computer using INPUTT4, the run will fatal in the INPUTT4 module with the following message in the .f06 file: RECEIVED SIGNAL 0, ABORTING ... BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR FILE INDEX (FILEX) = 0 SYSTEM FATAL ERROR 4276.RECOVR ERROR CODE 7777

Avoidance: An avoidance is to have the datablock output in the neutral format (OUTPUT4 in the ASCII format) and edit the datablock header to change the data type (the 4th word in record 1) from 2 (double precision) to 1 (single precision). @ 6949; 69; Documentation - RS/6000 IOI DBset ASSIGN Description format is i ncomplete. In Section 4.3 of the MSC/NASTRAN Installation And Operating Instructions, the description of the DBset ASSIGN statement is incomplete. The reference to the "action" options is not specified. Please see the MSC/NASTRAN Quick Reference Guide for a complete description of the ASSIGN statement. @ 6951; 70.5; Nonlinear Analysis - P2G - Wrong Answers, Null Load If P2G is used to import loads onto the residual structure (with or without superelements) in SOL 106, their contribution to the overall loading is ignored. Avoidance: Insert the following DMAP alter: compile nlstatic alter 'ssg1 ' MTRXIN CASESx2,MATPOOLS,EQEXINS,,/

P2J,,/LUSETS/S,N,NOP2J///2 $ ADD Pgres,P2J/Pgres1/CP1/CP2 $ equivx pgres1/pgres/-1 $ endalter or use a LOADSET callout above the SUBCASE level and a CLOAD within each residual subcase to include the P2G contribution as shown: LOADSET = 10 P2G = ExtLds SUBCASE 1 LOAD = 20 CLOAD = 30 BEGIN BULK LSEQ, 10,9999,20 CLOAD,30,1.,1.,9999

@ 6952; ; P-Elements/Superelements - GP4 - UFM 2101, SFM 3062 The output coordinate specification is not allowed for the FEEDGE or FEFACE entries when an FEEDGE or FEFACE is specified on the superelement boundry. Do not use global coordinates on superelement boundries in p-element substructures else the following fatal message will be encountered: *** USER FATAL MESSAGE 2101 (GP4) GRID POINT 101000048 COMPONENT 1 ILLEGALLY DEFINED IN SETS UB UM

@ 6953; 7 The following message is misleading in that it should be a USER FATAL MESSAGE and it does not indicate that the error is on a DRESP1 entry. SYSTEM FATAL MESSAGE 7127 (DOPR3B) AN ELEMENT WHICH REFERENCES PTYPE 52 20 PID ANALYSIS MODEL. @ 6956; 69; Nonlinear Analysis - NOLINi - Wrong Answers NOLINi loads produce wrong answers in nonlinear transient analysis if the loads are a function of velocity and the TSTEPNL entry uses methods "AUTO" or "TSTEP". Avoidance: Use the ADAPT option on the TSTEPNL entry. @ 6957; 69; Documentation - NOLIN2 Remark 4 under the MSC/NASTRAN Version 68 Quick Reference Guide description of the NOLIN2 entry should state "GK-CK" not "CK-CK". 34 IS NOT DEFINED IN THE 69; Optimization/Design Sensitivity - DRESP1 - Poor Message, SFM 712

@ 6960; 69; Optimization/Design Sensitivity - SUBSEQ, SDR2 - SFM 4276, EC 30 01, SWM 2092 If zero values are used for the SUBSEQ command for a SUBCCOM subcase in a solution 200 analysis the following messages will be encountered from the SDR2 Module: SYSTEM WARNING MESSAGE 2092 MODULE = SDR2 SDR2 FINDS A SYMMETRY SEQUENCE LENGTH = X AND AN INSUFFICIENT NUMBER OF VECTORS AVAILABLE = Y WHILE ATTEMPTING TO COMPUTE STRESSES AND FORCES. ALL FURTHER STRESS AND FORCE COMPUTATION TERMINATED. SYSTEM FATAL ERROR 4276.GREAD ERROR CODE 3001 PID= 0

Avoidance: The avoidance is to not use 0.0 coefficients in a SUBSEQ request when performing design sensitivity and optimization. @ 6963; 69; Documentation - SUBCOM, SUBSEQ - Unexpected Answers The documentation for the SUBSEQ command is misleading. Unless it is read carefully, the user might expect the first number to apply to the first SUBCASE and so on. Actually, the way it is applied is that the last number is applied to the most recent SUBCASE, the next to last number is applied to the SUBCASE preceding it, and so on. If scaling factors are not included for all preceding SUBCASEs, unexpected results may result. Avoidance: Include scaling factors for all preceding SUBCASEs. @ 6964; 69; Utilities - Intel Windows/NT Only If the utilities are executed on the Intel Windows/NT Version 68.2 release of MSC/NASTRAN without any argument the command line input format is not output to the console. This command line description is instead output the FOR000 file. The command line description is also documented in section 6 of Installation and Operation Instructions. @ 6965; 69; Miscellaneous - Include File Not Found MSC/NASTRAN on Windows NT can be run one of two ways. One is via command line and second is by clicking on an icon. If nastran is run via the second method and the input data deck contains an include file, the include file will not be found. Avoidance: 1. Run nastran using the command line. 2. Define the environmental variable in the MSCN4NT1.INI file specifying the location of the data deck. This also requires the same environmental symbol be included while specifying the include file. This method is described in the section 4.1 of the Installation and Operation instructions. 3. Add a complete path name to the include file. @ 6966; ers 69; Nonlinear Transient Heat Transfer - HBDY, FLUX, SET - Wrong Answ

If a set definition is used for a FLUX request, then wrong (zero) values in the radiation column of the heat flow into HBDY elements may be produced. This error causes all results in the radiation column to be incorrect. This error does not affect the temperature results. Avoidance: A FLUX=all request avoids the problem. @ 6968; 69; Miscellaneous - Bad File Names - Run Never Starts When using MSC/NASTRAN in either Windows95 or Windows/NT, the program will fail if any embedded blanks exist in either the filenames or directory names used. Avoidance: Do not include any embedded blanks in file or directory names. @ 6969; n 4.0 69; Installation - MSC/NASTRAN Will Not Install on Windows NT Versio For Windows NT Version 4.0, which is the Windows 95 interface on top of Windows NT 3.5, the installation of MSC/NASTRAN will not work. There is a problem with the installation software (Installshield). Avoidance: @ 6971; 69; Loads - RFORCE - Wrong Answers On one client model, RFORCE entries generated null loading vectors in MSC/NASTRAN Version 68.2. Correct answers were produced in MSC/NASTRAN Version 68.0 and V68.1. This error results in incorrect results. Avoidance: We have been unable to duplicate this problem on any other input files. For the client input deck, setting METHOD to 2 on the RFORCE entry resulted in the correct loads being generated. Our recommendation is whenever you have an RFORCE in your input file, include a SUBCASE which applies the RFORCE only and verify the OLOAD resultant is correct for that SUBCASE. If the OLOAD resultant for that SUBCASE is null set method to two on the RFORCE entry and please send the deck to MSC. @ 6972; 69; Installation - MSCID - UFM 3060 The MSC Authorization program (an icon) has, on some configurations (usually laptops with docking stations), shown a "0" for "MSC id". This is not correct. If you get a "0" while using the authorization program during installation or running from the icon, do not use this number to get an authorization code from MSC. To get the proper ID, run a MSC/NASTRAN job and the 3060 error message in the .f06 file will contain the correct ID. Use the ID in the .f06 file when calling MSC for an authorization code. @ 6976; 69; Acoustic Analysis - EMA, PSOLID - Poor Message When executing a fluid/structure analysis, if the FCTN field is not set to FLUID on the PSOLID data entry for a fluid element the execution will fail with the message: SYSTEM FATAL ERROR 4276.BLDPK ERROR CODE 211 Do not upgrade your operating system to 4.0.

Avoidance: Other than correcting the model, there is no avoidance. @ 6980; 69; Heat Transfer Analysis - QVECT - Wrong Answers for HBDY Output The QVECT load defines thermal flux from a distant source into a face of a CHBDYi boundary element. If a coordinate system(CE) for the thermal vector flux is rotated with respect to the basic coordinate system, the data recovery for the APPLIED LOAD under the HEAT FLOW INTO HBDY ELEMENTS is not correct. However, the actual temperature results are correct. The actual loading can be seen using the OLOAD=all output. There is no error if the coordinate system is merely translated, not rotated, with respect to the basic coordinate system. Avoidance: Ignore the incorrect output. @ 6981; 69; Nonlinear Transient Heat Transfer - RESTART, PARAM,SLOOPID - Wro ng Answers If PARAM,SLOOPID is specified for a nonlinear transient heat transfer (SOL 159) analysis restarting from a steady state thermal analysis (SOL 153), the initial temperature from the SOL 153 run is zero. Avoidance: Use the following DMAP ALTER.(V68.2 ALTER) $ DMAP alter to repair SLOOPID restart from sol 153 $ MSC/NASTRAN V68.2 $ compile nltcsh,souin=mscsou,nolist,noref $ $ bypass the restart deletion of estnl in case that a $ TEMPBC entry is added to the bulk with the sol 159 input $ alter 39 $ type db zuzr01 $ dbview xestnl = estnl (where nloop=sloopid) $ copy xestnl/zuzr01 $ alter 281,281 $ alter 300,300 $ copy zuzr01/estnl $ $ $ save the sol 153 temperatures to database $ alter 310 $ V68.2 equivx ulx/ulnt/noeset $ cend @ 6982; ge ; Heat Transfer Analysis - PARAM,LGDISP - Solution Does Not Conver PARAM,LGDISP is intended only for the nonlinear structural analysis. However, if it is used in heat transfer analysis (SOL 153 or 159) the solution will not converge. Avoidance: Delete the PARAM,LGDISP from any heat transfer analysis. @ 6984; 69; DMAP - MALTER - UFM 806

If the input file for a restart run contains a DMAP alter using MALTER commands and has no "compile" commands, the run will fail with User Fatal Message 806 stating that the MALTER statements were not found in the source files. Avoidance: This error may be avoided by either: 1) Inserting ACQUIRE NDDL in the FMS section of the restart deck. or 2) Inserting a COMPILE command for a subdmap in the deck. @ 6986; 69; Dynamics - DPD, SEQP - Poor Efficiency on Convex Models which have been resequenced cause the DPD module to run up to 5 times slower than models which have not been resequenced on MSC/NASTRAN Version 68.2 on the Convex. This also affects reading large numbers of FORCE bulk data entries. Avoidance: The combination of the sparse solver and the seqp module is almost always more than the sparse solver without resequencing. Thus, sequencing should be simply turned off. Remove resequencing by: 1. Remove any PARAM, NEWSEQ entry and replace it with PARAM,NEWSEQ,-1. 2. Remove any SEQGP entries which may be in the bulk data file. This avoidance will not help the reading of the FORCE bulk data entries. @ 6990; 69; Acoustic Analysis - PARAM,G - Wrong Answers If the structural damping (either PARAM,G or "ge" damping on the MATi entry) is the only change in a restart to a fluid-structure acoustic analysis, the change is ignored and the answers are incorrect, i.e., they remain the same as before the change. Avoidance: Execute the analysis as a cold start. @ 6991; 69; Heat Transfer - CHBDYG, CTRIAX6 - Wrong Answers Currently there is a design limitation using CHBDYG,REV to represent the higher order CTRIAX6 element because only two grid points are allowed on the REV option. Therefore, the weighting on the mid-side node is misrepresented if the two node REV option is used. Avoidance: Use CHBDYE to represent the CTRIAX6 element with midside nodes. Please note that without the midside node, the CHBDYG,REV can be used. Please note that the REV option G1,G2,G3 Where G1 and G2 are the edge grids and the G3 is the midpoint grid @ 6992; VECTOR 69; Heat Transfer Analysis - OLOAD, CTRIAX6 - Wrong Answers for LOAD

Applying a heat flux to the edge of the midside node of a CTRIAX6 element results in wrong answers for the LOAD VECTOR output at the midpoint. This error is only in the data recovery. The temperature results are not affected. Avoidance: Ignore the OLOAD output for the midside node of the CTRIAX6 element

@ 6994; 2001; Miscellaneous - ASET1 - UFM 2049 If an ASET1 entry references a non-existing grid as an end point for the "thru" option, then the following fatal message is issued: "USER FATAL MESSAGE 2049, UNDEFINED GRID POINT ** HAS A COORDINATE REFERENCED ON A ASET CARD." Besides the fact that this message should not be encountered at all, this message is poor for the following reasons: (1) The missing grid point specified in the message will always be the first grid point in the "thru" range, not the actual missing grid point. (2) The word "coordinate" is incompatible with other documentation which uses the word "component". (3) It uses the obsolete word "CARD" (4) It repeats the message twice. (5) If the user explicitly specified the grids on the ASET1 entry rather than using the "thru" option, they would have gotten User Warning Message 2049A instead of a fatal message. Avoidance: Don't specify "thru" ranges on the ASET1 entry that have a missing end point. @ 6996; ; Optimization/Design Sensitivity - Q4TAPER - UWM 5491, UFM 4297 If an auxiliary model is used to generate the shape basis vector for a shape optimization analysis and in the subsequent dblocate optimization run an element exceeds the maximum allowable taper value of .5 then EMG issues UWM 5491 that may identify elements with bad geometry that do not exist in the model. The run may also issue USER FATAL MESSAGE 4297 indicating that an element has an interior angle greater than 180 degrees. An example of the warning message follows: USER WARNING MESSAGE 5491 (EQD4D) BAD GEOMETRY FOR QUAD4 ELEMENT ID = 2600 THE TAPER OF 0.582639 EXCEEDS THE MAXIMUM VALUE OF 0.500000 USER ACTION: THE TOLERANCE MAY BE ADJUSTED WITH Q4TAPER KEYWORD ON THE NASTRAN STATEMENT. Avoidance: EMG is reporting the internal identification numbers for the perturbed elements. These element id's are incremented based on the base line finite elements. The model should be modified to a finer mesh or the Q4TAPER keyword should be set on the NASTRAN statement to a larger value. @ 6997; er 70.5; Acoustic Analysis - Superelements, Reference Manual - Wrong answ The Version 68 MSC/NASTRAN Reference Manual Section 15.4.2 states that the "grid points at the fluid-structure interface may be assigned to the residual structure only". This statement is not restrictive enough. The MSC/NASTRAN Common Questions and Answers Guide states the restriction correctly; "for coupling to occur, all fluid elements that are in contact with the structural elements must be in the residual structure, and all structural elements that are in contact with the fluid elements must also belong to the residual structure. @ 6998; 69; Heat Transfer Analysis - QBDY2 - Wrong Answers

QBDY2 (Boundary Heat Flux Loading) loading on CQUAD8, CTRIA6, CHEXA20, CPENTA15, CTETRA10 elements results in incorrect data recovery for the element. The temperature vectors are correct. The heat flow into the HBDY elements is wrong. Avoidance: Ignore the Heat Flow into HBDY element for the Applied Heat load. All other output is unaffected by this error. If the user wants to see the correct applied load under the Heat Flow into the HBDY element, they must use the QBDY3 bulk data entry to apply the loading. @ 7001; 70.5; Nonlinear Transient Heat Transfer - QRG, PARAM,AUTOSPC - UFM, Fa iled to Converge PARAM,AUTOSPC,YES will automatically constrain floating degrees of freedom. The floating grid is used in heat transfer to model ambient temperature as a function of time .Heat transfer analysis uses these floating grids to apply temperature as a function of time through the TEMPBC,TRAN. Use of this PARAM,AUTOSPC,YES in nonlinear heat transfer results in a run that fails to converge. This limitation is documented in the Version 68 Release notes but needs to be documented in the Quick Reference Guide. @ 7002; ; Sequencer - GP2 - UFM 2007 If an attempt is made to use the punched grid and element output of a p-element run in the bulk data the following error will occur and the run will fail in the SEQP module or GP2 modules: *** USER FATAL MESSAGE 2007 MODULE = GP2 ELEMENT ******** REFERENCES UNDEFINED GRID POINT 111001009 USER FATAL MESSAGE 2007, ELEMENT******** REFERENCES UNDEFINED GRID POINT********. This message occurs since grid ids larger than 100000000 are not valid on the grid card and the vu-grids have ids greater than 100000000. . Avoidance: Following system cells should be set in a p-element run to generate smaller vu-grid ids and vu-element ids. System cell values are beginning and maximum for vu-grids and vu-elements nastran system(175)=9000, system(178)=100, system(179)=150000, system(180)=150001, system(181)=2000000, system(182)=150, system(183)=2000000 @ 7006; entation 69; Documentation - DRESP1, DVPREL2 - Incorrect RTYPE, Missing Docum $ $ $ $ $ $ $ GRMAX GNSTRT GNMAX VGSTRT VGMAX VESTRT VEMAX = = = = = = = 100000000 101000000 200000000 201001001 300000000 1001001 2000000

The Version 68 MSC/NASTRAN Quick Reference Guide description of the DRESP1

entry indicates the RTYPE for SPCFORCE as FRSPC and TSPC, it should be FRSPCF and TSPCF. Remark 3 under the description for the DVPREL1 entry for PTYPE="PBEND" should also be included for the DVPREL2 entry. @ 7007; 69; Installation - NASTRAN FOR WINDOWS - FILE TRANSFER ERROR NO. -3 If long file names have been disabled by the following registry entry the NASTRAN FOR WINDOWS VERSION 68.2 Installation will fail with the message "FILE TRANSFER ERROR NO. -3". HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\FileSystem\Win31FileS stem Avoidance: Reset the registry entry and restart your computer. @ 7010; 2001; Axisymmetric Fluid Elements - CAXIF4, VECPLOT - UFM 5423 Models with CAXIF4 axisymmetric fluid elements fail in the VECPLOT module with the message: USER FATAL MESSAGE 5423 - ATTEMPT TO MULTIPLY INCOMPATIBLE MATRICES. Avoidance: Apply the following alter: compile selg nolist noref alter 63,63 endalter @ 7011; 69; DMAP - VECPLOT - Excessive CPU The VECPLOT module can consume a significant amount of CPU when processing RESULTANTS and MAXIMUMS. This information may be of no interest to the analyst yet the only way to turn it off is thru DMAP. MSC/NASTRAN Version 69 has been enhanced to give the analyst options to avoid these calculations. PARAM,PRTMAXIM (default=no) and PARAM,PRTRESLT (default=yes) have been added to control the calculation of MAXIMUMS and RESULTANTS respectively. @ 7012; 70.5; Executive - DIAG 14,64 - UFM 69 The presence of both DIAG 14 and DIAG 64 in the same run will cause the following message "USER FATAL MESSAGE 69, SUBDMAP STATEMENT PARAMETER APP NOT DEFINED". This error is only known to occur in the structured solution sequences (SOLs > 101) Avoidance: Remove Diag 14 and/or Diag 64. @ 7013; ries ; Miscellaneous - COMIFP, PUNCH - Produces Incorrect Bulk Data Ent The comifp module has been modified for Version 69 to support the part superelement concept. It can take ifp-like files from the sep2 module and punch bulk data files which may then be used as part files. At present, the RBE2 and RBE3 elements are punched incorrectly. Avoidance: An avoidance is to separate the elements not properly

processed by hand. @ 7014; 69; Heat Transfer Analysis - MAT4, CONVM, EMG - FATAL ERROR, Bad Dia gnostic Message If the user fails to define a dynamic viscosity in field 7 of the MAT4 entry for a forced convection analysis the job fails in the EMG module with no indication of the cause of the error. Avoidance: Make sure the dynamic viscosity is input on the MAT4 card if the problem has CONVM bulk data entries. @ 7015; 69; Material Nonlinear Analysis - TEMP(LOAD), CLOAD - Wrong Answers In Version 68.2 and SOL 106 only, thermal loads applied to upstream superelements results in incorrect stresses and forces. Avoidance: Use SOL 66 or apply the following alter to SOL 106: compile super3 noref nolist alter 'IF ( NOCASRS1>-1 AND SEID=0 )','' IF ( (NOCASRS1>-1 AND SEID=0) or app='nlst' ) THEN $

@ 7021; 69; Optimization/Design Sensitivity - DSCM2 - Wrong Ordering In design sensitivity analysis, there is a table printed just prior to the DSCM2 (design sensitivities matrix) which specifies the ordering of the columns in the matrix. If some of the design responses specified on the DRESP1 entry are the eigenvalues, then the table claims that the eigenvalue responses are in ascending order by DRESP1 entry ID. However, the matrix is actually in ascending order by mode number. Avoidance: The only avoidance is to be aware that the output is in ascending order by mode number. @ 7024; ; IDENT option on ACMODL entry If INTER=IDENT on the ACMODL data entry is used in acoustic analysis, then only two fluid grids per structural grids at the fluid/structure inter face. This precludes modeling structures such as t-sections stiffeners in tanks or t-sections immersed in fluid. @ 7040; 69; Documentation - DRESP1 - UFM 316 The documentation for the DRESP1 entry in the MSC/NASTRAN Version 68 Quick Reference Guide states that the SPC force component is chosen with the RTYPE=FRSPC or RTYPE=TSPC. However, if either of these RTYPEs are used, the following fatal message is issued, "USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA CARD DRESP1 331 SORTED CARD COUNT = ***". The documentation should say the the SPC force component is chosen with RTYPE=FRSPCF or RTYPE=TSPCF. Avoidance: Use RTYPE=FRSPCF or RTYPE=TSPCF not FRSPC or TSPC when the SPC force component is desired as a response type.

@ 7041; 69; Nonlinear Analysis - EMG - Poor Performance NO EXTERNAL ERROR REPORT When there are no linear elements in a model, the EMG module is unnecessarily executed. This appears to be a minor performance improvement, I see no need to publish to clients. @ 7044; s Scaling by other than 1. of a vector flux QVECT load by a DLOAD bulk data entry produces wrong results in the "APPLIED-LOAD" column of the "Heat Flow into HBDY Elements" table. The actual load vector generated is correct and therefore the temperatures are correct. Avoidance: Use only scale factors 1 on the DLOAD entry. @ 7045; 69; Case Control - SUBCASE - Poor Message The message " *** USER FATAL MESSAGE 609 (IFP1D) SUBCASE DELIMITER CARDS MUST HAVE A UNIQUE IDENTIFYING INTEGER " which does not accurately describe the input error is issued for the following situation: 1. When the subcase id exceeds 9999999. 2. In solution 200 when the subcase id of the preceding subcase is higher than the id of subsequent subcase. @ 7047; 69; Aeroelastic Analysis - Unwanted Output in the Flutter Summary In the pk flutter solution, all combinations of velocity, Mach number and density are used. In certain situations (e.g., if analyses are only required at 50 velocity-density combinations) it would be preferable to only do ordered combinations. Avoidance: The current avoidance is to put each required combination in its own subcase. For the example given above, this would require 50 subcases, but it would reduce the number of flutter analyses from 2500 to 50. @ 7048; 69; Submittal Script - node - Error With node Keyword When the "node" keyword is used along with the "user" keyword ( i.e. nastran ..... node=pelee user=ssp) then job will not get submitted to remote node and the following message will be displayed on the console: nast68: (W) node=pelee specified, setting app=yes bat=no scr=yes nast68: (E) node=pelee: no execute permission on remote node. Avoidance: The following line in the nastran shell script should be changed. The shell script is $MSC_INSTALLDIR/bin/nastran. 69.1; Nonlinear Transient Heat Transfer - QVECT, DLOAD - Wrong Answer

rx="$remcmd $tuser $node" should be changed to rx="$remcmd $node $tuser" @ 7050; h file 69; Multipoint Constraint - MPCFORCE Output - Labeled as SPCF in .pc MPCFORCE output when printed to a punch file, is mis-labeled as SPCF. The data is correct however. Confusion may arise when both SPCFORCE and MPCFORCE are printed to the punch file. The user must use care to read the correct results. The error is considered fixed in Version 69; however, setting of SYSTEM(210)=1 is required. Avoidance: Print and punch the data so that the output can be compared and correlated.

@ 7053; 69; Documentation - DVPREL1 - UFM 2042 Misleading The DVPREL1 entry is used to define a relationship between an analysis model property and a design variable. However, not all analysis model properties are valid choices. For example it is not valid to link the MID field on a property entry (which is an integer) to a design variable (which is always real). If this is done the MATi entry in the analysis model is overwritten with a nonexistent MATi and then the following misleading message is issued, "USER FATAL MESSAGE 2042, MISSING MATERIAL TABLE******** FOR ELEMENT xx". Invalid choices for the analysis property chosen on the DVPREL1 should be screened out prior to the start of optimization. Avoidance: Choose only valid analysis properties on the DVPREL1. @ 7055; 69; Documentation - Composites, Buckling, PSHELL, PCOMP - Missing Do cumentation The following remark should be added to the PSHELL bulk data entry description in the MSC/NASTRAN Quick Reference Guide: The use of MID4 will cause incorrect results in buckling analysis and differential stiffness. The following remark should be added to the PCOMP bulk data entry documentation in the MSC/NASTRAN Quick Reference Guide: An unsymmetric layup or the use of ZO to create an unsymmetric layup will cause incorrect results in buckling analysis and differential stiffness. @ 7059; 2001; Miscellaneous - DMIG - Psuedo Infinite Loop, Device Full If characters are input in the NCOL field of a DMIG header entry and the matrix is rectangular (IFO=9), then the MTRXIN module may be caught in a pseudo infinite loop which may be finally stopped by a "device full" message. The MTRXIN module interprets the characters in NCOL as an

integer and this is normally a huge number. The IFP module should stop the execution with "User Fatal Message 327" in case that the input for NCOL is not an integer. @ 7065; 69; P-Elements - Wrong Adaptivity Analysis - Poor Answers If a data deck contains different TYPEs of P-elements (e.g. a HEXA and a PENTA), the adaptivity error analysis likely to be in-accurate leading to poor answers. Avoidance: There is no avoidance. @ 7073; 69; Heat Transfer - PARAM,POST,-2 - SDRC Cannot Read TEMP Vector SDRC is not able to read the temperature vector in the output2 file produced with PARAM,POST,-2. Avoidance: Use EQUIVX instead. @ 7076; ; Timing Estimate - DECOMP - UIM 4157 The CPU time estimates issued in USER INFORMATION MESSAGE 4157 for sparse decomposition can be overstated by a factor of 7 to 15 times on some machines. This is due to the fact that the timing estimates do not take into account the high rank performance update enhancement which was introduced beginning in Version 68.1. By overstating the CPU time estimates the job may fail with a message that it does not have enough time to complete Decomposition. Avoidance: Include an extra large value for time on the TIME executive control entry or remove the TIME entry. @ 7085; 70.5.2; Cyclic Symmetric - Dihedral, LOADCYH, DSYM - UFM 4342 A cyclic symmetry analysis using dihedral symmetry and centrifugal loads may issue the following message: USER FATAL MESSAGE 4342, THE LOAD BULK DATA CARD GIVEN BELOW IS INCONSISTENT WITH SPECIFIED ANALYSIS HARMONICS OR SEGMENTS OR SYMMETRY TYPE. LOADCYH 20 0CSTAR 2000004 The message occurs when a SS (Symmetric=Symmetric) boundary condition is specified on the DSYM case control command and the LOADCYH bulk data requests an 'RLOAD' for the HTYPE field. In this condition the code generates too many load vectors. Avoidance: One avoidance is to change the HTYPE field to 'C' on the LOADCYH bulk data entry. A better avoidance is to eliminate the DSYM request in the case control. @ 7088; ned 70; Documentation - PSHELL, MID4, ZOFFS - MID4 and ZOFF may be combi Remark 6 under the description of the PSHELL entry in the Version 68 Quick Reference Guide states that "Only one of the options MID4 or ZOFFS may be

used". This is a guideline to prevent users from offsetting the element twice. Depending on the intent of the user these affects may be combined and correct answers achieved. @ 7093; 70.5; Complex Eigenvalues - CLAN - Wrong Answers For one model, the complex lanczos eigenvalue method in solution 107 produced erroneous eigenvalues on the CRAY computer. The same model produced consistent real eigenvalues across all machines so there does not appear to be anything wrong with the model. MSC/NASTRAN Version 70.5 introduces the new Block Complex Lanczos method. This method correctly solves this problem on CRAY and all other machines. This method must be selected in Version 70.5 by setting system(108)=4 on the NASTRAN entry or rc file. Avoidance: Until the error is better understood, use SOL 110, modal complex eigenvalues, with the Hessenberg Method. @ 7100; 69; Static Analysis - SPC - UFM 5423 A DMAP error exists in multiple boundary condition with non-zero SPC in SOL 101. The error is triggered when the same boundary conditions are not grouped together, i.e., the same SPC is selected again following a SUBCASE with a different SPC. The error is detected by the SSG2 module with a UFM 5423 "attempt to multiply incompatible matrices." This error is related to error 5977 which is detected when SPCD is used instead of non-zero SPC. Avoidance: 1. Shuffle the Subcases so that the same boundary conditions are grouped together (to appear consecutive Subcases). 2. Use SOL 106 with unsorted SUBCASE set-up and run it as a linear problem without any nonlinearity. @ 7108; 69; Acoustic Analysis - SDR1, MPYAD - Insufficient Time Acoustic Analysis in MSC/NASTRAN does not support the matrix method of data recovery. Thus, the data recovery performed by the modal displacement method for small sets of data is very inefficient. This deficiency is complicated by poor timing estimates in MPYAD. MPYAD, which is called by SDR1 during the data recovery, may yield a time estimate that overflows the machine. Version 69 will support data recovery in the DDRMM module, i.e., the matrix method. Avoidance: There is no simple avoidance. Perform modal reduction on the fluid and limit the number of frequencies in the analysis. The following alter can be applied to recover pressure only by the matrix method. compile semfreq souin=mscsou noref nolist alter 74 lamx ,,lamas/mlamas/-1 $ lamx ,,lamaf/mlamaf/-1 $ matgen ,/cp0/6/5/2/2/1 $ matgen ,/cp1/6/3/0/1/1/1 $ $

partn mlamas,cp0,/,,emats0,/1 $ merge ,,,emats0,,cp1,/emats1/1 $ trnsp emats1/emats $ $ partn mlamaf,cp0,/,,ematf0,/1 $ merge ,,,ematf0,,cp1,/ematf1/1 $ trnsp ematf1/ematf $ $ append emats,ematf/emat $ lamx emat,/lamax $ ofp lamax/ $ ALTER 112,112 IF (DDRMM>=0) APP1='MMREIG' alter 117 equivx lamax/lama/-1 $ $ compile super3 souin=mscsou alter 24,24 IF ( APP1='MMREIG' ) THEN $ compile sedrcvr souin=mscsou noref alter 61 if ( app1='mmreig' ) then $ equivx oes1/qes1/-1 $ equivx oef1/qef1/-1 $ delete /ougv1,oqg1,oes1,oef1, $ DDRMM CASEdr,UH1,OL1,qugv1,qQG1,qES1,qEF1,XYCDBdr/ OUGV1,OQG1,OES1,OEF1, $ EQUIVX OQG1/QQG1/ALWAYS $ EQUIVX OUGV1/QUGV1/ALWAYS $ endif $ alter 159,159 IF ( sort2 ) THEN $ SORT2 @ 7109; 70.5; DMAP - MATMOD OPTION 17 - Wrong Answers MATMOD option 17 generates a g-size partitioning vector from a user-defined set of grid and/or scalar points. If the 3rd parameter (SETFLG) is set to 1, it is supposed to look for the PARTN case control command and put a value of 1.0 in the vector for grid points listed in the set defined by the PARTN case control command. If there is no PARTN case control command, it is supposed to use the sets defined by the 2nd parameter (UBIT). However, the above operation returned a vector with all values set to 1.0 instead. Depending on subsequent operations, this can either lead to fatal errors or wrong answers. Avoidance: Include the appropriate PARTN case control command. @ 7110; 69; SSSALTERS - evds.v682 - Fails to Converge When using the eigenvector sensitivity dmap alter (evds.v682) delivered with the sssalter library in Version 68.2, and the model does not contain any elements which references a mass property entry, then the job may fail to converge. Avoidance: Replace the following dmap statement in evds.v682 subdmap fox: TYPE PARM,,I,N,NULLDM $

with TYPE @ 7118; 69; Documentation - ADUMi, PDUMi - Undocumented Limits, User Modifia ble NASTRAN The number of additional fields "NP" on the PDUMi entry is limited to 24 words. This limit is not documented in the description of th NP field of the ADUMi bulk data entry description. @ 7127; le TA1 69; Optimization/Design Sensitivity - PARAM,SNORM - UFM 2025 in modu If in SOL 200, parameter SNORM is greater than or equal to 0.0 and if a coordinate system is specified in the MCID field of the connectivity entry in the run then the following fatal message will be issued. *** USER FATAL MESSAGE 2025 MODULE = TA1 UNDEFINED COORDINATE SYSTEM 3 This error might also occur in SOL 101 and 103 if p-element adaptive analysis is performed. Avoidance: Two avoidances exists: 1. Remove the SNORM parameter. 2. Add the following alter. compile update souin=mscsou nolist noref alter 51,51 $ TA1 ,,ECT,EPTX,BGPDTSCR,SIL,,CSTMSCR,/ ESTJNK,,,GPECT0,/ LUSETS/-1/S,N,NOSIMP/1/S,N,NOGENL/0 $ @ 7136; t There is an error in the nddl for the MCSID field of the MAT2 entry. The nddl types the field as real when it should be an integer. If this field is changed in a RESTART the execution will fail. The message "time: command terminated abnormally" is issued in the log file. Avoidance: The run should be executed as a cold start. @ 7137; 69; Postprocessing - PARAM,POST,-2, SOL 112 - Extra Output In SOL 112 with PARAM, POST, -2, the geometry tables are written twice to the OUTPUT2 file. However, this does not seem to negatively affect the IDEAS translator. 69; Restarts - MAT2, IFPL, RESTART - DEC Alpha Only Aborts on Restar PARM,,I,N,NULLDM,NZWD $

Avoidance: There is no avoidance. @ 7138; 70; Documentation - Common Questions and Answers - Incorrect Problem Descriptions The problem description on page 103 of the Common Questions and Answers ,2nd Edition, does not match figure 40 on the same page. The acceleration should be negative, as is reflected in fig. 40 and the 3 in field 5 of the RLOAD2 entry should be moved to field 6. The title for figure 39 on page 101 of the Common Questions and Answers book refers to point loads when in fact the example discusses pressure loads (PLOAD4s). @ 7139; 69; Executive System - COMPILE - Segmentation or Access Violation NO EXTERNAL ERROR REPORT IS NECESSARY MAR VERIFIES THIS EVERY TIME HE REBUILDS THE DELIVERY DATABASES. An attempt to compile the NDDL and more than 75 subDMAPs in a single run causes a segmentation or access violation. The limit in Version 69 has been expanded to allow for at least 100 subdmaps. Avoidance: Split the job into two runs. @ 7141; 70.7.2; Nonlinear Transient Analysis - CBEAM - UFM 4551 (NL2CON) It has been found that in rare cases when then the a beam element is exactly normal to the axis of the exhibit a round-off error that leads to convergence ultimately USER FATAL MESSAGE 4551, STOPPED PROBLEM CONSECUTIVE DIVERGING TIME STEPS. @ 7142; 69; Optimization/Design Sensitivity - DOPR3 - SFM 4276 If the user provides constraints for multiple DRESP1 entries in a single flutter subcase, then SYSTEM FATAL MESSAGE 4276 GREAD ERROR CODE 3001 will occur in the DOPR3 module. Avoidance: If it is not possible to define all the flutter responses on a single DRESP1 entry, multiple flutter subcases are required, each with a single referenced DRESP1 entry. @ 7144; ; Nonlinear Transient Analysis - OTIME - Missing Output In nonlinear transient analysis (SOLs 129 and 159), the solution vectors (SDISP, SVELO, and SACCEL) should be output for each subcase independently. However, SDISP, SVEL, or SACCEL does not appear in the output unless the request is specified above the subcase. Avoidance: Specify the output request for SDISP, SVELO, or SACCEL above the subcase level. @ orientation vector for beam, the Cray may difficulties and DUE TO FIVE

7147;

69; Utilities - OPTCONV - Fails On Windows NT If the OPTCONV program delivered with the Windows NT version of NASTRAN is used to convert an old optimization deck to the current format, the following message is obtained in a pop-up window: "The instruction at 004147cc referenced memory at 0d4212b9 The memory could not be written Click on OK to terminate the application". Therefore, it is not possible to convert any old optimization decks on Windows NT. Avoidance: Convert the decks by hand or have them converted on a different machine variation.

@ 7156; 69; P-Elements - PARAM,USETPRT,0 - Incorrect Output PARAM,USETPRT,0 produces incorrect output for P-Elements. Avoidance: @ 7163; 69; Nonlinear Static Analysis - ASETi, OMITi, AUTOSPCR - UWM 5225, W rong Answers In nonlinear static analysis, any nonlinear degree of freedom in the o-set (via ASETi or OMITi entries) may result in a singularity in the o-set linear stiffness matrix (UWM 5225). Also, if PARAM,AUTOSPCR is used, all the nonlinear degrees of freedom will be constrained and the resulting solution will be incorrect. Avoidance: Include all the active nonlinear degrees of freedom in the A-set. @ 7165; 69; Optimization/Design Sensitivities - CTRIAX6 - Wrong Answers When CTRIAX6 elements without mid-side nodes are used in an optimization job which includes specification of temperature loads, the thermal displacement/stress sensitivities may show wrong signs which may result in strange convergence behaviors. Avoidance: Use CTRIAX6 elements with midside nodes. @ 7167; 69; Nonlinear Heat Transfer - UIM 6194, UIM 6193, UFM 4551 Nonlinear Heat Transfer jobs with more than one subcase may fail to converge and issue UIM 6194, UIM 6193 and USER FATAL MESSAGE 4551. This error was experienced in a job where the second subcase was the same as the first. The job should have converged immediately in the second subcase since there was no change. USER INFORMATION MESSAGE 6194, STOPPED ITERATIONS DUE TO REACHING MAXIMUM ITERATION LIMIT WITHOUT CONVERGENCE *** USER INFORMATION MESSAGE 6193, MAXIMUM NUMBER OF BISECTIONS OR MINIMUM LOAD STEP HAS BEEN REACHED. USER FATAL MESSAGE 4551, Use PARAM,USETPRT,1 instead.

STOPPED PROBLEM DUE TO FAILED CONVERGENCE @ 7170; 69; Heat Transfer - CHBDYP, LINE with midpoint node option Currently there is a design limitation using CHBDYP,LINE to represent thhigher order CQUAD8 or CTRIA6 for the edges with midpoint option. Therefore, the weighting on the mid-side node is misrepresented if the two node LINE option is used. Avoidance: Use CHBDYE to represent the CQUAD8 or CTRIA3 element with midsidenodes. Please note that without the midside node, the CHBDYP,LINE can be used. @ 7171; 2001; Documentation - Reference Manual - Large Field description There is an error in the Reference Manual documentation for the description of the Small/Large/Free field format of the bulk data entries as follows: pp 3.7.1-2 Bottom of the page, Large Field example. The first field of the parent card should be TYPE* not TYPE. The continuation of the parent card should be QED123 not QED124. pp 3.7.1-3 Top of the page, Large field with large field continuation example. This is shown as all small fields. There should only be 4 (double wide) fields between the first and last fields. @ 7173; 69; P-elements - CHEXA - UFM 9050 CHEX p-Elements can have singularities in the stiffness for the default integration options. This error only occurs in a model with a single layer of elements. In particular the default reduced integration order with rectangular elements with p values of [2,2,3] in the three directions may be singular. This error may lead to the following fatal message: USER FATAL MESSAGE 9050 (SEKRRS) - RUN TERMINATED DUE TO EXCESSIVE PIVOT RATIOS IN MATRIX KLL. USER PARAMETER BAILOUT MAY BE USED TO CONTINUE THE RUN. Avoidance: Specify a different p value.

@ 7175; 70.5; Documentation - CALL DBFETCH - DMAP Module Dictionary The description of the DBFETCH command in the DMAP Module Dictionary should enhanced to state "if you are using DBFETCH in a Sol 100 (USER DMAP) to get data blocks from a previous run, include a RESTART FMS statement to specify the version number on which the data blocks were written". @ 7176; 69; Direct Frequency Response - CHACAB - Wrong Answers The damping terms for the acoustic absorber element, CHACAB, are "lost" in solution 108 resulting in null damping from this element.

Avoidance: Place equivalent CDAMP or CVISC elements across the structural boundary points. @ 7179; 69; Documentation - QRG - Inconsistent E,G, NU with MAT1 Table 1 under the MAT1 description in the Version 68 Reference Guide indicates that "E" and "NU" are used properties for "Membrane and Bending" of the CQUADi, elements. This should be clarified to state "should "Membrane, Bending and Inplane Shear". @ 7182; 69; Post Processing - SOL 103 and 105, OUTPUT2 - Missing Output When multiple subcases are used in SOL 103 or 105 and PARAM,POST,-1/-2 is chosen, the output2 file is written over with each subcase and only the last subcase remains in the output2 file for post processing. Avoidance: Request different output2 units in each subcase using param,ounit2 or the following DMAP alters may be used for some post processors (PATRAN3 v5) For SOL 103: COMPILE SUPER3 SOUIN=MSCSOU NOLIST NOREF $ ALTER 'IF ( POST=-1 AND OUNITX\<\>OUN','IF ( ( ((DROPT=0 OR DROPT=4)' PARAML CASECC//'DTI'/1/1//S,N,SUBC1 $ PARAML CASEDR//'DTI'/1/1//S,N,SUBCNO $ IF ( POST=-1 AND SUBCNO=SUBC1 AND OUNITX<>OUNIT2 AND NOT(PEXIST) AND NOT((APP='NLST' OR APP='TRANRESP') AND PATVER>=3.) AND (((DROPT=0 OR DROPT=4) AND APP1<>'BKL1') OR APP='BKL0') ) OTAPE2=-1 $ IF ( ( ((DROPT=0 OR DROPT=4) AND APP1<>'BKL1') OR APP='BKL0') AND (OGEOM='YES' OR OGEOM=' ') AND SUBCNO=SUBC1 AND ((POST=-1 AND PATVER>=3. AND NOT(APP='NLST' OR APP='TRANRESP')) OR POST=-2 OR POST=-4) ) THEN $ $ ALTER 'IF ( ((DROPT=0 OR APP1=','IF ( ((DROPT=0 OR APP1=' PARAML CASECC//'TRAILER'/1/S,N,NSUBC $ PARAML CASECC//'DTI'/NSUBC/1//S,N,LASTNO $ IF ( (SUBCNO=LASTNO OR APP<>'REIG ') AND (((DROPT=0 OR APP1='BKL1') OR (PEXIST AND ADPTEXIT)) AND POST<0 AND POST>-4 AND NOT(POST=-1 AND PATVER>=3.0 AND SCRSPEC>-1)) ) OUTPUT2 //-9/OUNIT2//OMAXR $ ENDALTER $ For SOL 105: compile super3 alter '-9'(2),'' compile sebuckl alter 'dbdict1' type parm,nddl,i,y,ounit2 $ type parm,,i,y,omaxr $ OUTPUT2 //-9/OUNIT2//OMAXR $ @ 7184; 69; Postprocessing - PARAM,OUGCORD,BASIC, PATVER,3 - Incorrect Defor MSC/NASTRAN Quick to define the CTRIAi and CCONEAX be clarified to

med Shape PATRAN 3 will display an incorrect deformed shape when PARAM OUGCORD is set to BASIC and the gridpoint output coordinate systems (field 7, CD) is anything other than the basic (coordinate system ID 0). SubDMAP SEDRCVR transforms the displacement vector into the basic coordinate system based on the value of the parameter OUGCORD and PATVER. In Version 67.5 the DMAP IF test includes the value of the parameter PATVER which prevents transformation if PATVER=3.0 (default for PATVER is 3.0). 66 67 ELSE IF ( NOT(PATVER>=3.) AND OUGCORD='BASIC' ) THEN $ VECPLOT UGVS,BGPDP,EQEXINS,CSTMS,CASEDR,/UGVSB//0/1 $

The reason for preventing this transformation is that PATRAN 3 understands the CD field for the GRID entry. In Version 68 the DMAP test for PATVER is not present: 70 71 ELSE IF ( OUGCORDX='BASIC' ) THEN $ VECPLOT UG,BGPDP,EQEXINS,CSTMS,CASEDR,,MEDGE,SIP/UGB/ /0/1/' PARAM,'/'OUGCORD,'/' BASIC '/ALTSHAPE $

This causes the displacement vector to be transformed into the basic coordinate system for post processing by PATRAN 3. However, PATRAN 3 is expecting the output in the output coordinate system of the gridpoints. If the transformed displacement datablock is used by PATRAN 3, the displaced shape is wrong. The description in the Quick Reference Guide, Section 6 (parameters) also needs to be updated under the section 'POST,POST,-1: Results Data Blocks'. The PATVER description needs to include the interaction between parameter OUGCORD and PATVER. Avoidance: Do not specify the value of parameter OUGCORD as BASIC if you are using PATRAN 3. PATRAN 3 expects the displacements to be expressed in the CD value of the gridpoint (analysis coordinate system in PATRAN 3.) You can transform displacements to different coordinate systems inside P3. @ 7185; 66; Response Spectra - SDR2 - Wrong Answers The response spectra summation process for the NRL and SRSS methods handle the CBAR and CBEAM elements differently. Both the NRL and SRSS methods involve taking the square root of the sum of the squares of the value of interest, in this case stress. For the BEAM element the summation is made for the combined bending and axial stress, while for the BAR element the bending and axial stress is summed independently to coincide with the BAR's output which prints the axial stress separately. This leads to incorrect results for the output of the maximum and minimum stress. Avoidance: The CBEAM element should be used for Response Spectra Analysis whenever stress results are desired. @ 7186; 70.6; Documentation - INCLUDE - Include File Not Found The Installation and Operations manual Section 4.5 states that the default directory for an include file is the input file's directory. The default directory is actually the current directory.

Avoidance: Place the file to be included in the current directory or specify the complete path name of the file to be included. @ 7191; 69; Optimization/Design Sensitivity - TEMP(LOAD), TEMP(MATERIAL), SH PCAS - SFM 3001 When a TEMP(LOAD) and/or TEMP(MATERIAL) entry is used in the primary Subcase Control section of a shape optimization job using the analytic shape boundary shape approach in MSC/NASTRAN Version 68.2, the job will fail in module SHPCAS with the message "SYSTEM FATAL MESSAGE 3001". Avoidance: Apply the following DMAP alter. compile bndshp nolist noref souin=mscsou alter 'auxmid'(2) dbview ugx=ug (where mtemp=*) $ alter 'partn','partn' PARTN UGx,,AXMVEC/UGA,,,/1 $ alter 'equivx','equivx' EQUIVX UGx/UGA/ALWAYS $ @ 7193; wers ; Miscellaneous - Wrong continuation format - UFM 313 or Wrong Ans If column 1 of the continuation entry is blank, then the continuation entry is interpreted as a continuation of the last sorted bulk data entry. The error is likely to result in USER FATAL MESSAGE 313 ILLEGAL NUMBER OF WORDS ON CONTINUATION" or in rare cases possibly even wrong answers. An example follows: . Begin Bulk . . tic,100,101,1,0.,0. $ grid,1000,,0.,2.5,2.5 $ grid,101,,0.,0.,0. grid,102,,10.,0.,0. $ 2 3 cbar 1 1 +000015 +000015 $ . . enddata will be interpreted as BEGIN BULK . . CBAR 1 4 101 0.25 5 102 0. 6 1000 0. 7 0.25 8 0. 9 0. 10

101

102

1000

+000015 . GRID GRID . GRID . . TIC +000015 +000015 . . ENDDATA

101 102 1000 100 101

0. 10. 0. 1 0.25 0.

0. 0. 2.5 0. 0.

0. 0. 2.5 0. 0.25 0. 0.

Avoidance: Correct the offending continuation entry. @ 7196; 2001; Miscellaneous - GP2, GENEL - Poor Message, SFM 3003 When at least one entire line is missing from a GENEL element, GP2 issues the message: SFM 3003 ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK GEOM2S (FILE 101) IN SUBROUTINE GP2. Avoidance: Correct any offending GENEL elements. At a minimum the Reference Manual documentation should cite incorrect GENEL input as a possible cause of the error message. @ 7199; ; Nonlinear Transient Analysis - PARAM,LGDISP - Poor Answers In MSC/NASTRAN Version 68 nonlinear transient analysis, solutions 99 and 129, were improved to correct error 2665 which stated that in nonlinear transient analysis the damping matrix is not updated. The damping forces for the large displacement solution now use the current "updated" damping matrix, which may have been calculated at an unconverged solution. This approach was intended to provide better directionality to the damping forces when they rotate. However, every time the damping matrix is updated for convergence, a change will occur in the force balance which will cause an erroneous step in the converged accelerations. ( It causes very large errors. - DNH) @ 7201; ision 70; Element - CELASi - Forces/Stresses are Calculated in Single Prec Scalar spring and damper element's (ELASi and DAMPi, 1=1,2..4) stress and force outputs are calculated with only single-precision arithmetic. If a scalar element connects two points with small differences in displacements, relative to the total displacements, some numerical roundoff will occur. Since the stiffness and damping matrices are calculated in double precision the overall results will not be affected. The problem also does not occur on large-word, 64-bit computers such as CRAY and NEC. Avoidance: More precise results for the ELASi elements may be obtained from the GPFORCE output request. @ 7202; 69; Case Control - MPCFORCE=NONE - Extra Output

The mpcf=none command does not turn off mpcforce output in a subcase when it has been requested above the subcase level. Avoidance: Define set 1 = [x], where x is a nonexistent grid or scalar point and use mpcf=1 instead. @ 7206; 69; Heat Transfer Analysis - CHBDYE,TUBE PHBDY - Wrong Answers The convection area used by the CHBDYE for the structure TUBE element is incorrect whenever D2 is left blank. In Version 69 the code is changed so that if D2 is left blank then D2 is set to D1. Avoidance: Use CHBDYP,TUBE option to represent the tube convection area or explicitly set D2. @ 7207; 69; Restart - SPCFORCE - Missing Output SPC force output request is ignored during automatic restarts in dynamic structured solution sequences. See related Error 3119 which was fixed for SOL 101 only. Avoidance: Include the entry PARAM, SERST, SEDR in the bulk data section of the restart deck. @ 7213; ; DMAP - FBS, SYSTEM(72)=1 - Wrong Answers Left handed fbs solves the equation of the form: t t [X] [L][U]=[B] In addition left handed fbs has an option of solving T [X] L U = [B] This option is only documented in the MSC/NASTRAN Numerical Methods Guide and is selected by setting SYSTEM(72)=1. This option produces wrong answres when B is rectangular. The left hand fbs method is not used in the delivered solution sequences for non symmetric matrices. Thus, only user DMAPs may encounter this error. Avoidance: Use right hand fbs. @ 7215; ; Cyclic Symmetry - SECSET1, VAXW, GIV, MGIV - UWM 3034, UFM 4645 Massless degrees of freedom that are placed in the c-set are not properly removed before entering the READ Module in Cyclic Symmetry component mode synthesis. Attempts to remove these degrees of freedom by explicitly constraining them will also be ignored. This may lead to "USER WARNING MESSAGE 3034, ORTHOGONALITY TEST FAILED" followed by a fatal message, "USER FATAL MESSAGE 4645---- THE SHIFTED STIFFNESS MATRIX IS NOT POSITIVE DEFINITE IN MGIVENS METHOD" or "USER FATAL MESSAGE 4646---- THE MASS MATRIX IS NOT POSITIVE DEFINITE, USING THE GIVENS METHOD. USE MGIV INSTEAD". Avoidance: Do not put the degrees of freedom in the c-set, specify Lanczos instead of a transformation method or add a small amount of mass to the massless degrees of freedom. @

7219;

69; Miscellaneous - IFP - Infinite CPU Loop The following continuation entries while legal when using the NASTRAN MESH option result in an infinite cpu loop in IFP when the NASTRAN MESH command is removed. SPC1 = DESVAR =2 10 *1 20 *1 10 *-1 ABC = 11 *1 1. *1.

Avoidance: Add the NASTRAN MESH command or correct the continuation entries in error. @ 7220; 69; Postprocessing - PARAM,POST - Missing Output If the ASSIGN statement is not used or if the complete path is not specified on the ASSIGN statement for an OUTPUT2 file on the Windows NT Version of MSC/NASTRAN, the file is not written to the jid directory but is instead written to the "bin" directory. Avoidance: Specify a complete path on the ASSIGN statement. @ 7231; ut 69; Optimization/Design Sensitivity - XSORT - Incomplete Sorted Outp The XSORT Module may encounter errors for large input files, when it is executed more than once during a run. The messages encountered are various, but the sorted file will be incomplete in every case. The error occurs only on large files, and only on the second and subsequent bulk data files. The errors occur when the bulk data to be sorted does not fit into memory. In Version 68 this error occurs only in design optimization analysis. Avoidance: Increase the memory request so that the file will fit into memory, or reduce the size of the file. @ 7232; 69; Miscellaneous - IFP, RBEi- UFM 4276 EC 7777 When the RBE3 bulk data entries contains illegal data in the REFC field, USER FATAL MESSAGE 316 should be issued to identify the offending entry. USER FATAL MESSAGE 4276 ERROR CODE 7777 is issued instead, which does not identify the illegal data. Avoidance: Other than locating the input in error and correcting it, there is no avoidance. @ 7233; 69; Matrix Methods - FBS, MPYAD - UFM 4547 FBS may set the density of GOAT incorrectly for some large models. The density should be between 0 and 10000 (i.e., 100%). However, it sometimes overflows this field and is printed in the f04 as *******. In this case the run will fail in the MPYAD module with the message "USER FATAL MESSAGE 4547" Avoidance: Use the following alter to re-set the trailer to a value of 80% (or any value they desire).

compile sekr, souin=mscsou, nolist, noref alter 106 $ the line which creates GOAT in his run modtrl goat///////8000 $ @ 7234; 69; Loads - PLOAD1 - Missing Documentation, Unexpected Answers If on the TYPE field of the PLOAD1 entry, the element coordinate system direction (e.g. TYPE=FYE) option is selected, then the projection (i.e. SCALE=FRPR or LEPR) option is ignored and the result is the same as the SCALE=FR (or LE) option. Avoidance: There is no avoidance. @ 7237; 70.5; Documentation - DMAP Module Dictionary, OUTPUT2, INPUTT2 - SFE 4 276 EC 1401 MSC/NASTRAN Version 68 may not be able to handle an OUTPUT2 file of a matrix generated from an external client written computer program that follows MSC's documentation description of proper OUTPUT2 format. It is not the INPUTT2 step that is crashing, it is any subsequent operation on the datablock, such as TABPT, OUTPUT4, ADD, MATPRN or the like. Avoidance: Use V67.5 to convert the OUTPUT2 file to an OUTPUT4, and then read in the OUTPUT4 file into V68.

@ 7238; 69; p-Elements - DATAREC, OUTPUT - UWM 6514 (SDRPOP) When there are any OUTPUT bulk data entries, but no DATAREC case control keywords, the following warning is produced: *** USER WARNING MESSAGE 6514 (SDRPOP) THE DATAREC= 0 CASE CONTROL COMMAND REFERENCES UNDEFINED OUTPUT BULK DATA ENTRY. USER INFORMATION: THE DATAREC COMMAND WILL BE IGNORED. USER INFORMATION: THERE WILL BE NO ERROR-ESTIMATE OUTPUT FOR THE ADAPTIVITY ANALYSIS. Avoidance: To avoid the warnings, the OUTPUT entries should be removed. @ 7242; ; p-Elements - GMCORD - Inefficient Processing If a GMCORD coordinate system is used for the material, load, or displacement coordinate system, execution may be inefficient since the coordinate system is processed independently for each integration point. Avoidance: Other than not using GMCORDs there is no avoidance. @ 7243; 70.5; Documentation - TABPT, BGPDTS - DMAP Module Dictionary According to the DMAP module dictionary "TABPT may occasionally misidentify real numbers or character values on the IBM/MVS-like computers". This problem is not limited to the IBM/MVS-like computers.

Avoidance: Use TABPRT instead of TABPT. @ 7244; Answers 69; Optimization/Design Sensitivity - ANALYSIS=FLUTTER, DMIG - Wrong

For flutter design sensitivity and optimization, the sensitivity results will be in error if there are multiple subcases and a second or subsequent subcase includes a different M2PP, K2PP or B2PP request in the case control. The sensitivity analysis is based on the generalized matrices from the first subcase and will therefore be incorrect for the subsequent subcases. Avoidance: There is no avoidance. @ 7253; 69; Nonlinear Analysis - MPCFORCE - Missing Output MPCforces are not output in nonlinear analysis, solution 106. Avoidance: Insert the following DMAP alter in Version 68.2: compile nlstatic alter 'ADD DELCVECT,CVECT0'(,-1) add5 qqgni,qgni,,,/qmgni//-1. $ call dbstore qmgni,,,,//nloop/0/' '/0 $ alter 'DBVIEW CVECTH' call dbfetch /qmgni,,,,/solcuri/0/0/0/0 $ call dbstore qmgni,,,,//solcur/0/' '/0 $ alter 'APPEND QGNI,/QGNR' $ file qmgni=ovrwrt $ $ call dbfetch /qmgni,,,,/nloop/0/0/0/0 $ file qmgnr=append $ APPEND QmGNI,/QmGNR/2 $ APPEND MPC FORCE VECTORS alter 'EQUIVX QGNR/QGN' call dbstore qmgnr,,,,//0/0/' '/0 $ compile super3 alter 'ADD PG,PJ/DPJ' call dbfetch /qmgnr,,,,/0/0/0/0/0 $ equivx qmgnr/qmg/-1 $ COMPILE SEDISP ALTER 'MPYAD PSS,CRX'(2) MPYAD PG,CRX,/PGX $ $ ALTER 'IF ( NOQMG>-1 ','' $ IF ( NOQMG>-1 AND NOMSET>-1 AND NOT(DELTA) ) THEN $ $ ALTER 'UPARTN USET,PG','' $ IF ( APP <> 'NLST' ) EQUIVX PG/PGX/ALWAYS $ UPARTN USET,PGX/PM0,,,/'G'/'M'/'S'/1 $ $ COMPILE CFORCE ALTER 'MPYAD KXY,UY,PX','' $ IF ( APP='STATICS' or APP='NLST' OR APP='BKL0' OR (APP='CYC' AND APP1='STATICS') ) MPYAD KXY,UY,PX/QX///-1 $ @ 7260; 69.1; Nonlinear Transient Heat Transfer - DELAY, CHBDY - Incorrect Loa d Output

If the DELAY bulk data entry is used to define an offset in time for the applied load. The actual heat load is applied correctly, and the resultant temperature value is correct. However in the data recovery for the heat flux of the HBDY element the load is applied incorrectly. Avoidance: Ignore the applied load output for the HBDY elements. @ 7266; 69; DMAP - MALTER - Extra output The MALTER statement produces a listing of the altered subDMAP. This printout cannot be suppressed even if COMPILER NOLIST is specified. However, COMPILER REF and COMPILER DECK produce the correct results. Avoidance: There is no avoidance. @ 7270; 69; Acoustic Analysis - PARAM,WTMASS - Wrong Answers PARAM,WTMASS is applied to both the fluid and the structure, even though the mass of the fluid is in different units. Avoidance: a. use mass density for density values and ensure that all mass input values are in mass units. b. or insert the following DMAP alter in V68.2: compile semg noref nolist alter 'MJJX,/-1/WTMASS','' $ type parm,,cs,n,cwtmass,mwtmass $ IF (NOMGG>=0) then $ dbstatus VGFS//S,N,NOVGFS $ IF ( NOVGFS>0 and wtmass<>1.0 ) THEN $ EMA GPECT,MDICT,MELM,BGPDTS,SILS,CSTMS,MDICTP,MELMP/ MJJX1,/-1 $ MASS matgen ,/ig/6/lusets/0/lusets $ gx1 matrix cwtmass=cmplx(wtmass) $ add ig,vgfs/wtmass/cwtmass//3 $ matmod wtmass,,,,,/wtmassg,/28 $ gxg mpyad mjjx1,wtmassg,/mjjx $ ELSE $ EMA GPECT,MDICT,MELM,BGPDTS,SILS,CSTMS,MDICTP,MELMP/ MJJX,/-1/WTMASS $ MASS ENDIF $ endIF $ NOMGG>=0 @ 7276; ; Composites - PARAM,NOCOMP - Inconsistent Stresses Homogeneous stresses, output when using 0 or -1 for PARAM,NOCOMPS, for PCOMP-based elements are calculated inconsistently between linear and non-linear solutions. Since homogeneous stresses are based upon smeared laminate properties, both calculations yield only approximate the lamina stresses (see also current error list number 1818) but both solutions should use the same method. The two methods currently being used are: For linear analysis:

S(z) = G1 * E - z * G2 * C while for nonlinear analysis: S(z) = G1 * ( E - z * C ) where: C = curvature (chi) E = reference plane strain (epsilon) G = modulus matrix S = stress z = distance from reference plane @ 7284; (2)

(1)

69; Miscellaneous - DIAG - DIAGs Out of Range Not Checked I don't think this needs to be an external error report. Please read text below for error.

@ 7288; 69; Nonlinear Buckling - MPC, RBE*, RBAR - Wrong Answers If MPC's or rigid elements are used in a nonlinear buckling analysis, then there is an error in the nonlinear buckling load factor (alpha computed by DMAP in SOL 66 or 106). This error causes the critical buckling factor (UIM 9040) to be estimated incorrectly. The remainder of the solution (eigenvalues and displacements) is OK. Avoidance: Use the following DMAP alter: compile nlstatic noref nolist $ alter 'MPYAD DULV,DPL0,/NUM3/1','' $ mpyad dugni,dpgni,/num3/1 $ @ 7289; 3032 69; Optimization/Design Sensitivity - CELASi, TEMP(LOAD), SSG1 - UFM USER FATAL MESSAGE 3032 may occur in optimization solution sequence when temperature loads are selected. The error will occur when the following three conditions are true: 1) A structural model contains scalar elements such as CELAS1. 2) The grid points connected to these scalar elements are subjected to a temperature load. 3) Only these scalar elements are designed and no other structural elements are designed. Note that scalar elements do not support temperature loading. Avoidance: 1) Make a dummy design variable and related it to a non scalar structural element in the model. Then impose the same lower and upper limits on the dummy design variables.

2) Replace the scalar elements with non scalar elements, i.e., use CROD instead of CELAS1. @ 7290; ; Nonlinear Analysis - QUAD4, TRIA3 - Update Documentation The existing documentation , such as MSC/NASTRAN V68 Reference Manual, Section 15.3.9 Small Strain Elements, needs to be updated to indicate that the QUAD4 and the TRIA3 act differently than the other "small strain" elements in that they satisfy equilibrium in the deformed configuration. This difference can have an increase in accuracy of internal loads in nonlinear analysis if elements strains are significant. @ 7291; 69; Eigenvalue Analysis - REIGL, EIGRL - Poor Performance, The REIGL module, when solving for normal modes of a vibration problem, under certain rare circumstances will perform many negative shifts and ultimately fail to find the modes requested by the user. Although one negative shift often proves useful in picking up rigid body modes, additional negative shifts are unnecessary and wasteful of computer resources. One way to duplicate the problem is to specify on the EIGRL card, F1=blank, F2=very large, ND=blank. If the F2 value is several orders of magnitude greater than the largest eigenvalue in the model, the shift logic will usually fail with the multiple negative shifts. @ 7294; 2001; Miscellaneous - SUPORT1 - UFM 316 If a value of 0 (zero) is entered for the component number on a SUPORT1 entry USER FATAL MESSAGE 316 will result. USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA CARD SUPORT1 *** Avoidance: If you are referencing SPOINTs on the SUPORT1 entry, then use 1 (one) or blank. @ 7295; 70.5; Documentation - SYSTEM(196), SCRSAVE - No Documentation on Lancz os System Cell There is no documentation in the Quick Reference Guide on the use of SYSTEM(196) (a.k.a. SCRSAVE). This system cell control reuse of scratch files when segment logic is invoked. System(196) is useful only when multiple frequency segments are requested on a Lanczos run. (Multiple frequency segments can be requested via the NUMS field in the EIGRL entry and by system(197).) Each frequency segment requires a minimum of three scratch files. When multiple frequency segments are used on a single processor computer then each frequency segment is solved serially. In this case, it makes sense to let segment #2 use the scratch files which were used by segment #1 since work for segment #1 has been completed (otherwise it wouldn't be working on #2). Similarly, when work for segment #2 is finished, segment #3 should be able to use #2's scratch files. System(196)=1 allows such file reuse. Note that when multiple processors at the Lanczos segment level is selected, i.e. system(107) (a.k.a. parallel) is set greater than 1,

system(196)=1 may cause much grief. Parallel processing may be retained for other operations but deselected specifically for Lanczos segments by adding 32768 to the number of processors. Note also that MSC/NASTRAN Version 70 has parallelism at the segment level disabled. Therefore system(196)=1 is a safe default. @ 7298; 70.7; Hyperelastic Elements - QUAD4, MATHP, PLPLANE - UFM 4297 Small hyperelastic QUAD4 elements may incorrectly fail with the message: USER FATAL MESSAGE 4297 (EQDGED) QUAD ELEMENT WITH ID = XX HAS AN INTERIOR ANGLE POSSIBLY GREATER THAN OR EQUAL TO 180 DEGREES. event though the geometry is actually good. The error is caused by a bad test on elements with lengths less than 1.0E-3. Avoidance: The only avoidance is to use different units to make the distances larger. @ 7300; ; Nonlinear Analysis - CGAP - Poor Results If the adaptive GAP element indicates an update of an element stiffness AFTER the program has converged on a solution, the program FAILS to re-iterate upon that newly updated stiffness matrix hence the results are based upon the GAP stiffness prior to the update. This error has been encountered when the adaptive GAP gets the following message: *** NEW STIFFNESS MATRIX IS REQUIRED BY GAP FOR PENALTY VALUE ADJUSTMENT *** PENALTY VALUES ARE ADJUSTED BY YYY FOR ELID = xx AFTER the program has converged on the solution. Avoidance: Increase the number of increments, NINC on the nlparm entry, or adjust the value of the stiffness on the offending PGAP or both. @ 7304; 69; Nonlinear Analysis - Creep - UIM 6193, UIM 6194, UFM 4551, Faile d Convergence Glenn; What exactly was corrected for this error? All I know is that a creep problem didn't converge in Version 68. Is there an error in all creep problems? Does it have to do with specific elements, multiple subcases, DT or what? Regards Kevin Kilroy @ 7305; 70; Documentation - OTIME - Extra Output, OTIME Not Effective in Non linear Transient The OTIME case control command is used to limit output in transient analysis. This command is not effective for nonlinear transient stress output. @ 7306; 70; Online Documentation - Common Questions and Answers The line "NASTRAN SCRATCH(MEM=n)" on page 267 of MSC/NASTRAN Common

Questions and Answers (Second Edition) in the MSC/NASTRAN Encyclopedia, page 144 in the printed text, should be: "INIT SCRATCH(MEM=n)" @ 7307; 2001; Dynamic Analysis - UWM 4684 LOADSET ID NOT FOUND In a dynamic response analysis, when the bulk data contains LOAD entries that are not referenced by the CASE control and if STRESS output is requested, then the following warning message is issued. USER WARNING MESSAGE NUMBER 4684 LOADSET ID NOT FOUND FOR SUBCASE Avoidance: Remove the non referenced LOAD entries from the bulk data. @ 7312; 69; Optimization/Design Sensitivity - DOM11, GEOM1 - SFM 3001 When only scalar points (no GRID entries) are defined in an optimization model and the SEQP module is not executed to resequence the model, then there is no GEOM1 generated. The run subsequently fails in the DOM11 module when it tries to make a copy of GEOM1 with the message "SYSTEM FATAL MESSAGE 3001 MODULE = DOM11 GEOM1 AT POSITION 12 DOES NOT EXIST". Avoidance : Include a dummy GRID entry. @ 7314; 69; Performance - VECPLOT - Poor Performance On problems with more than 65,000 dofs, the vecplot module is inefficient, and may cost as much as the rest of the calculations. It is used in several places, primarily as a means of providing checks such as applied load resultants. Avoidance: If you observe that the amount of time being spent in this module is excessive you may insert an alter to delete this module call. The alter for all Version 68.2 structured solutions is: compile selg nolist noref $ alter 63,63 $ @ 7320; 69; Heat Transfer Analysis - CHBDYE, VIEW - SFM 3007 The message "SYSTEM FATAL MESSAGE 3007 MODULE=VIEW, ILLEGAL INPUT TO SUBROUTINE VNASIN" will occur if using the backside on the CHBDYP LINE element, for example: CHBDYP + + 1111 0 -1.0 1.0 0.0 101 11 LINE 111 2 3 1

Avodiance: Specifying the VIEWID on the frontside resolves this problem. @ 7326; 70; Documentation - FORCE(PLOT) - Missing Documentation The MSC/NASTRAN Quick Reference Guide does not document the FORCE command (PLOT) option. This option may be used if element force calculations are desired for post-processing without printing them. @ 7331; ; Postprocessing - PBCOMP - Formatted OUTPUT2 on DEC Alpha dies Wr

iting EPT If an EPT table containing entries for more than two PBCOMPs is written to a formatted OUTPUT2 file on a DEC Alpha system, then the job will terminate with either no message or a "time: command terminated abnormally" message in the .log file. Avoidance include: 1) Using PARAM, 2) Writing to a 3) Altering the the OUTPUT2 statement. 4) Running on a @ 7333; 69; Frequency Response and Random Analysis - TETRA, PSDF - UFM 3052 Random response analysis that requests PSDF xyplots of TETRA stress results may result in the following fatal message: USER WARNING MESSAGE 3052, A RANDOM REQUEST FOR CURVE TYPE - STRE -, POINT 1 COMPONENT - 12 -, SPECIFIES TOO LARGE A COMPONENT ID. THE LAST COMPONENT WILL BE USED. The XYPLOT summary will then indicate that the correct plot is used, but wrong results will be presented on the XYPLOT summary. In fact, if 6 plots are selected, with items codes 6,7,8,9,10 and 11, the summary will say they all have the same RMS value, even though they are asking for normalx, normaly, normalz, tauxy, tauxz, and tauyz, which cannot be the same. Running a similar problem with a HEXA instead of TETRAs, produces good results. @ 7335; 69; Optimization/Design Sensitivity - PBEAM - Wrong Answers Errors occur in design optimization of beam elements if the following two conditions are met: 1.) The beam section is constant (e.g., end b data is not supplied on the PBEAM entry) and 2.) End B data is referenced on a DVPREL1 entry. By not supplying end B data, the code is mislead to believe the section is constant, even if this is in conflict with information supplied in the design model. Avoidance: Supply end B data in the case of a tapered beam, and all necessary intermediate stations in the case of a variable beam section. @ 7338; 69; SSSALTER - resflexa.v682 An error in the resflexa.v682 alter in sssalter directory may result in loadings applied to the residual structure not being included in the residual vector calculation. Avoidance: Add the following line after the line containing 'type db,pss $' OGEOM, NO. formatted OUTPUT2 file. OUT2GEOM subdmap and removing EPT from different platform

TYPE DB,PL @ 7339; ; Nonlinear Transient Analysis - NLTRD2, ADAPT - SFM 4276 EC 1044 Nonlinear transient analysis fails in the module NLTRD2 with the message "SYSTEM FATAL ERROR 4276.OPEN ERROR CODE 1044". This error happens when the adapt method is used if 1. There are no nonlinear elements in the model. 2. Stress output is requested. 3. The NO field of the TSTEPNL entry is greater than 2. Avoidance: Use the default (1) for NO field of the TSTEPNL entry. @ 7340; ; DMAP - LTU - SFM 3001 User DMAP runs submitted with scr=yes may fail with SYSTEM FATAL MESSAGE 3001. This error may result if the LTU logic deletes a datablock even though it is still needed. Avoidance: Run the job using SCR=NO or create a copy of the datablock and use the copy in the DMAP alter. @ 7342; 70; Optimization/Design Sensitivity - STATSUB - Wrong Answers Wrong buckling sensitivities may occur when the following conditions are true: 1. DSNOKD is nonzero, 2. The Subcase Control command STATSUB does not refer to the first subcase and 3. The referred subcase does not have retained responses of its own. Avoidance: Define an artificial response with very large upper and lower limits and force retention of the response in the referred subcase using DRESP1, DCONSTR and DSCREEN entries. @ 7356; 2001; P-Elements - DATAREC - UFM 6352 If a model contains output requests for p-elements (i.e., a DATAREC and OUTPUT entry), but no p-elements are defined in the model, UFM 6352 may occur. Avoidance: Either create a p-element in the model or remove the p-element output requests. @ 7358; 69.1; Installation - User Modifiable NASTRAN - CRAY C90 Only The CD for User Modifiable MSC/NASTRAN Version 68.2 contains errors in the unicos and unicosc90 directories. 1) The unicos*/undynamic* files are misnamed. They are renamed as follows: mv mv mv mv unicos/undynamic_7X.o unicos/undynamic_70.o unicos/undynamic_8X.o unicos/undynamic_80.o unicosc90/undynamic_7X.o unicosc90/undynamic_70.o unicosc90/undynamic_8X.o unicosc90/undynamic_80.o

2) The unicosc90/umnld script is incorrect. All references to "unicos" must be changed to "unicosc90".

@ 7365; tion ; Heat Transfer Analysis - VIEW3D, CHBDYE - No VIEW Factor Calcula The CHBDYE element references the conduction element for the geometry. If the user attempts to calculate the view factor exchange between two plates by using a line element such as a bar or rod and the CHBDYE element which references the rod element for the surface information, then no view factor calculation will be performed on the CHBDYE line element. . Avoidance: Use the CHBDYP,LINE to model the surfaces instead of the CHBDYE element. @ 7367; 70; Documentation - Design Sensitivity/Optimization, RBEi, MPC's, DV GRID - UFM 6499 It is a limitation of MSC/NASTRAN, that the sensitivities are incorrect for shape optimization runs which have DVGRID entries which reference grid points on MPC or RBEi entries. Often, the sensitivities are 0.0 and therefore the following message is produced "ONE OR MORE VIOLATED CONSTRAINTS HAS A ZERO GRADIENT **** NO PROGRESS IS POSSIBLE. OPTIMIZATION TERMINATED *** USER FATAL MESSAGE 6499 (DOM9) THE DOT OPTIMIZATION ALGORITHM HAS ENCOUNTERED AN ERROR THAT MAKES FURTHER PROGRESS IMPOSSIBLE. SEE MESSAGES ABOVE". This limitation should be documented under the DVGRID entry. Avoidance: It may be possible to replace the RBEi or MPC relation with an element which has an extremely large stiffness. @ 7376; ; p-Elements - CORD3G - SFM 6333 (EQNINI) When a convective coordinate system (cord3g) is specified as a stress output coordinate system in OUTRCV, OFP crashes with SFM 6333 (EQNINI) Avoidance: Other than not referencing a cord3g entry, there is no avoidance. @ 7380; 69; Documentation - QUAD4, PARAM,S1M, PARAM,CURV,1 - User Output Coo rd. Systems QUAD4 corner stresses do not support output in a user defined coordinate system using PARAM,CURV,1 and PARAM,S1M,1 and a MCSID coordinate system on a MAT1 or MAT2 entry. The MAT1 and MAT2 Quick Reference guide documentation needs to be updated as follows: The MAT1 remark 7 needs to changed to: 7. MCSID must be nonzero if PARAM,CURV is specified to calculate stresses or strains at grid points on plate and shell elements only. This feature does not support QUAD4 corner stresses.

The MAT2 remark 6 needs to changed to: 6. MCSID must be nonzero if PARAM,CURV is specified to calculate stresses or strains at grid points. This feature does not support QUAD4 corner stresses. @ 7391; 69; Heat Transfer Analysis - RADCAV - SFM 4276, Infinite Loop In SOL 153, if global view factors are specified (i.e. the continuation entry on the RADCAV is used), then the run may go into an infinite loop or issue SYSTEM FATAL MESSAGE 4276 for some set definitions. Avoidance: Do not use global view factors or redefine which elements are contained in sets selected by the RADCAV entry. @ 7397; 69; Documentation - QRG, APP COUPLED - Incorrect Reference to Therma l User's Guide The MSC/NASTRAN Version 68 Quick reference guide makes the following refererence under the description of the APP Executive Control Command: APP COUPLED Indicates that a coupled analysis combining heat transfer and structural analysis in SOL 153. See the MSC/NASTRAN Thermal Analysis Users Guide. In fact there is no description of the coupled thermal analysis in the Version 68 Thermal Analysis User's Guide. @ 7403; 70.7; Optimization - PCOMP - UFM in SDRCOMP, Corner Stress Request The presence of composites in SOL 200 when corner stresses are requested will cause the following unnumbered fatal message in the SDRCOMP module: USER FATAL MESSAGE. INCONSISTENT DATA DETECTED IN THE OES1A LAMINATE STRAIN AND OEF1A ELEMENT FORCE DATABLOCKS In V70.5 and later, an error message similar to the following may be output instead from the SDRCOMP module: *** SYSTEM FATAL ERROR 4276.GREAD ERROR CODE 3001 PID= 0

Avoidance: Remvoe the CORNER, BILIN, SGAGE or CUBIC option from your STRESS or FORCE output requests. However if noncomposite load responses are not used, the following DMAP alters will allow noncomposite corner stress responses and composite center stress and failure index responses to be used. But it will cause non-composite force responses to fail. For V68.2: compile sdr2stat noref nolist alter 'oes1as' SDR2 CASEDS,CSTMS,MPTS,DIT,EQEXINS,sils,ETTDC,EDT,BGPDTS,, QGX,UGX,ESTDC,,,,,gpsnts/ ,,,,OEF1AS,/'STATICS'//-2 $ alter 'sdrcomp','' SDRCOMP CASEDS,MPTS,EPTS,ETTDC,ESTDC,OES1AS,OEF1AS,DIT/ ES1CS,EFITS,OEF1AA/0/1 $ alter 'sdrcomp'(2),''

SDRCOMP For v70:

CASEDS,MPTS,EPTS,ETTDC,ESTDC,OES1AS,OEF1AS,DIT/ EA1CS,EFITS,OEF1AA/1/1 $

compile sdr2stat noref nolist alter 'oes1as' SDR2 CASEDS,CSTMS,MPTS,DIT,EQEXINS,,ETTDC,EDT,BGPDTS,, QGX,UGX,ESTDC,,OINT,PELSETS,VIEWTB,GPSNTS,DEQATN,DEQIND,/ ,,,,oef1as,/'STATICS'//-2/////ACOUT/PREFDB/// 1/TRUE/FALSE $ forces alter 'sdrcomp','' SDRCOMP CASEDS,MPTS,EPTS,ETTDC,ESTDC,OES1AS,oef1as,DIT,BGPDTS/ ES1CS,EFITS,OEF1AA/0/1 $ alter 'sdrcomp'(2),'' SDRCOMP CASEDS,MPTS,EPTS,ETTDC,ESTDC,OES1AS,oef1as,DIT,BGPDTS/ ES1CS,EFITS,OEF1AA/0/1 $ @ 7404; ime 69; p-elements - GP4 - SFM 3008, Insufficient Memory, Excessive Runt When converting a large h-element model directly to a p-element model, the run may fail in GP4 with the following error messages: SYSTEM FATAL MESSAGE 3008 MODULE = GP4 0INSUFFICIENT MEMORY AVAILABLE FOR SUBROUTINE CSTRAN 0FATAL ERROR The runtime in GP4 will also be excessive. This error is most likely to occur when there are many grid points which reference non-basic output coordinate systems. Avoidance: Other than increasing the memory on the mem keyword, there is no other known avoidances. The amount of memory required for a large problem can be substantial in this case. @ 7409; 69; Heat Transfer Analysis - SOL 101 - SFM 4276. FREAD EC 3002 Solution 101 heat transfer (app heat) superelement runs will fail in solution 101 with the following message: SYSTEM FATAL ERROR 4276.FREAD ERROR CODE 3002

Avoidance: Other than no running superelements there is no avoidance. @ 7410; 69; Eigenvalue Analysis - EIGRL - Wrong Answers If the following EIGRL format EIGRL,SID,V1,,ND is used, and the first flexible mode is substantially higher than the rigid body modes, then modes below V1 may also be included in the eigenvalue table printout. Avoidance: Either increase V1 or include a shift scale that is close to

the estimated first desired mode. @ 7414; ; Optimization/Design Sensitivity - Poorly Labeled Output It is difficult to relate the output in solution 200 (DISPLACEMENT, STRESSES, SENSITIVITY, ETC ) to the design cycle. The output headers do not include the design cycle number for which the output is printed. Avoidance: Look for the "USER INFORMATION MESSAGE 9051" , which contains the design cycle information preceding the output. @ 7425; ; DMAP - CALL - Incorrect PARAM Value Passed Back If a parameter is passed into a subDMAP on the CALL statement, it is possible that the program may pass a value of 0.0 back to the calling subDMAP for that parameter. This has only been observed if the following conditions are met: 1) The parameter appears on the CALL statement without the 'S' feature 2) The parameter is defined as being able to be changed in the bulk data or case control ( 'Y' option on the TYPE statement in the called subDMAP') 3) The same name is used for the parameter in the calling and called subDMAPs 4) The value of the parameter is changed in the called subDMAP Avoidance: Change the name of the parameter in the called subDMAP or use the 'N' option on the TYPE statement in the called subDMAP. @ 7428; 70; Documentation/Elements - CGAP, CORDCi, CORDSi - Wrong Answers If a cylindrical (CORD1C/CORD2C) or spherical (CORD1S/CORD2S) coordinate system is specified for a gap element (CID field of CGAP entry), and one of the connected GRIDS is defined at the center of that coordinate system, wrong answers result. Wrong answers result, even if the GRID at the center of the coordinate system is defined in a cartesian system. Avoidance: Displace the GRID at the center by as little as 1.E-40 in the direction of the desired gap axis, or define the v vector and leave the CID field blank. @ 7452; from xdb 69; Postprocessing - Optimization, TRANS - Missing Datablock CHISTM

MSC/NASTRAN Version 68.2 on CRAY computers fails to write the CHISTM datablock to the .xdb file. An incorrect xdb will show the following when the resulting xdb file is written to neutral format with the TRANS utility: . . DATA BLOCK CHIST CQD4 CQD4 CQD4 CTAB PROJECT 1 1 1 1 1 SE ID 0 0 0 0 0 PATH 1 1 1 1 1 SUBCASE 0 0 0 0 0 SET ID APPROACH 0 0 0 0 0 0 0 0 0 0

CTAB2 DCYCT DHIST DINFO DISPR DISPR DISPR DOINFO DOINFO DVECT . . . .

1 1 1 1 1 1 1 0 1 1

0 0 0 0 0 0 0 0 0 0

1 1 1 1 1 1 1 0 1 1

0 0 0 0 10 10 10 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 1 1 1 0 0 0

Note that the correct xdb has a CHISTM datablock and two DINFOs, but the incorrect xdb has no CHISTM and three DINFOs. @ 7454; 7777 69; Nonlinear Analysis - PLSOLID, PLPLANE, TEMP(LOAD) - SFM 4276, EC Hyperelastic elements with temperature loads do not work on MSC/NASTRAN Version 68 Alpha OSF/1 system. All jobs with temperature loads and hyperelastic elements will fail with the message "SYSTEM FATAL MESSAGE 4276, ERROR CODE 7777" in the NLITER Module. Avoidance: @ 7460; 69; Optimization/Design Sensitivity - PCOMP, SUBCOM - SFM 3007 The message, "SYSTEM FATAL MESSAGE 3007 MODULE = DSAL" is produced if: (1)there are active DRESP1 responses associated with composites (CSTRESS,CSTRAIN,CFAILURE), (2) SUBCOM's are used and (3) One or more of the subcases used in the SUBCOM has no active responses. Avoidance: An avoidance is to add responses in the subcases referenced on the subcom and force their retention using DSCREEN. @ 7462; 69; Heat Transfer - CHBDYG, RADBC - Wrong Answers If a negative EID is specified on the RADBC entry to indicate that the boundary condition is to be applied to the back face of the element, then the load applied is incorrect and wrong answers result. Avoidance: Define another CHBDYG or CHBDYE on the backside and use it to apply the load. Other than running on another machine there is no avoidance.

@ 7463; 70.5; Stress Discontinuities - STDCON - Inefficient Processing The STDCON module can be very inefficient. For some problems, enhancements in Version 70.5 have improved the performance of this module by several orders of magnitude. @ 7464; able 69; P-Elements - CQUAD8, CTRIA6 - SFM 6398 (GPOI) Preliminary edge t

If there are h-element CQUAD8s and/or CTRIA6s next to p-elements, and there are midsides on the common edges, the following message may result: SYSTEM FATAL MESSAGE 6398, (GP0I), PRELIMINARY EDGE TABLE CREATED IN GP0B IS INVALID. USER ACTION: CONTACT CLIENT SUPPORT AT MSC. The error message should be more specific. The error is considered corrected in Version 69 by issuing the following message: USER FATAL MESSAGE 6544, (GP0I), THE EDGE WITH GRIDS ID = **, ** ON P-ELEMENT ID = ** IS SHARED BY AN H-ELEMENT WITH MID SIDE GRID.

@ 7469; 2001.0.1; Frequency Response Analysis - Random - SFM 4276 EC 3001, Poor Message In a random analysis, the J and K fields (fields 3 and 4) on the RANDPS entry should reference the subcase identification number, as documented in the Quick Reference Guide. However, if the J and K fields reference the subcase sequence number instead, the job will fail with the following unclear messages. BIOMSG: ERROR 3001 HAS OCCURRED IN ROUTINE FREAD 0 SYSTEM FATAL ERROR 4276.FREAD ERROR CODE , FILE INDEX (FILEX) = 0

3001 PID=

Avoidance: Modify the J and K fields on the RANDPS entry to reference the appropriate subcase identification numbers. @ 7470; 69; Geometric Nonlinear Analysis - ELAS3, ELAS4 - Wrong Answers CELAS3 and CELAS4 elements produce wrong answers in SOL 64. Avoidance: Use CELAS1 or CELAS2 elements. @ 7476; 69; Heat Transfer - EMG, FTUBE, MATT4 - Quits in EMG If the data on the TABLEM1 referenced by a MATT4 entry is poorly defined such that the extrapolated viscosity versus temperature results in a negative viscosity (which is physically impossible) then the run fails without message in the EMG module. Avoidance: Make sure that the viscosity versus temperature curve always remains in the (+,+) quadrant. @ 7480; ; DMAP - Delivery Database Build - SFE 4276 in XCSA When building a delivery database in more than one run, if any subDMAP fails compilation in the first run, which also compiles the NDDL, then a subsequent run may fail with BIOMSG error 3001 in GREAD and SYSTEM FATAL MESSAGE 4276 in XCSA.

Avoidance: If this message is encountered, the delivery database is corrupted and must be rebuilt. To avoid corrupting the database again, do not compile any subDMAPs in the first run. @ 7481; R When building a delivery database in more than one run, if a subDMAP which is called by a main subDMAP fails during compilation in one run and then a link of the main subDMAP is attempted in the next run, then the run will fail with BIOMSG error 3001 in GREAD and SYSTEM FATAL MESSAGE 4276 in LINKER. The run should have failed with USER FATAL MESSAGE 443 or 444. Avoidance: Recompile the offending subDMAP and relink the main subDMAPs which call the offending subDMAP. @ 7495; 69; P-elements - VIEWP, TETRA, FEEDGE - Excessive CPU For large models of TETRA p-elements with curved edges (i.e. FEEDGE bulk data entries), the VIEWP module takes excessive CPU time. The time spent in the VIEWP for TETRA p-elements with curved edges is approximately ten times a model that does not have curved edges. The overall run time may be increased approximately 30 percent. Avoidance: There is no avoidance. @ 7498; 69; Transient Heat Transfer - DPD - UFM 2066, 2079, SWM 2072 If the grid points are given TEMP for the initial conditions without the use of a TEMPD bulk data entry, then you the following error message is encountered: USER FATAL MESSAGE 2079, GRID OR SCALAR POINT ******** HAS AN UNDEFINED COORDINATE REFERENCED ON A NOLINI CARD. USER FATAL MESSAGE 2066, UNDEFINED GRID POINT 0 ON DAREA-, RLOAD- OR TLOAD- CARD. SYSTEM WARNING MESSAGE 2072, CARD TYPE 7107 NOT FOUND ON DATA BLOCK. Avoidance: Add a TEMPD bulk data entry. @ 7511; 69; Eigenvalue Analysis - REIGL, EIGRL - Wrong Answers In rare cases Lanczos may produce wrong answers and no warning message is issued. MSC/NASTRAN Version 69 corrects this situation by issuing appropriate messages. Avoidance: Prior to Version 69, it is the user's responsibility to verify that the actual eigenvalues found agree with all sturm checks. If the results do not agree then they can not be correct. The eigrl entry should be modified by increasing maxset to 15 from the default of seven and the job should be re-run. Changing the frequency range of interest such that f1 is blank, not 0.0, often corrects the problem. @ 7513; 70.5; Optimization/Design Sensitivity - DOM6, DRESP2, PARAM,OPTEXIT SFM 3001 When only DRESP2 responses are defined in an optimization job, the requested OPTEXIT=4 or -4 results in two different errors: ; DMAP - Delivery Database Build - Poor Message, SFE 4276 in LINKE

1. In Version 68.x, SYSTEM FATAL MESSAGE 3001 occurs in DOM6. 2. In Versions 69 and 70, the DSCM2 matrix is not generated in an f06 file. Avoidance: There is no avoidance for v68.x. A dmap avoidance is available for v69 and v70: compile exitopt alter 'if ( ((','if ( ((' if ( ((dsprint or dsunform or dsexport) and mdtrkflg=0) and (NORSP2>0 or nodscmg>0) ) THEN $ endalter @ 7515; 69.0.4; Direct Transient Analysis - CBAR,PLOAD1,LOADSET/LSEQ - Wrong Ans wers In a direct transient dynamic analysis, if a CBAR element is loaded by a PLOAD1 entry through the LOADSET/LSEQ combination, then the output titled "force distribution in bar elements" will be wrong. Avoidance: There are two possible avoidances: 1. Replace the CBAR element with a CBEAM element. 2. Replace the PLOAD1/LOADSET/LSEQ combination with equivalent point loads using the DLOAD/TLOADi/DAREA combination. @ 7518; utput 70.5; Post-processing - PARAM,POST,-5, PARAM,OUGCORD,BASIC - Missing O If PARAM,POST,-5 is specified with PARAM,OUGCORD,BASIC then PARAM,POST,-5 is effectively ignored and the expected output will be missing. Avoidance: Insert following DMAP alter: compile sedrcvr alter 'ELSE IF ( NOT(post','' ELSE IF ( post>-3 and PATVER<3. and OUGCORDX='BASIC' ) THEN $ @ 7539; 69; Optimization/Design Sensitivity - PARTN - UFM 2166 If a statics and a modes subcase are attempted in the same geometric shape optimization run, then the following fatal message is issued, "USER FATAL MESSAGE 2166, MATRIX TO BE PARTITIONED IS OF SIZE 22806 ROWS BY 7 COLUMNS. ROW PARTITION SIZE IS 22806 COLUMN PARTITION SIZE IS 5 (INCOMPATIBLE)." Avoidance: Include the following DMAP avoidance: compile bndshp nolist noref alter 'shpcas' type parm,,i,n,nt $ MDCASE CASEXXu/ CASESTAT,,,,,,,,,,,,,/ S,N,STATCC $ tabpt casestat/ $ paraml casestat//'trailer'/1/s,n,nstat $

paraml ygbndr//'trailer'/1/s,n,nyg $ nt=nstat+nyg $ delete /cvec,,,, $ matgen ,/cvec/6/nt/nstat/nyg $ @ 7542; in DOM9 69; Optimization/Design Sensitivity - DRESP1, DOM9 - Unexpected EOF

If a data deck has multiple weight or multiple volume responses that are referenced by the objective or are constrained, unexpected results occur and the answers are wrong. System Fatal Messages such as "Unexpected end of file in DOM9" may also occur. Avoidance: Remove the multiple DRESP entries. @ 7546; 69; Optimization/Design Sensitivity - DRESP1 - SFM 7127 In an optimization run, if a series of DRESP1 entries makes alternating references between different types (such as PSHELL and ELEM) in the PTYPE field, then the message "SYSTEM FATAL MESSAGE 7127 (DOPR3B) AN ELEMENT WHICH REFERENCES PTYPE **** ** PID *** IS NOT DEFINED IN THE ANALYSIS MODEL" is issued. Avoidance: Put all DRESP1's that reference a particular type in consecutive order, where the order is given by the DRESP1 ID. @ 7547; ; Documentation - Slip Ratio - Reference Manual Documentation that defines the meaning of "Slip Ratio" for gap and slideline elements does not exist. The following paragraph should be added to the Section 15.3.12 of the Reference Manual. The slip ratio is defined as the (shear force)/(coefficient of friction * normal force). Thus, a ratio of 0.0 indicates no shear (or no contact), a slip ratio of 1.0 indicates slipping, and a value in between 0.0 and 1.0 is a normalized value which indicates how close to impending slipping the shear force is." @ 7550; 69; Optimization/Design Sensitivity - PARAM,MODEPT - Parameter not s et Correctly The MODEPT parameter in SOL 200 is intended to indicate whether the design model EPT (element property table) differs from the analysis model. There are a number of implementations errors associated with this parameter: The DOPR1 module has an error which results in the MODEPT parameter being set to true only when the FINAL designed property differs from the analysis value. The DESOPT driver does not correctly pick up the MODEPT value set in subDMAP PREDOM because of an error in the call to PREDOM. The design intent should be that the EPTN datablock is only equivalenced to EPT when MODEPT is true. Currently it is always equivalenced with the results that the finite element analyses on the database gets destroyed on a restart when they should not.

@ 7552; stems 69; Installation - MSC/NASTRAN will not make the icons on Beta NT sy For Windows NT Version 4.0, which is the Windows 95 interface on top of Windows NT, the installation of MSC/NASTRAN will not create the icons or group for the Beta versions of this Operating System. Avoidance: Do not upgrade your operating system to 4.0 until the non Beta software is available. Or manually create the group (MSC) and create the icons in the group. @ 7553; 70; Restarts - - PBCOMP - SWM 2072 CARD TYPE 3002 If a model contains a pbcomp entry, and a restart is performed going from SOL 103 to SOl 111 on a DEC AlPHA OSF machine, then the job may terminate with the following messages in the f06 file. SWM 2072 - CARD type 3002 On some DEC ALPHA machines, no messages may appear in the f06 file. In all cases, the following messages appear in the log file: time: command terminated abnormally Avoidance: There are two possible avoidances: 1. Perform the analysis in one pass using solution 111 without using solution 103 for restart. 2. Replace the PBCOMP with an equivalent PBEAM entry. 3. Replace the PBCOMP record in the NDDL with the following: RECORD=PBCOMP(5403,55,349), PID,I, MID,I, A,RS, I1,RS, I2,RS, I12,RS, J,RS, NSM,RS, K1,RS, K2,RS, M1,RS, M2,RS, N1,RS, N2,RS, nsect(c),i, either,back(nsect)=0, undef,20, or,1, entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1, endentry,count=nsect, or,2, entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1, endentry,count=nsect, or,3, entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1, endentry,count=nsect, or,4, entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1, endentry,count=nsect, or,5, entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1, endentry,count=nsect, or,6, entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1, endentry,count=nsect, or,7, entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,

endentry,count=nsect, or,8, entry=section, Y,RS, Z,RS, endentry,count=nsect, or,9, entry=section, Y,RS, Z,RS, endentry,count=nsect, or,10, entry=section, Y,RS, Z,RS, endentry,count=nsect, or,11, entry=section, Y,RS, Z,RS, endentry,count=nsect, or,12, entry=section, Y,RS, Z,RS, endentry,count=nsect, or,13, entry=section, Y,RS, Z,RS, endentry,count=nsect, or,14, entry=section, Y,RS, Z,RS, endentry,count=nsect, or,15, entry=section, Y,RS, Z,RS, endentry,count=nsect, or,16, entry=section, Y,RS, Z,RS, endentry,count=nsect, or,17, entry=section, Y,RS, Z,RS, endentry,count=nsect, or,18, entry=section, Y,RS, Z,RS, endentry,count=nsect, or,19, entry=section, Y,RS, Z,RS, endentry,count=nsect, or,20, entry=section, Y,RS, Z,RS, endentry,count=nsect, endeith, EOR, @

C,RS, MID,I, undef,1, C,RS, MID,I, undef,1, C,RS, MID,I, undef,1, C,RS, MID,I, undef,1, C,RS, MID,I, undef,1, C,RS, MID,I, undef,1, C,RS, MID,I, undef,1, C,RS, MID,I, undef,1, C,RS, MID,I, undef,1, C,RS, MID,I, undef,1, C,RS, MID,I, undef,1, C,RS, MID,I, undef,1, C,RS, MID,I, undef,1,

7555; 68.2.9; Miscellaneous - PUNCH - SFM 4276, Invalid Floating Operation A request for PUNCH output may result in the following traceback on Fujitsu computers running MSC/NASTRAN Version 68.2: jwe0292i-e Invalid floating operation exception was detected. execution mode = advanced error occurs at ofppun_ line 156 loc 00857c10 offset 00001800 ofppun_ at loc 00856410 called from loc 0083be68 in ofp_ line 1340 ofp_ at loc 008366f0 called from loc 00006eb8 in xsem14_ line 124 xsem14_ at loc 00006850 called from loc 00000928 in MAIN__ line 108 MAIN__ at loc 00000500 called from o.s. taken to (standard) corrective action, execution continuing.

Avoidance: The error can be avoided by setting system(210)=1. It is recommended that this system cell be set in the system rc file of all Fujitsu systems. Please be aware that other machines have experienced problems when system(210) is set. These problems are very rare however, and are not expected to effect the Fujitsu system.

@ 7556; s When an excitation is made at zero frequency for a model with rigid body modes, the contribution from the rigid body modes (which is theoretically infinite) is discarded. Only the finite modal displacements from modal DOFs with non-zero natural frequencies are retained. An xy plot of element stresses from zero to small excitation frequencies will be continuous. An xy plot of displacements will grow rapidly as zero excitation frequency is approached, then drop to values of the same magnitude as response at frequencies away from natural frequencies. The solution at zero frequency is analogous to an inertia relief solution, while the other solutions are made in an inertial basis. The following remark should be added under the FREQi entries: Remark 4. In modal analysis, solutions for modal DOFs from rigid body modes at zero excitation frequencies may be discarded. Solutions for nonzero modes are retained @ 7557; 70; Documentation - PBCOMP, I12 - Incorrect Default The default value for the I12 field of the PBCOMP entry is 0.0 and not I1*I2-(I12)**2 as described in the MSC/NASTRAN Quick Reference Guide. The correct description for I12 is as follows: I12 Area product of inertia. See Remark 6. (Real, default=0.0, but I1*I2-I12**2 > 0.0). @ 7561; 70; Documentation - BUFFPOOL - UFM 1286 Remark 2d in the section "How Can I Reduce the Size of the SCRATCH DBset?" of the Common Questions and Answers Guide, Second Edition, should be changed from d. Increase the available memory to the functional modules by setting the "smem" and BUFFPOOL parameters to zero...... to d. Increase the available memory to the functional modules by setting the "smem" parameter to zero...... Starting in Version 68, the default solver is the sparse solver. BUFFPOOL cannot be turned off with the sparse solver. @ 7564; 69; Frequency Response/Random Analysis - RANDOM, K2PP - SFM 4276 FRE 70; Documentation - Frequency Response - Response of Rigid Body Mode

AD When K2PP matrices or FREQ looping are present in random analysis, the RANDOM module fails. The problem has been identified as the wrong case control data block being passed to the RANDOM module. The RANDOM module should get CASEXX instead of CASECC. The following is a DMAP avoidance. compile sol30 or SOL26 alter 'random','' RANDOM XYCDB,DIT,PSDL,OUGV2,OPG2,OQG2,OES2,OEF2,CASExx/PSDF,AUTO/ S,N,NORAND $ @ 7583; 68.2.8; Miscellaneous - INPUTT4 - Various Fatal Messages MSC/NASTRAN may not be able to read INPUTT4 files on certain machines. Known problem machines are Fujitu and DEC ALPHA. The Fujitsu systems gave the following traceback in the INPUTT4 module: error occurs pack_ line 359 inptt4_ line 125 xsem02_ line 84 MAIN__ error summary at pack_ line 113 loc 00e577b0 offset 00000360 at loc 00e57450 called from loc 00712bb0 in inptt4_ at loc 00711580 called from loc 000024c0 in xsem02_ at loc 00001e60 called from loc 00000868 in MAIN__ at loc 00000500 called from o.s. (fortran77 ex system)

While the DEC ALPHA OSF system completed but issued the following messages: Unaligned access pid=15155 <analysis> va=14011d924 pc=1201febe0 ra=1201feb4c type=ldt Unaligned access pid=15155 <analysis> va=14011d924 pc=1201febe0 ra=1201feb4c type=ldt @ 7591; 70; Databases - MATREDU - Wasted Space in DBALL The module MATREDU is using the DBALL as its scratch space instead of SCR300 or the regular SCRATCH space. Avoidance: There is no avoidance. @ 7599; 69; Documentation - PARAM,TSTATIC - Can Be Used in Case Control The MSC/NASTRAN Version 68 Quick Reference Guide states that param,tstatic can only be used in the Bulk Data section. It can be used in the Bulk Data and Case Control. @ 7600; 70.5; Complex Eigenvalue Analysis - EIGC, CLAN - Missing Roots A complex lanczos eigenvalue analysis of a model with zero damping will likely not find all the roots that are found by a real lanczos eigenvalue analysis with the same shifts. Likewise in models with damping, complex

lanczos is likely to have missing roots. Avoidance: The only avoidance is to add additional shift points. In Version 69.1 the epsilon has been adjusted so that correct behavior should be achieved for real world models. Thus, a Version 69.1 run is likely to find more roots then the same run in 69.0. @ 7615; 70.5; DMAP - ALTER - UFM 3063 If a DMAP contains a long alter followed by short alters followed by a long alter the following message may be encountered: SYSTEM FATAL MESSAGE 3063 MODULE = PREFACE MEMORY ALLOCATION FAILURE IN SUBROUTINE XALTER ERROR CODE: -1 MEMORY CORRUPTION FOUND. PREVIOUSLY ALLOCATED MEMORY AREAS HAVE BEEN INCORRECTLY MODIFIED. USER ACTION: CONTACT MSC CLIENT SUPPORT. Avoidance: There are two avoidances. The user can either re-arrange the alters so that the long and short alters are placed together or split alters that are more than 112 lines into smaller segments. @ 7616; 69; Aeroelastic Analysis - APD, PFAERO - No Message, Wrong Answers Specifying a theta-1 array and no theta-2 array for an aerodynamic interference body may yield results that do not include all the interference body effects. For example, if a PAERO2 specifies that an interference body has 3 elements, of which elements 1 and 2 specify a theta-1 array, and element 3 does not, the program will run and yield wrong results. Avoidance: The wrong results can be avoid by properly specifying all interference elements with a theta array. @ 7617; ; Loads - PLOAD2, PLOAD4 - SFM 2105, SFM 4577, Missing Elements If PLOAD2 or PLOAD4 entries specify a range of element ids with the "thru" option and an element does not exist for each id in the range, the job will fail with SYSTEM FATAL MESSAGE 2105 and SYSTEM FATAL MESSAGE 4577. This limitation is not clearly documented and limits the usefulness of the thru option and should be corrected. @ 7619; ; Documentation - Nonlinear Analysis Data Recovery Restarts The current documentation does not make it clear how one should prepare input files to perform data recovery restarts with a nonlinear analysis. If the user expects to obtain any data recovery from solution 106, 129, 153 or 129, the input file should be prepared to perform the complete solution and not rely on the database to contain the necessary datablocks. For example, flux output must be explicitly requested in the cold start if one wants to obtain it on the restart. The user may specify flux(plot) if he chooses not to see the printed output in the cold start. @ 7627; 69; Heat Transfer Analysis - VIEW - Poor Performance, SFM 4276 Heat Transfer analysis that involves third body shadowing calculations for

a large number of surface element in the VIEW module may experience poor performance or fail with the message SYSTEM FATAL MESSAGE 4276. Avoidance: There is no guaranteed avoidance; however, if parallel processing is being selected, deselecting parallel may avoid the problem. Furthermore, even though the error is possible on all machine types a specific problem may run on one machine but not another. Thus, running on another machine may avoid the problem. @ 7632; 70.7; Design Sensitivity - SENSITY, SOL 103 - UFM 4287 Material sensitivity analysis in SOL 103 fails on Cray computers with the following message, "USER FATAL MESSAGE 4287 , ILLEGAL THICKNESS FOR PLATE ELEMENT ID = xxx". Avoidance: Use SOL 200. @ 7633; mization 69; Optimization/Design Sensitivity - Premature Convergence for Opti

As explained in Section 3.4 of the MSC/NASTRAN User's Guide for Design Sensitivity and Optimization, a design task is considered converged if the objective is not changing and the design is feasible. This has been found to be sometimes unconservative in the case where the previous design was unfeasible. That is, if the design changes from being infeasible to feasible without a weight change, it may be that further iterations will produce a further improvement in the objective. Avoidance: If the next to last design was infeasible, as indicated by its maximum constraint value being greater that GMAX, it is worthwhile to start a new design task starting from the final design of the run using the DESVAR values from the punch file. @ 7637; 69; Case Control - ECHO=NOSORT - Missing Output If ECHO=nosort is used in the case control section, bulk data echo is not printed to the output. Avoidance: use ECHO=unsort. @ 7644; GL; Transient Heat Transfer - CELAS2, TEMPBC,TRAN - UFM 315 In transient thermal analysis, if the user has CELAS2 entries in the input test deck and there is also TEMPBC,TRAN bulk data entries, then the following fatal message will be encountered: 0*** USER FATAL MESSAGE 315 (IFPDRV) FORMAT ERROR ON BULK DATA ENTRY TEMPBC SORTED CARD COUNT = 761

102

This error occurs because the TEMPB,TRAN bulk data entry internally produce CELAS2 entries for the large conductance method. These entries then conflict with explicitly defined CELAS2 entries.

Avoidance: Instead of CELAS2 bulk data entry uses CELAS1,and PELAS to define the thermal conductance value. @ 7655; ors 70.6; Nonlinear Eigenvalue Analysis - PARAM,NMLOOP - Missing Eigenvect If a nonlinear normal modes analysis (param,nmloop) is requested in a SOL 106 run, and there are more than one subcase and the displacement output is not requested for all the subcases, then the nth eigenvector will be missing for the nth subcase that does not have a displacement output request. In the following example, eigenvectors 1,3,4,5,6,7,8,9, and 10 will be printed. Eigenvector number 2 will be missing. sol 106 . subcase 1 load 1 nlparm = 10 disp = all subcase 20 load = 2 nlparm = 20 subcase 35 load = 3 nlparm = 30 disp = all Begin bulk . param,nmloop,20 eigrl,10,,,10 . Avoidance: Request displacement output for all subcases. @ 7659; 69.1; DMAP - COMPILER - Various Failures If an "ELSE" statement appears in a DMAP without a matching set of IF and ENDIF statements, various random failures may occur. In one instance, a "FILPOS" error exit was taken upon entering the subdmap where the error was made. In another instance, a USER FATAL MESSAGE was issued indicating that there was "NO MASS MATRIX AVAILABLE FOR EIGENVALUE ANALYSIS"." Avoidance: There is no avoidance only the recommendation that the statements inside if-endif clauses be indented to insure that all else statements are bracketed by if-endif statements. DIAG 44 is also useful for pinpointing the location of the error when this type of error is suspected. @ 7660; ; P-Elements/Superelements - GP4 - UFM 2101A The following error message will be issued if a grid point appearing on a p-element edge is constrained and not in an upstream superelement. *** USER FATAL MESSAGE 2101A MODULE = GP4 GRID POINT xxxx COMPONENT y ILLEGALLY DEFINED IN SETS UM UB

For superelements, there is a general restriction in p-elements for boundary degrees of freedom ( exterior degrees of freedom in superelements). That is there is an internal MPC created for SPCs on edges and faces (GMSPC, GMBC). Consequently, the constrained degrees of freedom are subject to the rules for MPCs. The dependent degrees of freedom (m-set) have to be in the upstream superelement, they can not be exterior. @ 7668; 70; Documentation - PCOMP, TREF - Missing Default The PCOMP bulk data entry description in the MSC/NASTRAN Quick Reference Guide does not give the default for TREF field. The default is "0.0". In addition, remark 3 would be clearer if instead of: 3. TREF given on the PCOMP entry will be used for all plies of the element; it will override values supplied on material entries for individual plies. If the PCOMP references temperature dependent material properties, then TREF given on the PCOMP will be used as the temperature to determine material properties. TEMPERATURE Case Control commands are ignored for deriving the equivalent PSHELL and MATi entries used to describe the composite element. it said the following: 3. The TREF specified on the material entries referenced by plies are not used. Instead TREF on the PCOMP entry is used for all plies of the element. If not specified, it defaults to "0.0". If the PCOMP references temperature dependent material properties, then the TREF given on the PCOMP will be used as the temperature to determine material properties. Note that MAT8 data entries do not support temperature dependent material properties. TEMPERATURE Case Control commands are ignored for deriving the equivalent PSHELL and MATi entries used to describe the composite element. @ 7678; 69; Matrix Methods - Sparse DECOMP - UFM 6137 RANK DEFICIENT Sparse decomposition in Version 69 has been changed to all rank deficient, singular matrices, to go through. The action taken is now the same as the old active column decomposition, i.e., it puts a very small term on the diagonal and proceeds with the solution. The user is then warned about the degrees of freedom causing the high maxratio. @ 7680; rs 69.1; Rigid Elements and Multipoint Constraints - RSSCON - Wrong Answe The RSSCON element produces wrong answers if the shell grid is not on the straight edge connecting the two solid grids. This is evident because the element in this condition will not pass a rigid body check. The magnitude of the error is related to the magnitude of the offset. In Version 69 an offset up to 5% of the height of the solid is allowed. The mpc relations generated by the element are adjusted as if the grid point was moved onto the straight line but the actual geometry is not adjusted.

Avoidance: Other than manually moving the grid points so that they line along the straight line there is no avoidance. In V69.1, the shell grid point is moved on to the straight edge connecting the two solid grids if the shell grid lies close to the edge within a tolerance. The tolerance is user modifiable with PARAM, TOLRSC, eps where eps is 0.05 by default. The absolute value of the tolerance is then eps times the length of the solid edge. Shell grid points outside the tolerance will cause a fatal error. @ 7682; 69.0.2; Elements - RSSCON - UWM 6695, References Incorrect GRID IDs If a shell grid point lies outside the straight line between the two connecting solid grid points when using the RSSCON element, then the following warning messages will be issued. *** USER WARNING MESSAGE 6695 (CONSSD ) RSSCON NUMBER aaa WILL MPC ONLY GRID bbb OF QUAD ELEMENT ccc TO HEXA ELEMENT ddd IT WILL NOT MPC GRID ddd; THE CORNER SHELL GRIDS MUST LIE ON OPPOSITE EDGES OF A FACE OF THE SOLID ELEMENT. This is the proper message to issue; however, the error message may reference an incorrect offending shell grid point "bbb". Avoidance: Modify the shell grid point to lie on a straight line between the two connecting solid grid points. @ 7689; 69.0.2; Optimization - PBARL - Various Fatal Messages The new capability in MSC/NASTRAN Version 69 which allows users to define their own beam cross sections has two undocumented limitations. There is a limit of 250 user defined beam cross section types, and there is a limit of 7 dimensions, including the Non-Structural Mass (NSM) on user defined sections. If either of these limits are exceeded various error types may be generated. In optimization, the fatal message may be issued from the module DOPR1H, or it may be issued from the API with the phrase buffer overflow. Avoidance: Do not exceed the limits. The limits have been raised to 2000 beam cross section types and unlimited dimensions on a beam section for the next release. @ 7690; 70.5; Documentation - DMAP Module Dictionary - MATMOD Option 5 Example 3 of the MATMOD option 5 documentation is incorrect. This example shows how to transform KGG from a global to a basic coordinate system. Following the documentation results in the wrong matrix. In the SMPYAD shown, the last and not the first TRANSG matrix should be transposed. However, SMPYAD will not let you transpose the last matrix of a triple multiply, so you must transpose TRANSG before SMPYAD. The incorrect and corrected documentation follow: Incorrect:

matmod cstm,sil,bgpdt,,,/transg,/5//-1 $ smpyad transg,kgg,transgt,,,/kggbasic/3////1 $ Correct: matmod cstm,sil,bgpdt,,,/transg,/5//-1 $ trnsp transg/transgt $ smpyad transg,kgg,transgt,,,/kggbasic/3//// $ No transpose flags set @ 7692; 70; Superelements - MATGPR - Overflow (******) When using the enhanced superelement features (parts), the boundary points of the upstream superelements may be sent downstream. Internally, these boundary points are assigned very large ID numbers, numbers that are much larger than the user can specify on the grid entries. These large numbers do not fit in the format for the MATGPR module. If MATGPR is used with the enhanced superelement, the IDs for these grids will be printed as (*******). The matrix values are, however, correct. Avoidance: Use MATPRN instead of MATGPR and use PARAM,USETPRT and the boundary sequence table to identical the external degree of freedom. @ 7697; 70; Heat Transfer - SPC - Poor message, UFM MATMOD-4 Currently the transient heat transfer analysis capability does not support multiple SPC conditions. An analysis can only have one SPC condition apply through all subcases. If multiple boundary conditions are attempted the message "USER FATAL MESSAGE MATMOD-4" will be issued. Avoidance: Multiple boundary conditions must be executed in separate runs. @ 7699; 69.0.2; P-Elements - GP0, GMLOAD, FEFACE - UFM 6380 If the GMLOAD entry references an FEFACE entry that is undefined the following fatal message is correctly issued but the FEFACE ID is not properly identified. USER FATAL MESSAGE 6380, (GP0F), THE FEFACE ID = 0 REFERENCED BY GMLOAD IS UNDEFINED.

Avoidance: There is no avoidance to make the code identify the incorrect entries. The user must find and correct the input error. @ 7705; 70; Documentation - EIGRL - Performance/Crashes The MSC/NASTRAN Version 69 Quick Reference Guide description of the EIGRL entry documents a new parallel Lanczos capability. This new capability was not documented in the Version 69 Release Guide because it did not meet MSC's strict standards for quality. The capability and documentation has several errors: 1. The Quick Reference Guide fails to mention that the new capability is only available on the CRAY and SGI systems. 2. The following FATAL messages have occurred while using this capability: *** SYSTEM FATAL ERROR 4276.SFUPFI *** USER FATAL MESSAGE 815 (RDINIT)

(Note that bad answers have never been encountered as a result of this capability.) 3. The purpose of this capability was to reduce the elapsed time. It has been found that parallel lanczos helps reduce elapsed time on a very small set of problems, and that elapsed time actually INCREASES for many runs. Avoidance: Do not set the NUMSEG keyword (a.k.a. SYSTEM(197). Do not set NUMSEQ on the EIGRL entry. @ 7706; 69.1; SSSALTER - evds200a.v69 - UFM 149 If the evds200a.v69 sssalter is used, the job will fail with the following messages: *** USER FATAL MESSAGE 149 (LINKAR ) CALL STATEMENT IFPS IN SUBDMAP IFPL HAS AN ARGUMENT LIST OF DIFFERENT LENGTH THAN THE CALLED SUBDMAP *** SYSTEM INFORMATION MESSAGE 480 (SEMINT) JOB TERMINATED DUE TO ABOVE ERRORS IN THE DMAP/NDDL LINKAGE EDITOR (LINKER) Avoidance: Replace the following statements: type parm,,logical,n,noedom2,ambs9,nogoifp9,ngoxsrt9 type parm,,char8,n,dball9 $ call IFPS ,,bulk2/BULK9, $ GEOM1Q,EPT,MPT,EDT,DIT,DYNAMIC,GEOM2,GEOM3,GEOM4, GEOM1Q9,,MPT9,EDT9,DIT9,DYNAMIC9,GEOM29,GEOM39,GEOM49, $ MATPOOL,AXIC,DEFUSET,DMI,DMINDX,DTI,DTINDX,PVTX, MATPOOL9,AXIC9,DEFUSET9,DMI9,DMINDX9,DTI9,DTINDX9, edom2,DEQATN9,DEQIND9,CONTACT9,OINT9,EPTA9/ DBALL9/AMBS9/SEID9/AUXMID9/NOGOIFP9/NGOXSRT9/noedom2 $ $ with ,

type parm,,logical,n,noedom2,ambs9,nogoifp9,ngoxsrt9,selist9 type parm,,char8,n,dball9 $ call IFPS ,,bulk2/BULK9, $ GEOM1Q,EPT,MPT,EDT,DIT,DYNAMIC,GEOM2,GEOM3,GEOM4, GEOM1Q9,,MPT9,EDT9,DIT9,DYNAMIC9,GEOM29,GEOM39,GEOM49, $ MATPOOL,AXIC,DEFUSET,DMI,DMINDX,DTI,DTINDX,PVTX, MATPOOL9,AXIC9,DEFUSET9,DMI9,DMINDX9,DTI9,DTINDX9, edom2,DEQATN9,DEQIND9,CONTACT9,OINT9,EPTA9/ ,

DBALL9/AMBS9/SEID9/AUXMID9/NOGOIFP9/NGOXSRT9/noedom2/ selist9 $ $ @ 7707; 69.1; Nonlinear Analysis - CBUSH - Incorrect SPCFORCE Output Nonlinear static analysis does not include the effect of the CBUSH element forces in the forces of single point constraint output. Avoidance: There is no avoidance.

@ 7711; 70; Frequency Response - SEUPPLOT - SFM 4276, Time Command Terminate d Abnormally IF an SEUPPLOT request is made in solution 108, the job may fail with the message time command terminated abnormally or, SYSTEM FATAL MESSAGE 4276. Avoidance: Use SEPLOT instead of SEUPPLOT. @ 7713; 69; P-elements - FEFACE, CIDBC, MODGM4 - Stops With No Message In MSC/NASTRAN Version 68.2 on the DEC/ALPHA, if a secondary coordinate system is referenced on a FEFACE (CIDBC field), then the job may terminate in the MODGM4 module without any error messages. Another symptom of this error is the generation of a core file in the submittal directory. Avoidance: If feasible, do not reference a secondary coordinate system on the CIDBC field. @ 7715; 69.0.2; P-Elements - GMINTC, GMBNDC - SFM 6425 (EMA) On machines that enforce boundary alignment such as the HP 9000 the following fatal message may be issued when the GMINTC entry is used: *** SYSTEM FATAL MESSAGE 6425 (EMA). THE KDICTP TABLE AND KDICTP TRAILER ARE INCONSISTENT. USER ACTION: CONTACT MSC CLIENT SUPPORT. This error does not occur on the IBM/RS6000 or the CRAY machines. Avoidance: There is no avoidance. @ 7716; 69.0.2; Miscellaneous - SYSTEM(210)=1 - Excessive Output SYSTEM(210) can be used to correct punch errors on many systems; however when SYSTEM(210)=1 is used on the Sun/Solaris machines, excessive output is generated causing the .f06 and/or .pch file to be very large. Avoidance: There are no known problems that require Sun/Solaris systems to set system(210)=1. Thus, on this system do not specify SYSTEM(210)=1. @

7717;

69.1; Database - Neutral File Transfer with p-elements, GP2 - SFM 3001 The following message is issued when an attemt is made to load the neutral database (created by dbunload) via sol DBTRANS. This error occurs only if the database datablocks are related to p-element analysis. SYSTEM FATAL MESSAGE 3001 MODULE = GP2 THE INPUT DATA BLOCK NAMED EQEXINS AT POSITION 2 DOES NOT EXIST.

Avoidance: There is a rather complex avoidance which requires an NDDL recompilation; however, it is not deemed practical. @ 7718; 69.1; Virtual Mass - MPRES - Wrong Answers, SFM 4276 EC 7777 If MPRES output is requested and PARAM,DDRMM,-1 is not specified wrong answers or a SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777 may appear. Avoidance: In Version 68 simply specify PARAM,DDRMM,-1. For Version 69 the following alter from the sssalter library will improve efficiency and correct the error. $ FILE VMEA.V69 $ LAST REVISED 17 DEC 96 M. A. GOCKEL $ VIRTUAL MASS EFFICIENCY IMPROVEMENTS, FOR SOLS 103,111,112. $ FOR NAST69 DATED 10/18/96, WITH DEL. DB TIME STAMP OF 961018160119 ECHOOFF $ COMPILE SEMRM nolist NOREF $ ALTER 'CALL MGEN1','' $ $ MODERNIZE NAMES EQUIVX MCHI/CEE/ALWAYS $ EQUIVX MCHI2/BEC/ALWAYS $ EQUIVX MLAM/MEE/ALWAYS $ EQUIVX MLAM2/MEC/ALWAYS $ EQUIVX GEG/GGE/ALWAYS $ CALL DBSTORE CEE,MEE,BEC,MEC,MAR//SEID/0/' '/0 $ FOR MGEN2 CALL FDRMGEN1 USET,GM,GOA/RGA $ IF (NOA>-1) THEN $ MPYAD GGE,RGA,/GEA/1 $ ELSE $ TRNSP GGE/GEA $ ENDIF $ VECPLOT ,,BGPDTS,EQEXINS,CSTMS,,,,/RBGRT/GRDPNT//4 $ RB VECTORS FROM $ GEOMETRY. For weight summary. Always made, but summary output $ must be requested by parm, grdpnt TRNSP RBGRT/RBGR $ RBGR HAS G ROWS, 6 COLUMNS MPYAD GGE,RBGR,/GER/1 $ CALL DBSTORE GEA,GER,,,//SEID/0/' '/0 $ FOR MODERS IF (SEID>0 AND NOOSET>-1) THEN $ CALL MGEN2 CASES,/MVMAA/SEID $ ADD MLAA1,MVMAA/MLAA $ ENDIF $ COMPILE MODERS nolist NOREF $ ALTER 'REIGL KXX'(,1) $ AFTER ENDIF FOR READ OR REIGL CALL MGENRS PHIx,USET,LAMA,CASES// $ PHIAZ AND LAMA ARE UPDATED COMPILE SEDISP NOLIST NOREF $ REMOVE FOR V69.1 ALTER 293,299 $ OLD MGEN BRANCH MOVED TO SEDRCVR COMPILE SEDRCVR noLIST NOREF $ REVISE FOR V69.1 ALTER 43 $ MOVE MDATA FROM SEDISP, AFTER SDR2 FOR NON-DDRMM $ CHECK FOR VIRTUAL MASS PRESSURE OUTPUT REQUESTS. LIFTED FROM NAST69T2 $ REMOVE MOST OF THIS WHEN NAST69T1 SUPPORTED

CALL DBFETCH /MEA,MAR,,,/SEID/0/0/0/0 $ AVOID CALL CHANGES DBSTATUS Mea//S,N,NOMGEN $ IF ( NOMGEN>0 AND (APP='FREQRESP' OR APP='TRANRESP' OR APP='REIG ' OR APP='CEIGEN ') ) THEN $ PARAML USET//'USET'//////'A'/S,N,NOASET $ IF ( LUSETS=NOASET ) THEN $ EQUIVX UG/UL1/-1 $ ELSE $ UPARTN USET,UG/UL1,,,/'G'/'A'/'S'/1 $ ENDIF $ MDATA CASEDR,,MAR,MEA,UL1,OL2/OEP1/APP $ ENDIF $ NOMGEN>0 ALTER 166,166 $ ADD OEP1 SDR3 OUG1F,OQMG1 ,OEP1,,,/OUG2F,OQMG2,OEP2,,, $ ALTER 226 $ INSERTIONS OFP OEP1//S,N,CARDNO $ ALTER 184 $ BEFORE OTHER SORT2 OFPS OFP OEP2//S,N,CARDNO $ ALTER 196 $ IF ( NOMGEN>0 AND (APP='FREQRESP' OR APP='TRANRESP' OR APP='REIG ' OR APP='CEIGEN ') AND APP1<>'MMREIG' ) THEN $ XYTRAN XYCDBDR,,,,OEP2,/XYPPLOT/ APP/'DSET'/S,N,PFILE/S,N,CARDNO/S,N,NOPPLOT $ IF ( NOPPLOT>-1 ) XYPLOT XYPPLOT// $ ENDIF $ NOMGEN>0 COMPILE FDRMGEN1 NOLIST NOREF $ SUBDMAP FDRMGEN1 USET,GM,GO/RGA $ TYPE PARM,,I,N,NOOSET,NOMSET,NOA,NOSSET $ $ $ PRODUCES REDUCTION FROM G-SET TO F-SET $ PARAML USET//'USET'//////'G'/S,N,NOGSET/'A'/S,N,NOASET/ 'F'/S,N,NOFSET/'N'/S,N,NONSET/'O'/S,N,NOOSET $ IF (NOGSET=NOASET) NOA=-1 $ IF ( NOA=-1 ) THEN $ MATGEN, /RGA/1/NOGSET $ RETURN $ ENDIF $ IF ( NOMSET>-1 ) THEN $ MATGEN ,/INN/1/NONSET $ UMERGE1 USET,GM,INN,,/RGN/'G'/'M'/'N'/1 $ ENDIF $ NOMSET>-1 $ IF ( NOSSET>-1 ) THEN $ MATGEN ,/IFF/1/NOFSET $ UMERGE1 USET,,IFF,,/RNF/'N'/'S'/'F'/1 $ ENDIF $ NOSSET>-1 $ IF ( NOMSET>-1 AND NOSSET>-1 ) THEN $ MPYAD RGN,RNF,/RGF $ ELSE IF ( NOMSET>-1 ) THEN $ EQUIVX RGN/RGF/-1 $ ELSE IF ( NOSSET>-1 ) THEN $ EQUIVX RNF/RGF/-1 $ ENDIF $ IF ( NOOSET>-1 ) THEN $ MATGEN ,/IAA/1/NOASET $ UMERGE1 USET,GO,IAA,,/RFA/'F'/'O'/'A'/1 $

IF ( NOMSET>-1 OR NOSSET>-1 ) THEN $ MPYAD RGF,RFA,/RGA $ ELSE $ EQUIVX RFA/RGA/-1 $ ENDIF $ ELSE $ EQUIVX RGF/RGA/-1 $ ENDIF $ NOOSET>-1 RETURN $ END $ FDRMGEN1 COMPILE MGEN2 nolist NOREF $ SUBDMAP MGEN2 CASES,PHIAZ/MavZZ/SEID $ $ COMPUTES VIRTUAL MASS IN REDUCED BASIS, ALSO RIGID BODY MASS OUTPUT $ OLD NOTATION NEW NOTATION $ MCHI CEE $ MCHI2 BEC $ MLAM MEE $ MLAM2 MEC TYPE PARM,,I,N,NF,N1,N2,NX,NZP6,ny,nw $ TYPE PARM,,I,Y,GRDPNT=-1 $ TYPE PARM,,RS,Y,WTMASS=1.0 $ TYPE PARM,,CS,N,MASSWT $ TYPE PARM,NDDL,I,N,SEID $ type parm,,char8,y,solvit='no' $ CALL DBFETCH /cee,,,,/seid/0/0/0/0 $ paraml cee//'presence'////s,n,nocee $ if (nocee=-1) return $ CALL DBFETCH /ger,MEE,BEC,MEC,GEA/seid/0/0/0/0 $ PARAML PHIAZ//'PRESENCE'////S,N,NX $ PARAML CASES//'DTI'/-1/194//S,N,MPRESID $ MAKE MEA ONLY IF MPRES REQUESTED $ this logic somewhat flawed for superelements. if mpres anywhere, $ it goes to all se subcases, including se0. It shouldn't. IF (NX>-1) THEN $ MPYAD GEA,PHIAZ,/PRHEZ $ ELSE $ EQUIVX GEA/PRHEZ/ALWAYS $ ENDIF $ PARAML CEE//'TRAILER'/1/S,N,NE $ NUMBER OF FLUID ELEMENTS PARAML BEC//'PRESENCE'////S,N,NC $ NUMBER OF CAVITIES PARAML PRHEZ//'TRAILER'/1/S,N,NZ $ NUMBER OF MODES or a-set PARAML GEA//'TRAILER'/1/S,N,NA $ SIZE OF A-SET NF = NE $ equivx prhez/pez/always $ temp EQUIVX PRHEZ/PRHFZ/NC $ EQUIVX GER/GFR/NC $ append pez,ger/pey $ Ny = NZ + 6 $ IF (MPRESID<>0) THEN $ APPEND GEA TO THE MODAL SHAPES nw = Ny + NA $ MATGEN ,/vwya/6/Nw/Ny/NA $ APPEND Pey,GeA/Pew $ ELSE $ EQUIVX Pey/Pew/ALWAYS $ ENDIF $ IF (NC>-1) THEN $ CLOSED VOLUME OPTION PARAML BEC//'TRAILER'/1/S,N,NC $ NUMBER OF CAVITIES NF = NE + NC $ N1 = NE - NC $ N2 = 2*NC $

MATGEN, /VFEC/6/NF/NE/NC $ CALL MAKVE1C BEC/VE1C/NE/N1/NC $ MATGEN, /VC/6/NC/0/NC $ UNIT VECTOR MATGEN, /V2CC/6/N2/NC/NC $ MERGE VE1C,VC,,,,VFEC/VF12/0 $ partn pew,,ve1c/p1w,pcw,,/0 $ nx = 2 * nc $ matgen,/icc/1/nc $ ENDIF $ IF (NC=-1) THEN $ DO THE FREE SURFACE METHOD if (solvit><'yes') then $ putsys (0,209) $ use old solver solve cee,pew,,,/uew $ putsys (1,209) $ restore new solver else $ matmod pew,,,,,/va01,/12/s,n,nonull/1 $ 1-set is null columns if (nonull>-1) then $ partn pew,va01,/pe0,,pe1,/0 $ remove null columns else $ equivx pew/pe0/always $ endif $ type parm,,i,n, unit6 $ unit6 = getsys(unit6,2) $ see where .f06 going now. v69 only putsys(50,2) $ send .f06 to unit 50. makes large, useless file solvit cee,pe0,,,,,,,,,/ue0,,//1//////-1 $ jacobi preconditioner putsys(unit6,2) $ reset .f06 to original file if (nonull>-1) then $ merge ue0,,,,va01,/uew/0 $ insert null columns else $ equivx ue0/uew/always $ endif $ endif $ mpyad mee,uew,/mew $ ELSE $ DO THE ENCLOSED FLUID METHOD merge pcw,,,,,v2cc/p2w/0 $ PARTN BEC,,VE1C/B1C,BCC,,/1 $ TRNSP B1C/B1CT $ TRNSP BCC/BCCT $ PARTN CEE,VE1C,/C11,CC1,C1C,CCC/-1 $ type parm,,i,n,old209 $ old209 = getsys(old209,209) $ putsys (0,209) $ use old solver DECOMP C11/L11,U11, $ putsys (old209,209) $ restore new solver FBS L11,U11,C1C/X1C $ trnsp x1c/x1ct $ trnsp c11/ct11 $ decomp ct11/lt11,ut11, $ trnsp ccc/ccct $ MPYAD X1C,X1C,ICC/IXCC/1 $ SOLVE IXCC,x1ct,,,/xxcc $ trnsp cc1/cc1t $ MPYAD C11,Cc1t,/AX1C $ MPYAD C1C,CCCt,AX1C/A1C $ MPYAD CC1,Cc1t,/AXCC $ MPYAD CCC,CCCt,AXCC/ACC $ MERGE A1C,,B1C,,V2CC,/A12/0 $ TRNSP A12/A21 $ MERGE ACC,BCCT,BCC,,V2CC,/A22/-1 $ DECOMP A22/L22,U22, $

CALL FBSX L11,U11,lt11,ut11,XxCC,x1c,A12,P1w/UX2,UA21w $ MPYAD A21,UA21w,P2w/PA2w//-1 $ MINUS PRODUCT FBS L22,U22,PA2w/UB2w $ MPYAD A21,Ux2,/PB2 $ FBS L22,U22,PB2/UC2 $ MATGEN, /I22/1/N2 $ ADD I22,UC2/UD2//-1.0 $ DECOMP UD2/LD22,UD22, $ FBS LD22,UD22,UB2w/U2w $ MPYAD A12,U2w,P1w/PC1w//-1 $ MINUS PRODUCT CALL FBSX L11,U11,lt11,ut11,xXCC,x1c,PC1w,/U1w, $ partn u2w,,v2cc/ucw,qcw,,/0 $ merge u1w,ucw,,,,ve1c/uew/0 $ mpyad cee,uew,/cuw/1 $ mpyad mee,cuw,/mcuew $ mpyad mec,qcw,mcuew/mew $ ENDIF $ IF (MPRESID<>0) THEN $ PARTN mew,VwyA,/mey,,meA,/0 $ ELSE $ EQUIVX mew/mey/ALWAYS $ ENDIF $ CALL DBSTORE MEA,,,,//SEID/0/' '/0 $ MATGEN ,/VyZ6/6/Ny/NZ/6 $ PARTN mey,VyZ6,/meZ,,me6,/0 $ MPYAD PeZ,mez,/MZZ/1////6 $ IF (GRDPNT>-1) THEN $ DO WEIGHT SUMMARY mpyad ger,me6,/m66/1 $ TRNSP m66/m66t $ MASSWT = (0.5,0.0)/CMPLX(WTMASS) $ ADD m66,m66t/VMRR/MASSWT/MASSWT $ MESSAGE //'RIGID BODY VIRTUAL MASS MATRIX, IN WEIGHT UNITS' $ MATPRT VMRR // $ ENDIF $ WTMASS TRNSP MZZ/MZZT $ ADD MZZ,MZZT/MAVZZ/0.5/0.5 $ CAVALIER AVERAGING RETURN $ END $ MGEN2 COMPILE MGENRS noLIST NOREF $ SUBDMAP MGENRS PHIAZ,USET,LAMA,CASES// $ PHIAZ AND LAMA ARE UPDATED $ TO IMPROVE THE EIGENSOLUTION WITH VIRTUAL FLUID MASS EFFECTS TYPE PARM,,CHAR4,Y,DEBUG='NO' $ PARAML CASES//'DTI'/-1/166//S,N,MFLID $ if (mflid=0) return $ this doesn't work for se.s $ for some unknown reason mflid gets into se 0 case control CALL DBFETCH /cee,,,,/0/0/0/0/0 $ paraml cee//'presence'////s,n,nocee $ if (nocee=-1) return $ belt and suspenders paraml phiaz//'trailer'/2/s,n,nz $ number of modes CALL MGEN2 CASES,PHIAZ/MVMZZ/0 $ LAMX, ,LAMA/LMAT/-1 $ MATMOD LMAT,,,,,/MZ,/1/4 $ 4TH COL IS GEN. MASS MATMOD LMAT,,,,,/KZ,/1/5 $ 5TH COL IS GEN. STIFFNESS MATMOD KZ,,,,,/KZZ,/28 $ DIAGONALIZE MATMOD MZ,,,,,/MZZ1,/28 $ ADD5 MZZ1,MvmZZ,,,/MZZ $ IF (DEBUG='YES') THEN $

MESSAGE //'ROOTS WHEN ONLY STRUCTURAL MASS'/ ' PRESENT, DEBUG OUTPUT' $ OFP LAMA// $ ENDIF $ REIGL KZZ,MZZ,,,,,,,,,/LAMAZ,PHIZZ,MIZ,EIGVMATZ,/ 'MODES'/S,N,NEIGV////-1///NZ/0 $ FIND THE SAME NUMBER OF MODES MESSAGE //'IMPROVED ROOTS FOR VIRTUAL MASS' $ OFP IS AFTER ALTER MPYAD PHIAZ,PHIZZ,/PHIAZ1 $ EQUIVX PHIAZ1/PHIAZ/ALWAYS $ OVERWRITE INPUTS EQUIVX LAMAZ/LAMA/ALWAYS $ RETURN $ END $ MGENRS1 COMPILE MAKVE1C nolist NOREF $ SUBDMAP MAKVE1C BEC/VE1C/NE/N1/NC $ $ select a c-set which produces stable c11 matrices TYPE PARM,,I,N,NE,N1,NC,NX,II=1,JJ=1,NEm1,CI,IIM1,NCOMP, prec,non,nonm1,lasti $ FILE besum=save/ve1c=save $ IF (NC=1) THEN $ NEM1 = NE - 1 $ MATGEN ,/VE1C/6/NE/0/1/NEMRI $ RETURN $ ENDIF $ matgen ,/besum/6/ne/ne $ a null vector copy besum/ve1c $ DO WHILE (JJ<NC+1) $ MATMOD BEC,,,,,/BEJ,/1/JJ $ EXTRACT A COLUMN if (jj>1) then $ partn bej,,besum/bok,bng,,/0 $ bng used up on prior passes paraml bok//'presence'////s,n,boksiz $ if (boksiz>-1) paraml bok//'trailer'/5/s,n,boksiz $ if (boksiz<1) then $ non=0 $ jump usedup $ endif $ endif $ ADD BEJ,Besum/BESUMJ $ equivx besumj/besum/always $ PARAML BEj//'TRAILER'/5/S,N,NZWDSB $ PREC = PRECISON() $ NZWDSB = NzWDSB/PREC $ IF (NZWDSB<Ne) THEN $ PARTN BEj,,BEsum/BENULL,BENON,,/0 $ separate the null, non-null terms PARAML BENON//'TRAILER'/2/S,N,NON $ NUMBER OF NON-NULL TERMS ELSE $ NON = 0 $ endif $ label usedup $ if (non>0) then $ paraml bej//'trailer'/5/s,n,nzbj $ nzbj = nzbj/prec $ nx = nzbj - 1 $ matgen, /vnic/6/nzbj/0/1/nx $ 1.0 in first row, null elsewhere merge, ,vnic,,,,bej/ve1i/0 $ else $ do a dmap binary search until you find the first non-zero term partn bej,,ve1c/beok,beused,,/0 $ don't use a c-set already used paraml beok//'trailer'/2/s,n,nok $ ii = 1 do while (ii<nok+1) $

paraml beok//'dmi'/1/ii/s,n,beji $ if (beji<>0.0) then $ iim1 = ii - 1 $ nx = nok-ii matgen, /veoki/6/nok/iim1/1/nx $ merge veoki,,,,,ve1c/ve1i/0 $ jump loopiend $ endif $ ii = ii + 1 $ enddo $ label loopiend $ endif $ add ve1i,ve1c/ve1cj $ equivx ve1cj/ve1c/always $ JJ = JJ + 1 $ ENDDO $ JJ return $ END $ MAKVE1C COMPILE FBSX nolist NOREF $ SUBDMAP FBSX L11,U11,lt11,ut11,xXCC,x1c,P1,P2/U1,U2 $ $ SOLVE FOR TWO BLOCKS OF VECTORS ON ONE PASS $ [C11*C11 + C1C*CC1]*U = P, WHERE L11 IS FACTOR OF C11 $ P1 MUST BE PRESENT. P2-P4 MAY BE PURGED TYPE PARM,,I,N,N1,N2,nt,NTM2 $ PARAML P1//'TRAILER'/1/S,N,N1 $ PARAML P2//'TRAILER'/1/S,N,N2 $ NT = N1 + N2 IF (N2>0) THEN $ NTM2 = NT - N2 $ MATGEN ,/VN12/6/NT/Ntm2/N2 $ APPEND P1,P2/P12 $ ELSE $ EQUIVX P1/P12/ALWAYS $ ENDIF $ FBS L11,U11,P12/P $ MPYAD XXCc,P,/XYCZ $ MPYAD X1C,XYCZ,P/Y//-1 $ FBS Lt11,Ut11,Y/U $ IF (N2>0) THEN $ PARTN U,VN12,/U1,,U2,/0 $ ELSE $ EQUIVX U/U1/ALWAYS $ ENDIF $ RETURN $ END $ FBSX $ END OF VME ALTER PACKAGE ECHOON $

@ 7719; 70.7; Material Nonlinear Analysis - HEXA, PENTA, TEMP - Wrong Answers In non-linear analysis the HEXA and PENTA elements will give incorrect results if the element references a MATS1 or MATT1 or both and has a non-constant temperature distribution. The displacements are correct. The center stresses are correct. The Gauss/Grid Point Stresses have an incremental error which is proportional to the temperature gradient. Avoidance: There is no known avoidance.

@ 7720; ; Documentation - MSC/NASTRAN Reference Manual - Wrong Dimension The dimensioning of the channel-section as shown on page 5.2.1-8 of the Version 68 MSC/NASTRAN Reference Manual is incorrect. The b value should only extend to the median line of the web instead of all the way to the outside edge of the web. @ 7728; 69.1; Database - DBUNLOAD, DBLOAD - Inefficient Restart When a neutral database created by DBUNLOAD is migrated to CRAY and is dbloaded for the purpose of a subsequent restart, the restart job goes through unnecessary steps and recalulates the stiffness matrix. @ 7734; 69.0.2; Elements - BEAM, PBAR1 - Incorrect K2 Formulation for Beam Libra ry I-beam. The PBAR1 and PBEAM1 area shear factor term, K2, for the I-beam cross section is incorrectly calculated. It is formulated to be the area of flanges divided by the total area, when it should be 5/6 of this amount. Avoidance: The error can be avoided by specifying the correct properties on a PBAR or PBEAM entry or the user could add their own section. The error has been corrected in MSC/NASTRAN Version 69.0.2. If you have already been delivered 69.0.0 or 69.0.1, then 69.0.2 must be explicitly requested. @ 7735; 69.0.2; Heat Transfer - QVECT, CHBDYE or CHBDYG with TYPE REV - Wrong An swers When QVECT loads are applied to CHBDYE or CHBDYG elements with surface type REV and midside nodes, the orientation vectors are not calculated. This error results in zero directional heat loads for this type of element. Avoidance: There is no avoidance. The error has been corrected in Version 69.0.2. Version 69.0.2 will not be automatically delivered to clients who have received 69.0.0 or 69.0.1, it must be requested. @ 7737; 69.0.2; Optimization/Design Sensitivity - DOM9 - SFE 4276 EC 3001 In certain cases solution 200 jobs that contain DRESP2 entries may fail in the DOM9 module with the following error: *** SYSTEM FATAL ERROR 4276.GREAD ERROR CODE 3001 PID= 0

Avoidance: There is no simple avoidance. The error is corrected in MSC/NASTRAN Version 69.0.2. Version 69.0.2 will not be automatically delivered to clients who have already received Version 69.0.0 or 69.0.1. The client should request this corrected version. @ 7738; 70; Documentation - PBARL - Wrong Information Figure 1 under the PBARL description on page 793 and PBEAML on page 810 of the MSC/NASTRAN Version 69 Quick Reference Guide indicates TYPE=BOX where it should be BAR and vice versa.

@ 7740; 69.1; Superelements - SEP1X, RBAR, RBE2 - BIOMSG 3002 Superelements with RBAR and RBE2 elements may cause an error in the sep1x module which results in the message "BIOMSG 3002" being output. Avoidance: Replace the RBAR and RBE2 elements with other types of rigid elements. @ 7742; ; Superelements - MIRROR, REPEAT - Possible Wrong Answers When using the enhanced superelement features, if the mirror or repeat option is used to generate a copy of a primary superelement, and any of the GEOM4-type data (e.g., MPC, RBEi, SPCi, ASETi, OMITi, BSETi, CSETi, etc.) for the mirrored or repeated superelement is modified as compared to the primary superelement, then the answers may be wrong. Avoidance: If any of the GEOM4-type data for the mirrored or repeated superelement must be different than the primary superelement, create a new primary superelement instead of using the secondary superelement. @ 7743; GL; Submittal - notify=yes May Not Appear to Work. If a user has set the message state to off, e.g., with the 'mesg n' command, the notification message cannot be sent. This is most noticable when a job is run in the background, as the notification message is the only positive indication the job has completed. The user may receive no indication at all, or may receive the cryptic message "permission denied" when the notification is actually attempted. Avoidance: The user must enable messages for the notification message to be received under batch=yes. @ 7748; 70; Miscellaneous - Authorization - UFM 3060 If the memory specified on the 'mem' parameter is too high, then MSC/NASTRAN may abort with a UFM 3060. The MSCID printed will be 0: 0*** USER FATAL MESSAGE 3060, SUBROUTINE MODEL - OPTION NAST NOT IN APPROVED LIST. SYSTEM DATE (MM/DD/YY): 11/11/96 SYSTEM MSCID: 0 (DECIMAL) 00000000 (HEXADECIMAL) SYSTEM MODEL NUMBER: -1 , SYSTEM OS CODE: 1 Avoidance: Reduce the memory request. @ 7757; 69.1; Documentation - 'symbol' Keyword Documentation Error The documentation for the symbol keyword in Section 4.1.2 and in Appendix B.1 of the MSC/NASTRAN Version 69 Configuration and Operations Guide (UNIX) is incorrect. The correct syntax for the keyword is: symbol=<symbolic_name>=<string> where <symbolic_name> is the name of the symbol to be defined and <string> is the value of the symbol. The specific example in 4.1.2 should read:

SYMBOL=DATADIR=/dbs/data The specific example in Appendix B.1 should read: SYMBOL=DBSDIR=/dbs @ 7773; 69.0.2; Miscellaneous - Job hangs on HP-UX 10.20 when submitted through at(1). On HP-UX 10.20, an MSC/NASTRAN V69 job may hang if the job was run using the at(1) facility. There are two ways to run a job using at(1), the most obvious is to use the after keyword. Another is to create a queue command such as submit=at -q %queue% -m -f %job% %after% Once the job is submitted, it will hang in the limits special function with the following error message displayed on your window or sent in an email: sh[29]: ulimit: The specified value exceeds the user's allowable limit. Avoidance: This problem is corrected by inserting the line unset MSC_ARGV just before the line %0% limits in the file <install_dir>/bin/nast69.lcl @ 7775; 69.0.2; Plotting - PlotPS - Contour lines are scrambled Contour plots are incorrect on the Solaris Version 69 system. These plots have the contour lines going to the origin instead of to each contour point. The effect is a set of vectors going from the origin to the contour points. Avoidance: Use a post-processor such as MSC/PATRAN to view stress contours. The error has been fixed in Version 69.0.2 and Version 69.1. @ 7776; ; Superelements - BSET - UFE 4156, Poor Message When using the enhanced superelement capabilities introduced in Version 69, only the boundary points can be assigned to the BSET degree of freedoms, and this is the design intent. However, if any interior point on a part (superelement) is assigned to the BSET, then the job will fail with the following unclear error messages. 0*** USER FATAL ERROR 4156--- DIMENSIONS OR TYPE OF DATA BLOCK KLAAV , SUPERELEMENT y TRAILER=( xx xx x x x xxx) ARE INCONSISTENT WITH SUPERELEMENT MAP. Avoidance: If an interior degree of freedom on a part is desired to be placed in the BSET, then assign that grid point to the ASET. Note that

when using parts (superelements), if a degree of freedom at a grid point is assigned to the ASET, then all six degrees of freedom for that grid point will also be assigned to the ASET. @ 7780; 70.7; Executive System - ACQUIRE, DBLOAD, ENDJOB - UFM 453 If you specify the ACQUIRE NDDL, DBLOAD, and ENDJOB statements and the "del=" keyword on the submittal command then UFM 453 is issued. Avoidance: Do not use the "del=" keyword. Add ASSIGN statements for MASTERA, MSCOBJ, and MSCSOU. @ 7789; 70; Superelements - PARAM,METHCMRS - UFM 3032 If superelements are present and PARAM,METHCMRS is specified for the residual structure, then UFM 3032 will be issued stating that the program is unable to find the selected set (EIGRL entry). Avoidance: Verify that the referenced EIGRL entry exists in the input file. If it does, then include the following DMAP alter in the input file: $ $ for V69 $ compile ifpl alter 'call ifps' TYPE PARM,,I,N,DOIT=0 $ CALL DBSTORE DYNAMIC,,,,//0/0/'MASTER'/S,DOIT $ $ compile semr3 alter 'REIGL.*CMKXX,CMMXX','REIGL.*CMKXX,CMMXX' TYPE PARM,,I,N,OK $ CALL DBFETCH /DYNAMIC,,,,/0/0/0/0/S,OK $ REIGL CMKXX,CMMXX,DYNAMIC,CASES,,,,USET,EQEXINS,SILS,VYXW/ CMLAMA,CMPHIXZ,CMMI,EIGVMATS,/ 'MODE'/S,N,NOZSET////METHRS////////CMSET $ $ $ DMAP to fix error 7789 - V70.5 $ compile ifpl alter 'call ifps' TYPE PARM,,I,N,DOIT=0 $ CALL DBSTORE DYNAMIC,,,,//0/0/'MASTER'/S,DOIT $ $ compile cmpmode alter 'MATMOD.*CASES,DYNAMICS.*METHTYP.*METHC','' $ $ TYPE PARM,,I,N,OK $ CALL DBFETCH /DYNAMIC,,,,/0/0/0/0/S,OK $ MATMOD CASES,DYNAMIC,,,,/,/23/S,N,METHTYP//METHC $ alter 'READ.*CMKXX,CMMXX.*DYNAMICS','' READ CMKXX,CMMXX,,,DYNAMIC,USET,CASES,VYXW,SILS,USET,,,EQEXINS/ CMLAMA,CMPHIXZ,CMMI,CMOEIGS,EIGVMATS,/ 'MODE'/S,N,NOZSET//METHF/CMSET/METHCMRS $ @ 7791; 70.5; Databases - DBLOAD, GP1 - SFM 307 in IFPSIZ

If a neutral database is DBLOAD'ed and GP1 is executed without first running IFP, then the job will fail in GP1 with: 0*** SYSTEM FATAL MESSAGE 307 (IFPSIZ) THE NAME OF BULK DATA ENTRY CORD1C IS INVALID. PROGRAMMER ACTION:IF THE ABOVE NAME IS A VALID BULK ENTRY, THEN CHECK THE HASHING FUNCTION TABLE(S), AND MAKE SURE THERE ARE NO "LOST" ITEMS. 0*** SYSTEM FATAL MESSAGE 3007 MODULE = GP1 0ILLEGAL INPUT TO SUBROUTINE IFPSIZ 0FATAL ERROR Avoidance: Always run DBTRANS when DBLOADing a database. @ 7792; ; Documentation - Reference Manual, PBEND, FSI=3 - UWM 4260 The documentation in Volume 3 of the MSC/NASTRAN Reference Manual for USER WARNING MESSAGE 4260, BENDING CHARACTERISTIC PARAMETER EXCEEDS RANGE OF APPLICABILITY FOR ELEMENT ** indicates an acceptable range of .5 to 1.0. In fact the acceptable range is .05 to 1.0. @ 7796; GL; Superelements - PLOTS - Missing Features, Missing Plots The "LABEL GSPC" and "LABEL EPID" are not supported with the new superelement features in Version 69. @ 7798; by Default ; Documentation - Parallel Processing EMA, SYSTEM(107) - Disabled

On most of the systems supporting parallel operation with MSC/NASTRAN, the Parallel EMA function shows little or no performance gain over the non-parallel EMA. Because of this, by default, Parallel EMA will be disabled on all platforms for MSC/NASTRAN Version 69. If a you wish to enable PEMA, you may do so by setting the MSC_PEMA environment variable to a value of YES (the test for "YES" is case-insensitive). In this mode, the standard SYSTEM(107) enable/disable tests for PEMA are active. @ 7799; 69; Performance - Infinite Cpu Loop of Cray - Missing Insufficient S pace Message On the Cray, if there is insufficient disk space in the directory that you submit your job, then the job may be trapped in an infinite loop when attempting to report the error message and shut down. Messages similar to what is shown below, including "not enough space" or "no space left on device", should appear in your log file. In version 69, the job will terminate with similar messages as shown below, except the "not enough space" or "no space left on device" message may or may not appear. Process Id = <pid> Attempting to bypass eie layer : Not enough space for cache pages apparent state: unit 6 named <path>.f06 last format: (13X ,I8, 1H-, 8X, 10A 08 ) lately writing sequential formatted external

SIGNAL - Signal SIGABRT(6) caught, job aborting. See signal(2), CRI manual SR-2012 for further information. TB001 - BEGINNING OF TRACEBACK - $TRBK WAS CALLED BY RECOVR AT 561544a (LINE NUMBER 33) - RECOVR WAS CALLED BY sysrcv AT 561764a (LINE NUMBER 112) - sysrcv WAS CALLED BY __handlr AT 1074600a - __handlr WAS CALLED BY killm AT 1070722a - killm WAS CALLED BY abort AT 775403d (LINE NUMBER 22) - abort WAS CALLED BY _ferror AT 1170241d (LINE NUMBER 46) - _ferror WAS CALLED BY _errhndl AT 1167630b (LINE NUMBER 22) - _errhndl WAS CALLED BY $ioerp AT 1174051b (LINE NUMBER 31) - $ioerp WAS CALLED BY $WFF AT 1203202c - $WFF WAS CALLED BY XSOWRT AT 3705646a (LINE NUMBER 322086) - XSOWRT WAS CALLED BY XSOPIV AT 3700566c (LINE NUMBER 32290) - XSOPIV WAS CALLED BY XSORT AT 3674767d (LINE NUMBER 321937) - XSORT WAS CALLED BY $SEGCALL AT 12323d (LINE NUMBER 316417) - $SEGCALL WAS CALLED BY XSEM01 AT 3667237b (LINE NUMBER 316417) - XSEM01 WAS CALLED BY $SEGCALL AT 12323d (LINE NUMBER 11778) - $SEGCALL WAS CALLED BY MAINAL AT 4636a (LINE NUMBER 11778) - MAINAL WAS CALLED BY main AT 32c (LINE NUMBER 15) - main WAS CALLED BY $SEGRES AT 12270d (LINE NUMBER 369) - $SEGRES WAS CALLED BY $START$ AT 330525a (LINE NUMBER 369) TB002 - END OF TRACEBACK lib-1208 /nast/msc682/exe/unicos/analysis: UNRECOVERABLE library error An I/O statement was invoked while another was active. apparent state: unit 101 named last format: 5001(' BIOMSG: ERROR',I6,' HAS OCCURRED IN ROUTINE ',2A4) lately writing sequential formatted external lib-1208 /nast/msc682/exe/unicos/analysis: UNRECOVERABLE library error An I/O statement was invoked while another was active. apparent state: unit 101 named last format: 5001(' BIOMStime: command terminated abnormally. Avoidances: There are two possible avoidances. 1. Increase the size of the directory that you are submitting your job from. 2. Send the output files to a different directory that has plenty of space (e.g., the scratch directory) using the "out=..." keyword on the submittal command, and then copy the files back to your submittal directory afterward. @ 7808; 69.1; Optimization/Design Sensitivity - PARAM,DESPCH - Missing Output If PARAM,DESPCH is set to a value greater than one, then there should be DESVAR and GRID output in the punch file for every design cycle. However, in Version 69 the GRIDS are not punched. Avoidance: Insert the following DMAP alter: compile desopt alter 'dom12,','' DOM12, XINIT,XO,CVALO,PROPI,PROPO,OPTPRMG,HIS,DESTAB,

geom1,coordo,edom,mtrak/ HISADD,NEWPRM,,NEWDES/ DESCYCLE/OBJIN/OBJOUT/S,N,CNVFLG/1/OPTEXIT// MDTRKFLG/DESPCH/MODETRAK/eignfreq $ Grid output for the last design cycle can be obtained by setting PARAM,DESPCH,0. @ 7814; ; P-Elements - GMBCs - Wrong Answers When boundary conditions in shell p-elements are specified on an edge using a global coordinate system, the answers become wrong for p=9 and the program crashes for p=10. The answers are correct up to the order of p=8. @ 7818; essage ; Optimization/Design Sensitivity - TRD2, NOLINi - Quits With No M If a modal transient analysis is performed in Sol 200 (ANALYSIS=MTRAN) and the job contains NOLINi entries, the job may abort in module TRD2 with no message except a machine dependent message in the log file. Avoidance: NOLINi entries are not supported in solution 200. The modal transient analysis may be run in solution 112. @ 7829; 70.7; Restart - DBLOCATE read-only restart - UFM 1127 If the LOGICAL keyword is specified on the RESTART statement and the corresponding database on the ASSIGN statement specifies a path for the MASTER DBset which is different from the path in which the job is run then the DBALL and any other associated DBsets will not be auto-assigned. For example, if an input file is located at /home/zzz/test.dat, and the desired databases are /tmp/fender.MASTER and /tmp/fender.DBALL, and test.dat contains: ASSIGN abc='/tmp/fender.MASTER' RESTART LOGICAL=abc then only the MASTER DBset is attached and all other DBsets will have the status of "OFFLINE" echoed in the .f04 file. Also, UFM 1127 will probably be issued. Avoidance: Move the DBsets to the same path that the job is run in or specify ASSIGN statements for all DBsets. @ 7832; ; Elements - PBAR - Missing Torsional Mass Moment The BAR element in MSC/NASTRAN does not include a term for the torsional mass moment of inertia in its formulation (even in the coupled mass matrix). This is documented in the Basic Dynamics User's Guide, but not in the Quick Reference Guide. Avoidance: If torsional mass moments of inertia are needed, use BEAM elements or provide the values using CONM elements. @ 7839; 69.1; P-Elements - DBUNLOAD - SFM 4276, EC 7777, Segmentation Fault A DBUNLOAD of a database with p-elements may produce SYSTEM FATAL MESSAGE

4276 or a segmentation fault followed by a traceback. @ 7841; 69.0.2; DMAP - REIGL - UFM 6413, Insufficient Memory The REIGL module was enhanced in Version 69 to allow a dmap writer to call it without the input of a dynamics or casecc input data blocks. The data they contain is input on some new parameters added to the end of the REIGL call. If REIGL is run by itself, it quits with an "insufficient memory" error: USER FATAL MESSAGE 6413 (DFMSA) INSUFFICIENT MEMORY FOR THE SYMBOLIC PHASE OF SPARSE DECOMPOSITION. USER ACTION : INCREASE MEMORY BY AN ESTIMATED 9401 WORDS. Avoidance: An avoidance is to run reigl without case control inputs after a call with the case control input. The first call will initalize a variable properly for the second call. The first call could contain a small set of matrices from matgen, for instance. @ 7844; 69.0.2; Acoustic Analysis - ACMG - SFM 4276, Segmentation Fault, Access Violation Acoustic jobs may fail unexpectedly in ACMG. The error was discovered by looking at the code, to our knowledge the error has never been encountered. Avoidance: There is no avoidance. The error has been corrected in MSC/NASTRAN Version 69.0.2. This release is not being shipped to clients who have already received Version 69.0.0, it must be specifically requested. @ 7851; 70; Buckling Analysis - SUPORTi - SFM 3011 When a SUPORT entry is present in sol 105 the following error message appears: *** SYSTEM FATAL MESSAGE 3011 MODULE = RBMG4 ATTEMPT TO WRITE A TRAILER ON FILE 201WHEN IT HAS BEEN PURGED Avoidance: Replace the SUPORT entry with a selected SPCi entry. In Version 70 the suport entry is treated as an SPC which is consistent with solution 101 when param,inrel is not selected. This error occurs in sol 103 also. @ 7852; 70; Inertia Relief - DMIG, UACCEL - Wrong Answers When using inertia relief without using the rigid body acceleration (DMIG,UACCEL), and any one of the following entries BNDFL, DMIAX, ELIST, MFLUID, RADCAV, RADSET, RADLIST, RADMTX, DMIG is present in the input file, then a conventional static analysis is performed instead of inertia relief. A symptom of this error is the printout of the URA intermediate matrix will be missing. A review of the spcforce output will also reflect that a conventional static analysis is performed instead of inertia relief. Avoidances: There are two possible avoidances.

1. Remove or comment out the above offending entries since they are not used in a static analysis anyway. or 2. Add the following dmap avoidances: compile statrs $ alter 'paraml.*matpool' $ nompool = -1 $ endalter $ @ 7854; ; DMAP - TYPE - Possible Wrong Values for Parameters If a parameter is defined in the NDDL, and a DMAP "TYPE" statement declares it as changeable in CASE CONTROL or BULK DATA ("Y") but not as an "NDDL" parameter, then the program will incorrectly use the NDDL default value and TYPE for the parameter. If the parameter is listed as a qualifier (QUAL) in the NDDL, the default value works correctly in the DMAP. For example, the following subDMAP in SOL 100 SUBDMAP TEST input/output type parm,,i,y,maxit=10 $ parameter is in the NDDL, NDDL default = 4 type parm,,i,n,modacc=10 $parameter is in the NDDL, but "N" is used on the TYPE type parm,,i,n,maxratio=1 $ parameter is in the NDDL as RS, but "N" is used on the TYPE type parm,,i,y,mesh=1 $ different type than NDDL type parm,,i,y,seid=345 $ test a qualifier $ RETURN END will result in the following values for the parameters: maxit = 4 - NDDL default value in place of user-requested default modacc = 10 - user requested default maxratio = 1 - integer rather than RS as defined in NDDL mesh = 4, but UFM 1053 indicating wrong parameter type SEID = 345 - qualifiers work ok Avoidance: If a parameter is defined in the NDDL, either define it as TYPE NDDL or do not allow it to be changed by case control or bulk data. NOTE: this only affects the default value, changes in DMAP or values in BULK DATA or CASE CONTROL will work. i @ 7855; 69.1; Buckling - TAFF, PARAM,FOLLOWK - Traceback If a PARAM,FOLLOWK,YES is requested to include follower force affects to the differential stiffness and loads that do not exist are referenced the job will stop in the TAFF module and the following traceback will be issued in the log file. Traceback:

Offset Offset Offset Offset Offset @ 7856;

0x000000ac 0x000000e8 0x00000684 0x000002d0 0x00000950

in in in in in

procedure procedure procedure procedure procedure

tafest_ tafldt_ taff_ xsem10_ mainal

Avoidance: Remove the reference to the loads that do not exist. 69.1; ESTIMATE Utility - Missing GRID Entry The ESTIMATE utility may flag GRIDs as missing when in fact the GRID entries are present. The error occurs when a line that exceeds 80 characters precedes an otherwise valid line. Under certain circumstances, the valid line may be skipped. Avoidance: As a work around, delete all trailing blanks from the data file, expand all tabs to spaces, and truncate any lines exceeding 80 characters. @ 7857; 69.0.4; Superelements - SELA - SFM 4276, EC 1200 Superelement jobs may fail in module SELA with the message SYSTEM FATAL MESSAGE 4276 ERROR CODE 1200. Encountering the error is dependent upon the size of the load vector and the amount of memory requested. Avoidance: Increase memory and re-run the job. @ 7860; 68.2.8; Optimization/Design Sensitivity - FRRD2 - SFM 4276, EC 7777 Solution 200 with analysis=dfreq may fail in the FRRD2 module with the message: SYSTEM FATAL ERROR 4276.RECOVR ERROR CODE 7777

The log file may contain a traceback from subroutine frrd2. Avoidance: The error has been corrected in Version 69. It has also been corrected in Version 68.2.8, a re-release for Fujitsu systems. The following DMAP avoidance may also be applied that replaces FRRD2 with FRRD1: compile fea list noref $ $ 285 FRRD2 CKDD,CBDD,CMDD,,PDF,FOL1/DETA/0.0/0.0/0.0 $ $ FRRD1 CASES,DIT,CKDD,CBDD,CMDD,K4DD,PDF,FRL,FOL/ $ UDVF/SOLTYP/0 $ RF8 - FRRD /UDVF alter 'FRRD2 CKDD','' $ frrd1 casers,dit,ckdd,cbdd,cmdd,,pdf,frl,fol1/deta/'direct'/0 $ @ 7871; 70; Heat Transfer Analysis - GPFORCE - Missing Output GPFORCE output is missing for steady state heat transfer analysis, SOL 153, in Version 69.0. Avoidance: Place the following alter in Version 69.0. compile sedrcvr list$ alter 299,299 $ V69.0.2 IF ( GPFDR ) THEN $ GRID POINT FORCE

@ 7874; 2001; Miscellaneous - Extra plot request - Punch File If an output request has two plot requests, e.g., disp(plot,plot) = all then in addition to the plot file, a punch file containing the referenced quantities will also be generated. Avoidance: Remove the extra plot request. @ 7875; 69; P-Elements - GP1, GMCORD, GMLOAD - SFM 3001 If a GMLOAD references a GMCORD in Version 68.2, then the job will fail with the following error messages. *** SYSTEM FATAL MESSAGE 3001 MODULE = GP1 THE INPUT DATA BLOCK IN POSITION 5 DOES NOT EXIST. Avoidance: Use a CORDij instead of GMCORD when using GMLOAD. @ 7878; ; Optimization/Design Sensitivity - TEMP - Wrong Answers An optimization job may yield wrong sensitivity results when temperature loads are applied but certain designed elements (either DVGRIDs or DVPRELi) are not temperature loaded. Avoidance: A possible avoidance is to add a small temperature load on unheated elements in Version 68.2. @ 7881; b file 69.1; Postprocessing - PARAM,POST,0 in SOL 200 - Missing tables in .xd Design sensitivity coefficient tables ( R1SEN and R2SEN datablocks) are missing from the XDB. Avoidance: 1. Copy the nddl.dat file from the MSC/NASTRAN DEL directory. 2. Replace the DATABLK r1MAPR and R2MAPR statements with the following. DATABLK R1mapR TYPE=TABLE EOF DATABLK R2mapR TYPE=TABLE EOF PATH=geomi, PATH=geomi, LOCATION=DSOr, LOCATION=DSOr,

3. Insert the following in your executive control section above all compile statements: compile nddl noref nolist incl nddl.dat

@ 7892; 69.1; Matrix Methods - SOLVIT - Iterative Solver Gives Wrong Answers If the solvit module is given a diagonal matrix it returns a null output matrix. Avoidance: Add at least one small off-diagonal term. @ 7893; 69.1; Plotting - SEPLOT - UFM, Problems Parsing the SEPLOT Command When plotting superelements, the SEPLOT module may have problems parsing when the command extends multiple lines. This may lead to a USER FATAL MESSAGE pointing at the SEPLOT command. Avoidance: If possible put the plot command all on one line. If not possible, rearranging the command may help. @ 7903; ation 70.5; Documentation - QRG, SYSTEM(148), DBVERCHK - Misleading Document The MSC/NASTRAN Quick Reference Guide documentation for SYSTEM(148), a.k.a. DBVERCHK, states that this system cell permits restart from a pre-Version 69 database into Version 69. This system cell simply bypasses the check in MSC/NASTRAN that ensures the Version of the database is the same as the Version being executed and to that extent it is applicable to all Versions of MSC/NASTRAN. The system cell does nothing to make the databases compatible. Avoidance: In general databases are not compatible between Version 68, 69 or 70. The only way to restart is to perform a DBUNLOAD and a DBLOAD: Step 1: Unload database in V68 assi dbunload='______.u68' dbunload endjob Step 2: Load V68 database into V69 assi dbload='______.u68' dbload $ sol dbtrans compile dbtrans alter 'PARAML BULK'(,-1) dbequiv datablk=bulk where (seid=0 and peid=0) , convert(seid=-1;peid=-1) $ alter 'EQUIVX XGEOM1.4/XGEOM1Q','' type db geom1s,geom1q $ EQUIVX XGEOM1.4/GEOM1/-1 $ EQUIVX XGEOM1.4/GEOM1q/-1 $ alter 'if ( rsonly'(2),'call setq'(,-1) IF ( RSONLY ) THEN equivx geom1/geom1s/-1 $ ELSE IF ( NOT(IMAGE OR EXTER) ) then $ SEP2 GEOM1,GEOM2,,,,,SLIST,EMAP,,,/ GEOM1s,t1,,,,,t6,t7,t8,,/SEID $

ENDIF $ (RSONLY) GP1 GEOM1s,GEOM2s,,,,/ g1,g2,GPDTs,CSTMs,BGPDTs,g3,/ S,N,LUSETs/0/0 $ OUTPUTS ARE for SEP1 cend enddata @ 7904; 103 ; Documentation - Reference Manual - sensitivity of materials, sol

@ 18854; y

; Postprocessing - PARAM,POST,-1 CBAR - Incorrect Margins of Safet PATRAN P3 will display margin of safety results for bar elements even though no allowables were specified in the analysis. Avoidance: Either specify allowables or ignore the results.

@ 25363; reen

69.1; Miscellaneous - NOTIFY=YES - Extraneous Output Printed to the Sc Extraneous output will be written on Cray (and others?) when submitting a job. The text "is y" or "is n" may appear on the screen just after the system ID line is written as shown below: mwali[136]: nast69t2 d10101d bat=n app=y scr=y MSC/NASTRAN V69.0t2 (CRAY UNICOS 8.0.4.2) Tue Dec 17 09:12:20 1996 is y MSC/NASTRAN beginning job d10101d. This is the output from the mesg(1) command. This output should have been captured by the nastran driver, and will result in the nastran driver not being able to warn to user if they set notify=yes and their current message state is "n".

@ 25365;

70; Statics - DCMP, SGI R4400 - Segmentation fault In Version 69, when running a large problem on the SGI R4400, if there is insufficient memory allocated for SPARSE DECOMP, then the job may fail with no error messages in the f06 file. A symptom of this error is the printout of the following unclear message in the log file. Receiving signal number 11: Segmentation fault. Note that in Version 69 sparse decomp is designed with an enhanced spill logic so that it should be able to execute in significantly less memory. Avoidance: Increase the memory substantially to avoid spill.

@ 25375; 69.0.2; DMAP - Complex Double Precision Functions - SFM 4276, EC 7777 Double precision complex functions in DMAP cause the following error on

the HP 9000/800 Version 69.0.0 release of MSC/NASTRAN. SYSTEM FATAL ERROR 4276.RECOVR ERROR CODE 7777 PID= 0

Avoidance: Set nastran system(100)=255 or request Version 69.0.2 where this error has been corrected. @ 25377; ; Documentation - MATMOD option 25 matmod option 25 dmap dictionary article incomplete This option is designed to produce two output vectors for only unsymmetric matrices (form 1). For symmetric matrices (form 6) only the O1 vector is output. This condition can be detected by the sym output parameter, which is set to -1 for symmetric matrices, and 0 for unsymmetric matrices. @ 25378;

69.1; ACCESS - Cannot Build ACCESS Sample Programs Any attempt to build the ACCESS sample programs will fail with the error message: make: Cannot find a rule to create target sol301.o from dependencies. This error can be corrected by taking the revised makefile from Scopus tan 2523, or editing the file <install_dir>/msc69/access/makefile: Change the line: build: nastran emas ddl to build: nastran ddl Remove the trailing backslash in SECOND line of the following two: install: $(EXE)/demo1 $(EXE)/demo2 $(EXE)/smplr $(EXE)/d

dlprt $(EXE)/ddlqry And delete the immediately following line: $(EXE)/sol301 $(EXE)/sol302

@ 25379;

70; Matrix Methods - DECOMP - UWM ZERO DIAGONAL TERMS ENCOUNTERED In MSC/NASTRAN Version 69, if the decomp module runs out of memory and the DECOMP parameter is set to 4 or 32, incorrect warning messages may be issued that indicate "ZERO DIAGONAL TERMS ENCOUNTERED". The message that should be output is: USER FATAL (or WARNING) MESSAGE 6413 (DFMSA) INSUFFICIENT MEMORY FOR THE SYMBOLIC PHASE OF SPARSE DECOMPOSITION. USER ACTION : INCREASE MEMORY BY AN ESTIMATED XXXXX WORDS. Avoidance: Increase the memory and re-run.

@ 25422; rgence

69.1; Static Nonlinear Analysis - PARAM,KDIAG - UFM 4551, Failed Conve In nonlinear analysis (SOL 106), if the model has zero stiffness in the undeformed configuration, then the solution may fail to converge or it may converge to a wrong answer. This error first appears in Version 69, and is associated with an error in the "KDIAG" processing. Avoidance: Put PARAM,KDIAG,-1.-30 in the input file to avoid the error or apply the following alter: in NLSTATIC replace statement ADD5 KAAL,KLLTX,KLLRH,,/KAALX///-1. $ with the statement ADD5 KAAL0,KLLTX,KLLRH,,/KAALX///-1. $

@ 25423;

69.1; Installation - SGI IP19 is Incorrectly Identified The Silicon Graphics IP19 system is an IRIX64. SGI had categorically claimed on R8/R10 systems. As a result of the attempt to run the incompatible R8/R10 R44 system, but is able to run that IRIX64 would *only* be found presence of IRIX64, M/N V69 will executable.

An IP19 system can only run the "irix" executable. The work-around for V69.0.2 is to: 1) Modify the shell script "<install_dir/bin/msc69" and change the case label "IRIX" to "IRIX*". Note, this will leave the succeeding "IRIX64" case active, but ineffective. 2) Modify the file "<install_dir>/bin/nast69.ini" to include the line: s.arch=irix This error should not occur in V69.0.0 or V69.0.1. If it does, the solution is identical. @ 25425; 69.0.4; Optimization/Design Sensitivity - Beam Library Optimization is I nefficient Beam Library optimization is inefficient. For a model with 1000 designed properties, a speedup of a factor of 70 has been demonstrated. The poorly performing modules are DOM6 and DOM9. MSC/NASTRAN Version 69.1 corrects this inefficiency. Avoidance: There is no avoidance. @ 25439; 69.1; Optimization/Design Sensitivity - Beam Library - Incorrect Sensi tivities If the design optimization feature of the beam library is utilized with multiple designed stress recovery points, the sensitivities may be in error. The error is most likely to show up when the stress recovery coordinate (e.g., C1 or C2) is not a function of all the designed dimensions. This error can occur for both the MSC provided sections and for user provided sections.

Avoidance: There is no avoidance. @ 25451; 69; Submittal - Windows NT - Input File Not Found Version 68 of MSC/NASTRAN on Windows NT must have .dat extensions specified for input files. This error has been fixed in Version 69.

@ 25456;

69.1; Optimization/Design Sensitivity - DRESP2 - Zero Gradients It is possible to generate the message "Zero Gradients of Violated Constraints" in error if both the lower and upper bounds of a constraint are active for the DOT optimization algorithm and if the underlying response was generated using a DRESP2 entry. This condition will produce a fatal message and the job will terminate. In other cases where both constraints are active, the optimizer may become confused, but it will not terminate. Avoidance: Be careful when specifying tight limits on a DRESP2 entry and be aware that this may lead to difficulty. The error has only been diagnosed when using the SQP algorithm (METHOD=3 on the DOPTPRM entry).

@ 25470;

69.1; Optimization/Design Sensitivity - PARAM,INREL - Poor Performance Calculation of design sensitivities for static subcases with inertia relief can be CPU intensive. Avoidance: A DMAP avoidance is available.

@ 25472; 69.1; Inertia Relief - SOL 200, INREL - Various fatal messages or wro ng answers If the value of PARAM,INREL is specified differently in one or more subcases, then various fatal messages will be issued or wrong answers may result. Avoidance: Do not specify PARAM,INREL in the Case Control. Always specify in the Bulk Data. @ 25501; 70; Superelements - SEPLOT - Missing Plots If there are no SEMOVES, REPEAT, or MIRROR superelements in the deck and if there is a CSTM in a substructure there is a possible memory error causing the substructure not to get plotted or if plotted not in its correct orientation. Avoidance: If the user is missing plot output of a substructure or suspects that it is in the wrong location put in a dummy semove substructure and the problem should go away. @ 25531; 70; Submittal - Remote Execution via 'node=' May Fail The ability to run an MSC/NASTRAN job on a remote system using the 'node' command line keyword will fail in Version 69.0 if the command is not named 'nastran' on the remote system. The nastran command is supposed to use

the actual command name used to run the job on the local system when running the job on the remote system. The V69.0.X release will only use the generic 'nastran' command name on the remote system. Avoidance: To correct this problem, the generic nastran name must be available on the remote system. @ 25532; Answers 69; Miscellaneous - Large field format for TABLED1 - Possible Wrong

The correct specification of a long-field format entry requires a blank entry to be included for blank fields. For example, fields 4 thru 9 of the TABLED1 entry are not used, thus the correct specification of a long-field formatted TABLED1 entry requires the second entry to be blank. In Version 68, if this rule is not followed for the TABLED1 and x-y pair data is specified on the second entry the data is ignored. The program performs a linear interpolation from the data on subsequent entries for the ignored region and this is not likely to be the intent of the user. Version 69 corrects this error by issuing "USER FATAL MESSAGE 315 (IFPDRV)". Avoidance: An example of an incorrect entry and the correct specification follows: TABLED1*1 *AO11000 *AO11000 -1398.00*AO11001 *AO11001ENDT

.0000

-1397.00

.0007

A blank line must be included after the first entry as shown below. TABLED1*1 *A1 *A1 *AO11000 *AO11000 -1398.00*AO11001 *AO11001ENDT @ 25545;

.0000

-1397.00

.0007

69.1; Eigenvalue Analysis - REIGL, EIGRL - SFM 4276, EC 7777 In rare circumstances, Lanczos eigenvalue analysis may fail in the REIGL module and the message "SYSTEM FATAL MESSAGE 4276, ERROR CODE 7777" will be issued. Avoidance: The error is more likely when a large buffsize is set. The error may be easily avoided by either increasing or decreasing the buffsize.

@ 25555;

69.1; Miscellaneous - SSSALTER - Old Released Alters The SSSLATER library released for V69.0.2 is not the current library intended for release. Avoidance: Contact your MSC representative for the correct versions of the alters.

25595;

69.1; Submittal - Command Line Option "cpu" Causes a Memory Fault For Alpha/Unix Version 69.0.2 the use of the command line option cpu causes a memory fault. Avoidance: Set ppcdelta to .1 in the system rc file.

@ 25673; 70; Modal Transient Analysis - TRD1 failure with SYSTEM(146)=2 and u nsym. DMIG input If in a modal transient analysis uses Lanczos with SYSTEM(146)=2 and an un-symmetric direct matrix input with B2PP is activated, then the TRD1 module fails on an SGI R4400 machine with the log-file message: Receiving signal number 10: Bus Error Avoidance: Do not set SYSTEM(146)=2. @ 25675; 70; Documentation - CTRIA3, NASTRAN system cell T3SKEW - Missing Doc umentation The MSC/NASTRAN Version 68.2 Release Notes document the which allows the user to control the the minimum vertex elements at which USER WARNING MESSAGE 5491 is issued. should have been but was not transferred to the Version Guide. @ 25681; T3SKEW system cell angle for TRIA3 The documentation 69 Quick Reference

70; Documentation - SETREE - Unexpected answers If SETREE entries are used to define a multi-level processing tree, then according to Remark 6 under the Quick Reference Guide description of the SETREE entry superelements that are not specified on the SETREE entry will not be included in the analysis. However, if the non-listed superelements are attached to grid points which belong to the residual structure, they will be placed in the tree immediately above the residual structure and included in the analysis.

@ 25688;

69.1; SSSALTER - rflagb.v69 - Wrong Answers for Modal Analysis The rflagb.v69 alter gives correct answers for motion at the enforced motion DOFs, but wrong answers for all other DOFs. The cause of the error is in the assembly of the modal matrices. This error applies to the alter dated November 96. The version delivered with Version 69.1 dated 10 Jan 97 has this error corrected. Avoidance: If you do not have access to Version 69.1, you may request a copy of the updated alter through client support.

@ 25691;

69.1; Nonlinear Analysis - OFP - SFM 4276, EC 7777 A nonlinear analysis that requests stresses for plane strain elements in MSC/NASTRAN Version 68 may fail in the OFP module with the message SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777 on CONVEX computers.

25700; 69.1; Restart - Mini-database Restarts with Superelements - Missing Ou tput, UWM 4410 If superelements are present in a mini-database restart using a dynamic solution sequence such as SOL 103, 107--112, 129, 145, and 146, or dynamic analysis in 200, then data recovery output for the residual structure will be missing and UWM 4410 will be issued. Avoidance: Insert following DMAP alter: compile phase0 alter 'IF ( (DYRD','' IF ( (DYRD<>0 OR (SEID<>0 AND METH<>0)) AND (NOT(STATICS) OR AERO) and not(saverstl) ) THEN $ @ 25713; 69.0.2; Rigid Body Elements - RSSCON - Wrong Answers In Version 69.1, the shell-solid connector RSSCON does not allow the connection of midside noded elements with non midside noded elements when the ELEM option is used. For example, a CQUAD8 shell element can not be connected to a CHEXA with 8 nodes. A fatal message is issued. In Version 69, non matching elements were connected with the ELEM option. Spurious zero energy modes appeared as a result of the incomplete connection. In Version 69.0.2 the fatal message: USER FATAL MESSAGE 6689 (GP4RS2) BULK DATA RSSCON 201 ELEMENT 1 CONNECTED TO ELEMENT IS AN INCOMPLETE ATTACHMENT. ONLY A SUBSET OF THE SHELL GRID AND EDGE DEGREES OF FREEDOM ARE CONSTRAINED is issued, wrong answers are avoided and thus the error is considered corrected. Avoidance: If non matching elements need to be connected, the GRID option must be used. 5

@ 25715; 70; Rigid Body Element and Multipoint Constraints - RSSCON - UFM 669 2 (SECNSD) If more than one shell element is connected to the same solid element with the RSSCON element using the ELEM option, conflicts in the multi-point constraints may occur. A message of the form: USER FATAL MESSAGE 6692 (SECNSD ) RSSCON NUMBER *** CANNOT CONNECT AN EDGE OF TRIA ELEMENT ELEMENT ** TO TETRA

will be output for those shell grids which can not be connected. The Version 70 Quick Reference Guide documents this limitation closing this report. Avoidance: In general, it is NOT recommended to connect more than one shell element to the same solid element.

@ 25724;

70; Documentation - CELAS1, CELAS2 - UFM 316 The MSC/NASTRAN Version 69 Quick Reference Guide description of the CELAS1 and CELAS2 indicates that multiple degrees of freedom may be attached on a single entry. This is not the case and if attempted will result in the message: *** USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA ENTRY CELASi * *** SORTED ENTRY COUNT =

@ 25730;

70; Postprocessing - DBC - Excessive CPU Time DBC can take an inordinate amount of time to write to the .xdb file in some sort1 jobs. For instance, a 130000 dof dynamics job that is writing 100 modes spent 50 minutes cpu time to calculate the modes, but 16 hours writing to the .xdb file. Avoidance: Use the following alter which forces sort2 only for writing to xdb. compile sedrcvr souin=mscsou list, noref $ force sort2 output in a sol 103, normals modes run alter 329,329 $ SDR3 OUGV1,OPG1,OQG1,OEF1X,OES1X,OSTR1/ OUGV2,OPG2,OQG2,OEF2,OES2,OSTR2 $ SORT2 OUTPUT DBC OPG2,OUGV2,OEF2,OES2,OQG2,GPLS,EGPSTR,EGPSF,GPDCT,ELDCT, ONRGY1,OGPFB1,ERROR1,,,,,,,// 'OPG'/'OUG'/'OEF'/'OES'/'OQG'/'GPL'/'GPS'/'SVF'/ 'GPDCT'/'ELDCT'/'OESE'/'OGPF'/'PERROR'//////// -1/DBCPATH/S,N,CP/APP1/ICYCLIC/GEOMU/LOADU/POSTU/ DBCDIAG/DBCPROG/DBCOVWRT/DESITER/////ADPTINDX/LUSETS $

@ 25746;

; Matrix Methods - MPYAD - Segmentation Violation A client deck fails with a segmentation violation while performing a sparse MPYAD in the SDR1 module. Avoidance: The cause of this is under investigation, but the following alter avoided the problem by deselecting sparse MPYAD before the module where the error occurred. compile sedisp, souin=mscsou, nolist, noref alter 217 type parm,,i,n,sys126=0 $ sys126 = getsys(sys126,126) $ putsys(24,126) $ alter 218 putsys(sys126,126) $ put original value back

@ 25747; 70.5; Documentation - PARAM,POST,0 - DBCMSG Warning Message 3 from Sub routine CHEXA The DBC module which creates the graphics database may issue unclear

messages such as those that follow. CHEX ELEMENT 19 HAS BEEN UPDATED.

This message is simply an indication that an xdb existed prior to the run and the element cited has been updated or changed. @ 25773;

; Rigid Elements - RSSCON - UFM 6689 RSSCON elements do not correctly connect warped solid elements to shell elements. In the case of a CHEXA with midside nodes out of alignment with the corner grids by more than TOLRSC*(element edge length), the shell attachment to the midside nodes results in the following improperly issued FATAL message: USER FATAL MESSAGE 6689 (CONSSD) RSSCON NUMBER 200 CANNOT CONNECT HEXA ELEMENT 3 USER ACTION: CHECK GEOMETRY 1 TO QUAD ELEMENT

Avoidance: Other than modifying the geometry to eliminate the warped elements there is no avoidance. @ 25814; 70.5; DMAP - AUTOSPC Incorrectly Constrains USET. If a user places a degree of freedom into the user sets (U1-U6), and the degree of freedom is singular, then when GPSP moves it to the s-set it also removes it from the user sets. The user sets are mainly used by DMAP writers Avoidance: Use the USET table produced by the GP4 module rather than the updated version produced by the GPSP module. @ 25817;

70; Postprocessing - Stress sort PARAM SRTOPT, 1 and 3 Do Not Work. Module STRSORT which is activated by PARAM,SRTOPT does not work as documented for values 1 and 3. The module does not sort in ascending order. Avoidance: There is no avoidance.

@ 25819;

69.1; Superelements - Mini-Database (scr=mini) - UFM 4252 The mini-database option only supports data recovery restarts, i.e., the coldstart run must be executed for all superelements plus the residual structure; i.e., SEALL=ALL, before the data base is useful. In other words, superelements cannot be generated in separate runs with SEALL=n referring to one or more superelements under the mini-database option. If this is attempted then UFM 4252 will be issued in module SEMA. Avoidance: None.

@ 25821;

70; Online Documentation - Installation of MSC/NASTRAN Encyclopedia

When performing a custom install of the MSC/NASTRAN Version 69 Encyclopedia for WorldView 2.2 Architectures, the following choices for systems are presented which contain duplicate menu choice labels for HP and SGI systems: ---------------------------------------------------------------MSC Installation and Setup - Interleaf WorldView V2.2 Architectures Sel 1. 2. 3. 4. 5. 6. 7. 8. Architecture Digital Alpha UNIX . Hewlett Packard 9000 Hewlett Packard 9000 IBM RISC System/6000 Silicon Graphics . . Silicon Graphics Sun SPARC Solaris . Sun SPARC SunOS Mode Mbyte . . . . . . . . . . . . . . . . . . . . No 9 No 9 . . . . . . . . . . . . . . . . . . . . No 9 No 8 . . . . . . . . . . . . . . . . . . . . No 12 No 12 . . . . . . . . . . . . . . . . . . . . No 8 Yes 7 Total space required 7

C. Continue with installation. X. Completely exit this installation now. Enter the number of the architecture to select/deselect, C to go to the next menu, or X to exit this installation. Enter 1-8, C or X (no default)? ---------------------------------------------------------------There is no explanation as to the difference between the two HP 9000 systems (items 2&3) nor the two SGIs (items 5&6). Avoidance: Choose either one. @ 25824; 70; SSSALTER - sm03a.v69 - SIM 6470 If the lanczos module is given a mass matrix which contains terms further from the diagonal than occur for the stiffness matrix, various errors may occur. One symptom is a fatal message mentioning "spurious modes". Another is a message of the form *** SYSTEM INFORMATION MESSAGE 6470 (SSMR0D) MORE EFFICIENT PROCESSING IN LANCZOS EIGENSOLUTION MAY BE POSSIBLE BY INCREASING THE BUFFER SIZE OR THE VALUE OF SYSTEM CELL 199 SUCH THAT THE PRODUCT OF THESE TWO VALUES IS AT LEAST 2266 WORDS. This above message is actually a symptom of a fatal error which will occur later, of the form 0*** SYSTEM FATAL MESSAGE 3001 MODULE = LANCZOS THE EXECUTIVE FILE 0 DOES NOT EXIST. This problem most commonly occurs with component mode synthesis, a capability supported only by dmap alters in sol 3, the solution sequence in which the lanczos module is delivered. The error was encountered when using an alter from the sssalter library, which replaces the reigl module with the lanczos module for sol 103.

An avoidance is to replace the stiffness matrix with a slightly shifted matrix, as shown below. This will perturb the solutions very slightly while avoiding the error. Knew = Kold + a*M, where a is a small number If SIM 6570 appears, it is also necessary to increase buffsize on the nastran command. This feature will be added to alter package sm03a.v69 on the MSC web site. The factor a will be used only when the user sets the "a" parameter described in the readme file for the alter. @ 25836; 70; Documentation - V69 Online Encyclopedia Installation Problems The V69 MSC/NASTRAN ENCYCLOPEDIA (MNE) Installation does not work if the user chooses to keep both collections on the CD-ROM. This is the minimum installation option. Avoidance: If you want to install only the WorldView browser and access both V69 and V68.2 documentation collections from the CD-ROM (minimum installation), you must do the following before you start the installation. 1. Decide on the <install_dir> directory in which to install the MNE. 2. Before installation, create the directory <install_dir>/mne69 After installation, edit the <install-dir>/iview22/Iview file and change the line Iview*welcomeDoc:<install-dir>/mne69/v69docs/69doc/start69.pl to Iview*welcomeDoc:<install-dir>/mne69/v69docs-cd/69doc/start69.pl Also, on some IBM RS6000 systems, when iview (MNE) is started, the following error message may be issued: Error: Can't open display: :0.0 In order for the iview browser to display properly on IBM RISC System/6000 Computers systems, the DISPLAY environment variable must contain the node name. If the DISPLAY variable set by default on your system does not contain the node name, include the following line in your .cshrc file (for csh users): setenv DISPLAY <nodename>:0.0 or the following line in your .profile file(for Bourne or Korn shell users): DISPLAY=<nodename>:0.0; export DISPLAY @ 25845; 70; Eigenvalue Analysis - EIGRL, REIGL - UFM 6578 A new feature in V69 allows the use of parameters on the REIGL DMAP statement to be used to control Lanczos eigenvalue solutions. An error exists that if this approach is used, then the a value must be provided for parameter 9 (NDES) or the module will fail with UFM 6578 indicating that the value for parameter 9 is out of bounds.

Avoidance: If parameters on REIGL are used to control Lanczos, always include a value for parameter 9 (NDES). @ 25846; bounds 70; Eigenvalue Analysis - EIGRL, REIGL - SFM 6578 parameter 9 out of

In MSC/NASTRAN Version 69, it is possible to provide the control parameters for the eigenvalue solution on the REIGL statement. If this is done and no value is provided for ND (number of roots desired, the run will fail with SFM 6578 stating that parameter 9 is out of bounds on the REIGL module. Avoidance: Provide a large integer in the field for ND on the REIGL module statement. @ 25851; 26 70; Optimization/Design Sensitivity - SCR=NO, ANALYSIS=AERO - UFM 11 If aeroelastic optimization is attempted using scr=no, errors in the NDDL dependencies cause the run to fail on the second design cycle with the message: USER FATAL MESSAGE 1126 (GNFIST) DMAP MODULE AMP IS ATTEMPTING TO OUTPUT DATABLOCK 202 NAME = QKH , WHICH ALREADY EXISTS. USER ACTION: DELETE THE DATABLOCK, OR USE FILE QKH =OVRWRT If a flutter analysis is included, the QKH datablock causes the initial error. If this datablock problem is corrected and/or there is a static aeroelastic subcase, the failure is in the DBSTATUS module when it is checking for the presence of the AUG datablock. Avoidance: Make the following NDDL changes: depen qkh /qhh $ DEPEN AUG /IFPA(VP),lsall $ @ 25852; 3007 70; Optimization/Design Sensitivity - ANALYSIS=FLUTTER, DSAD14 - SFM If flutter optimization is used in conjunction with the PKNL flutter method, SYSTEM FATAL MESSAGE 3007 is issued which is of limited usefulness to clients. Avoidance: The PKNL method is not currently supported in SOL 200. Use the PK method. @ 25861; 70; Superelements - SCR=MINI - SFM 4276, EC 1044 In version 69 a new "minimum database" capability has been introduced for MSC/NASTRAN. This capability is implemented by specifying "SCR=mini" on the submittal statement. If this feature is used with the split database capability of superelements, the following error may occur: BIOMSG: ERROR 1044 HAS OCCURRED IN ROUTINE OPEN , FILE INDEX (FILEX) = 0 Avoidance: If this occurs, either use SCR=NO or SCR=YES on the submittal line. @ 25865; 70; Superelements - BUSH - Wrong Answers

If the BUSH element is used in a model with superelements, the frequency-dependent capability will not work. Avoidance: Either include the following DMAP alter in the input file: compile phase1a alter 'TA1.*,,ECTS,EPTS,BGPDTS'(,-1) if(seid=0)then $ type db,ept $ dbview epttemp=ept (where wildcard=true) $ TA1 ,,ECTS,EPTTEMP,BGPDTS,SILS,ETT,CSTMS,DIT,ECTAS,EHT/ EST,,GEI,GPECT,,VGFD,DITID/ LUSETS/-1/S,N,NOSIMP/1/S,N,NOGENL/0///S,N,FREQDEPS $ else $ alter 'TA1.*,,ECTS,EPTS,BGPDTS' endif $ or: place all superelements in partitioned bulk data sections and place the residual structure in the main bulk data. The frequency dependent capability will work if the main bulk data does not have any superelement-related entries. @ 25866; 70; Optimization/Design Sensitivity - DOM6 - SFM 4276 EC 7777 Sig 8 The overflow message, "SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777" has been observed on the CRAY in the DOM6 module and has been identified as being caused by a combination of requesting prints of sensitivity data (OPTEXIT = 4 or 7 the DSAPRT case control command) when both DLINK and DVPREL2 entries are present. Avoidance: The only avoidance is to eliminate one of the three contributing factors. The Beam Library enhancement in Version 69 alleviates the need for DVPREL2 entries and may be a desirable avoidance in many cases. @ 25873; ements 70; Superelements - SCR=MINI - No Data Recovery for External Superel In MSC/NASTRAN Version 69 a minimum database capability was introduced. If external superelements are used and SCR=MINI is used for the residual structure, then data recovery is not possible for the upstream external superelements. Avoidance: Add the following line to the FMS of the external superelement data recovery run: dblocate param=(saverstl), where (peid=0), logical = rsdb (where RSDB is the logical name used for the database with the residual structure solution) @ 25884; 70; Superelements - SESEB - Missing UFM 6794, Possible Wrong Answers Version 69 on CRAY computers may not correctly detect duplicate boundary points and thus may fail to issue: USER FATAL MESSAGE 6794 (SEPSEB) DUPLICATE BOUNDARY POINTS DETECTED CONNECTING TO SUPERELEMENT ** AT GRID

** FROM DOWNSTREAM

**.

The failure to issue this message may result in an upstream superelement being connected to an unintended degree of freedom resulting in wrong answers. Avoidance: When using the automatic superelement connection logic, be careful in maintaining unique grid locations for boundary points. @ 25888; 69.1; Eigenvalue Analysis - GIV - Wrong Answers, UIM 4158 Using the GIVENS method to calculate modes of a structure which has a non positive definite mass matrix, should result in the following error message: "USER FATAL MESSAGE 4646---- THE MASS MATRIX IS NOT POSITIVE DEFINITE,USING THE GIVENS METHOD. USE MGIV INSTEAD" However, the run may proceed and incorrect answers may be produced. Since MSC/NASTRAN automatically omits massless degrees of freedom, this error only occurs with models involving nearly massless mechanism. The user is warned of the poor answers by a very large value for the factor diagonal ratio in the following message: USER INFORMATION MESSAGE 4158 ---STATISTICS FOR CHOLESKY DECOMPOSITION OF DATA BLOCK MXX FOLLOW MAXIMUM RATION OF MATRIX DIAGONAL TO FACTOR DIAGONAL = The user may also be warned of the poor results by the issuance of a machine dependent message indicating a floating point overflow. Avoidance: Use MGIV or LANCZOS method of extracting eigenvalues for the structures of these nature. @ 25891; 70.5; Rigid Elements - RBE3 - Poor Answers If an RBE3 element is input with insufficient independent degrees of freedom to define all 6 possible rigid body modes of the element, a set of constraint equations may be produced that do not reflect the weighting factors input by the user, and can also cause internal constraints associated with stiffness and mass terms associated with degrees of freedom attached to the RBE3 element. This error was observed on a model which has an RBE3 element of the form $ refc wti ci RBE3, [eid], , [refgrid],123456, 1.0, 123, gida, gidb As the two attached grid points have only translations coupled, (123 input for ci), the element is not capable of reacting the component of moment applied at refgrid that is parallel to a line drawn between gida and gidb. The element was processed without warning or fatal messages, and produces different answers on different machine types. A method to detect this type of error is to use param,checkout,yes in sols 101 or 103, and look for non-zero terms in EMH. A fix for this particular problem is to add a third grid point, non-collinear with gida and gidb. In Version 70.5, an element of this type will cause a fatal error, with the first offending RBE3 element identified. The test performed to

identify a poor solution is find the largest number in [C*CINV-I], where C is the weighting matrix used to normalize the constraint equations for the element, CINV is its computed inverse, and I is the identity matrix. All numbers in the matrix of the brackets should be computational zeros. If any is greater than 1.0D-10 a fatal error exit is taken. @ 25893; 70; Superelements - External Superelement Data Recovery in SOL 103 UFM 1126 In external superelement data recovery restart runs in which there are multiple normal modes solutions, the run may fail with UFM 1126 on data block PHA. Avoidance: Insert the following alter: compile extdrin alter 1 file ul=ovrwrt/ol=ovrwrt $ @ 25909; 70; RESTARTS - PBCOMP - ERROR 7777 HAS OCCURRED IN ROUTINE tbk,FILE INDEX (FILEX)=0 Restarts of decks with PBCOMP entries may fail the message: ERROR 7777 HAS OCCURRED IN ROUTINE tbk,FILE INDEX (FILEX)=0 Avoidance: The simplest avoidance is to execute the analysis as a cold start. @ 25916; 2001; p-Elements - CBEAM - UFM 6367 Undefined Grid Point A client p-element model using the adaptive beam failed with the following message; USER FATAL MESSAGE 6367, (POLYD), GRID OR POINT ID =******** FOR ELEMENT ID = *** IS UNDEFINED even though the grid points were correctly defined for all elements in the model. Avoidance: The cause of this error is still under investigation; however, modifying the sixth field of the PVAL entry from explicitly specifying "0" to blank as follows caused the job to run successfully. PVAL to PVAL @ 25939; 1 1 1 1 SET 2 1 1 1 1 0 SET 2

70; Optimization/Design Sensitivity - OFREQ - SFM 4276 EC 3002 If an OFREQ entry is specified in the case control of a solution 200, modal frequency response analysis the following error message is issued: SYSTEM FATAL ERROR 4276.FREAD ERROR CODE 3002 PID= 0

Avoidance: Remove the OFREQ entry from the case control. A DMAP avoidance is under investigation. @ 25948; ystem 70; Miscellaneous - CORD1i - UFM 2006 (GP1C), Undefined Coordinate S In MSC/NASTRAN Version 69, If two coordinate systems are defined on a CORD1i entry, then the second coordinate system definition will be ignored. This error will likely lead to the message: USER FATAL MESSAGE 2006 (GP1C) GRID POINT *** REFERENCES UNDEFINED COORDINATE SYSTEM *** Avoidance: Other than editing the input bulk data to place the CORD1i definitions on separate entries, there is no avoidance. @ 25950; rs 70.5; Eigenvalue Analysis - SUPORT, MGIV, GIV, MHOU, HOU - Wrong Answe For a two degrees freedom system, if one of the tridiagonalization eigenvalue methods (e.g., MGIVENS) is used and the model contains a SUPORT entry, then the answers will be wrong. A symptom of this error is the existence of an extra rigid body mode and zero generalized mass for the flexible body mode. Avoidance: There are three possible avoidances. 1. Use the Lanczos eigenvalue solver. 2. Add an extra degree of freedom. 3. Remove the SUPORT entry if feasible. @ 25970; 70; Frequency Response Analysis - CBUSH, PBUSHT - Wrong Answers If the KN field is used on the PBUSHT entry in a frequency response solution, incorrect answers will result. Avoidance: If BUSH elements are used in a frequency response solution, do not use the KN entries. @ 26022; 67.5; DMAP - INPUTT2/OUTPUT2 - FORTRAN I/O ERROR 933, Unclear Message When using INPUTT2 and ALL of the following conditions are met: 1) Multiple rewind/reads off same INPUTT2 file (as would happen in a looping situation) 2) Using a neutral (formatted) INPUTT2 file 3) An EOF marker (OUTPUT2, ,,,,//-9/iunit $) was NOT used when writing out the output2 file then the job will terminate in the f06 file without any messages. The log file file may contain any one of the following messages: 1)FORTRAN I/O ERROR 933 - Job aborts reading INPUTT2 2)1525-001 The READ statement on the file xxxxx cannot be completed because the end of the file was reached. The program will stop.

Avoidance: Add an EOF marker by adding an output2 statement similar to the statement shown below: output2, @ 26060; ,,,,//-9/junit $

70; Submittal - Memory Allocation Error on HP-UX - UFM 3060 MSC/NASTRAN V69.X on HP-UX systems may be unable to acquire necessary shared memory or queue resources due to lack of available resources. These errors will be manifested as the inability to allocate memory or UFM 3060 errors citing NUSR. The problem is that MSC/NASTRAN's shared memory and queues may not get properly cleaned up if MSC/NASTRAN does not exit normally, e.g., is terminated via kill -9. As a result, system resource remain committed to jobs that have terminated, and are unavailable later on when needed for another job. Rebooting the system will clear up these resource, but we don't recommend this as the fix as it's more than a little disruptive. The following changes to the local template will give MSC/NASTRAN an additional opportunity to clean up after itself. MSC/NASTRAN V69.0: In <install-dir>/bin/nast69.lcl, just after the line: %s.arch=hpux:%%debug=n:% add the lines: %s.arch=hpux:%pid=`grep 'Process Id' %out%.log awk '{print $4}'` %s.arch=hpux:%test -n "$pid" && { %s.arch=hpux:% ipcs -mqp grep 0x while read type id key mode owner group pid1 pid2 %s.arch=hpux:% do %s.arch=hpux:% test "$pid1" = "$pid" && { %s.arch=hpux:% case "$type" in %s.arch=hpux:% m ) ipcrm -m "$id" ;; %s.arch=hpux:% q ) ipcrm -q "$id" ;; %s.arch=hpux:% esac %s.arch=hpux:% } %s.arch=hpux:% done %s.arch=hpux:%} MSC/NASTRAN V69.1, the variable "s.arch" was changed to "MSC_ARCH", and the change becomes: In <install-dir>/bin/nast691.lcl, just after line (247): %MSC_ARCH=hpux:%%debug=n:% add the lines: %MSC_ARCH=hpux:%pid=`grep 'Process Id' %out%.log awk '{print $4}'` %MSC_ARCH=hpux:%test -n "$pid" && { %MSC_ARCH=hpux:% ipcs -mqp grep 0x while read type id key mode owner group pid1 pid2 %MSC_ARCH=hpux:% do %MSC_ARCH=hpux:% test "$pid1" = "$pid" && { %MSC_ARCH=hpux:% case "$type" in $tcmd %a.solver% %a.rel=68:% $asgf %mem% $tcmd %a.solver% %a.rel=68:% $asgf %mem%

%MSC_ARCH=hpux:% %MSC_ARCH=hpux:% %MSC_ARCH=hpux:% %MSC_ARCH=hpux:% %MSC_ARCH=hpux:% %MSC_ARCH=hpux:%}

m ) ipcrm -m "$id" ;; q ) ipcrm -q "$id" ;; esac } done

These changes are completely safe as they only clean up system resources that were in use by the MSC/NASTRAN executable that just finished execution. If the generated script is terminated, the above code still won't get a chance to do its work. In that case, the following script can be used: -------------------- ipc_clean ------------------#!/bin/sh if [ $# -eq 1 ] ; then USER=${1} else USER=`whoami` fi if [ "$USER" = echo "Error, exit elif [ "$USER" echo "Error, exit fi "root" ] ; then will not remove ipcs for user: root" = "daemon" ] ; then will not remove ipcs for user: daemon"

#echo "setting user = <${USER}>" ipcs do # echo if [ if grep ${USER} while read type id addr perms owner group

"$type $id $owner" "$USER" = "$owner" ] ; then [ "$type" = "m" ] ; then echo "removing mem segment: $id" ipcrm -m $id elif [ "$type" = "s" ] ; then echo "removing semaphore: $id" ipcrm -s $id elif [ "$type" = "q" ] ; then echo "removing message queue: $id" ipcrm -q $id fi

fi done -------------------- ipc_clean ------------------This script can be invoked as: ipc_clean to remove all shared memory segments, queues, and semaphores associated with the current username, or ipc_clean _user_

to do the same for the specified username. This script will remove *all* shared memory, semaphores, and queues associated with a specific user. Make sure the user isn't actually using these resources, e.g., is currently running MSC/PATRAN or MSC/NASTRAN. @ 26069;

70; Restarts - RANDOM - Inefficient Processing In MSC/NASTRAN Version 69, if a random analysis is executed as a restart to a frequency response analysis the FRRD1 module will be unnecessarily re-executed. Avoidance: Apply the following DMAP avoidance. compile sedfreq alter 'if ( not noph2'(2),'' if ( false ) then $

@ 26076; ay Fail

70; Superelements - DBLOCATE - Multi-Master Data Recovery Restarts m

If the database for a superelement has changed from the original run, then a multi-master data recovery restart may fail in DBLOCATE.

@ 26078;

70; Static Analysis - PARAM,CURV1, PARAM,OG - Traceback in CURV1 If PARAM,CURV,1 is specified to request that the CTRIA3 and CQUAD4 element stress and strain output be computed in the material coordinate system and PARAM,OG,-1 is specified to avoid calculation of the stresses at the grid points, the run will fail in the CURV module with the following traceback printed in the log file or other machine dependent symptoms such as excessive times or stopping without any messages. Traceback: Offset Offset Offset Offset 0x00000a78 0x000000d4 0x000000e4 0x0000097c in in in in procedure procedure procedure procedure curv1_ curv_ xsem14_ mainal

Avoidance: The CURV module grid output in the user defined coordinate system is obtained by using PARAM,S1G,1 for stresses and PARAM,S1AG,1 for strains. This grid output can either be turned off by removing these parameters, setting them to "0", or using PARAM,OG,-1. Since, PARAM,OG,-1 is not working, use one of the other two methods to turn off the CURV module grid output. @ 26111; 70.5; Optimization/Design Sensitivity - MODTRK - UFM 6656 (MODTRK) PHG REF Not Found If mode tracking is attempted with multiple boundary conditions and if two or more DRESP1 entries identify the same mode, mode tracking will fail and

the message following message will appear. USER FATAL MESSAGE 6656 (MODTRK) PHGREF CANNOT BE FOUND Avoidance: Do not list the same mode number on multiple DRESP1 entries. If the same mode number is being tracked in multiple boundary conditions, a single DRESP1 entry can be invoked by separate DCONSTR entries for each of the boundary conditions. @ 26127; on 70; Miscellaneous - EMG - SIGNAL 10: Bus Error, Segmentation Violati MSC/NASTRAN Version 69 stops in EMG with no messages in the F06 file when materials with duplicate IDs exist. The log file indicates either: SIGNAL 10: Bus Error or: SIGSEGV - Segmentation violation. In Version 68.2, USER FATAL MESSAGE 2017 MATERIAL PROPERTY IDENTIFICATION xxx WAS DUPLICATED was correctly issued. @ 26128; 70; Differential Stiffness - STATSUB, BAILOUT, DECOMP - Unexpected F atal Message PARAM,BAILOUT has the incorrect default in solutions 103, 105 and 107 when the STATSUB command is specified; i.e., differential stiffness is requested. This may cause jobs to stop and issue fatal messages in DECOMP. Also, PARAM,BAILOUT,-1 is ignored in SOLs 103, 105, and 107 when the STATSUB command is specified. Avoidance: Insert the following DMAP avoidance: compile statics2 alter 'call phase1c'(,-1) bailout=-1 $ endalter @ 26132; GL; Optimization/Design Sensitivity - STATIC, SAERO - UFM 4574 Analysis types STATIC and SAERO, do not work together in solution 200. If they are attempted in the same run the following message is issued, "USER FATAL MESSAGE 4574, DATA RECOVERY STARTED BUT NO DISPLACEMENT VECTORS AVAILABLE" Avoidance: There is no avoidance. @ 26173; 2001; Documentation - GPFORCE(PLOT)=ALL - Missing Documentation Although it is not documented in the QRG, the plot option is supported for the GPFORCE entry. @

26180;

70; P-Elements - TEMPF, ADAPT - Traceback TETRA and PENTA p-elements with TEMPF thermal loading fail in the ADAPT module with the following traceback. Traceback: Trap encountered: SIGTRAP - Trace trap Traceback: Offset Offset Offset Offset Offset Offset Offset 0x00000414 0x00000b80 0x00000900 0x00000c70 0x0000134c 0x000003ac 0x0000094c in in in in in in in procedure procedure procedure procedure procedure procedure procedure gxstrn_ adpegd_ adpeld_ adp_ adapt_ xsem14_ mainal

--- End of call chain --Avoidance: Use another method to define the thermal loads. @ 26187; ntation ; Documentation - DMAP Module Dictionary, MAKEOLD - Missing Docume

In the Release Notes for Version 69 it states that the MAKEOLD module may be needed to convert datablocks to the old format. This module is not documented anywhere and there is no list of the datablocks it works on. Avoidance: A brief description follows: Description: Used to convert datablock GEOM1, GPDT, BGPDT, CSTM, AXIC and EST to old format for OUTPUT2 module. Format: MAKEOLD IDB1,IDB2,IDB3,IDB4,IDB5/ODB1,ODB2,ODB3,ODB4,ODB5/P1/P2/P3/P4/P5/P6/P7/P8 P9/P10 $ IDBi = Input Data Block, targeted datablocks are GEOM1, GPDT, BGPDT, CSTM, AXIC and EST. ODBi = Output Data Block. P1:P5 Character input, default is blank. Then generic name of the corresponding input data block; e.g. P3 corresponds to IDB3, etc. P6:P10 Character input, default is blank. The generic name of the corresponding output data block; e.g. P8 corresponds to ODB3, etc. Examples: MAKEOLD BGPDTS/BGPDT,SIL,GPL,EQESIN/'BGPDT'////'BGPDT'/'SIL'/'GPL'/'EQEXIN' $ MAKEOLD ESTS,BGPDTS/EST/'EST'/'BGPDT'////'EST' $

@ 26190; ; Heat Transfer Analysis - RADBC, RADM - Wrong Answers, Missing Wa rning Message If the user intends to use RADBC, radiation to space, the absorptivity value on the RADM bulk data entry must be specified. Otherwise, the ambient space temperature is ignored and unexpected answers will result. Avoidance: Make sure that when you have RADBC in the problem, you specify the absorptivity value. @ 26204; tput 70; Heat Transfer Analysis - CTRIAX6 - No Warning Message, NaN in Ou No warning message is issued the axis of rotation for the impossible physical boundary for the "HEAT FLOW INTO HBDY if convection or radiation is specified on axisymmetric CTRIAX6 element. This is an condition that may lead to "NaN" being output ELEMENTS".

Avoidance: The user should check whether he is applying this ill-posed boundary condition. @ 26221; 70; MSC/ACCESS - DBCTA1 Datablock Description The MSC/ACCESS sample programs need to be updated to the same version of MSC/NASTRAN that they are delivered with, including having the most recent versions of MSC/NASTRAN source in them and updating the program so they work. They currently contain v67.5 element definitions and, thus, cannot be used as a guide to MSC/ACCESS users on how to access information on modern elements. Avoidance: Send the latest version of the MSC/NASTRAN source file "dbctbd.f" to the client, and have the client update the "sampler.f" sample program as practice and then update their own program in a similar way. @ 26230; 70.5; Buckling and Differential Stiffness - GPFDR, TEMP(LOAD) - Wrong Answers Grid point force data recovery for static analysis in SOL 105 is incorrect when thermal loads are applied. In version 69, the error exists also in SOLs 103 and 107. In version 70, the error exists also in SOLs 103 and 107 through 112. Avoidance: A DMAP avoidance is given below for Versions 68.2, 69, and 70. compile phase1a alter 'call selg' equivx ptelem/ptel/-1 $ call dbstore ptel,,,,//seid/0/' '/0 $ compile sedrcvr alter 'GPFDR CASEDR','' call dbfetch /ptel,,,,/seid/0/0/0/0 $ GPFDR CASEDR,UG,KELMx,KDICTx,ECTS,EQEXINS,GPECT,PG1,QG, BGPDTS,SILS,CSTMS,VELEM,PTEL/ ONRGY1,OGPFB1/APP2/TINY//CYCFLG $

@ 26231;

70; Documentation - RESTART - UFM 146 (LINKTP) Version 69.1 cannot RESTART from a Version 69.0 database. If attempted the the message: USER FATAL MESSAGE 146 (LINKTP ) A 'TYPE' STATEMENT DEFINES A DATABLOCK PVTBC IS NOT DEFINED IN THE 'NDDL' will be encountered. IN SUBDMAP SEMFREQ WHICH

Avoidance: It is possible to perform a DBLOCATE restart. @ 26244; 70.7; Optimization/Design Sensitivity - PARAM,POST - Optimized Soluton not on OP2 File Solution 200 has been enhanced in MSC/NASTRAN Verison 70.7 to output the optimized solution to the op2 file. Prior to this enhancement it was required to update the bulk data file with the punch file solution results and rerun a static solution to obtain an updated op2 file. @ 26250; 70.5; Documentation - DBDELETE Module Not Documented The new DMAP module DBDELETE in Version 69 is not documented. This module is useful for those writing looping dmaps. It was introduced late in Version 69 development, and is not in current documentation. Examples of its use may be seen in the fea subdmap. Format: DBDELETE [DATABLK=(datablk-list)] WHERE (where_expr) The where_expr rules are similar to those for the dbdict module. A wildcard option is available. Example: DBDELETE KJJ WHERE (WILDCARD) Function: All copies of KJJ will be erased from the database, regardless of the values of their qualifiers. @ 26267; 70; Heat Transfer Analysis - RESTART, TEMP - UFM 2166 If enforced temperatures are applied in a transient heat transfer analysis cold start, then a subsequent restart may fail with the message: USER FATAL MESSAGE 2166, MATRIX TO BE PARTITIONED IS OF SIZE BY *** COLUMNS. ROW PARTITION SIZE IS ** COLUMN PARTITION SIZE IS *** (INCOMPATIBLE). ** ROWS

Avoidance: Insert the following alter into the Executive Control section: compile nltran alter 'copy ysnt/ysnr','' endalter @

26274;

70; Documentation - CBARAO - Plotting of Intermediate CBAR Results In order to get a CBAR element with a plot code of 100 which provides for data recovery at intermediate locations on the bar, a PLOAD1 must be specified for the element in question. Remark 1 under the Quick Reference Guide description on the CBARAO incorrectly indicates that this a requirement for the SCALE=LE option only.

@ 26278;

70; Heat Transfer Analysis - CHBDYG,AREA8, VIEW - Wrong Answers Radiation heat transfer using the backside of the CHBDYG,AREA8 or CHBDYG,AREA6 for view factor calculation yields incorrect results. Avoidance : Always use the front side for the CHBDYG,AREA8 and AREA6 elements for view factor calculations. The normal directions can be reversed on the CHBDYG,AREA8 by changing the grid connectivities.

@ 26283; ut

70; Documentation - Transient Heat Transfer - Incorrect Punched Outp The PARAM,CURVPLOT,1 is used to force SORT1 output for transient thermal analysis in MSC/NASTRAN Version 68. However, the DMAP has been changed in Version 69 so that PARAM,CURVPLOT,1 can no longer force SORT1 output in transient thermal analysis. To request SORT1 output in Version 69 specify THERMAL(SORT1) or FLUX(SORT1) and remove the OUTPUT(XYPLOT) section from the case control section. Avoidance: If you remove the OUTPUT(XYPLOT) section, then version 69 can output the SORT1 in transient thermal analysis

@ 26285;

70; Matrix Methods - Parallel Decomposition - Job Hangs Eigenvalue analysis using the Housholder method may hang on the VAX/VMS Version 69 system if parallel decomposition is selected. No sparse parallel decomp has been blocked on all systems in Version 70. Avoidance: Do not select parallel processing.

@ 26294;

; Utilities - tabtst.f - illegal structure for unformatted file In the MSC/NASTRAN Version 68.2 Installation and Operating Instructions Guide for Windows NT for INTEL, Chapter 2 How to install MSC/NASTRAN, Section 2.1, Installing MSC/NASTRAN, section 2.1.1, Before You Begin, it says: "If you are going to build the Utility Programs using the supplied source, you must also have a suitable compiler. The supplied source has been compiled and tested using: WATCOM FORTRAN 77/32, Version 9.5c WATCOM C/C++ 9.5c . .

Note: Later versions of the WATCOM compilers should work but have not been tested. The supplied source should work with other compilers available for Windows NT but has not been tested." This should be changed to indicate that if the customer's program reads binary files, it probably won't work due to binary file format differences between most Window's NT Fortran compilers. The following is an example of an error message received when running a binary file reading program that has an incompatible binary file format with the WATCOM compiler used on MSC/NASTRAN: "illegal structure for unformatted file" @ 26298; 70; Response Spectra - DDRMM, PARAM,OPTION,NRL - Incorrect von Mises Stresses In a response spectrum analysis, if the NRL summing option is used, then the Von Mises stresses are calculated differently than what is described in the NAVSEA 0908-LP-000-3010 specification. In the NAVSEA spec, the von Mises stresses are calculated for each mode based on the sigma-x, sigma-y, and tau-xy for that mode. The von Mises stress for each element is then summed using the NRL criteria based on the von Mises stresses for the modes. In MSC/NASTRAN, the component stress (e.g., sigma-x) for each element is calculated using the NRL criteria based on the component stresses for the modes. The von Mises stress for the element is then calculated using the component stresses. Avoidance: There is no avoidance. @ 26304; 70; Aeroelastic Analysis - ZONA51 - Wrong Answers The supersonic aerodynamics code (ZONA51) used in the AERO II option in MSC/NASTRAN has been observed to produce incorrect results when the z distance between boxes on interfering non-coplanar surfaces is less than half the width of a spanwise strip of boxes in the aerodynamic mesh. Avoidance: A possible avoidance is to decrease the width of the maximum strip. An alternative avoidance is to artificially modify the geometry of the aerodynamic model by either making the interfering surfaces coplanar or by increasing the z separation of the two surfaces. Another option is to make the surfaces non-interfering by putting them in interference groups (IGID on the CAERO1 entry). @ 26315; 70; Optimization/Design Sensitivity - CORD2i - SFM 4276, EC 7777 A shape optimization job may produce a numerical overflow (division by zero in tranms/d) or produce the wrong updated mesh when it uses a chain of CORD2i entries. The error may occur when the following conditions are true: 1) At least two CORD2i entries define a parent and child coordinate systems where the child system is a non-basic coordinate system and, 2) The non-basic child system has a different type of coordinate system from its parent.

For example, the following two entries will trigger the error: CORD2R 1 CORD2S 2 2 1 ... ...

where the first entry is the parent and the second entry is the child. Note that the second condition above is met because the parent is of type rectangular while the child is spherical. Avoidance: Use the same type of coordinate system for the parent and all the child coordinate systems when a chain of CORD2i's are used. Else use CORD1 type definitions.

@ 26329;

70.5; Optimization/Design Sensitivity - AEROS, REFC - Wrong Answers Solution 200 supports the ability to simultaneously analyze and design responses associated with static aeroelasticity and flutter. If the user inputs a reference chord on the AEROS entry that differs from the reference chord input on the AERO entry (both fields are designated REFC), an incorrect flutter analyses will be performed with unpredictable results. That is, the run may terminate with an unrelated error message or it may run to completion, with wrong answers.

Avoidance: Make sure that the two REFC values are the same. @ 26331; 70; Nonlinear Analysis - RFORCE, NLCOMB - Segmentation Violation, Tr aceback Nonlinear Analysis that combines RFORCEs with other loads will fail with the following traceback in the NLCOMB module: Traceback: Offset Offset Offset Offset Offset @ 26340; 0x0000099c 0x00000210 0x00000234 0x0000017c 0x0000098c in in in in in procedure procedure procedure procedure procedure nlcslb_ nlcslt_ nlcomb_ xsem15_ mainal

Avoidance: There is no practical avoidance. 70; Heat Transfer - RADBC, SPCFORCE - Missing Output SPCFORCE output is missing in heat transfer. In solution 159, any SPCFORCE output associated with an element with nonlinear capability is overwritten by noise at all points. In both solutions 153 and 159 the RADBC element gives no heat to ambient points. Avoidance: There is no avoidance. @ 26342; 70; Miscellaneous - CORDi - SFM 2005, UFM 6754 In Version 69.0, if there are multiple nested coordinate systems (e.g.,

CORD a references CORD b and CORD b references CORD c, etc.), then the job may fail with the following error messages on some machines. This error has so far been observed on the SUN and HP machines when the nested coordinate system is cylindrical and the z-axis lines up with one of the axes of the basic coordinate system. *** SYSTEM FATAL MESSAGE 2005 (GP1C) INCONSISTENT COORDINATE SYSTEM DEFINITIONS DETECTED IN SUPERELEMENT x *** USER FATAL MESSAGE 6754 (GP1C) NO GRID POINTS DEFINED FOR CORD1I ENTRIES IN SUPERELEMENT x Avoidance: Changing the location of the z-axis slightly in the nested cylindrical coordinate system may avoid the problem. @ 26349; 2 70; Optimization/Design Sensitivity - DVGRID - Poor Message, SFM 306 When a shape optimization job has DVGRID entries that refer to non-existing grids, Version 69 produces the following message: 0*** SYSTEM FATAL MESSAGE 3062 MODULE = DOPR2 FAILURE OCCURRED IN UTILITY MAPXTI CALLED IN SUBROUTINE DSAJ1 REQUESTED EXTERNAL ID(S) NOT FOUND. Version 68 produced the following message: 0*** USER FATAL MESSAGE 5492 (DSAJ1 AN UNDEFINED GRID ID = 9999 HAS BEEN REFERENCED BY A DVGRID BULK DATA CARD WITH A DVID = 1 USER ACTION: CHECK THE DVGRID AND GRID BULK DATA CARDS REQUESTED EXTERNAL ID(S) NOT FOUND. Avoidance: Correct the offending DVGRID entry. @ 26351; 70; Optimization/Design Sensitivity - DLINK - UWM 7001 (DOPR1B), Wr ong Answers If the DLINK ID and the DDVID across multiple DLINK bulk data entries are not in the same ascending order, then the message USER WARNING MESSAGE 7001 (DOPR1B) may be issued in error and the subsequent results will be incorrect. Avoidance: Make sure the lowest DLINK ID corresponds to the lowest DDVID and that subsequent ID, DDVID pairs are in monotonically ascending order. @ 26352; 39 70; Optimization/Design Sensitivity - DRESP2, Superelements - UFM 71 If an optimization run with ANALYSIS=SAERO or ANALYSIS=FLUTTER has superelements, then one of the following fatal messages are issued: "USER FATAL MESSAGE 7139 (DOPR3C) ON DRESP2 entry ID = xxxx, RESP1 ID = yyyy DOES NOT REFERENCE AN EXISTING DRESP1 ENTRY." "SYSTEM FATAL MESSAGE 3007 MODULE = SEDR ILLEGAL INPUT TO SUBROUTINE SEDR" .

"SYSTEM FATAL ERROR 4276.GETSTR @ 26354;

ERROR CODE

3021 PID=

0"

Avoidance: The run must be made without superelements. 70; Superelements - CSUPEXT, SEELT - THRU Option is ignored The CSUPEXT entry with the "THRU" option does not function correctly. Only the first grid point is made exterior to the superelement and all other points are ignored. The SEELT and the RELEASE entries also have problems processing the "THRU" range. It has been experienced that the last element of the range may be dropped. The "THRU" problem may apply to other entries also. Avoidance: Do not rely on the "THRU" range for the CSUPEXT, RELEASE or SEELT entries. Explicitly specify the list of grid points or element ids for these entries. @ 26356; 70; Utilities - BUILDSSS The buildsss.sh file in the <ins-dir>/msc691/util uses 'which', which is not part of the standard UNICOS bsh environment. This may result in the following error: buildsss.sh[122]: test: argument expected Avoidance: Comment out the lines: test -x test -x echo exit } @ 26376; IST) $progexe progexe=`which $PROG 2>/dev/null` $progexe { "$tool: prog=$PROG: executable not found." 1

70; Optimization/Design Sensitivity - PARAM,CDIF,YES - UFM 1126 (GNF If the central difference method is selected in a solution 200 run by setting PARAM,CDIF,YES with Transient response the following error message will be issued. 0*** USER DMAP = EGKGB USER FATAL MESSAGE 1126 (GNFIST) MODULE DSVG1 IS ATTEMPTING TO OUTPUT DATABLOCK 201 NAME , WHICH ALREADY EXISTS. ACTION: DELETE THE DATABLOCK, OR USE FILE EGKGB =OVRWRT

Avoidance: Insert the following DMAP alter: compile pslgdv alter 'egkgb'(,-1) delete /egkgb,egmgb,egbgb,, $ delete /egkgc,egmgc,egbgc,, $ @ 26377; rs 70; Optimization/Design Sensitivity - ANALYSIS=FLUTTER - Wrong Answe

If a design sensitivity run is made in SOL 200 that includes multiple flutter subcases that use the same FMETHOD case control entry to invoke the same FLUTTER bulk data entry, results will be repeated from the first subcase rather than extracted from each subcase in turn. Avoidance: Specify unique FMETHOD entries in each ANALYSIS=FLUTTER subcase in SOL 200. @ 26404; 70; Virtual Mass - MFLUID - UFM 4258 In Version 69, when you select PLANE1 symmetry on an MFLUID bulk data entry, MSC/NASTRAN incorrectly checks for PLANE2 symmetry instead. This may incorrectly cause the following fatal message to be output if your model violates the PLANE2 symmetry: USER FATAL MESSAGE 4258. ELEMENT xxxx WHICH IS ON FLUID STRUCTURE BOUNDARY, CROSSES A PLANE OF SYMMETRY Avoidance: Put a coordinate system on the MFLUID bulk data entry such that PLANE2 symmetry is not violated (even though it doesn't matter for the problem). Note that PLANE1 symmetry is not checked. @ 26410; 70; Optimization/Design Sensitivity - DOM9 - Unexpected Zero Gradien ts, UFM 6499 The optimization algorithm will stop if a violated constraint is found to have a zero gradient. It has been found that these zero gradients can be produced in error whenever the amount of space required to hold the active solution vectors (that is those that are associated with active constraints) exceeds the amount available in the run. For example, 100 double precision solution vectors, each having 250,000 degrees of freedom, would require 50 megabytes of memory to hold all this information in core. If less than this is available, all the sensitivity results will be in error and many will have zero gradients. Avoidance: Estimate the amount of memory required and request this amount when running the job, or reduce the amount required by reducing the number of solution vectors or the number of degrees of freedom or both. @ 26424; 69.1.1; Elements - PSOLID, MAT9, CORD*, TEMP, TEMPD - Wrong Answers If thermal loads are applied to a structure and all of the following conditions exist in the model: 1. A PSOLID property entry that references a MAT9 material property 2. The CORDM field on the PSOLID entry references a non-basic coordinate system then the answers will be incorrect. This error occurs on virtually all Version 69.0 and Version 69.1; however, the error has been corrected in Version 69 system that were built and released after April 1997. The corrected Version 69.0 systems are the DEC VAX and DEC Alpha/NT Version 69.0.4 systems and the HP Exemplar SPP S CLASS (PA2) and NEC SX-4 Version 69.1 systems. This error applies to all CHEXA, CPENTA, and CTETRA elements.

Avoidance: Transform the Gij matrix on the MAT9 entry into the basic coordinate system and references the basic coordinate system on the CORDM field of the PSOLID entry. @ 26425; 69.1; Optimization/Design Sensitivity - TEMP(LOAD) - Wrong Answers In MSC/NASTRAN Version 69, when an optimization job has TEMP load entries and different types of elements in the model are subjected to the thermal loads, wrong sensitivities may result. Avoidance: There is no practical avoidance. @ 26427; 70; Geometric Nonlinear Analysis - ELAS3, ELAS4 - Wrong Answers CELAS3 and CELAS4 elements may cause geometric nonlinear analysis, sol 4 in Version 69, sol 64 in Version 68, to stop in the GNFM module on DEC Alpha systems. Avoidance: Use CELAS1 or CELAS2 elements. @ 26429; 70; Optimization/Design Sensitivity - PARAM,OPTEXIT - Incorrect Term ination Message When param,optexit,4 is specified for an optimization job, the the correct job termination message is: RUN TERMINATED DUE TO PARAMETER OPTEXIT = 4.

However, when DESMAX on the DOPTPRM entry is set to 1 in the same deck, the following job termination message will be produced: RUN TERMINATED DUE TO MAXIMUM NUMBER OF DESIGN CYCLES = 1.

Avoidance: Remove DESMAX = 1 from the DOPTPRM entry, or simply ignore the message. @ 26436; ormats ; Superelements - PARAM,USETPRT PARAM,OPGEOM - Overflow of Print F PARAM,USETPRT and PARAM,OPGEOM request a print of the user set table for the g-set and the full set. When SENQSET entries are present in the main bulk data deck, or when CAEROi entries exist, leading to k-set variable, the degrees of freedom automatically generated have large ids which may overflow the print formats or these parameters. An avoidance is to use param, usetprt, 1 instead of ,0. @ 26438; R5000 70; Installation - Version 69.1 SGI IRIX 64 level 6.2, R4400, R4000, MSC/NASTRAN V69.1 installation will fail on SGI R4XX and R5000 machines running the IRIX64 operating system. A clue to this problem will show up when the mscsetup installation script incorrectly identifies the machine as type R8000/R10000. A "uname -a" command will also echo IRIX64. Avoidance: Prior to installation set the environment variable MSC_ARCH to irix. In

c-shell for example: setenv MSC_ARCH irix After this, the mscsetup script should correctly indicate the machine type as R4XX/R5000 and the installation should complete normally. @ 26440; 70; Parallel Processing - OLOAD - Erroneous Output If parallel processing is selected on SGI 4D computers, then erroneous output that contains "nan" may appear. Avoidance: Turn off parallel processing by removing the "parallel=x" from the nastran statement. @ 26448; ; Superelements - LOADSET, LSEQ - UFM 4390, Poor Design When the LOADSET command is used in partitioned bulk data superelements, all superelements for which the LOADSET applies must have LSEQ entries. If not, UFM 4390 will result. Avoidance: Either have LSEQ entries for all partitioned superelements using the LOADSET command or use the LOADSET command only in SUBCASEs applying to the superelements which have LSEQ entries. @ 26450; 01 70; Plotting - OUTPUT(PLOT) with superelements - Bad Message, SWM 30 If there are superelements present along with an OUTPUT(PLOT) section but there is no SEPLOT or SEUPPLOT command in the OUTPUT(PLOT) section then SWM 3001 will be issued. Avoidance: Specify an SEUPPLOT or SEPLOT command. @ 26455; ; Online Documentation - IVIEW - Can Not Print from Version 69 MNE Printing using the print command defaults will not work in the Version 69 MSC/NASTRAN online encycopedia for the UNIX BSD4.2 operating system. The UNIX System V default command messages: Avoidance: edit as root the file <NASTRAN-install-dir>/iview22/Iview as follows: ! example of printer definition for UNIX BSD4.2 default: Iview*printerConfig: grz13p2 PS-doublesided, LaserWriter+, lpr -P%s %s; no-default, LaserWriter+; @ 26464; DIN A4) ; Online Documentation - IVIEW - Printing to European Paper Size ( lp -s ... gives the following error

Printing pages from the NASTRAN V69 Encyclopedia to DIN A4 paper (european paper size) there is no reduced scaling. This means the contents of the page is cut on the upper and right border. This error exists since the first release of the on-line documentation.

Avoidance: Due to the non-function of the <NASTRAN-install-dir>/iview22/Iview entry Iview*europeanPaper: True use the following Workaround to print NASTRAN V69 Encyclopedia pages to european paper size: -------------------------------------------------------------------------$ su $ cd /msc/iview22/ps $ cp -p psprolog.600 psprolog.600.org Edit the standard postscript prolog file: $ vi psprolog.600 Edit original line 49: % modified by A. Hopf MSC/Germany 20-Mar-1997 ---------------------------------%/ia4{statusdict/a4tray known{rcignore statusdict begin a4tray end rcrestore}if}bdf /ia4{statusdict/a4tray known{rcignore statusdict begin a4tray 0.88 0.88 scale 50 -7 translate end rcrestore}if}bdf %/ia4{statusdict/a4tray known{rcignore statusdict begin a4tray 0.88 1.04 scale 50 -80 translate end rcrestore}if}bdf %----------------------------------------------------------------------------This creates a print with the default size of the xscale equal to the yscale. If you want full paper size, increase the y-scaling. For instance in the example above, comment the current /ia4 statement with % and uncomment the third /ia4 statement. If an extra page appears with the print message Interleaf::pl2ps 6.0.0 you can prevent this extra printing by editing the file psprolog.600 as follows: original lines 71 and 72: % modified by A. Hopf MSC/Germany 20-Mar-1997 ---------------------------------% (\tInterleaf::pl2ps )print % SUB MINOR MAJOR vp %----------------------------------------------------------------------------!!! Important: Print Encyclopedia pages with @ 26475; "Paper Type" A4 !!!

70; Submittal - MSC/NASTRAN Always Executes on first subcomplex

The MSC/NASTRAN submittal command provides the 'subcomplex' keyword for HP CONVEX Exemplar system to select a specific subcomplex to run the job on. If this keyword is not set, the job is supposed to run on the current subcomplex. An error in the driver always forces the job to run on the *first* subcomplex, usually this is 'System'. There are three choices: 1) Accept this restriction. 2) Force the submittal command to always choose the current subcomplex by make the following changes to the local template file(s): a) MSC/NASTRAN V69.0.X: In <ins-dir>/bin/nast69.lcl: i) remove the line: %s.arch=sppux:% echo "Subcomplex: %subcomplex%"

ii) Remove the two words "-sc %subcomplex%" from the line %s.arch=sppux:% $tcmd /bin/mpa -sc %subcomplex% -DATA -STACK ... b) MSC/NASTRAN V69.1: In <ins-dir>/bin/nast691.lcl: i) remove the line: %MSC_ARCH=sppux:% echo "Subcomplex: %subcomplex%" ii) Remove the two words "-sc %subcomplex%" from the line %MSC_ARCH=sppux:% $tcmd /bin/mpa -sc %subcomplex% -DATA -STACK ... 3) Install the "S-Class" (a.k.a PA-RISC 2.0) version of MSC/NASTRAN V69.1, this includes the correction to the submittal command and adds supports for the PA-RISC 2.0 processor. @ 26500; GL; Optimization/Design Sensitivity - DBLOCATE - SFM 1033 (XIO) When using DBLOCATE to continue any analysis into a subsequent optimization analysis, the following error occurs: 0*** SYSTEM FATAL MESSAGE 1033 (XIO) FILE 0 NOT FOUND IN FIST 0********* DUMP OF XIO PARAMETERS 0XIO NAME = 0 ATTEMPTED OPERATION = OPEN (UPDATE) FILEXX = 0 . . . 0*** SYSTEM FATAL ERROR 4276.XIO ERROR CODE 4276 PID= 0 0*** USER INFORMATION MESSAGE: TO OBTAIN A NASTRAN DUMP, RESUBMIT JOB WITH DIAG 44 INSERTED IN THE EXECUTIVE CONTROL SECTION. This is caused because the DBLOCATEd file is read only, but optimization analysis requires the desiter qualifier to be set on the datablocks Avoidance: Change (desiter=n)" where DBLOCATE statement instance, if three the DBLOCATE statements to include "convert n is the design iteration number. Include a separate for every design iteration that will occur. For design iterations will occur the DBLOATE statements may

look like the following: ASSIGN BB='beambit.master' DBLOCATE DATABLK=*, WHERE(SEID=9), LOGICAL=BB, CONVERT(DESITER=1) DBLOCATE DATABLK=*, WHERE(SEID=9), LOGICAL=BB, CONVERT(DESITER=2) DBLOCATE DATABLK=*, WHERE(SEID=9), LOGICAL=BB, CONVERT(DESITER=3) where beambit.master is the database that the second analysis is starting from. @ 26508; 4698 70; Miscellaneous - PARAM,BAILOUT Ignored - UWM 6137, UIM 4258, UIM It has been observed in a model which has obvious singularities that Version 69 may continue with the solution after flagging these singularities using UWM 6137 (rank deficient matrix) and UIM 4258 and 4698 indicating that MAXRATIO has been exceeded. This is in spite of explicitly setting PARAM,BAILOUT,1. Avoidance: There is no known avoidance for this error. Search the output file for message 6137. If it occurs look for UIM 4258 and 4698. The most obvious evidence is large displacements (possibly as large as 1.E=12) with low stresses. @ 26527; 70; PK-Method Flutter Solution - FLFACT - Unexpected Results when Ve locity = 0.0 If a PK flutter analysis is attempted at zero velocity, unpredictable results can occur, ranging from bad answers to floating point divide checks. This error has been corrected in Version 70 by issuing a user fatal message whenever a zero velocity is specified on a FLFACT entry. Avoidance: do not specify velocities of 0.0 of the flfact bulk data entry. Very low velocities should be avoided in general, since they may cause the aerodynamics to be extrapolated to reduced frequencies above those regarded as acceptable for the aerodynamic theory. @ 26530; 70; Multipoint Constraints - MPC, RBEi, MCE1 - Wrong Answers Models that contain MPC equations or RBEs may produce wrong answers in Version 69. The symptoms which can cause this error usually occur when the dependent points for the equations are widely spread over the columns of the matrix, resulting in a need to pivot when solving the equations. In this condition the pivoting may discard terms, leading to a plausible but incorrect solution. Avoidance: Models with rigid elements should be run through the checka alter delivered in the sssalter library. SYSTEM(247) may be set to 1 to avoid this error. @ 26547; 70; Rigid Elements - RSSCON - UFM 6694 Doesn't Identify Grid Point In Version 69.1 the following fatal messages about RSSCON are properly issued, but instead of printing the grid ID which is causing the error, the printout is *** . The grid IDs are properly printed in Version 69.0.2. *** USER FATAL MESSAGE 6694 (GP4RS4)

RSSCON CAUSES CONFLICTING ROTATIONAL CONSTRAINTS AT SHELL GRID POINT *** @ 26557; 70; Nonlinear Analysis - PARAM,KDIAG - Undocumented Differences In MSC/NASTRAN Version 68 and prior releases, in non-linear statics if the nonlinear stiffness matrix contained nulls on the diagonal then if kdiag<0 (the default) the program would automatically add a 1.0 to the 0. diagonal term(s). If kdiag>0 the program would always add a user specified value to the all diagonal terms. If kdiag=0, nothing was done. In Version 69 a change was made: First the addition of a term to the diagonal of the stiffness matrix was based on the total stiffness Knl+Kslide-line+Kdiff-stiff and second terms were added to the right hand side. If KDIAG<0 these additional terms will automatically be discarded (though in Version 69 there was a variable error-see alter below to fix) If kdiag>0, then these terms need to be turned off after the first subcase since they can lead to incorrect convergence. They are turned off by a subcase with KDIAG=0 in it. The advantage of these terms is a great speed up in convergence. However the structure is grounded until kdiag is turned off. The kdiag parameter is used in the case where at the start of a nonlinear iteration there is no stiffness in the structure. In the case of a 3-dimensional structure. Terms of zero stiffness are often not computed zeroes but rather small noise zeros. Thus the need for the KDIAG>0 option Avoidance: In Version 69 the following alter will correct the dmap problem: COMPILE NLSTATIC LIST NOREF alter 'CFACT=CMPLX'(1,-2) $ IF ( KDIAG=0. ) EQUIVX KAAL0/KAAL/-1 $ FOR (NEWP>-1 AND MSCHG>-1) FALSE alter 'IF ( KDIAG>0.)','ENDIF $ + ONE' $ IF ( KDIAG>0. AND NOT(NLHEAT)) ADD ILL,KAAL0/KAAL/CFACT//KOPT $ + ONE ALL DIA alter 'ADD5 KAAL,KLLTX,KLLRH','' $ ADD5 KAAL0,KLLTX,KLLRH,,/KAALX///-1. $ ENDALTER @ 26560; 70; Rigid Elements - RBE3 - Wrong Answers In MSC/NASTRAN Version 69, if the 'UM' option is used on the RBE3 entry and the GMi terms are not in ascending order then wrong answers result. Avoidance: Specify the GMi terms in ascending order. This error may be avoided by setting system(247)=1. Please also see error number 26530. @ 26571; 70; DMAP - TYPE, RD - BIOMSG: ERROR 7777 If a double precision parameter is specified as follows:

type parm,,rd,,(dumvar=0.0) then either incorrect values or garbage (e.g., #.#####-315) will be assigned to this variable on the Digital ALPHA OSF/1 and INTEL NT Versions of MSC/NASTRAN. Depending on the subsequent operations, a core dump similar to the following messages may be produced. BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE tbk, FILE INDEX (FILEX) = 0 STATUS = 0

********* NASTRAN FILE TABLE ********* Avoidance: Modify the type statement as follows: type parm,,rd,,(dumvar=0.d0) @ 26574; s 70; External Superelement - CSUPER, DBLOCATE - Possible Wrong Answer When using the external superelement feature in Version 69 with the DBLOCATE method, if data recovery for a superelement is requested in a separate run from the residual structure run (as shown in the superelement seminar notes), then the answers may be incorrect. Note that this error report does not apply to the enhanced superelement features introduced in Version 69, as external superelement using the enhanced superelement technology is not supported in Version 69. Avoidance: to: dblocate db=bgpdts, where (peid=0), logical = rsdb @ 26575; 70; Miscellaneous - RCFILE - Enhancement Request, UFM 1019 Blank characters are significant in the symbol definitions in an RC file or on the command line. For example, in the following line: SYMBOL = MTF = /u/jfc/csr Change the following:

dblocate db=eqexins, where (peid=0), logical = rsdb

The symbol name will be taken as "MTF and the value as " /u/jfc/csr" Write the symbol definition as symbol=mtf=/u/jfc/csr @ 26586; Output 70; Heat Transfer Analysis - VIEW, CHBDYG,REV - Missing View Factor "

Although the MSC/NASTRAN Quick Reference Guide description of the CHBDYG element states that View factor calculations are not supported with the REV option, Version 68 and prior releases supported the calculations when no third body shadowing existed. In Version 69, even for this limited case, View factor output for the CHBDYG element using the REV option is missing. Avoidance: There is no avoidance. @ 26609;

70; Post-processing - PARAM,POST,-2 - Missing Output PARAM,POST,-2 does not properly output the geometry data blocks if superelements exist. Avoidance: Insert the following alter: compile phase0 alter 'POST=-2 OR ','' if ( false ) then $ compile super3 alter 'RSPECTRX=' CALL OUT2GEOM CSTM,GPL,GPDT,EPT,MPT,GEOM2,GEOM3,GEOM4, GEOM1,DIT,DYNAMICS,EQEXIN,BGPDT,CASECC// OUNIT2/0 $

@ 26620; 70; Transient Response Analysis - SOLs 109 and 112 - Inefficient Loa d Calculation In SOLs 109 and 112, transient response loads are computed inefficiently and, depending on the number of time steps, may require significant CPU. Avoidance: Insert the following DMAP alter in SOL 109 or 112: compile sedtran alter 'trlg ','trd1 '(,-1) copy usetd/usett $ paraml usett//'trailer'/4/s,n,usetr $ PARAML USETT//'USET'//////'m'/S,N,NomSET/ 's'/S,N,NosSET/ 'o'/S,N,NooSET $ if ( nomset>-1 ) usetr=usetr-1 $ turn off m-set bit if ( nosset>-1 ) usetr=usetr-2 $ turn off s-set bit if ( nooset>-1 ) usetr=usetr-4 $ turn off o-set bit modtrl usett// / / /usetr $ TRLG CASEt,USETt,DLT,,BGPDTS,SILS,CSTMS,TRL,DIT, ,,rpd,EST,MPTS,MGG/ ppt,,,,pdt,tol, $ compile semtran alter 'trlg ','trd1 '(,-1) copy usetd/usett $ paraml usett//'trailer'/4/s,n,usetr $ PARAML USETT//'USET'//////'m'/S,N,NomSET/ 's'/S,N,NosSET/ 'o'/S,N,NooSET $ if ( nomset>-1 ) usetr=usetr-1 $ turn off m-set bit if ( nosset>-1 ) usetr=usetr-2 $ turn off s-set bit if ( nooset>-1 ) usetr=usetr-4 $ turn off o-set bit

modtrl usett// / / /usetr $ TRLG CASEt,USETt,DLT,,BGPDTS,SILS,CSTMS,TRL,DIT, ,,rph,EST,MPTS,MGG/ ppt,,,,pht,tol, $ @ 26624; s 70.7; DMAP - VECPLOT, option 1 - Output Matrix Wrong for Fluid Problem In a fluid problem (CD=-1 on GRID entry), if option 1 is used in module VECPLOT, then the output matrix is bad. A symptom of this error is indicated by a different matrix trailer size in the f04 file as compared to the matrix size printed in the f06 file. Sometimes "BIOMSG: ERROR 211 HAS OCCURRED IN ROUTINE BLDPK" fatal message occurs. This output matrix is not used in any standard SOL sequence, so only user-written DMAP may encounter this error. @ 26633; 2001; DMAP - BNDSPC, APPEND - BIOMSG: ERROR 1404, Wrong Answers The BNDSPC module is a new module introduced for partitioned bulk data superelements in Version 69. Its function is to take spcs and spcds on upstream superelement boundary points and transfer them to the downstream superelement with the same boundary point. While this module has passed many tests, different bad behaviors have been detected on some input files, including: [The s-set variables appear out of sequence when printed by use of param, usetprt, 0. If the input datablocks (YSB and YSD) to the BNDSPC module are null, then corrupted output datablocks may be generated which may lead to failure in a subsequent module with the following messages: this may be wrong. delete this sentence from error report] BIOMSG: ERROR 1404 HAS OCCURRED IN ROUTINE GETSTR , This error has been observed to occur in the APPEND module. However, it could appear in other downstream modules, if the APPEND module is branched over and the corrupted data block are used elsewhere. Also the YS1 vector is described as a table rather than a matrix when printed with MATPRN, and its values are typically very small numbers such as 1.e-330. At other times these data blocks may cause MATPRN to crash. The alter listed below avoids this problem for null YS vectors. This error is most likely to occur when user-dmap is present. This error is still under investigation, and its effects may not all be known at this time. The bad Ys vector symptoms were observed when diag 20 was turned on before the module. An avoidance is to prevent this module from moving terms by doing so yourself when you suspect it is malfunctioning. That is, put spcs on boundary points at the lowest level possible. After running the model, and confirming that you moved all of the spcs that were on the boundary, you can equiv the inputs to the outputs of this module, and branch over it. Another avoidance which sometimes works is to remove diag 20 from the input file. @ 26634; GL; Postprocessing - PARAM,POST, ANALYSIS - Missing Plot Output Optimization jobs which have multiple analysis types and requests to the MSC/NASTRAN plotter can cause problems with the OUTPUT2 file for

MSC/PATRAN. The OUTPUT2 file may be missing data and contain extra data. Avoidance: Do not combine NASPLOT requests with PARAM,POST,-1 in solution 200. @ 26677; 3001 70; Single Point Constraints - SPC, SPCD - UWM 2051A, UFM 2052, SFM If enforced displacements are specified on the SPC or SPCD entries on grid points that are not in the structure, then the following messages: *** USER WARNING MESSAGE 2051A (GP4) UNDEFINED GRID POINT x IN SPC SET OTHERS ALSO IGNORED) y IGNORED. (ANY

*** USER FATAL MESSAGE 2052 MODULE = GP4 UNDEFINED GRID POINT x IN SINGLE-POINT CONSTRAINT SET z are issued. These messages are the correct messages and the job should have terminated in GP4. MSC/NASTRAN is designed to continue when spcs are specified on non-existent grid points, but not on spcd values. However, the job will continue without producing a uset table. The next module that needs the uset table will then fatal with the following error message: *** SYSTEM FATAL MESSAGE 3001 MODULE = xxxxx THE INPUT DATA BLOCK NAMED USET0 AT POSITION EXIST. a DOES NOT

Avoidance: Look at the grid point that UFM 2052 is pointing to in GP4 and remove all illegally specified SPCD values. @ 26728; 70; Rigid Elements - RSSCON - UFM 4296, UFM 6692 The shell-solid connector RSSCON fails for certain irregular geometries when the ELEM option is used. In the example, the following fatal messages are issued *** USER FATAL MESSAGE 4296. (SECNSD) ILLEGAL GEOMETRY FOR QUAD4 ELEMENT WITH ID = 129131 *** USER FATAL MESSAGE 6692 (SECSR ) RSSCON NUMBER 129780 CANNOT CONNECT AN EDGE OF SHELL ELEMENT 129131 TO SOLID ELEMENT 129079 The failure occurs when the shell element shares an edge with the solid element and the shell plane is tilted to the solid face. Avoidance: Use the GRID option to make the connection. @ 26735; 70; Eigenvalue Analysis - EIGRL - UFM 17 SYMBOL MEMSAVE IS UNDEFINED If the System Cell MEMSAVE=1 is specified instead of SYSTEM(229)=1, the run results in a UFM 17.

@ 26739; ntation

70; Documentation - EIGRL, NASTRAN SYSTEM(229)=1 - Misleading Docume

The documentation for SYSTEM(229), a.k.a. MEMSAVE, indicates that setting this system cell to 1 reduces scratch space usage in Lanczos by 67%. This is the maximum possible savings. Jobs that execute three shifts or less will experience no savings. @ 26747; 70; Superelements - MATREDU - SFM 4276 EC 3012, Wrong Answers In superelement analysis, if rigid elements or mpcs exist for a superelement but no spcs exist, and damping elements exist that do not intersect the m-set DOFs the damping effects are discarded. This may lead to SYSTEM FATAL MESSAGE 4276, ERROR CODE 3012 or wrong answers Avoidance: There are two avoidances: 1. Add a token grid point to the superelement. Autospc will cause the dummy grid point to generate spcs. or 2. Insert the following DMAP alter: compile semrb alter 'matredu','' paraml uset//'uset'//////'f'/s,n,nofset/ 'a'/s,n,noaset/ 'm'/s,n,nomset/ 'o'/s,n,nooset/ 's'/s,n,nosset $ matgen ,/iff/1/nofset $ equivx iff/rfa/nooset $ if ( nooset>-1 ) then $ matgen ,/iaa/1/noaset $ umerge1 uset,iaa,goa,,/rfa/'f'/'a'/'o'/1 $ endif $ equivx rfa/rna/nosset $ if ( nosset>-1 ) umerge1 uset,rfa,,,/rna/'n'/'f'/'s'/1 $ equivx rna/rga/nomset $ EQUIVX xGG/xNN/NOMSET $ if ( nomset>-1 ) then $ equivx gm/gmf/nosset $ if ( nosset>-1 ) upartn uset,gm/gmf,,,/'n'/'f'/'s'/2 $ equivx gmf/rma/nooset $ if ( nooset>-1 ) then $ upartn uset,gmf/gma,,gmo,/'f'/'a'/'o'/2 $ mpyad gmo,goa,gma/rma $ endif $ nooset>-1 umerge1 uset,rma,rna,,/rga/'g'/'m'/'n'/1 $ MCE2 USET,GM,xGG,,,/xNN,,, $ endif $ nomset>-1 IF (NOSSET>=0) UPARTN USET,xNN/,xSF,,/'N'/'F'/'S' $ smpyad rga,xgg,rga,,,/xaa/3////1////6 $ @ 26752; 70; Frequency Response - SOLs 11, 111 and 200; SDAMP - Inefficient P rocessing

For modal frequency response (SOLs 11,111, and 200), if only modal damping (SDAMP) exists in the model, then the uncoupled solution algorithm should be used. However in Version 69.1, the coupled solution algorithm will be used even though only modal damping exists in the model. Avoidance: Insert the following dmap alter for SOLs 111 and 200: $ compile freqrs souin=mscsou list $$$$ alter ' frrd1 ','' $ FRRD1 CASES,DIT,CKDDF,CBDDF,CMDD,K4DDF,PDFI,FRLI,FOLI, EDT,SILD,USETD,/ UDFI,FOLI1/SOLTYP/-1/////S,N,FIRSTBD0/FRRDSET $ $$$$ alter ' frrd1 '(2),'' FRRD1 CASES,DIT,CKDD,CBDD,CMDD,K4DD,PDFx,FRL,FOL, edt,sild,usetd,/ UDF,fol1/ SOLTYP/-1/////s,n,firstbad/frrdset $ $

For SOL 11 (or 30) insert the following alter: compile sol11 alter 'frrd1','' FRRD1 CASExx,,KHH,BHH,MHH,,PH,FRL,OLST,edt,sild,usetd,/ UHV,olst0/ FORM/-1/////S,N,FIRSTBAD $ @ 26762; 70; Eigenvalue Analysis - MGIV, GIV, HOU, MHOU - Missing Eigenvector , UIM 5218 On long word machines like CRAY, if the following user information message is encountered in an eigenvalue analysis using a transformation method such as Givens or Househoulder, then some of the eigenvectors requested may not be output. USER INFORMATION MESSAGE 5218 EIGENVALUE APPROACHING INFINITY AT ** MODE. EIGENVECTORS WILL NOT BE COMPUTED BEYOND THIS POINT. The error only occurs when ND on the EIGR entry is less than the number of eigenvalues calculated. In this case there will be one missing eigenvector for each mode that UIM 5218 is output for. Avoidance: Increase ND by the number of missing eigenvectors. @ 26772; GL; Nonlinear Transient Heat Transfer - TEMPBC,TRAN - Wrong Answers If there are unreferenced type TRAN, TEMPBC entries in the test deck, then different answers occur. This is a design limitation. TEMPBC entries internally generate CELAS2 elements and DAREA input. The CELAS2 elements

are generated whether the TEMPBC is referenced or not and this may cause unexpected results. Avoidance: Remove any unreferenced type TRAN, TEMPC entries. @ 26775;

70; Superelements - DEFORM - Wrong Answers If DEFORM bulk data entries are included in a part superelement file (that is, after a BDGIN SUPER= entry) they will be ignored. Avoidance: Place DEFORM entries in the main bulk data deck.

@ 26778;

70; Nonlinear Transient Heat Transfer - EMG - Wrong Answers When multiple subcases are specified in a nonlinear heat transfer analysis in MSC/NASTRAN Version 69, module EMG does not generate the nonlinear heat transfer capacitance matrix BGGNL at the first iteration. This leads to wrong answers. Avoidance: Do not specify subcases.

@ 26818; 3001

70; Optimization/Design Sensitivity - ANALYSIS=SAERO, UMERGE1 - SFM The design intent in solution 200 is that all static aeroelasticity subcases should have the same boundary condition but that other analysis types can have different boundary conditions specified. The sensitivity analysis for static aeroelasticity may produce a fatal error in UMERGE1 of the form below if different boundary conditions are applied to static aeroelastic subcases: SYSTEM FATAL MESSAGE 3001 MODULE = UMERGE1 THE INPUT DATA BLOCK NAMED USET AT POSITION 1 DOES NOT EXIST. USER INFORMATION: THIS ERROR IS CAUSED BY ONE OF THE FOLLOWING: Avoidance: This following is a dmap avoidance for the problem: compile pslgdv alter 'desaerdr'(1,-1) CALL SETQ CASES//SEID/PEID/S,MTEMP/S,K2GG/S,M2GG/S,B2GG/S,MPC/ S,SPC/S,LOAD/S,DEFORM/S,TEMPLD/S,P2G/S,DYRD/S,METH/ S,MFLUID $ PARAML CASES//'DTI'/1/255//S,N,SUPORT $ CALL PMLUSET USET//S,NOASET/S,NOBSET/S,NOCSET/S,NOGSET/S,NOLSET/ S,NOOSET/S,NOQSET/S,NORSET/S,NOSSET/S,NOTSET/ S,NOVSET/S,NOA/S,NOSET/S,NORC/S,NOMSET/S,NOUSET $ compile fea alter 'saersens'(1,-1) CALL PMLUSET USET//S,NOASET/S,NOBSET/S,NOCSET/S,NOGSET/S,NOLSET/ S,NOOSET/S,NOQSET/S,NORSET/S,NOSSET/S,NOTSET/ S,NOVSET/S,NOA/S,NOSET/S,NORC/S,NOMSET/S,NOUSET $

@ 26825;

; Nonlinear Slide-Line Analysis - BLSEG, BCONP - Wrong Answers The statement below is in error. Replace unsymmetric with symmteric and

vice-versa. Unsymmetric is the default and use the deafult option. Rakesh Allahabadi Wrong answers may occur for nonlinear slide-line analysis when the unsymmetric search option, PTYPE=2 on the BCONP bulk data entry is selected. Avoidance: Use the symmetric option which is the default. @ 26842;

70.5; Superelements - ASETi, BSETi, CSETi Entries Ignored

When part superelements are present, aseti entries in the main bulk data section are ignored. That is all free degrees of freedom are placed in the a-set, and none in the omitted set. In a non-superelement model all free degrees of freedom not placed on aseti entries are moved to the omitted set. Avoidance: An avoidance is to use BSETi entries in place of ASETi entries @ 26844; ; Aeroelastic Analysis - SPLINEi, OMITi, ASETi - Wrong Answers In SOL 144 and for ANALYSIS=SAERO in solution 200, wrong answers will result if the user splines the aerodynamics to omitted degrees of freedom. Avoidance: Do not spline to omitted structural degrees of freedom @ 26846; 70; Submittal - XNASTRAN - Job will not run with batch=yes If a job is submitted using the XNASTRAN GUI with "batch=yes" and the "after" keyword blank, the job will not execute. Avoidance: Set the "after" keyword to a non-blank value. Setting "after=now" will cause the job to run as soon as it's submitted. See the at(1) man page for additional information and acceptable keyword values. @ 26855; ile 70; Postprocessing - PARAM,POST - Wrong Results Written to OUTPUT2 F If PARAM,POST,-1 or -2 is specified in a modal response run (e.g. SOL 111 or 112), and XYPLOTS are requested, or SORT2 output is requested, then the displacements written to the OUTPUT2 file are the modal eigenvectors instead of the frequency or transient response solutions. Avoidances: Two avoidances follow, however, the first avoidance is inefficient and the second avoidance is recommended. 1) Insert PARAM,DDRMM,-1 or 2) Insert the following DMAP alter (V69.x) compile sedrcvr alter 'if ( app1=\'mmreig\' and not(sort2) )','' if ( app1='mmreig' ) then $

alter ' sdr3 ','' SDR3 qUGV1,OPG1,qQG1,qEF1,qES1,qSTR1/ OUGV2,OPG2,OQG2,OEF2,OES2,OSTR2 $ SORT2 OUTPUT @ 26866; 70; Miscellaneous - CURV - SFM 6806 In MSC/NASTRAN Version 69, if the CURV module is used with elements other than QUAD4s or TRIA3s the run may abort with the following message: *** SYSTEM FATAL MESSAGE 6806 (CURV) RETURNED ERROR CONDITION -3, LOC CODE = CURV1, FILE NUM = 104 (NOTE ADDITIONAL MESSAGES.) 645, IN SUBROUTINE

No CURV module output will occur. The non-CURV module output will be available in the f06 file and is unaffected. Avoidance: There is no avoidance. @ 26867; 70; Optimization/Design Sensitivity Analysis - SELA, INREL - UIM 457 0, UFM 4575 If inertia relief is requested in solution 200 with superelements then the run may fail with UIM 4570 and UFM 4575. Avoidance: Insert the following DMAP alter: compile pslgdv alter 'null egx' type parm,,i,n,nopgs $ NOPGS=NDVAR*6 $ matgen ,/egpg6/7/lusets/nopgs $ @ 26869; 70.0.1; DMAP - OUTPUT4, Sparse - Quits No Message If an OUTPUT4 request is made for a sparse, formatted matrix, then the job quits in the OUTPUT4 module with no message, no "END OF JOB" and the OUTPUT4 file is empty. Avoidance: Do not request sparse output. @ 26879; e 70; Eigenvalue Analysis - SOL 3, 10, 11, 12, EIGRL - Poor performanc A lanczos eigenvalue analysis, EIGRL entry, may require more memory in solutions 3, 10, 11, and 12 than in Version 68.2 or the corresponding structured solution sequence. Avoidance: Insert the following DMAP alter: malter 'cmset=\'v\'' cmset='a' $ @ 26899; ; Composites - Buckling, PCOMP - Spurious Modes Spurious modes may be observed in linear buckling (Sol 105) of composite shells that have extremely soft transverse shear material properties. The

error is under investigation but it is possible that the spurious modes are be due to a limitation in Mindlin shell theory. @ 26953; t deck 70.5; Documentation - V69 Release Notes - Error in sample problem inpu The input deck for the V69 Release Notes sample problem "blexam.dat" contains an error on the continuation card for PBARL 1 (should be "A" instead of "B"). This error is in the file itself as well as the book (pp. 55-57). If run as is, the input gives a fatal error. The corrected test deck has been put on the Version 70.5 delivery in the archive directory, the documentation can not be updated. @ 27029; 70.5; Loads - DEFORM, GRAV - Wrong Answers If DEFORM entries are combined with GRAV entries with the same SID, wrong OLOAD resultants and wrong answers result. It is documented that DEFORM entries require unique SID; however, the following fatal messages should be issued not wrong answers: USER FATAL MESSAGE 628 (IFP1D) THERMAL, DEFORMATION, AND EXTERNAL LOADS CANNOT HAVE THE SAME SET IDENTIFICATION NUMBER. or: USER FATAL MESSAGE 2134 MODULE = GP3 LOAD SET 1 DEFINED FOR BOTH GRAVITY AND NON-GRAVITY LOADS. @ 27036;

70; Loads - PLOAD1, FRPR - Erroneous Output When applying a PLOAD1 with a SCALE of FRPR, and the load direction is parallel (or near parallel) to the beam element, then a zero load will be generated in versions prior to Version 69. In Version 69, an undefined load is generated which leads to output that contains the following character "NANQ". Avoidance: Since the load is parallel to the axis, use the SCALE of FR instead.

@ 27048; 2001; Optimization/Design Sensitivity - FREQ, DRESP1 - SFM 4276, Trap Encountered If a FREQ response type is used on the DRESP1 entry and the associated eigenvalue is negative, the code attempts to compute the frequency by taking the square root of the negative eigenvalue, subsequently a trap occurs in subroutine dsad11 and the job terminates without a meaningful explanation. Avoidance: A negative eigenvalues typically occurs with unsupported rigid body modes. It is probably not the intent of the user to place design conditions on these roots. Therefore, review the design intent and force the eigenvalue analysis to only find non-negative eigenvalues. @ 27055; 70; Static Analysis - SNORM - UFM 316 (IFPDRV)

Even though GRID ids may be up to one hundred million, the GRID ids referenced on a SNORM bulk data entry must be less than one million. If larger GRID ids are referenced the message: USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA ENTRY SNORM will be issued. Avoidance: Other than renumbering the model so that the grid id's are less than one million, there is no avoidance. @ 27065; ; DMAP - Inconsistent lusets - Possible Infinite Loop When using dmap, the lusets parameter of the TA1 module must be set consistent with the values in the SIL table. If these values are not set consistently, the results are unpredictable. In one instance, when lusets was set to 1 rather than the value of 150 set in the SIL data block, the TA1 module produces a corrupted gpect data block. When this data block is subsequently input to the EMA module, the job is trapped into an infinite loop and terminates only after it runs into a resource limit (e.g., insufficient disk space). Avoidance: Set the luset value equal to the second word in the trailer of sil @ 27139; 70; DMAP - DBDICT - Missing Output The datablk(local) option of the dbdict DMAP command which should list the non-nddl datablocks contained in a subdmap does not work. Avoidance: DBDIR ////64 $ can be used to provide the same information. @ 27147; 70; Nonlinear Analysis - CBUSH, PARAM,LGDISP - SFM 2085 Nonlinear jobs with PARAM,LGDISP,1 and CBUSH elements may fail with the message: SYSTEM FATAL MESSAGE 2085 MODULE = NLITER ELEMENT TYPE 102 NOT SUPPORTED FOR NONLINEAR ANALYSIS. SYSTEM FATAL MESSAGE 3007 MODULE = NLITER ILLEGAL INPUT TO SUBROUTINE NLEMG Avoidance: Other than not selecting PARAM,LGDISP there is no avoidance. @ 27247; 70.5; Static Analysis - SOL 1, LSEQ - Missing Output Solution 1 accepts the LOADSET case control command and the LSEQ entries it references. It calculates loads and solutions for all of the load conditions listed on LSEQ entries, but outputs loads only for the first loading condition. Avoidance: Use subcases with LOAD commands instead. @ 27255; ; DMAP - TABPT, CASECC - SFM or Nan in Output When a user dmap uses the TABPT module to print the CASECC datablock, a system fatal message may occur on certain platforms if a parameter is

specified in the case control. Other platforms may not give a system fatal message, but instead print "nan" or "Nan" in the output. Avoidance: Put the parameter in the bulk data if possible. @ 27257; 70.7; Parallel Processing - SYSTEM(107) - SWM 6299 (PARCPU) In NASTRAN Version 69 through Version 70 on HP computers with pa2.0 chips and on all NASTRAN versions on SUN Solaris Version 7, the following message in the execution summary will indicate that no parallel processing is occurring, when in fact it is: *** SYSTEM WARNING MESSAGE 6299 (PARCPU) NO PARALLEL PROCESS CREATED Avoidance: @ 27286; ile Ignore the message.

2001; Transient Response Analysis - STRAIN(PUNCH) - Mislabeled Punch F When strain output is requested for transient analysis, the punch file labels it "POINT ID" instead of "ELEMENT ID". The element ID numbers and the "ELEMENT TYPE" labels are correct.

@ 27291; 70; Eigenvalue Analysis - REIGL - Incorrect Hiwater Memory Output, P oor Performance The REIGL module does not keep track of the maximum amount of memory it uses so the HIWATER value reported in the .f04 file may be underestimated for eigenvalue jobs that go through REIGL. An eigenvalue run using mem=X, where X is the HIWATER value from a previous run of the same deck, may run slower, use a smaller MAXSET (Lanczos block vector size), require many more solves, or may not have sufficient memory to complete. @ 27368; ; Executive - ASSIGN - Cryptic Message, SFM 1019 If a file exists and the keyword NEW is indicated on an ASSIGN statement SYSTEM FATAL MESSAGE 1019 will be printed. Included in this message is text returned by the operating system. It should be noted that under some operating systems this text does not always clearly indicate the nature of the file open problem and may include the somewhat cryptic message "Not a typewriter (errno=25)". @ 27503; 70; Superelements - Split Database - UFM 1127 When using a split database with superelement in Version 69, if the upstream database for a superelement is offline during a residual structure run, then the job will fail with the following error messages: USER FATAL MESSAGE 1127 (GNFIST) DMAP MODULE XPURGE IS ATTEMPTING TO ACCESS DATA BLOCK 201 NAME = BTOPO WHICH IS ON AN UNAVAILABLE DBSET Avoidances: 1. Leave the upstream database online. 2. Modify the NDDL. Replace "PATH NLOOP ..." with: " PATH NLOOP PEID,APRCH,HIGHQUAL,NLOOP $" @

27515;

; SSSALTERS - rflagb.v69 - Restarts Restarts using the rflagb.v69 alter in the sssalter directory have not been tested. While some restarts may work correctly, this can not be determined for any particular run without testing, and perhaps modification of the alter. Restarts are not supported for production work, and should be regarded as an experimental technique when using the alter. Avoidance: Do not use the alter for restarts on production runs.

@ 27516;

70; SSSALTER - effmassa.v69 - UFM 2025 The effmassa.v69 dmap alter included in the sssalter directory for V69 may fail with the message: USER FATAL MESSAGE 2025, UNDEFINED COORDINATE SYSTEM Avoidance: The cause of this is under investigation. If you encounter this error, contact MSC client support for an updated version of the alter.

@ 27524;

2001; Superelements - DMIG - UFM 327 If a DMIG matrix is input after two or more BEGIN SUPER or BEGIN AUXMODEL bulk file delimiters, and the DMIG matrix in each partitioned file has the same name, an error message of the form

*** USER FATAL MESSAGE 327, BAD DATA OR FORMAT OR NON-UNIQUE NAME. DMIG AXGGT SORTED CARD COUNT = 1 is produced. Avoidance: An avoidance is to place a second dmig matrix in each file, with a different name. Only a header entry is needed. The code keeps only one name from the dmig entries in the bulk data deck stored between repeated calls to ifp. @ 27533; 70; Matrix Methods - SPARSE - Wrong Answers on SGI R4400 Computers Very Serious In Version 69 on SGI R4400 computers, if the memory specified with the default memory or a mem= on the command line causes the MEMORY AVAILABLE to be between the MINIMUM MEMORY REQUIREMENT and the SPARSE DECOMP SUGGESTED MEMEORY as listed in USER INFORMATION MESSAGE 4157 of the f04 file, bad results may be calculated. One symptom of the bad results is a very large epsilon in the f06 file. Larger problems are more likely to be affected by this. Sometimes under these conditions the problem will terminate with a Segmentation fault. Avoidance: Specify more memory or set SYSTEM(205)=1 and SYSTEM(198)=1. The latter avoidance can cause up to a 30% increase in decomposition times. MSC RECOMMENDS adding the two system cell settings to install_dir/conf/irix/nast69rc for V69 and install_dir/conf/irix/nast691rc for V69.1 BEFORE continuing production runs. These are the architecture specific rc files (see section 3.6 of the UNIX Configuration and Operations Guide). @ 27561; 70; Post Processing - CHEXA, CPENTA, CTETRA - Wrong Answers in XDB

The nonlinear stress values for the center point of any solid element are wrong in the .XDB graphical database. The stress in the x direction is swapped with the stress for the y direction. All other data is correct (including nonlinear strains for the center point). The data blocks affected are NHEXR, NTETR, and NPENR. Avoidance: There is no avoidance. @ 27572; 69.1; Documentation - Logical Symbol, RC file - UFM (decode_keyword) In Version 69, you can define a logical symbol in the RC file. However, if the logical symbol is defined in the rc file as shown on page 43 of the MSC/NASTRAN Configuration and Operations Guide, then the job will fail with the following error messages: *** USER FATAL MESSAGE (decode_keyword) SYMBOL XXX=yyy Set by ./.nast691rc[1]. This keyword is not recognized. Avoidance: Replace the following statement SYMBOL DATADIR=/dbs/data with SYMBOL=DATADIR=/dbs/data @ 27580; 70.0.1; PLOT - SEUPPLOT - Incorrect Plots If a superelement run is sent to the plotter and the residual contains no plot information, the resulting plots may be bad for SEUPPLOT 0. Avoidance: Include at least one grid/scalar which has a finite displacement. @ 27597; 70.5; Postprocessing - PARAM,POST,-1 or -2 - Mislabeled Datablocks Solutions 1-16 are new in Version 69. If PARAM,POST,-1 or -2 is used in these solution sequences, the output datablocks are mis-labeled. For example, GEOM1 will be mis-labeled as IGEOM1 on the output file and I-DEAS or MSC/PATRAN will not interpret the data properly. Avoidance: Other than using solution 101 thru 200 there is no avoidance. @ 27598; 70.5; Heat Transfer Analysis - Restarts with APP COUPLED - UFM 9086 A restart into a structural analysis with APP COUPLED from a previous heat transfer run fails with the message: USER FATAL MESSAGE 9086 (GETLOOP) THE DATA BASE IS INSUFFICIENT FOR PARAM,LOOPID RESTARTS. CHECK FOR BULK DATA CHANGES RELATED TO GEOMETRY, CONNECTIVITY, AND PROPERTIES, WHICH CAUSE THE RECALCULATION OF NONLINEAR DATA. Avoidance: Insert the following DMAP alter:

compile super1 alter 'PARAML CASECC',''

type parm,nddl,i,y,subid $ PVT PVT,CASECC/PVTX $ PARAML CASECC//'DTI'/SUBID/251////S,N,PATH $ compile getloop alter 1 type parm,nddl,char8,n,aprch $ if ( GETSYS(1,21)=5 and APRCH=' ' ) APRCH='HEAT @ 27611;

' $

70; Optimization/Design Sensitivity - SPC, SPCD, BC - Wrong Answers In Design Optimization (Sol 200) with multiple boundary conditions, an enforced displacement of 0.0 is not preserved after the initial design cycle when one subcase puts a non-zero enforced displacement (using SPC or SPCD) on the same degree of freedom that a different subcase puts an enforced displacement of 0.0. When these conditions apply, the initial design cycle will be correct, but the subsequent design cycles will put the nonzero enforced displacements on all subcases. Avoidance: Use the following DMAP avoidance: compile update nolist, noref $ alter 129 $ V69.0 $alter 132 $ V69.1 paraml ysb//'null'////s,n,nullysb $ if (nullysb < 0) then $ matgen ,/yg1/7/lusets/1 $ else $ alter 130 $ V69.0 $alter 133 $ V69.1 endif $

@ 27617;

70; SSSALTERS - alter9ga.v69 - Infinite cpu Loop If alter9ga.v69 is used in a frequency response solution, an infinite cpu loop will result. This problem did not occur in previous versions of the alter. Avoidance: Call MSC and request an updated version of the alter.

@ 27623;

70; Superelements - SELA - Wrong Answers If there are more than 5 superelements at a single level, then the static load matrices in the downstream superelement will be wrong. Avoidance: Increase BUFFPOOL on the NASTRAN statement or insert the following DMAP alter: compile fea noref nolist alter 'return'(,-1) dbdelete datablk=(zuzr11) where(zname='pav' and zuzr1=*) $ $ compile sela1 souin=mscsou noref nolist ALTER 'TEMP=SEID','END $ SELA1'(1,-2) $ V68.2.3 and up call dbfetch /pav,,,,/s,seid/0/0/0/0 $ add5 pav,pjc,,,/pg $ $ compile phase0 noref nolist

$ $ The alter and GP1 call are for V68.2.3 and up $ ALTER 'IF ( IMAGE OR EXTER ) THEN'(1,+1) $ V68.2.3 and up GP1 GEOM1I,GEOM2Scr,,GDNTAB,MEDGE/ xGPLS,xeQEXIN,xGPDT,xCSTM,xBGPDT,xSIL,xVGFS/ S,N,xLUSET/0/NOGPDT $ call dbstore xeqexin,xsil,,,//s,seid/0/' '/0 $ $ $ compile phase1dr noref nolist alter 1 $ All versions from V68.1 and up type parm,,i,n,iseid=0,isetyp $ $ generate 6x6 mirror transformations matgen ,/i1/6/6/0/1 $ matgen ,/i2/6/6/1/1 $ matgen ,/i3/6/6/2/1 $ matgen ,/i4/6/6/3/1 $ matgen ,/i5/6/6/4/1 $ matgen ,/i6/6/6/5/1 $ add5 i1,i2,i3, ,/t001x/-1./ 1./ 1. $ add5 i4,i5,i6,t001x,/t001y/ 1./-1./-1. $ add5 i1,i2,i3, ,/t002x/ 1./-1./ 1. $ add5 i4,i5,i6,t002x,/t002y/-1./ 1./-1. $ add5 i1,i2,i3, ,/t003x/ 1./ 1./-1. $ add5 i4,i5,i6,t003x,/t003y/-1./-1./ 1. $ matmod t001y,,,,,/t001,/28 $ matmod t002y,,,,,/t002,/28 $ matmod t003y,,,,,/t003,/28 $ mpyad t001,t002,/t012 $ mpyad t001,t003,/t013 $ mpyad t002,t003,/t023 $ smpyad t001,t002,t003,,,/t123/3 $ ALTER 'SEP2DR SLIST,EMAP' $ V68.2.3 iseid=iseid+1 $ superlement counter $ ALTER 'RETURN'(1,-2) $ V68.2.3 $ if ( seid>0 ) then $ if ( image or exter ) then $ call dbfetch /xeqexin,xsil,,,/s,seid/0/0/0/0 $ isetyp=(iseid-1)*7+7 $ word location of type paraml emap//'dti'/1/isetyp//s,n,setyp $ type setyp=setyp/64 $ right shift by six bits if ( exter ) setyp=0 $ if ( setyp=1 ) then $ x matmod t001,xsil,,,,/tcd,/5 $ else if ( setyp=2 ) then $ y matmod t002,xsil,,,,/tcd,/5 $ else if ( setyp=4 ) then $ z matmod t003,xsil,,,,/tcd,/5 $ else if ( setyp=3 ) then $ xy matmod t012,xsil,,,,/tcd,/5 $ else if ( setyp=5 ) then $ xz matmod t013,xsil,,,,/tcd,/5 $ else if ( setyp=6 ) then $ yz matmod t023,xsil,,,,/tcd,/5 $ else if ( setyp=7 ) then $ xyx matmod t123,xsil,,,,/tcd,/5 $ else $ external or identical w/no reflections

paraml xsil//'trailer'/2/s,n,xluset $ matgen ,/tcd/1/xluset $ endif $ dbstatus maps//s,n,nomaps $ if ( nomaps>0 ) then $ mpyad maps,tcd,/tcd1 $ equivx tcd1/tcd/-1 $ endif $ nomaps>0 else $ equivx eqexins/xeqexin/-1 $ for call sela0 PARAML USET0//'USET'//////'A'/S,N,NOASET $ matgen eqexins/intext/9/0/lusets $ upartn uset0,intext/tcd,,,/'g'/'a'/'o'/1 $ a(int) by g(ext) endif $ dbstatus pav//s,n,nopav $ if ( nopav>0 ) call sela0 tcd,eqexins,sils,xeqexin/pav / sedwn $ endif $ $ compile sela0 noref nolist subdmap sela0 tcd,eqexins,sils,xeqexin/ pav/ sedwn $ type parm,,i,n,sedwn,temp $ mpyad tcd,pav,/paup/1 $ matmod paup,xeqexin,,,,/dmigpa,/16 $ delete /pav,paup,,, $ temp=sedwn $ dbview eqexind=eqexins where (peid=temp) $ dbview sild =sils where (peid=temp) $ paraml sild//'trailer'/2/s,n,lusetd $ mtrxin ,,dmigpa,eqexind,sild,/paup,,/lusetd $ call dbfetch /pav,,,,/s,sedwn/0/0/0/0 $ add5 paup,pav,,,/paupd1 $ equivx paupd1/pav/-1 $ call dbstore pav,,,,//s,sedwn/0/' '/0 $ delete /pav,,,, $ return $ end $ sela0 $ compile dbmgr noref nolist alter 1 putsys(1,109) $ alter 'return '(,-1) putsys(0,109) $ compile dbfetch noref nolist alter 1 putsys(1,109) $ alter 'return '(,-1) putsys(0,109) $ compile dbstore noref nolist alter 1 putsys(1,109) $ alter 'return '(2,-1) putsys(0,109) $ @ 27629; 70.5; Buckling - EIGB,SINV - UWM 6278 A buckling analysis in Version 69.1 that specifies the SINV method will

issue the warning message: *** USER WARNING MESSAGE 6278 (MTMD23) THE BULK DATA ENTRY SELECTED BY THE CASE CONTROL COMMAND METHOD NOT FOUND = xx WAS

If there is no fatal message printed in the f06 file, this warning message can be safely ignored. Avoidance: There are two possible avoidances: 1. Ignore the extraneous warning message. 2. Use Lanczos method. @ 27647; 70; Post Processing - CBEAM - Orientation Vector Information not Wri tten to XDB When the GRID option is used to define the orientation of the CBEAM element, the orientation information is not written to the XDB file. Avoidance: Other than using the vector option, there is no avoidance. @ 27659;

70; Optimization/Design Sensitivity - BC, STATSUB - Wrong Answers Eigenvalue sensitivity analyses (normal modes and buckling) require the associated eigenvectors. In Version 69, the appropriate set of normal modes eigenvectors is selected based on the BC case control command. For buckling, the STATSUB command is used. If BC or STATSUB does not result in a unique eigenvector set, wrong answers can result. Avoidance: Use BC and STATSUB requests in a way that ensures uniqueness.

@ 27674; VED LIST

69; Security - UFM 3060, SUBROUTINE MODEL - OPTION NUSR NOT IN APPRO

The message "USER FATAL MESSAGE 3060, SUBROUTINE MODEL, OPTION NUSR NOT IN APPROVED LIST" will appear when running on Windows NT systems that allow multiple users to run concurrently, e.g., Windows NT 4.0 with "remote shell" support. The problem has been fixed in MSC/NASTRAN V69.0.4 on the updated Windows NT CD released with the Digital Alpha NT system. @ 27677; 70; Documentation - SEMPLN,NORMAL - Unsupported Features When defining the mirror plane for a partitioned, image superelement, only the MTYPE=PLANE option on the SEMPLN entry is supported. The Version 69.1 Release Guide incorrectly states that the MTYPE=NORMAL option is also supported. Avoidance: Use the MTYPE=PLANE option on the SEMPLN entry. @ 27685; 70; Post Processing - CTRIAX6 - Missing Information in XDB The "negative" key is not written to the STX6R Object to distinguish between Von Mises and Octahedral Shear stress results for the CTRIAX6 element.

@ 27704; 2001; Nonlinear Transient Heat Transfer - CP is Temperature Dependant - Wrong Answers If CP, specific heat is a function of termperature, in a nonlinear heat transfer transient analysis then wrong answers will be produced. Avoidance: There is no avoidance @ 27717; 70.5; Postprocessing - BAR, BEAM, ROD, CONROD - Wrong Margins of Safet y in XDB The Bar, Beam, Rod and Conrod elements all get wrong answers for the margins of safety in the .XDB file. Avoidance: There is no avoidance. @ 27722; 70; Documentation - QRG, PARAM,PRETEQXIN - UFM 5199 The MSC/NASTRAN Version 69 Quick Reference Guide incorrectly documents PARAM,PRETEQXIN, the parameter should be PARAM,PRTEQXIN. If PRETEQXIN is specified USER FATAL MESSAGE 5199 may result. @ 27734; 70; Documentation - QRG - ESE With Thermal Loads Prior to Version 70 element strain energy output was incorrect in the presence of thermal loads. This limitation was documented in the Quick Reference Guide under Remark 4 of the ESE Case Control Command. This limitation has been removed in Version 70 but the Version 70 QRG had already been sent to press, so Remark 4 was not removed. @ 27739; 70; Nonlinear Heat Transfer - SDR3, XDB - UFM 4276 Nonlinear Heat Transfer jobs that request an XDB (i.e. param,post,0) may fail with USER FATAL MESSAGE 4276 on HP 9000/700 systems. The error appears to be dependent on problem size, machine and operating system. The error has been corrected in Version 70. Avoidance: Other than running on a different system there is no avoidance. @ 27747; 70.5.1; Heat Transfer Analysis - CHBDYi, SSG1 - SFM 3007, Wrong Answers If the model has nonunique element IDs among with heterogeneous type of CHBDYP,CHBDYG, or CHBDYE , MSC/NASTRAN does not know which CHBDYx element ID should be used in the convection or radiation. This confusion in the code may lead to wrong answers or SYSTEM FATAL MESSAGE 3007 in the SSG1 module. The fix to the code allows the program to flag the non-unique CHBDYP and CHBDYG in the GP2 module. Avoidance: Change the Element IDs of CHBDYP and CHBDYG so that the IDs are unique. Duplicate CHBDYi element ids may result in the message:

@ 27755;

; Eigenvalue Analysis - EIGRL, REIGL - SFM 5299 Small models with many repeated modes may require smaller values for MAXSET else the following message may be encountered: *** SYSTEM FATAL MESSAGE 5299.UNRECOVERABLE TERMINATION FROM LANCZOS ITERATION INTERNAL ERROR CODE = -3 USER ERROR RETURN = -33 Avoidance: Reduce the value of MAXSET on the EIGRL entry.

@ 27757; 1305

70; P-Elements - FEFACE,SURFID, EMA, GPSP - UWM 4398, BIOMSG ERROR If p-elements have FEFACE records that use SURFID to specify the geometry of the edges of the face, bad answers or one of the following messages may result. Module EMA may give the message: BIOMSG: ERROR 1305 HAS OCCURRED IN ROUTINE ENDPUT, FILE INDEX (FILEX) = 0 Module GPSP may give the message: USER WARNING MESSAGE 4398 (GPSP) GRID POINT nnnnnnnnn FOR TRANSLATION DEGREES OF FREEDOM DOES NOT PASS SYMMETRY CHECK. NO SINGULARITY TESTS MADE ON THESE TERMS. Avoidance: This error occurs when several elements are connected to the edge with conflicting geometry. The only known avoidance is to identify the edges which are failing, and use consistent geometry for the FEFACEs which are associated with the edge.

@ 27761;

70; Submittal - Can't Modify the Input File Suffix In Version 69, there is no simple method to specify a user or site-dependent default input file suffix for the nastran command. The V70 nastran command allows the keyword 'j.dat' in the '.ini' file to define the default input file suffix. For example, 'j.dat=.bdf' sets the default suffix to '.bdf'. Avoidance: There is no simple avoidance. The nastran script could be modified by the user to add a suffix whenever one is not provided. Since the order of input to the nastran command does not matter, the script would have to search for the input with no equal sign and then append the desired suffix to that input.

@ 27762;

2001; P-Elements - EMG - SFM 3008 A p-element run may terminate with a fatal message if a shell p-element

and a solid p-element share a face and the p-order of the edges of the face is higher than 2. There is no avoidance to this errror. The following fatal message is issued 0*** SYSTEM FATAL MESSAGE 3008 MODULE = EMG 0INSUFFICIENT MEMORY AVAILABLE FOR SUBROUTINE PSHELD 0FATAL ERROR @ 27774; 70; Documentation - EXTRN - Incorrect Format The format for the EXTRN entry as shown in the Version 69.1 Release Guide is incorrect. The correct format is as follows: EXTRN GID1 GID5 C1 C5 GID2 -etcC2 GID3 C3 GID4 C4

Avoidance: Correct the entry as shown above. @ 27775; 70; Documentation - SEBULK - Incorrect TYPE When defining the superelement boundary search options for the external superelement using the partitioned bulk data method, the TYPE field on the SEBULK entry should be specified as EXTERNAL. The V69.1 Release Guide incorrectly states that the TYPE field is either EXTRNLG or EXTRNLA. The documentation will be corrected in the Version 70 Quick Reference Guide. Avoidance: Use TYPE is equal to EXTERNAL. @ 27789; 69.0.4; Matrix Methods - PARALLEL, SYSTEM(107) - Wrong answers If parallel processing is selected by setting SYSTEM(107) or the PARALLEL keyword on MSC/NASTRAN Version 69.0.2 for Windows NT wrong answers may result. Avoidance: Other than not selecting parallel there is no avoidance. A new version of MSC/NASTRAN Version 69 (V69.0.4) is in the process of being released and the problem will be corrected in this version. Also, the new executable (analysis.exe) is available on the MSC FTP site, in file mscnv69.exe in the /pub/outgoing directory. This is a self-extracting ZIP file that contains a single file, analysis.exe. To update an installation, copy this mscnv69.exe file to the install_dir\msc69\i386 directory and run it by entering "mscnv69". Reply "y" to the request to overwrite the existing executable (or rename the existing executable before running mscnv69). This updated version also corrects some UFM 3060 problems that exist with multiple concurrent users (i.e., NUSR failures). @ 27804; 3060 ; Installation - Authorization File not Updated in .ini File - UFM If during an authorization default, the authorization @ 27811; installation of MSC.Nastran on an NT system an existing file is selected and that file's name is different from the .ini file will not be updated with the location of the file. This will lead to USER FATAL MESSAGE 3060.

70.5; Documentation - PBARL, QRG - Missing Information In the MSC/NASTRAN Quick Reference Guide it states that elements which use the PBARL entry to define their properties have the origin of the element coordinate system at the shear center of the cross-section. The manual

does not state that the PBARL does not account for offsets between the neutral axis and shear center. Therefore, the CHAN, CHAN1, and CHAN2 cross-sections may produce incorrect results. Avoidance: If a cross-section does not have coincident shear center and neutral axis, use a PBEAML. @ 27819; s 2001; Miscellaneous - GPSTRESS, Von Mises, Plane Strain - Wrong Answer In a plane strain analysis, the von Mises stress output using the grid point stress option is incorrect. The component stresses using the grid point stress option are, however, correct. Avoidance: If feasible, use the von Mises stress output from the element stress output. @ 27820; ; Nonlinear Analysis - PELAST - Possible Wrong Answers When the nonlinear property PELAST is used for the scalar spring, CELAS1 or CELAS3, the linear stiffness, Ki, must be greater than zero. If Ki is non-zero the actual value is replaced by the PELAST function. However, if 0.0 is entered then the value of 0.0 is used. @ 27821; 70.5; ESTIMATE - Estimate Utility does not search standard MSC/NASTRAN RC files The ESTIMATE utility does not examine the standard nast*rc files to determine default settings for any of its keywords. Avoidance: In order to set default values, you must set them in the $MSC_BASE/$MSC_VERSD/$MSC_ARCH/estimate.ini, $HOME/.estimaterc, or the ./.estimaterc files. ESTIMATE V70.5 will use the nastran command to search the standard MSC/NASTRAN RC files. In order to do this properly, (i) the nastran command executable must present in the same directory containing the ESTIMATE executable, and (ii) ESTIMATE must be passed the same values for <version> and <rcf> (if any) that will be passed to the nastran command. Items set in the ESTIMATE RC files and on the ESTIMATE command line will override the MSC/NASTRAN RC files. @ 27822; 70.5; Optimization/Design Sensitivity - MODTRAK - UFM 316 The Quick Reference Guide appears to allow the user to leave the HIGHRNG field blank with the algorithm defaulting to the highest extracted mode. However, if the user has set LOWRNG and leaves HIGHRNG blank, the following fatal message is issued: USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA ENTRY MODTRAK Avoidance: Either remove the LOWRNG value or specify the HIGHRNG value. @ 27824; 70; Miscellaneous - SDR3 - SFE 4276 EC 7777, SIGNAL 10 ABORTING On the HP that uses the PA2.0 processor, if the SDR3 module is being exercised (whether it is through user-dmap or executed automatically by the SOL sequence, e.g., SOL 111), the job may fail with the following error messages:

***** RECEIVED SIGNAL 10, ABORTING ... BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX (FILEX) = 0 0*** SYSTEM FATAL ERROR 4276.RECOVR ERROR CODE 7777 PID= 0 The module SDR3 is used to convert sort1 output to sort2 output. Avoidance: If feasible, use sort1 output instead of sort2 output to avoid execution of the SDR3 module. @ 27826; UFM 6677 70; Optimization/Design Sensitivity - Mode Tracking and Acoustics -

If fluid elements are present in SOL 200 and mode tracking is requested the run may fail with: USER FATAL MESSAGE 6677 (DOM12E) RUN TERMINATED DUE TO MODE TRACKING FAILURE. SEE USER FATAL MESSAGE 6677 (MTFTRD) PRINTED ABOVE FOR INFORMATION. Avoidance: Insert the following DMAP alter: COMPILE FEA ALTER 'MODTRK','MODTRK' MODTRK CASEM,EDOM,LAMAS,MGG,MAA,PHG,PHSA, PHGREF0,PHAREF0/ MTRAK,NEWLAMA,NEWPHG,NEWPHA,PHGREF,PHAREF/ DESCYCLE/S,N,NOTRACK $ @ 27828; 70; Heat Transfer Analysis - RADMT - UFM 4658 Large models with RADMT bulk data entries may fail in NLITER with the message: USER FATAL MESSAGE 4658 . DIFFICULT GEOMETRY PREVENTS FURTHER COMPUTATIONS FOR ELEMENT WITH ID = *** Avoidance: The only avoidance is to increase memory. @ 27833;

; Documentation - Getting Started with MSC/NASTRAN The "Getting Started with MSC/NASTRAN User's Guide" states that "In any coordinate system, all angular input is in degrees, but output ... is in radians". This statement needs clarification. The "input" refers only to geometric locations. If an angular enforced displacement is specified the input is in radians.

@ 27867; TKG

70; Superelements - SEMA - SFM 3007, ILLEGAL INPUT TO SUBROUTINE MPG Superelement jobs with SETs with large thru ranges may encounter the following fatal message: SYSTEM FATAL MESSAGE 3007 MODULE = SEMA 0ILLEGAL INPUT TO SUBROUTINE MPGTKG

0FATAL ERROR Avoidance: There is no practical avoidance. @ 27872; 70; Dynamics - PARAM,APHAi - Raleigh Damping Ignored Raleigh damping is ignored when using param,alpha1 (proportional to mass) and alpha2 (proportional to stiffness). Avoidance: Add the following dmap avoidance. compile gma $ alter 'call.*xadd5.*mhh.*khh.*bhh','' $ add5 mhh,khh,bhh,,/bhhx/alpha1/alpha2/cu $ alter 'call.*xadd5.*mdd.*kdd.*bdd','' $ add5 mdd,kdd,bdd,,/bddx/alpha1/alpha2/cu $ @ 27879; m XDB 70; Post Processing - Optimization, Superelements - Missing Data fro In solution 200 with superelements, requested data may be missing from the XDB for any design cycle after the initial analysis. Avoidance: Insert the following DMAP alter: compile update alter 'do while'(,-1) descycle=descycle-1 $ alter 'descycle=descycle-1','' @ 27880; 70; Superelements - Postprocessing, PARAM,POST,0 - SWM 6065, 6066 Superelement jobs that set PARAM,POST,0 to produce an .XDB file may encounter the following messages when the new partitioned bulk data superelements are used: SYSTEM WARNING MESSAGE 6065 (DBC) *** DBCMSG ERROR MESSAGE 11 FROM SUBROUTINE DBCEQG PREMATURE EOR FOR MSC/NASTRAN LOGICAL FILE GEOM1S 0, DBSET 0.

FOR SEID

*** SYSTEM WARNING MESSAGE 6066 (DBC) *** DBCMSG ERROR MESSAGE 34 FROM SUBROUTINE DBC THE DBC MODULE HAS BEEN TERMINATED BECAUSE OF THE ABOVE FATAL DBC ERROR. USER INFORMATION: SUBSEQUENT DBC CALLS WILL NOT BE EXECUTED, AND THE DATABASE CONVERSION IS INCOMPLETE. @ 27881; 70; Miscellaneous - IFPDRV - UFM 316 Large input files (<1 million entries) can cause overflow in the sorted entry count of an error message. For instance: *** USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA ENTRY MPC 1 SORTED ENTRY COUNT = ****** ^^^ USER FATAL MESSAGE 9002 (IFPS) ^^^ ERROR(S) ENCOUNTERED IN THE MAIN BULK DATA SECTION

^^^ SEE MESSAGES ABOVE. ERROR ENCOUNTERED IN MODULE @ 27917;

IFP,

70.5; Superelements - External Superelement, SEP1X - SFM 3001 In Version 69.1, when using the parameter EXTOUT=dmigop2 option for external superelements, and more than one external superelement is brought in, then the job will fail with the following error messages: *** SYSTEM FATAL MESSAGE 3001 (SEP1X ) THE EXECUTIVE FILE 0 DOES NOT EXIST. PROGRAMMER INFORMATION: SUBROUTINE SEPMAB 0FATAL ERROR Avoidance: If more than one external superelement is to be brought in, use the dmigpch option. Or insert the following DMAP avoidance: compile phase0 alter 'dbview geom1f','dbview dbview geom1f=geom1qs where ( dbview geom2f=geom2s where ( dbview geom4f=geom4qs where ( dbview sgpdtf=sgpdts where ( geom1f'(,3) peid<>0 ) $ peid<>0 ) $ peid<>0 ) $ peid<>0 ) $

@ 27918;

70; Miscellaneous - DEQATN - SFM 6048 A DEQATN entry is validated for syntax by evaluating the equation with all variable set to zero. If this evaluation results in an invalid operation, like the square root of a negative number, then the following message may be unnecessarily issued: SYSTEM FATAL MESSAGE 6048 (XQFNC7) Avoidance: The error should be able to be avoided by adding appropriate absolute value operators to the equation.

@ 27927;

70; Documentation - QRG, PARAM,ACOUT - Poor Description The description of the ACOUT parameter in the Version 69 Quick Reference Guide states; To obtain sound pressure level in units of dB and dBA given by the FORCE, DISPLACEMENT, and PRESSURE commands, a peak reference pressure must be specified with PARAM, PREFDB. The sentence should read: To obtain sound pressure level in units of dB and dBA given by the FORCE command, a peak reference pressure must be specified with PARAM, PREFDB.

@ 27930;

; Acoustic Analysis - Modelling Error - SFM 6169, Poor Message When a plate is immersed in a fluid, then the structural grid points will be attached to up to eight fluid HEXA elements (more if PENTAs and/or TETRAs are use) and two fluid/structure interfaces, one for each side of

the plate structure. The fluid grids on each side of the plate are on different interfaces and thus must have unique coincident grids. Thus, at a point in space there should be two fluid grids and one structural grid. If only one fluid grid is erroneously used, that is, the fluid elements on both sides of the interface use the same grid, then the following error will occur: *** USER FATAL MESSAGE 6169, FOUND TWO OR MORE STRUCTURAL GRID POINTS WITH IDENTICAL COORDINATES WHICH LIE ON THE FLUID/STRUCTURE BOUNDARY. USER ACTION: CHECK GRIDS AND/OR SET1 CARDS WHICH BELONG TO PANEL OR ACMODL CARDS. GRID ID = nnnn GRID ID = xxx @ 27936;

70; Documentation - External Superelement - Missing FMS statements Item 1 of Step 3 of the data recovery procedure as described on page 8 of the MSC/NASTRAN Version 69.1 Release Guide is incomplete. Avoidance: Add the following FMS statements. Assign sexx='step1.MASTER' Restart logical=sexx where step1.MASTER is the database from the first step.

@ 27937;

70; Miscellaneous - NASTRAN - NASTRAN Command Ignored. The NASTRAN command must be the first entry in an input file, disregarding comment lines. If the NASTRAN command is mixed in with FMS commands such as ASSIGN statements, the command may be ignored.

@ 27938;

70; Cyclic Symmetry - EMG - UFM 2016, Various UFM APP HEAT will be ignored in cyclic symmetry jobs. The job may fail in various ways, one likely way is: USER FATAL MESSAGE 2016 MODULE = EMG NO MATERIAL PROPERTIES EXIST. FATAL ERROR Avoidance: Apply the following dmap avoidance. COMPILE SUPER1 NOLIST NOREF ALTER 'IF (GETSYS(1,56)=1)'(,-1) $ PUTSYS(1,56) $ HEAT TRANSFER

@ 27944;

70; Rigid Elements - RSSCON - UFM 6692 In nonlinear statics (Sol 106), the following user fatal messages are issued when a bulk data entry RSSCON with option

"ELEM" is in the residual structure. *** USER FATAL MESSAGE 6692 (GP4RS1) BULK DATA ENTRY RSSCON CAN NOT FIND ELEMENT IDS, USER ACTION: CHECK SUPERELEMENT STRUCTURE. *** SYSTEM FATAL MESSAGE 3001 (GP4 ) THE INPUT DATA BLOCK IN POSITION 11 DOES NOT EXIST.

Avoidance: Use the following DMAP alter which provides the element connection GEOM2S to the GP4 module. compile nlstatic souin=mscsou, nolist, noref alter 'GP4 CASESX2,GEOM4S','GP4 CASESX2,GEOM4S' $ GP4 CASESX2,GEOM4S,EQEXINS,SILS,GPDTS,BGPDTH,CSTMS,,,,GEOM2S,/ RGNL,YS1H,USET1,/ LUSETS/S,N,MPCF1/S,N,MPCF2/S,N,NOSSETX/S,N,NOOSETX/ S,N,NORSET/0/0/S,N,NOSETX/S,N,NOLSET/S,N,NOMSO/0 $ endalter $

@ 27964; ; Old Design Sensitivity - DSVG1, PARAM,EIGD - SFM 4276 EC 1409 IN ROUTINE GETSTR When param,eigd,yes is used in solution 103 to calculate the eigenvector derivatives, SYSTEM FATAL MESSAGE 4276 ERROR CODE 1409 is issued from module DSVG1 in subdmap DSAMODES. Avoidance: Use the alter evds103a.v691 provided in the SSSALTER library. @ 27968; ; Submittal - PSL: NUSR: unable to initialize (0 expired keys) The following error is issued in the log file when an abnormality occurs in the NUSR queues in the system. (i.e. if a nastran job gets killed some of the NUSR queues will not be removed from the system and following message will be issued when a subsequent job is started. PSL: NUSR: unable to initialize (0 expired keys). Avoidance: Flush the queues by rebooting the machine. or, Remove nastran related queues by the command ipcrm -q PROCESSID of the queue. @ 27996; 70; Superelements - Plotting, PLTSET - SFM 3062, Subroutine CNSTRC Use of a plot set defining a list of elements and grids to be used by the structure plotter cause a SFM 3062 whenever the exclude option is used. For example:

SET 1001 =110106 THRU 110452 110553 220101 THRU 220120 120274 THRU, 120292 120134 THRU 120178 230274 THRU 230286 220208 220218, . . 5033 5080 241 5034 5078 240 209 218 520 545 208 219, exclude grid points 1 thru 999999 results in the SFM 3062 from the PLTSET module. Avoidance: Commenting out the exclude set from the set definition allows the run to proceed but the plots are not consistent with the request of the set. That is, more information is plotted than requested. @ 28008; 2001; DMAP - MATMOD Option 2 - SFM 4276, Traceback This option of matmod filters small numbers from the input matrix. When given a complex matrix, a floating point overflow occurs at 1.E19 on rs/6000 computers, and other relatively small numbers on other computer types. Since the range of this computer type is greater than 1.E30, this is not a reasonable limitation. The diagnostics from such a crash are machine dependent. On the rs/6000, the following appears in the .log file: SIGTRAP - Trace trap Traceback: Offset 0x00000028 in Offset 0x0000042c in Offset 0x000028e4 in FP overflow procedure pack_ procedure mtmd02_ procedure matmod_

Avoidance: Use the NORM module to first find the largest number in the matrix, scale it with the inverse of this number before using matmod(2), then re-scale the results. The NORM module can handle numbers larger than matmod(2) can. @ 28009; 70; Miscellaneous - TABLED1, TABLEM1 - SFM 4276, EC 7777 Extrapolation of data from either TABLED1 or TABLEM1 entries beyond the end points of the table may result in numerical overlfow. Version 70 has been enhanced to trap machine overflow due to extrapolation and will issue USER FATAL MESSAGE 6917 to address this error. Avoidance: Add additional data points at each end of the curve to make the curve flat so that extrapolation to any x value can not be unbounded. @ 28011; 70.5; Documentation - CHEXA, CPENTA, CTETRA, Plot Codes - Missing Defi nitions The stress item codes 8, 16 and 22 in the Quick Reference Guide are incorrectly labeled for the CHEXA (page A-7), CPENTA (page A-9 and A-10), and CTETRA (page A-17). Modern textbooks arrange the principal stresses in this order: First Principal> Second Principal> Third Principal. MSC/NASTRAN arranges the principals as: First Principal> Third Principal> Second Principal. MSC/NASTRAN arranges the principals this way because the principal stresses are the roots of a cubic polynomial and the this is the order that the method that extracts the roots puts out. The stress item codes 8, 16, and 22 for the solid elements should have the labels changed to the

following, 8 - First principal 16 - Third principal 22 - Second principal A note should also be added to the element descriptions that indicate that the principal stresses are output such that A>B>C. @ 28027;

70; Elements - TETRA, TEMP(LOAD), MAT9 - Wrong Answers In MSC/NASTRAN Version 69 on the IBM RS6000 power2 computer, the 10 noded TETRA element in conjunction with MAT9 and thermal loads produces wrong answers whenever there is also HEXA elements in the model. Avoidance: The non-power2 executable can be selected with the exe submittal line command but this executable will run significantly slower.

@ 28047;

70.5; Material Nonlinear Analysis - TABLES1 - Wrong Answers If stress-strain data is entered on the TABLES1 entry in descending order wrong answers may results. Avoidance: Enter data in ascending order.

@ 28052;

70; Dynamics - OFREQ and OTIME, SOLs 108 and 109 - Wrong answers The OFREQ and OTIME case control commands will produce wrong answers in SOLs 108 and 109 in MSC/NASTRAN Version 69.1. Avoidance: Insert following DMAP alter to avoid this error: compile disprs alter 'upf1' equivx upf1/upf/-1 $

@ 28063;

70.5; Acoustic Analysis - SDR2,OUG1F - Poor Performance Acoustic analysis jobs that request data recovery sets with the thru option may experience very poor performance in the SDR2 module. Avoidance: There is no avoidance.

@ 28070;

70; Postprocessing - PARAM,POST,0 - XDB File Left on Remote Node When using the <node> keyword, the XDB file is left on the remote node. This file should have been deleted after it was copied from the remote node to the local node. Avoidance: Add the following line: %trans=[ya]:% %s.rsh% %node% -l %username% /bin/rm -f %r.out%.xdb

just before the following line in the <install-dir>/bin/nast*.rmt file:

%trans=[ya]:%} @ 28073; 2 70; Optimization/Design Sensitivity - QUAD4, CTRIA3, SNORM - SFM 306 In Version 69.1 of MSC/NASTRAN, optimization and design sensitivity runs with shell normals may fail with SYSTEM FATAL MESSAGE 3062. @ 28087; 70; Miscellaneous - ESTIMATE - SMEM and BPOOL Parameter Not Passed t o Utility If the ESTIMATE utility is being used by the nastran command to estimate the memory requirements of a job, the nastran command fails to pass in either "smemory" or "bpool". If set to non-default values, they could have a significant impact on ESTIMATE's accuracy. Avoidance: Either use the default "smemory" and "bpool", or set the values in the following files: UNIX: $HOME/.estimaterc or <jid-dir>/.estimaterc NT: %HOME%\estimate.ini or <jid-dir>\estimate.ini VMS: SYS$LOGIN:estimate.ini or <jid-dir>:estimate.ini @ 28088; t Used 70; Miscellaneous - ESTIMATE, SYMBOL - SYMBOLS Defined in RC File No The nastran command will not use SYMBOLs defined in the RC files on the command line. For example, if the following line exists in one of the system RC files or the $HOME RC file, symbol=MYDIR=/my/main/project The following will fail msc69 nastran MYDIR:myjob ... Avoidance: Set the SYMBOL values that will be used in the command line as environmental variables, e.g., export MYDIR=/my/big/project MYDIR=/my/big/project; export MYDIR setenv MYDIR /my/big/project # Korn shell # Bourne shell # C-shell

and use the environment variable on the command line. msc69 nastran $MYDIR/myjob ... @ 28102; 70; Submittal - System RC Files on Remote Node Cannot Contain 'nastr an' Statements The system RC files on a remote system cannot contain a 'nastran' statement. For example $ msc691 nastran myjob node=remotenode ...

The system RC files on remotenode cannot contain a 'nastran' statement. Note: This restriction *only* occurs if the "node" keyword is used. Avoidance: Remove the 'nastran' keyword and set system cells by number. For example, change nastran system(196)=1, massbuf=10 to system(196)=1 system(199)=10 @ 28104; ; DMAP - TABPT of CSTM and BGPDT - Wrong Answers on Double Precisi on Machines If the TABPT module is used to print the CSTM or BGPDT tables on double precision, short-word machines, then incorrect data for the tables is printed. Avoidance: Use TABPRT (Formatted print of selected table data blocks) to print the CSTM and BGPDT table. TABPRT TABPRT TABPRT TABPRT $TABPT @ 28123; CSTM/////1$ CSTM//'CSTM'$ BGPDT/////1$ BGPDT//'BGPDT'$ CSTM,EQEXIN,BGPDT/ $

70; Installation - Invalid Media for V69.1 (R3) CD for HP The MSC/NASTRAN V69.1 (r3) CD has an error when attempting to install on an HP system with a local CD. The installation script, mscsetup, will incorrectly report the following error: MSC Installation and Setup - Invalid Media The specified CD pathname /CDROM does not contain any of the installation filesets. Please ensure that the pathname is correct, and the correct installation media is mounted. Press ENTER or RETURN to continue... Avoidance: There are two avoidances: 1) Copy the installation script to a local directory cp /CDROM/MSCSETUP.\;1 mscsetup Change line 3475 from

iv_file="$iv_dev/${iv_file};1" to iv_file="${iv_file};1" Install the CD using the modified script mscsetup 2) Use the installation script in TAN #2828.

@ 28124;

70; Nonlinear Analysis - PELAST - UFM 4551, No Convergence Nonlinear analysis with CELAS elements that reference PELAST may encounter difficult converging. The problem is caused by an error in the element's coordinate transformation. Avoidance: Replace the nonlinear spring by a ROD element with nonlinear elastic material.

@ 28128;

70.5; Documentation - CMASS1, CMASS2 - Incorrect Description in QRG The description of the CMASSi bulk data entries in the Version 69 and 70 Quick Reference Guide indicates that multiple degrees of freedom, "up to six unique integers", may be specified for the component degree of freedom. This is incorrect, only a single degree of freedom may be specified.

@ 28158; 70; Submittal - Remote Execution will fail if MSC_BASE is not Proper ly Defined. The remote execution capability of the nastran command will fail if the MSC_BASE on the remote machine is not properly defined. The command will fail even if the nastran command on the remote machine was able to determine the proper MSC_BASE value. Avoidance: Make sure the MSC_BASE value is properly defined in the <install_dir>/bin/nast70 script. @ 28169; 70; Miscellaneous - Cray T90 Hardware Error May Cause Wrong Answers Cray Research has reported a T90 hardware error that has the potential to return random results for certain vector operations. The specific MSC/NASTRAN routines that may be affected by this error have been identified. It is very important to note that: 1) In our testing, and in specific testing of MSC/NASTRAN by Cray Research, the effects of this error have NOT been seen. 2) This error is ONLY present on T90 systems.

3) No client has ever reported a problem related to this error. The specific MSC/NASTRAN executables that are affected by this error are: MSC/NASTRAN Version --------------------V68.0 V68.1 V68.2 V68.2.3R2 V69.0 V69.1 V70 BETA Executables --------------------------------------unicos/analysis.C90 unicos/analysis.C90 unicos/analysis.c90 unicosc90/analysis, unicosts/analysis unicosts/analysis unicosts/analysis unicosts/analysis

In summary, this affects all C90 and IEEE T90 executables for V68.X, and only the IEEE T90 executables for V69.X and V70. We have recompiled the IEEE T90 V70 executable with a new compiler that prevents this error from occurring. Avoidance: There are two avoidances: 1) Do not use one of the above executables. 2) Disable bi-directional memory using the cpu(8) utility. This may result in up to a 15% performance penalty. See the man page for details as the command may vary at each installation. @ 28184; 70.5; Documentation - PARAM,COUPMASS - Incorrect Description The definition of PARAM,COUPMASS in the Version 69 and 70 MSC/NASTRAN Quick Reference Guide is incorrect. It says: COUPMASS Default=-1 COUPMASS>=0 requests the generation of coupled rather than lumped mass matrices for elements with coupled mass capability, ... It should say: COUPMASS Default=-1 COUPMASS>0 requests the generation of coupled rather than lumped mass matrices for elements with coupled mass capability, ... Note that the action of the code has never changed. To obtain coupled mass a setting of 1 or greater has always been required. @ 28186; em. 70.5; Submittal - Job will fail if <sdir> is set to non-local filesyst Jobs on an SGI R8K, R10K system will fail if the <sdirectory> keyword or TMPDIR environment variable is set to a non-local, i.e., NFS-mounted, filesystem.

Avoidance: There are two avoidances: 1) Do not set <sdirectory> or $TMPDIR to an NFS-mounted filesystem. 2) In <install-dir>/bin/nast*.lcl, change the line "TMPDIR=%sdir%; export TMPDIR" to "TMPDIR=%TMPDIR%; export TMPDIR" @ 28194; 70; Documentation - QRG, CQUADR, CTRIAR, PCOMP - Missing Output The Quick Reference Guide indicates that a PCOMP entry may be referenced by the CQUADR and CTRIAR element. It should be noted that composites are not supported for these elements and it is not possible to obtain lamina stresses and strains. @ 28199;

; External Superelement - SEBULK, EXTRN - Possible Wrong Answers When using the enhanced external superelement capability in Version 69.1, if more than one external superelement is dblocated in a single assembly run using the matrixdb option, then the answers may be wrong. Also see error 27917. Avoidance: Attach one external superelement at a time in the assembly run. Attach additional external superelements by making successive restarts--one additional restart for each additional external superelement.

@ 28205;

69; Post Processing - Special Version of HP-UX dbiolib.a for Catia Catia Versions 4.1.6 and 4.1.7 require a shared dbiolib.a to link with MSC/NASTRAN Version 68.0 and Version 68.1 on HP 9000 systems. The standard dbiolib.a delivered with V68.0/V68.1 was not built this way (all subsequent releases were). This standard library is available from the anonymous ftp site under the URL ftp://ftp.macsch.com/pub/msc-products/NASTRAN/V68/v68-hp9000-dbiolib.a+z This file is NOT required for MSC/NASTRAN V69 and later versions.

@ 28210;

70.5; Documentation - RFALTER and RFINCLUDE - Commands No Longer Exist Starting with Version 69, the RFALTER executive control statement and the RFINCLUDE FMS statement no longer work. This is due to the removal of RFALTERs from the program. The Quick Reference Guide incorrectly still lists these as valid commands. Avoidance: Use the INCLUDE command.

@ 28218;

70; External Superelement - SOL 101, NO SPC AND SCR=NO - UFM 9050 If the reduced external se is unconstrained and scr=no is used, then the job will fail with UFM 9050 (run failed due to excessive pivot ratio). The subsequent residual restart run will fail with the following messages:

SYSTEM FATAL MESSAGE 6551 (INPTX2) AN END-OF-FILE WAS DETECTED ON UNIT 30 DURING BINARY INPUTT2 PROCESSING PROGRAMMER INFORMATION: 1. A KEY VALUE WAS BEING READ 2. THE LAST KEY VALUE READ: 0 FATAL ERROR. 0 NO. FORTRAN RECORDS READ = 0 NO. FORTRAN WORDS READ = 0 0FATAL ERROR Avoidance: There is no known avoidance. @ 28219; 70.5; Documentation - QUADR, PCOMP - UWM 4385 The Quick Reference Guide indicates that the QUADR element may reference a PSHELL or PCOMP property entry. Although referencing a PCOMP entry may not cause a fatal error, technically the QUADR element does not support composites and referencing a PCOMP will lead to the following message: USER WARNING MESSAGE 4385, STRESS PROCESSING FOR NON-COMP ELEMENTS IN SUBCASE XX NOT COMPLETE BECAUSE OF INSUFFICIENT DATA IN DATA BLOCK OES1A @ 28226; GL; Elements - QUAD4 - Incorrect Corner Stresses and Forces with Var iable Thickness The corner stresses and forces are incorrect for the QUAD4 element if variable thickness is specified. Avoidance: Do not request corner stresses, request stresses at the centroid (the default). @ 28227; 70.0.1; Optimization/Design Sensitivity - DSAE - UFM 3003, Wrong Answers Beginning in Version 69 a shape optimization deck with superelements whose grids are designed may either fail with the message "USER FATAL MESSAGE 3003 MODULE = DSAE, READ PAST DATABLOCK TABEVS" or result in wrong answers. Avoidance: Either do no use superelements or use Version 68. @ 28234; PU 70.5; Acoustic Analysis - SOL 111 with CAABSF elements - Inefficient C Use of frequency dependent material property elements such as the CAABSF element may make frequency response analyses take an order of magnitude more cpu time. Avoidance: Insert the following DMAP alter, in Version 69 or V69.1 solutions 108 or 111, to improve performance. compile sedfreq alter 'call freqrs'(,-1) call dbstore vgfs,,,,//0/0/' compile semfreq alter 'call freqrs'(,-1) call dbstore vgfs,,,,//0/0/'

'/0 $ '/0 $

compile freqrs alter 'do while'(,-1) call dbfetch /vgfs,,,,/0/0/0/0/0 $ trnsp vgfs/vgfst $ matmod vgfst,,,,,/vgsf,/12//1 $ alter 'ckddf1','' call fd3mpy kggf1,vphfs,vgsf,vgfs,rgdfd/ckddf1 $ alter 'cbddf1','' call fd3mpy bggf,vphfs,vgsf,vgfs,rgdfd/cbddf1 $ alter 'k4ddf','' call fd3mpy k4ggf,vphfs,vgsf,vgfs,rgdfd/k4ddf $ compile fd3mpy subdmap fd3mpy kgg,vphfs,vgsf,vgfs,rgx/kdd $ matmod kgg,,,,,/nzk,/12//1 $ add5 vgsf,nzk,,,/vgnzf $ partn kgg,vgnzf,/kfgg,,, $ add5 vgfs,nzk,,,/vgnzs $ partn kgg,vgnzs,/ksgg,,, $ paraml ksgg//'null'////s,n,noksgg $ if ( noksgg>-1 ) then $ partn rgx,VPHFS,Vgnzs/rsgx,,,/1 $ smpyad rsgx,ksgg,rsgx,,,/ksdd/3////1////6 $ endif $ paraml kfgg//'null'////s,n,nokfgg $ if ( nokfgg>-1 ) then $ partn rgx,VPHFS,VgnzF/,,rfgx,/1 $ smpyad rfgx,kfgg,rfgx,,,/kfdd/3////1////6 $ endif $ MERGE ksdd,,,kfdd,VPHFS,/kdd///6 $ return $ end $ fd3mpy

@ 28275;

70; Documentation - Superelements - UFM 6794 The example jet101a.dat in the MSC/NASTRAN Version 69 Release Notes section describing superelement enhancements contains an error that will lead to: UFM 6794 (SEPSEB) The error is caused by duplicate boundary points connecting to superelement 4 at grid 26. The input requires the following changes to work: $GRID,4001,,-1.975,-3.4125,0. $GRID,4002,,-1.925,-3.9875,0. GRID,4003,,-1.925,-3.9875,1. $SELOC,5,1,3,6,4001,4002,4003 SELOC,5,1,3,6,4101,4102,4003

@ 28281; put

70.5; Documentation - QRG, PARAM,PRTRESLT PARAM,PRTMAXIM - Missing Out The MSC/NASTRAN Quick Reference Guide indicates that PARAM,PRTRESLT and PARAM,PRTMAXIM are supported for solution 106. They are not supported and their specification will yield no output. These parameters on supported

in solution 106 starting in Version 70.5. @ 28304; ; Documentation/Online Encyclopedia - Missing Footnotes The element summary table (Table 1. Element Summary - Small Strain Elements) in the Encyclopedia Reference Manual chapter 5.1 pg 158-164 is missing important footnotes. The symbols appear in the table but the corresponding footnotes are missing. The critical footnote that is missing is that that midside nodes are not allowed for elements in geometric nonlinear analysis. Avoidance: See the printed version of the Reference manual for the proper footnotes. @ 28338; 70.5; Optimization/Design Sensitivity - CBUSH - UFM 2042 When a design optimization analysis has a DRESP1 entry with RTYPE=WEIGHT and CBUSH elements, it may result in the following error message: *** USER FATAL MESSAGE 2042 MODULE = WEIGHT MISSING MATERIAL TABLE ******** FOR ELEMENT xxx.

Avoidance: Specify VOLUME instead of WEIGHT as the RTYPE. @ 28348; 70.5; Postprocessing/Buckling Analysis - PARAM,POST - Missing Output In buckling analysis, solution 5, the static results are not written to the xdb file. The buckling results are written properly. Avoidance: Insert the follwing DMAP alter: compile sol5 alter 'EQUIVX CASECC\/CASEXX','' if ( post=0 ) soltype='statics' $ CASE CASECC,/CASEXx/'TRAN'/1 $ alter 'DBC OSTR1' soltype='bkl0' $ @ 28373;

70.5; Installation - mscsetup hangs when starting FLEXlm As documented in the INSTALL.html and INSTALL.txt files, and in the mscsetup output, installing FLEXlm requires root privileges. Specifically, root is required to modify the /etc/inittab file. If non-root attempts to install FLEXlm AND modify /etc/inittab AND start lmgrd before the next system boot, mscsetup may appear to hang after it writes the following message to the screen: Starting /etc/inittab command now... The system is actually waiting for the password of the account that owns the lmgrd process. Avoidances: There are two avoidances-

1) Install FLEXlm as root. 2) Enter the password for the user that owns the lmgrd process if mscsetup appears to hang as described above. @ 28383; 70.5; Limitation on Problem Size - Version 70 on HP-UX Doesn't Support Large Files MSC/NASTRAN V70 on HP-UX does not support large files. Both the PA-RISC 1.1 and PA-RISC 2.0 executables should have supported large files, i.e., files larger that 2GB. This support is missing. Avoidance: Use FMS commands to partition the database into multiple files, each of which is smaller than 2GB. @ 28395; 69; Nonlinear Normal Modes - PARAM,NMLOOP - SFM 3002 When performing a nonlinear normal modes analysis (param,nmloop,x), if the load increment is anything other than one, then the job will fail on the NT system in Version 68.2.5 with the following error messages: *** SYSTEM FATAL MESSAGE 3002 MODULE = CASE EOF ENCOUNTERED WHILE READING DATA BLOCK CASEI (FILE 201) IN SUBROUTINE CASE FATAL ERROR Avoidance: If the number of load increments is more than one, break the subcase into multiple subcases--one load increment for each subcase. @ 28421; 70.0.1; Nonlinear Analysis - CBUSH, COR1S, CORD2S, CORD1C, CORD2C - Wron g Answers If the CBUSH entry references a non-cartesian coordinate system, wrong answers result. Avoidance: Use a cartesian system. @ 28483;

70.5; P-Elements - GMSURF, GMCURV - UFM 4659 If an element edge being mapped to a GMSURF or GMCURV has tangents which are anti-parallel at the endpoints, the mapping is incorrect, possibly causing the following fatal message in EMG. *** USER FATAL MESSAGE 4659 (EMGPSD) DURING INTEGRATION FOR ELEMENT WITH ID = BAD GEOMETRY HAS BEEN DETECTED. 1

An example of an element that will cause this error is an element edge spanning a semicircle. Avoidance: Move one of the endpoints or split the element into multiple elements, such that the tangents are not anti-parallel. @ 28508; 70; SSSALTERs - alter1ga.v691 - UFM 5423 (DUMERG)

Use of alter1ga.v691 from the sssalter library may result in USER FATAL MESSAGE 5423 point to the trailers for matrices GOAQ and AIDENT matrices. Avoidance: If this error occurs, call your local MSC office for an updated version of the alter. If you wish to make the correction yourself, change the line: UMERGE USET,QIDENT,/AIDENT/'A'/'Q'/'T' $ to: umerge1 uset,qident,,,/aident/'A'/'Q'/'T' $ @ 28509; 70; Parallel Processing - SYSTEM(107) - SFM 427, EC 7777 It has been reported on a Cray Version 69 system that when more than one processors is requested, then the job may fail with the following error messages: 0*** SYSTEM FATAL ERROR 4276.RECOVR ERROR CODE 7777 PID= 0

Avoidance: We have been unable to duplicate this error at MSC. To avoid the error simply do not select parallel processing. @ 28525;

70; Documentation - V70 COG: HP Remote CD Install, V70 No such file The "MSC/NASTRAN V70 Configuration and Operations Guide UNIX Edition", Section 2.2.2, for remote CD installation of V70 on HP systems specifies file_set = "/CDROM/MSCSETUP.TAR;1" It should specify file_set = "'/CDROM/MSCSETUP.TAR;1'"

@ 28533;

; Acoustic Analysis - VELO(PHASE) - Wrong Answers If the user requests VELO(PHASE) output zero, then NASTRAN will not convert the will print the real/imaginary result in example of the case control request and VELO(PHASE,SORT2) = 100 X-VELOCITY 0.0 -0.0159 Avoidance: Correct E format results may be obtained by selecting VELO(REAL,SORT2), the phase option is not applicable. X-VELOCITY 0.0 1.591795E-02 and the resultant velocity is near result to the phase format and an F rather than an E format. An the incorrect output is:

@ 28554;

; Matrix Methods - Iterative Solver, ITER - SFM 6723 The iterative solver may issue the message "SYSTEM FATAL MESSAGE 6723" NON POSITIVE DIAGONAL ELEMENT IN SCALAR ROW INCREASE IPART VALUE ON BULK DATA CARD". Avoidance: IPAD may be increased by 1, the current value is given in SYSTEM INFORMATION MESSAGE 4157 in the f04. However, SFM 6723 is often an indication of an ill-conditioned model. Thus, the model should be reviewed by plotting and/or by using the checka alter in the sssalter library.

@ 28557; 70.0.1; Frequency Response - MPYAD - SFM 4276, EC 1305, EC 32 In Version 70, if a large number of frequencies is requested in a frequency response run, then the job may fail with one of the following error messages: = BIOMSG: ERROR 1305 HAS OCCURRED IN ROUTINE ENDPUT , FILE INDEX (FILEX) 0 STATUS = 0 0*** SYSTEM FATAL ERROR 4276.ENDPUT ERROR CODE 1305 PID= 0 BIOMSG: ERROR 32 HAS OCCURRED IN ROUTINE GETSTR , FILE INDEX (FILEX) 0 0*** SYSTEM FATAL ERROR 4276.GETSTR ERROR CODE 32 PID= 0

Avoidance: Reduce the number of frequencies. In modal frequency response, the new freq3, freq4 and freq5 entries may be used to guarantee that a solution is requested at each of the modes without specify a large number of frequencies. @ 28558; 70.5; Installation - rld: Fatal Error: cannot successfully map soname 'c_code.so' The "Minimum" installation of MSC/NASTRAN Version 70 on the SGI R8000/R10000 will fail to load a file required by the MSC/NASTRAN solver executable, as a result, MSC/NASTRAN will fail with the following error message: $ msc70 nastran myjob MSC/NASTRAN V70.0 (Silicon Graphics IRIX64 6.1) Tue Jul 29 08:52:34 1997 MSC/NASTRAN beginning job myjob 21692:/scratch1/dnl/msc/msc70/irix64/analysis: rld: Fatal Error: cannot successfully map soname 'c_code.so' under any of the filenames /scratch1/dnl/msc/msc70/irix64/c_code.so:/usr/lib64/c_code.so:/lib64/c_co e.so:/usr/lib/64bit/c_code.so:/lib/64bit/c_code.so: $ This file is loaded with the "standard", "full", and "custom" installations. Avoidances: There are two avoidances: 1) Use the "standard" or "full" installation. 2) If a minimum installation is required, or has already been performed, the following commands will unload the

missing file directly from the CD-ROM: $ cd _install_dir_ $ tar xvf /CDROM/msc70/irix64.tar msc70/irix64/c_code.so.gz x msc70/irix64/c_code.so.gz, 522805 bytes, 1022 blocks $ msc70/irix64/gunzip -v msc70/irix64/c_code.so.gz msc70/irix64/c_code.so.gz: 61.4% -- replaced with msc70/irix64/c_code.so $ @ 28565; 70.0.1; Design Optimization/Aeroelastic Analysis - FLFACT - UFM 7087 The message "USER FATAL MESSAGE 7087" is issued when a velocity requested in a flutter design task does not correspond to one used in the analysis. This could be a user input error, but it has been found that it can also be encountered when the user uses the thru option on the flfact entry. An equality check is being used to make the comparison and the numerical values that appear to be the same can differ slightly when they are generated from the thru data. Avoidance: Do not use the thru option on the flfact entry to specify analysis and/or design flutter velocities in design optimization (SOL 200). @ 28566; 70.0.1; Optimization/Design Sensitivity - DSDJ2D, DSDJ2S - SFM 3007 The message "SYSTEM FATAL MESSAGE 3007" can appear when performing adjoint sensitivity analysis if the adjoint solution vector is null. The adjoint solution vector will only be null if the resultant sensitivity is null. Thus, the message indicates a poorly posed design task in that a critical response is unaffected by the design variables. An example is asking that an acceleration response be greater than zero at zero hertz. Since this response can only be zero, the error is triggered. Avoidance: Remove the responses that trigger the message since they cannot affect the design. @ 28573; ; Normal Modes - REIGL with auto-omit - SFM 3007 When computing normal modes using Lanczos, if the number of generalized coordinates assigned is larger than the actual number of component modes available or auto-omit is performed, then the job will fail with the following messages: *** SYSTEM FATAL MESSAGE 3007 (REIGL ) ILLEGAL INPUT TO SUBROUTINE MERGE Avoidance: Remove the extra generalized coordinates or insert the following DMAP alter avoidance: compile moders alter 'EQUIVX MKAA\/KXX\/NOARED'(,-1) noared=-1 $ @ 28581; stems 70.5; Security - FLEXlm - Unable to serve licenses on SGI R8K, R10K Sy If the FLEXlm license server is installed on an SGI R8K, R10K system and

mscsetup installs the license file, the DAEMON line will incorrectly reference the irix64 architecture (which isn't used on FLEXlm). As a result, the license server will not function. Avoidances: Either of the following two avoidances can be used: 1) Modify the <install_dir>/flexlm/licenses/license.dat file and change the "irix64" reference on the "DAEMON" line to "irix" 2) Create a link in the <install-dir>/flexlm directory from irix64 to irix with the command ln -s irix irix64 @ 28582; PER3 69; DMAP - SUPER3 - OPG1, OQG1, etc do not Exist Upon Return from SU In SOL 101, the tables containing containing the results in "OFP" format are not returned from subDMAP SUPER3. This means if a client DMAP alter tries to use these tables (OPG1, OQG1, etc) in SESTATIC, they will not exist. There is no effect on the solution or output. Avoidance: Insert the following DMAP alter in runs using DMAP in SOL 101: COMPILE SEDRCVR, SOUIN=MSCSOU, NOLIST, NOREF ALTER 2,2 @ 28604; 70.5; Miscellaneous - PUNCH files Cannot Exceed 2gB on HP Punch files can not exceed the 2 giga-byte limit on HP and possibly other machines. The 2 giga-byte limit still applies to all FORTRAN I/O files. Avoidance: The only avoidance is to break-up the output requests so that the punch files remain under the 2 giga-byte limit. @ 28605; ; Matrix Method - HP-Examplar PA-RISC 1.1 - SFE 4276, SFM 1013 If a large job is run on a HP-Exemplar PA-RISC 1.1 system in Version 70, then the job may fail in DCMP with one of the following error messages: *** SYSTEM FATAL MESSAGE 1013 (GBKMGR) CANNOT FIND CORRECT CLUSTER POINTER RECORD BIOMSG: ERROR 4276 HAS OCCURRED IN ROUTINE GBKMGR , FILE INDEX (FILEX) = 0 *** SYSTEM FATAL ERROR 4276.GBKMGR ERROR CODE 4276 PID= 0

***** RECEIVED SIGNAL 2, ABORTING ... BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX (FILEX) = 0 STATUS = 0 Avoidance: Use the iterative solver instead of the sparse solver. The

iterative solver can be activated by adding the following nastran statement to the top of the input file. nastran iter=yes @ 28615; n Matrix 70; SSSALTER - alter1ga.v691 - Wrong Answers, Error in Transformatio

An error in the transformation matrix used in alter1ga.v691 may result in wrong answers for displacements during data recovery or in applied loads. Avoidance: Modify the following line in the alter: UMERGE USET,GOAQ,/GOGT/'G'/'O'/'A' $ to: UMERGE USET,GOAT,/GOGT/'G'/'O'/'A' $ or request an updated copy of the alter from MSC. @ 28648; ; Nonlinear Transient Heat Transfer - CHEXA - Wrong Answers Transient heat transfer - Convection to a negative temperature - wrong answers If a model has a center node ,for example 8 elements within a cubic. The center node temperature actually will increases the temperature despite the fact the whole solid should be cool off by convection to an ambient temperature odf -20 degree. Avoidance: use a single CHEXA element to represent the model. If I increase the timestep from 100 seconds to 1000 seconds,then my center node temperature will eventually comes down for my 8 hexa problem. However,I have no problem with a single CHEXA using a timestep of 0.1 seconds. When I double the mesh by changing it to 64 elements, then the center node temperature does go down. See the test deck called /net/kodiak/app/etl/s28648_64hexa.dat @ 28650; 70.7; Aeroelastic Analysis - USPARSE=0 - Wrong Answers If unsymmetric sparse is turned off using NASTRAN system cell 209 (USPARSE=0) and a complex left-handed fbs is required, wrong answers are produced and there is no warning message. This error affects all SOL 145 and 146 runs that use doublet lattice aerodynamics. Also any solution 200 run that uses doublet lattice aero in a flutter analysis. Avoidance: Do not select USPARSE=0. @ 28655; 70.5; Nonlinear Slide-Line Analysis - BGP - MN4W Poor Performance The performance suffers on the NT running slide-line contact problem using SOL 106 as memory is increased. Avoidance: Limiting the memory to 4 million words will improve performance.

@ 28662;

70.5; Frequency Response/Random Analysis - TABRND1, IFP - UFM 315 If the TABRND1 entry uses the log-log option with very small data the following fatal message may result: USER FATAL MESSAGE 315 (IFPDRV) FORMAT ERROR ON BULK DATA ENTRY TABRND1 Avoidance: This message has been encountered for data on the order of e-12. The only avoidance is to either use the linear option or to use different units to scale the data up.

@ 28669;

2001; Eigenvalue Analysis - PLOAD1, FORCE - Missing Output If element force output is requested in SOL 103 and one or more PLOAD1 entries exist in the bulk data, there will be no element force output provided. Avoidance: Remove the PLOAD1 entries from the input.

@ 28673; S

70.6; Eigenvalue Analysis - READ, EIGR - SFM 4276, EC 1401 Using GIVEN If GIVENS or MODIFIED GIVENS method is used for an eigenvalue solution and spill is required, it is possible that the run may terminate with: SYSTEM FATAL ERROR 4276.GETS68 ERROR CODE 1401

Avoidance: Either provide additional memory so that spill is not required or use the Lanczos method. @ 28675; 70.5; Documentation - Plotting, QRG - MAXIMUM DEFORM The documentation for the MAXIMUM DEFORM plot entry is deficient. Remark 2 is not true, and should be replaced with the remark below. Remarks 3, and 4 should be added: 2. If you wish the FIND command to utilize the d value, a MAXIMUM DEFORMATION command should precede the FIND command. 3. If you wish the plot deformation scaling to be different then the FIND scaling a different MAXIMUM DEFORMATION command can appear first before the PLOT command. 4. For nonlinear plotting MAXIMUM DEFORMATION d and the MAXIMUM DEFORMATION field on the PLOT command should have the same value. @ 28679; 70.5; Dynamic Analysis - Multiple FREQ, TSTEP, DLOAD Subcases - Ineffi cient Processing In SOLs 111 and 112, if multiple FREQ or TSTEP and/or DLOAD subcases are specified and default data recovery (PARAM, DDRMM, 0) is performed then SDR1 and SDR2 are re-executed every subcase even though their outputs do not change. Avoidance: If no superelements are present, then use SOLs 11 or 12. @ 28680; 70.5; Acoustic Analysis - PARAM,RESVEC, MERGE, MPYAD - SFM 3007, UFM 5

423 When a fluid/structure model with rigid body modes (requiring a SUPORT card) is run with PARAM,RESVEC,YES the following FATAL errors are issued: *** SYSTEM FATAL MESSAGE 3007 (REIGL ) ILLEGAL INPUT TO SUBROUTINE MERGE 0FATAL ERROR or *** USER FATAL MESSAGE 5423---ATTEMPT TO MULTIPLY OR ADD INCOMPATIBLE MATRICES, (SMPYAD) Avoidance: Apply the following DMAP avoidance: compile modefsrs alter 'call resvec'(,-1) call dbstore vafs,,,,//0/0/' '/0 $ compile resvec alter 'px1','' call dbfetch /vafs,,,,/0/0/0/0/0 $ partn pab,,vafs/px1,,,/1 $ alter 'px3,,,'(,-2) merge pxf,,,,,vafs/pxfs/1 $ equivx pxfs/pxf/-1 $ alter 'phux1'(3,-1) partn ux2,,vafs/ux2s,,,/1 $ equivx ux2s/ux2/-1 $ endalter @ 28684; utput 70.5; Dynamic Analysis - STRAIN and SPCF output in SOLs 11 - Missing o In SOLs 11 with default dynamic data recovery (PARAM, DDRMM,0) and multiple FREQ subcases, element strain and SPCforce (if PARAM,DYNSPCF=NEW) output will be missing in the 2nd and subsequent subcases. Avoidance: Insert the following alter: compile sol11 alter 1 file mus=append/kus=append/bus=append/istr1=append $ file mum=append/kum=append/bum=append/istr2=append $ @ 28685; 70.5; Superelements - CASE, Multiple Dynamic Subcases - SFM 3002 In SOLs 108, subcases, if then the run CASE module. disp=all subcase 10 super=10 subcase 100 109, 111 and 112 with multiple FREQ and TSTEP (or DLOAD) a superelement has fewer subcases than the residual structure will fail with the message SYSTEM FATAL MESSAGE 3002 in the For example,

freq=1 subcase 200 freq=2 Avoidance: Specify SUPER=ALL or the same number of subcases for the superelements and the residual structure or insert the following DMAP alter: compile super3 alter 'casedrm','' paraml casedr//'trailer'/1/s,n,norec $ if ( dloop>norec ) then $ CASE CASEDR,/CASEDRM/'tran'/norec $ else $ CASE CASEDR,/CASEDRM/APP/DLOOP $ endif $ @ 28686; 70.5; Acoustic Analysis - Modal Participation Factors - Missing Outpu t, Wrong Answers In solution 111, modal frequency response, modal particicipation factor output is wrong or missing when PARAM,ACSYM,YES, which is the default. Avoidance: Insert PARAM,ACSYM,NO or the following DMAP alter: compile mpf alter ',ah,' type parm,,char8,y,acsym='yes' $ if ( acsym='yes' ) then delete /ah,,,, $ PARTN bHH,VPHFS,/,mAH,, $ add5 mah,,,,/ah/-1. $ endif $ alter 'z2x' if ( acsym='yes' ) then $ add5 z2x,,,,/z2xx/-1. $ equivx z2xx/z2x/-1 $ endif $ @ 28687; 70.5; Eigenvalue Analysis - ASETi, SPOINT, EIGRL, EIGR, SUPORT - SFM 3 007, UFM 4405 In an eigenvalue solution, if null dofs (e.g., SPOINTS) are placed in the aset, and a SUPORT entry exists, then depending on the eigenvalue solver requested, the job will fail with one of the following error messages: *** SYSTEM FATAL MESSAGE 3007 (REIGL ) ILLEGAL INPUT TO SUBROUTINE MERGE FATAL ERROR or USER FATAL MESSAGE 4405,0 NO EIGENVECTORS COMPUTED FOR COMPONENT MODE SYNTHESIS OR SYSTEM SOLUTION. Avoidance:

1. Do not place the unreferenced SPOINTs in the ASET. 2. Remove the SUPORT entries if feasible. This avoidance only works for the Lanczos method. @ 28688; 70.5; Acoustic Analysis - EMG, CAABSF - SFM 4276, EC 7777 When a solution at zero frequency is requested and frequency dependent acoustic absorber elements, CAABSF, are specified, then EMG fails with a divide by zero. At zero frequency, the frequency dependent acoustic absorber element should put out zero impedance. Avoidance: Do not request a solution at zero frequency. @ 28697; 2001.0.1; Acoustic Analysis - ACSRC, FORCE, DPD6 - SFM 3007 If an ACSRCE Bulk Data entry has the same DAREA ID (field 3) for loads applied to structural grid points and the loading applies to more than one degree of freedom at a given structural grid point then the run will fail in the DPD module with SFM 3007. The run should fail with UFM 6825 (see CSR 4618). Avoidance: See CSR 4618. @ 28710; 70.5; Heat Transfer - MAT4, CONVM - UFM 6715, Wrong Answers If UFM 6715 shown below is issued in SOL 153 (or 106) then the run will not terminate but will continue and produce wrong answers. *** USER FATAL MESSAGE 6715 (ECNVMD) THERE IS NO DYNAMIC VISCOSITY SPECIFIED ON THE MAT4 ENTRY FOR CONVM ELEMENT 11. USER INFORMATION: FORCED CONVECTION ELEMENTS REQUIRE SPECIFICATION OF DYNAMIC VISCOSITY ON A MAT4 ENTRY IF ON THE PCONVM ENTRY THE REYNOLDS NUMBER CONVECTION EXPONENT IS GREATER THAN THE PRANDTL NUMBER CONVECTION EXPONENT.

Avoidance: Insert the following DMAP alter: compile nlstatic alter 'call super1' putsys(0,82) $ @ 28712; 70.5; Eigenvalue Analysis - REIGL, DECOMP - Insufficient Disk Space MSC/NASTRAN Version 70 introduces a new re-ordering method in sparse decomposition for CRAY and SGI computers. While this method has been found to have performance advantages in nearly all cases, there may be cases where the improved performance is at the expense of requiring more memory. If the memory is not available, which is more likely the case on very large jobs, additional spill may be required and the job may fail due to insufficient space. Avoidance: In the single problem case encountered to date the problem was resolved by increasing NQS memory. An error was found in the error handling when NQS memory was insufficient. Although we don't think it

should be necessary for the user to ever have to deselect the new ordering method, SYSTEM(206) controls the re-ordering method as follows: 0: 1: 2: 3: 4: @ 28713; method selected automatically in symbolic factoring phase minimum degree ordering modified minimum degree ordering for indefinite matrices no ordering (uses given sequence) Extreme ordering (Cray and SGI only; version 70)

70; Security - NT Only - UFM 3060 A problem with MSC/NASTRAN Version 69 NT system causes authorization codes that would normally produce expiration warning messages to issue USER FATAL MESSAGE 3060. Authorization expiration warning messages should be produced under the circumstances listed below. The only avoidance is to request a new auth code for a longer duration. 1. If the duration of an auth code is four months or longer, an expiration warning message is displayed when the code is within 3 months of expiring. 2. If the duration of an auth code is three months, an expiration warning message is displayed when the code is within 2 months of expiring. 3. If the duration of an auth code is two months, an expiration warning message is displayed when the code is within 1 month of expiring. If the duration of an auth code is one month, no expiration warning messages are displayed. The expiration threshold is calculated based on the end of the month of expiration, not on the actual end date. For Version 69, the following options can be used to circumvent this problem: 1. Generate auth codes with at least a four month term, and add 3 additional months to the code. For example, if you want a code to be valid from August, 1997 thru December, 1997, use the start date of 8/97, and an end date of 3/98. If this method is used, I suggest editing the license file and license report delivered to the customer to use the original expiration date. 2. Generate a series of one-month codes for each of the months in the contract period. The customer would add each of these codes to a single license file, to bypass the error.

@ 28716; on

70.5; SSSALTERS - rflagb - Wrong Answers for Fluids with Enforced Moti Wrong answers have been observed when the rflagb.v70 alter is used on fluid-structure problems, and enforced pressures are input via the Lagrange Multiplier Technique. The problem has been observed only on short word machines, and does not occur on long word computers such as Cray. The alter package has been modified to test for wrong answers, and

to switch to the avoidance procedure automatically when wrong answers are found. An investigation is underway to determine if this type of wrong answer could be achieved in any other manner, in particular when no dmap alters are present. The only cases found to date were for the circumstances given above. Avoidance: An avoidance is to cause the unsymmetric matrix solver to run instead of the symmetric solver. There is some economic penalty for this method, but it has been shown to work on all problems encountered to date. One method to cause the unsymmetric solver to be selected is to put in a k2pp matrix with one very small term on the diagonal, and with a FORM of 1 on the dmig header card. This says the matrix may be unsymmetric (even if it isn't as in this example), and forces the unsymmetric solver to be selected. @ 28730;

70.5; Miscellaneous - GPFORCE, BUFFSIZE - Possible Longer Runtime A significant efficiency improvement has been added to the GPFDR module in Version 70 which can cause GPFORCE output to run much faster than previous versions. This improvement is most significant for large models with multiple subcases and open sets (using the thru option). A 20 to 1 cpu improvement has been observed on a client model. However, for smaller models, the GPFDR time may take longer than the previous versions, especially when using a large buffsize is specified. Avoidance: Use a smaller buffsize and the run time should then be comparable to previous versions. As the model gets larger, the runtime will become faster as compared to previous versions.

@ 28757; 70.5; Frequency Response - SOLs 108 and 111, TPL, K2PP, M2PP, B2PP - W rong answers If the Case Control commands TF, K2PP, M2PP, or B2PP, change between subcases in SOLs 108 and 111, then the frequency response solution is repeated but the answers will not be correct. Avoidance: There is no avoidance. @ 28762; 70.0.5; Buckling - CBAR, MATi, ALPHA, TREF - Possible Wrong Answers In a buckling analysis, if the thermal coefficient of expansion on the MAT1 entry is non-zero, and no thermal loads are applied, then the differential stiffness for the BAR element will be incorrect. Avoidance: If there are no thermal loads then remove the thermal coefficient of expansion since this quantity is not needed for a mechanical load. @ 28765; ; Documentation - Aeroelastic User's Guide Equation 8-43 in Section 8.10 of the Aeroelastic User's Guide indicates that the Imaginary part of Q31 is used to calculate the yawing moment coefficient due to sideslip rate. This is a typographical error. It should be using the Real part of Q31. See the Version 65 Aeroelastic Handbook equation 18 on page 6.2-190. @ 28778; 70.7; SSSALTERs - rflagba - Segmentation Fault

A frequency response run on an SGI R8000 using the sssalter rflagba.v691 may fail in FRRD1 with "signal 11 : segmentation fault" in the log file. Avoidance: The problem is under investigation at this time. Try to use the large mass approach in place of the alter. @ 28782; 70; User Modifiable NASTRAN - XGPIDG - SFM 52 When adding new user modules to the MSC/N user modifiable in version 69.1, the job may fail with the following error messages: ** SYSTEM FATAL MESSAGE 52 (XGPIDG )

NAMED COMMON /XLINK/ IS TOO SMALL This error is most likely to occur when adding more than one new module. Avoidance: Contact your local support for the corrected source to semdbdf.f. @ 28798; 2001; Nonlinear Analysis - SPCD - Wrong Answers Applying an enforced displacement with SPC or SPCD while using the CONV field set to UPW causes wrong answers. @ 28813; 6011

; DMAP - PARAM - CHAR8 Variables on PARAM Entries Must Be BCD, UFM The Quick Reference Guide states that character parameters are allowed as inputs on the PARAM entry. However, the full range of character values are not allowed. The Old Exec construct of "BCD" parameters is enforced by the preface. This construct states that only alphanmumeric values are used, that is, the first character must be a letter of the alphabet. An input of param, char, -1, where char is typed as a char8 variable in dmap, results in the fatal diagnostic: USER FATAL MESSAGE 6011 (IPVDFT) THE MAIN SUBDMAP VPS TYPE OF PARAMETER CHAR DIFFERS FROM THE TYPE DETERMINED FROM PARAM CARD DEFINITION. USER INFORMATION: VPS TYPE IS CHARACTER. PARAM TYPE IS INTEGER.

@ 28817; 70.5; Documentation - PARAM,BAILOUT - Incorrect Default Value for BAIL OUT in QRG The V70 Quick Reference Guide states the default for BAILOUT is -1. In Version 70 of MSC/NASTRAN, the default is 0. @ 28829;

2001; Superelements - SEP2D, SUPER - UFM (SEP2D), Poor Message When the SUPER=i,j feature is used to input different load ids in subcases for different superelements, messages of the form *** USER FATAL ERROR ---- (SEP2D)

CASE CONTROL RECORDS FOR SUPERELEMENT ZERO AND NON-ZERO LOAD SEQUENCE NUMBERS.

2 CONTAIN A MIXTURE OF

may occur. The error message does not cause the program to exit, however. In all cases studied to date it continues to completion, and produces correct answers. Avoidance: The error message may be ignored. Using a fully expanded case control instead of using SUPER=ALL above the subcase level and SUPER=i,j within the subcase will also avoid the problem. @ 28840; UFM 476 ; File Management Section - ASSIGN - DELETE Option does not work,

If an ASSIGN statement is used to define the database and the ASSIGN statement specifies the DELETE option, a pre-existing database may not be deleted and the run may fail with the message: USER FATAL MESSAGE 476 (GPFILX) THE LOGICAL NAME DBALL_1 ASSIGNED TO A PHYSICAL FILE NAME OF THE PRIMARY DATA BASE, DOES NOT EXIST. USER INFORMATION: PHYSICAL FILE NAME = Avoidance: Delete unwanted databases before executing the run @ 28846; 70.5; Matrix Methods - FBS - Wrong Answers for Complex FBS with 2 by 2 Pivots FBS of complex symmetric, non-positive semi-definite matrices may produce wrong answers. Standard structural elements do not produce non-positive semi-definite matrices so this error is unlikely to occur in most runs but it is possible in poorly conditioned models and users dmaps that solve Ax=b, where A is complex symmetric and not positive semi-definite. The error has been observed on runs using the rflagb.v70 alter, which uses Lagrange Multiplier Method techniques which cause indefinite systems of equations. Avoidance: Use the unsymmetric solution method. In the current version of the rflagb.v70 alter, the model is run first with the symmetric solver, usually the least expensive option. A check of the solution is made immediately after the frrd1 module, and if the accuracy is poor, a second solution is generated with the unsymmetric solver. If it is suspected that this error is being encountered in a standard solution, an avoidance to add a K2pp matrix with a form of 1 (square, but not symmetric). This data is input in the IFO (field 4) of the dmig entry used for K2pp effects. A very small number may be placed on a diagonal term, so that this change does not modify the model significantly. This will cause the unsymmetric option to be selected. @ 28848; lt 70.5; Inertia Relief - PARAM,INREL, MERGE - UFM 6143, Segmentation Fau On Digital UNIX Version 69.1 only, if inertia relief is requested in some subcases, but not all, then the run will fail with a "segmentation fault" in the MATGEN module. On all other machines in Version 69.1 and 70, if inertia relief is requested in some subcases, but not all, in the presence of enforced displacements then the run may fail with UFM 6143 in the MERGE module.

Avoidance: Insert the following DMAP alter: compile selr alter 'IF ( INREL0','' IF ( INREL=-1 and not(delta) ) THEN $ @ 28853; d 70.5; Installation - mscsetup - Not all the files could be decompresse On Version 70 HP systems, mscsetup may incorrectly report the total or available disk space on a file system during the query or installation phases. If the error occurs during the installation phase, the installation may terminate with the following error message: Not all the files could be decompressed. <req> MB free space are required. <avail_insdir> MB were available in the installation directory <insdir>. <avail_tmpdir> MB were available in the work directory <tmpdir>. Please make more space available and redo this installation using the command ./mscsetup -t <work_directory> <insdir>/mscsetup.pbk where <work_directory> is an alternate directory with at least <req> MB free space that can be used for scratch purposes during the installation process. Avoidance: The mscsetup script can be corrected by copying the script to a temporary directory and changing line 3269 from $awk 'NR==1{avail=$3} NR==2{printf("%d %d\n",avail/MULT,$1/MULT)}' MULT=$gs_mult -` to $awk 'NR==1{avail=$(NF-3)} NR==2{printf("%d %d\n",avail/MULT,$1/MULT)}' MULT=$gs_mult -` This will select the fourth-to-last item in the line instead of the third item in the line. In a "well-formed" line, the third item *is* the fourth-to-last item, but if the filesystem's actual device name is short, it won't be. You can redo the failed installation with the command mscsetup <insdir>/mscsetup.pbk where <insdir> is the installation directory specified in the failed attempt.

@ 28857; 1

GL; Optimization/Design Sensitivity - Multi Master - SFM 4276 EC 300 The multi-master technique, allows the user to develop a design task in pieces. That is, an automotive design task might have a chassis in one model, the body in another and doors and windows in other models. As part of the adjoint sensitivity development, an efficiency was introduced, which skips the generation of pseudo-loads if there are no active responses for a particular subcase. Unfortunately, this breaks the random order capability of the multi-master technique which would allow the user to have pseudo-load information available on a standby basis in case a later job submittal requires it. If a later run needs the pseudo-load information from a previous run, which did not calculate it, then the message "SYSTEM FATAL ERROR 4276.GREAD ERROR CODE 3001 PID= 0" is issued. Avoidance: The avoidance is to calculate the responses prior to (or in the same run as) the pseudo-load calculation.

@ 28891;

70.5; Documentation - MATHP - Equations in QRG Incorrect Remark 1 under the Version 70 Quick Reference Guide description of the MATHP entry shows the equation 2D1=K2D1=K, it should read 2D1=K.

@ 28892;

70.5; Optimization/Design Sensitivity - AERO - Wrong Answers If, in design optimization, a static aeroelastic subcase is present and their are no static aeroelastic specific responses (TRIM or STABDER on a DRESP1), the mass effects in the sensitivity analysis are ignored and wrong answers can result. Avoidance: Include a token static aeroelastic response. It could be constrained with very generous limits so that it is never needed for the design task, but at the same time the DSCREEN parameters should be set to force retention of the unimportant constraint.

@ 28893;

70.5; Documentation - MODTRL - Various UFM, Missing Documentation In Version 70, the capability to modify the type and/or precision of a matrix with the MODTRL module was removed. DMAPs that use MODTRL in this way will fail with a message of the form: USER FATAL MESSAGE (MODTRL) AN ATTEMPT TO MODIFY MATRIX 101 TYPE OF TERMS HAS BEEN MADE. USER INFORMATION: USE MATMOD OPTION 34 TO CREATE TWO OUTPUT REAL MATRICES FROM A COMPLEX INPUT MATRIX. USE ADD OR ADD5 TO CREATE A COMPLEX OUTPUT MATRIX FROM ONE OR TWO REAL INPUT MATRICES. Unfortunately there is no documentation in the Version 70 Release Guide and the diagnostic messages from the MODTRL module above may be misleading. Avoidance:

1. In order to change the precision of a matrix, use ADD5. If the new precision does not match the machine precision specify PUTSYS(new precision,55) before ADD5. For example, on a double-word machine: a. single to double ADD5 SINGLE,,,,/DOUBLE $ b. double to single PUTSYS(1,55) $ ADD5 DOUBLE,,,,/SINGLE $ PUTSYS(2,55) $ or MPYAD DOUBLE,,DOUBLE/SINGLE///1 $

2. In order to change the type (complex or real) of a matrix, use ADD to convert real to complex and MATMOD(34) for complex to real. For example, a. real to complex ADD REAL,/CMPLX//(0.,1.) $

b. complex to real MATMOD @ 28894; swers CMPLX,,,,,/REAL,/34 $

70.5; Eigenvalue Analysis - Residual Vectors (PARAM,RESVEC) - Wrong An If there is a c-set and/or r-set and component modes are augmented with residual vectors (PARAM,RESVEC,YES or PARAM,INRLM,-1) then the answers will be wrong. Avoidance: Insert the following DMAP alter: compile semr3 alter 'call resvec','' call resvec pg,uset,gm,kfs,ys,bgpdts,scstm,cstms,medge,mff, cmpho0,loo,koo,cmlama0,moo,got,koo,cmphrc/ cmpho,cmlama/ nozset/norc/seid/'o ' $ compile resvec alter 'if ( nobset','' if ( nobset>-1 ) upartn uset,got/,,gorc,/'t'/'b'/'c'/2 $

@ 28917; 70.5; Nonlinear Analysis - MATHP - Poor Answers Using MATHP with Isotr opic Materials Hyperelastic elements have been designed to allow incompressible solutions. The elements yield poor results when used with isotropic material properties.

1. In the Version 70 QRG entry for MATHP, Remark 1. Wrong: "2D1 = K2D1 = K" Correct: "2*D1 = K" 2. The "MEAN PRESSURE" is wrong for the hyperelastic element. @ 28919; 70; Matrix Methods - DCMP, Spill- SFM 4276.SFPCKI EC 62 The message, "SYSTEM FATAL ERROR 4276.SFPCKI ERROR CODE 62 PID= 0", is supposed to be issued when the matrix to be decomposed exceeds the maximum allowable size. However, prior to MSC/NASTRAN Version 70, it may be issued erroneously, when the DCMP module starts to spill. This is because the spill logic may overwrite the variable that contains the matrix size. Avoidance: If possible, increase memory to avoid spill. Otherwise, change the available memory, so the spill logic will work differently, and perhaps not overwrite the variable with a number larger than allowed. @ 28936; 70.5; Elements - CQUAD4, OLDQ4K - Unexpected Answers The NASTRAN system cell OLDQ4K is ignored in Version 69.1 and 70 systems. Avoidance: The avoidance is to set SYSTEM(173)=1. @ 28939;

70.5; Composites - PCOMP, PSHELL - SFM 3002 MODULE = SDRCOMP If a model requests STRESS output for some plate elements referencing PCOMPs and other plate elements referencing an equivalent PSHELL/MAT2 combination generated in a prior MSC/NASTRAN run, then no composite ply stress output or homogeneous stress output occurs and the following error message is issued: 0*** SYSTEM WARNING MESSAGE 3002 MODULE = SDRCOMP 0EOF ENCOUNTERED WHILE READING DATA BLOCK EPTS (FILE 103) IN SUBROUTINE SDRCMA Avoidance: MID1s, MID2s, MID3s, MID4s that are created on the equivalent PSHELL/MAT2s are equal to 100000000 plus the PCOMP PID. This is a flag to NASTRAN that they are from PCOMPs. If the MIDis in the punched output are change to smaller numbers, then composite ply stresses for the PCOMP elements and homogeneous stresses for the PSHELL elements may be obtained.

@ 28959;

70.5; Superelements - EXTRN - UFM 6894 The Quick Reference Guide documentation for the EXTRN bulk data entry states that a thru option is available for listing connected points. When this option is used for scalar points the following fatal error messages are output: FOR EXTERNAL SUPERELEMENT ID= ILLEGAL COMPONENT DESIGNATED ON THE EXTRN BULK DATA ENTRY. 3, SCALAR POINT ID= 712 HAS AN

Avoidance: Remove the thru range and list each point explicitly. @ 28965; ; Nonlinear Analysis/Hyperelastic Elements - Lumped Mass, MATHP Wrong Answers The 20-noded hyperelastic hexa element produces wrong answers when used

with lumped mass. The mean pressure output is wrong for all hyperelastic elements. Avoidance: Use consistent mass, ignore the mean pressure output. @ 28972; 70.5; Optimization/Design Sensitivity - AERO, FLFACT, DSAD - SFM 3007 If the velocity specification for a flutter design response is a positive number on the associated FLFACT entry while the FLFACT velocity specification for the analysis is negative (to indicate that eigenvectors are required at that velocity), the above fatal message will occur. Avoidance: The simple avoidance is to not specify negative velocities on either the analysis or the design FLFACT entries. A slightly more complex avoidance is to not design at velocities at which flutter eigenvectors are desired. @ 28987; 70.5; Submittal - V68.X and earlier used nice(1) with bat=yes, V69 and later doesn't. The nastran command, introduced with Version 69, removed the nice(1) command from the "batch=yes" submittal. This can be returned by setting the "-bg" queue command to "nice %job%" in any RC file. For example, the line submit=-bg=nice %job% can be placed in any RC file. @ 29002; 70.5; Frequency Response, Flutter, Complex Eigen - DLOAD, CMETH - UWM 6498, Signal 11 In Version 70, if the DLOAD Case Control command references a non-existing DLOAD or RLOADi entry, or a CMETHOD Case COntrol command references a non-existing EIGC entry, othen depending on the machine, the job may fail with one of the following unclear messages: *** USER WARNING MESSAGE 6498 (MsgGetStr) Message number is not found PROGRAMMER INFORMATION: Check call to MSGPOP from FRLG with the message number 3031 Severity is USER WARNING Parameter 1 is 2 Parameter 2 is DLT Parameter 3 is FRRD USER ACTION: Please contact MSC Client support Signal 11 (in the log file) Avoidance: Correct the offending DLOAD/RLOADi entries or CMETHOD/EIGC entries. @ 29005; 70.5; Miscellaneous - nastxxrc, symbol - Overwrite Ignored Once a symbol is defined in an rc file, it cannot be overwritten by a later RC file that you would expect to take precedence. This error prevents a site from providing a default value in a system RC file and the

user from overriding that value in their local RC file. Similarly, a user cannot provide a default in the HOME RC file and then override it in the local RC file. Avoidance: There are two possible avoidances: 1. Use a different symbol or 2. Override the RC file value with an environment variable as follows (assuming the symbol name is MYDIR): export MYDIR=/my/big/project MYDIR=/my/big/project; export MYDIR setenv MYDIR /my/big/project @ 29011; ut # Korn shell # Bourne shell # C-shell

70.5; Optimization/Design Sensitivity - ANALYSIS=MODES - Missing Outp A solution 200 shape optimization analysis that contains multiple subcases and ANALYSIS=MODES will not output the eigenvectors for the final design cycle. Avoidance: Apply the following DMAP avoidance: compile presens alter 'deseig=false',''

@ 29013;

70.5; Documentation - DRESP2 - Questions about Optimization Input Remark 2 under the DRESP2 entry should be changed to read: DRESP1 entries can not span analysis types or superelements.

@ 29016;

70.5; Documentation - QRG, EXTRN - Blank Field Feature not Documented. The EXTRN bulk data entry, first documented in the Version 70 Quick Reference Guide, allows pairs of blank fields. The pair is read as if it weren't there. This makes it easier to repair or modify EXTRN entries. For example, suppose that an EXTRN entry were of the form: $EXTRN GID $ GID EXTRN, 12 etc. CID GID CID (ETC) 123456 14 CID GID CID GID CID 123456, $

123456 22

123456 24

and the model is changed so that GID 13 should be added to this list. Since the sequence of points on the list is of critical importance, it is necessary to move the input entry for every point after GID 13 to meet the sequencing requirements. This change can be accomplished as shown: $EXTRN GID CID GID CID $ GID CID (ETC) EXTRN, 12 123456 , 13, 123456, 14 123456 22 GID CID GID CID

123456 24

123456, $ etc.

@ 29043; 70; Installation - Folded license.dat Files may cause Auth Code Erro r, UFM 3060 Various mail tools may fold text that exceeds a pre-determined line length. Generally, the tool will break the line between words, i.e., at whitespace. In normal text this is, at worst, annoying. In a license.dat file, this can result in the failure of lmgrd. Avoidance: Use the mscsetup script in TAN #2778 to install a license.dat file. If the file has already been installed, but has been folded, copy the file to another location and reinstall it using the revised mscsetup. @ 29044; AS 2001; Documentation - PELAS, PELAST - Requires Initial Stiffnes on PEL When a CELAS1 element is used with a PELAST property entry to represent a nonlinear spring in nonlinear analysis, an initial, non-zero stiffness must be specified on the PELAS entry. Other capabilities in MSC/NASTRAN require the flexibility to allow the stiffness field on the PELAS entry to be blank so the code will not be changed to trap this condition. A remark will be added to the PELAST documentation stating the requirement. @ 29047; 70.5; Optimization/Design Sensitivity - AERO - SWM 6911 (QCBFIN) The warning message: *** SYSTEM WARNING MESSAGE 6911 (QCBFIN) MORE THAN ONE KHH appears when a solution 200 run is restarted from a solution 145 run. The message is, most likely, benign, but does have the potential for producing wrong answers. Avoidance: Apply the following DMAP avoidance: compile respsen alter 'khhx','qhhx' type parm,,i,n,ncycl dbview khhx=khh dbview mhhx=mhh dbview bhhx=bhh dbview qhhx=qhhl alter 'dscmg'(2) ncycl = desiter $

$ where where where where

(wildcard (wildcard (wildcard (wildcard

and and and and

desiter=ncycl) desiter=ncycl) desiter=ncycl) desiter=ncycl)

$ $ $ $

@ 29050;

70; SSSALTERs - resflexa, RESTART - SFM 3003 It is possible that the resflexa.vxxx alter will not automatically remove static vectors which are a linear combination of the elastic modes. If this error occurs, the program may encounter SYSTEM FATAL MESSAGE 3003, indicating an attempt to read past the end of a record. Avoidance: The alter has been updated to correct the error. Either request an updated version of the alter or include the following in the alter: mpyad phiajc,pfac,/stamod/0 $ <== existing line

add stamod,uarsflx/resvec//-1. $ <== existing line $ filter null columns if they exist - Sept, 1997 matmod resvec,,,,,/nullcol,/12/s,n,nonull/1 $ if(nonull>0)then $ message //'at least one residual vector can be represented' $ message //'by the elastic modes. the following matrix has 1.0' $ message //'in the column associated with those vectors.' $ message //'these columns will be removed automatically' $ message //' ' $ matprn nullcol// $ partn resvec,nullcol,/restemp1,,,/1 $ equivx restemp1/resvec/always $ nloads=nloads-nonull $ endif $ @ 29052; 70.5; P-Elements - ADAPT, PSHELL - Negative Eigenvalues, Wrong Answers , UFM 9050 In eigenvalue analysis, negative eigenvalues may occur in shell p-elements if bubble functions are present and the distribution of p-values is non uniform. In static analysis wrong answers or USER FATAL MESSAGE 9050 may occurBubble functions may be present whenever the p-level is less than 3. Avoidance: To avoid this error, bubble functions must be turned off by setting system cell 225 to 1. @ 29053; 70.5; Rigid Elements - RSSCON - Wrong Answers on SGI The RSSCON (rigid shell to solid connector element) has been found to produce wrong answers for one problem on SGI computers. Avoidance: The problem is not yet fully understood. In the case where the answers were wrong the oload and spc resultants did not agree. Thus, the user is advised to closely check these results. If they do not agree, then if possible run the job on another computer or use another method such as rbe2s to connect the shell elements to the solids. @ 29064; 70.5; Nonlinear - Modes, Preload - Wrong Stresses for Prestressed Eige nvectors If a Sol 106 run calculates a preloaded condition and then solves for the modes, the stress output for the modal portion of the analysis will be wrong whenever the stiffness matrix has been updated because the stresses are calculated using the initial geometry. The nonlinear stresses are still correct. Avoidance: A DMAP avoidance for Version 70 is: $ $ s29064.v70 11-13-97 $ $ ALTER FOR SOL 106 (VER 70) FOR NMLOOP DATA RECOVERY $ AVOIDS LARGE SOURIOUS STRESSES (CSR 29064) ECHOOFF COMPILE NLSTATIC $ ALTER 'APPEND *UGNI' $ IF ( NLOOP = NMLOOP ) CALL DBSTORE UGNI,,,,//0/0/' $ COMPILE SEDRCVR $ ALTER 'OPG1,OQG1,OUGV1','' $

'/0 $

TYPE PARM, ,I,Y,NMLOOP=0 $ TYPE PARM,NDDL,I,Y,LGDISP $ IF ( APP1 = 'REIG' AND NMLOOP > 0 AND LGDISP > 0 ) THEN $ CALL DBFETCH /UGNI,,,,/0/0/0/0/0 $ VECPLOT UGNI,BGPDTS,,CSTMS,CASEDR,,,/UGVBAS/0/0/3 $ MATMOD UGVBAS,BGPDTS,,,,/BGPDTU,/11 $ ELSE $ EQUIVX BGPDTS/BGPDTU/-1 $ ENDIF $ SDR2 CASEDR,CSTMS,MPTS,DIT,EQEXINS,,ETT,OL2,BGPDTU,PJ1,QG,UG,EST, XYCDBDR,OINT,PELSETS,VIEWTB,GPSNTS,DEQATN,DEQIND,DITID/ OPG1,OQG1,OUGV1,OES1,OEF1,PUGV/APP1/S,N,NOSORT2/NOCOMPS///// ACOUT/PREFDB/-1./-1./ADPTINDX/ADPTEXIT/FALSE $ ENDALTER $ @ 29068; FRRD2

70; Matrix Methods - UDSFA, FRRD2 - UFM 6136, Insufficient Memory in It has been observed that for some frequency response problems that an incorrect estimate is made for the amount of core required for the incore decomposition. This error results in the message: *** USER FATAL MESSAGE 6136 (UDSFA) insufficient memory in FRRD2 regardless of how much memory the job is given. Avoidance: The avoidance is to set system(166)=1 so that if this error is encountered the job will revert to standard decomp and avoid this error.

@ 29077;

; Documentation - CORD3G - Incomplete Feature The Quick Reference Guide for Version 70 contains information on the CORD3G Bulk Data entry. This feature is still being developed and is not ready for use.

@ 29082;

70.5; Installation - mscsetup - noclobber Causing mscsetup to Crash If the noclobber flag is set, mscsetup will fail when it starts. Avoidance: Insert the line set +C as the first executable line of mscsetup and rerun.

@ 29087; 70.5; Aeroelastic Analysis - MKAERO - Lack of precision in Dynamic Aer oelasticity The aerodynamic matrices required in SOL 146 are interpolated from aerodynamics calculated at reduced frequencies input on the MKAEROi bulk data entries. The calculations are performed in single precision and it has been observed that this can produce bad answers when a poor choice of MKAERO data is entered. The problem has not shown up on long word

machines. Avoidance: Restrict the number of k's input on the MKAERO to no more than eight. Avoid k's that are close to one another (certainly no closer than 0.005). Check answers on a long word machine if possible. @ 29096; s 70.5; Superelements - MPCFORCE - Answers do not Include Inertial Force If MPCFORCEs are requested for MPCs or R-type elements (RBAR, RBE2, etc) in a transient response or frequency response solution, the resulting output will not include the inertial forces due to attached masses. Avoidance: Insert the following alter into the input file: compile cforce, souin=mscsou, list, noref alter 'IF ( APP=','' IF ( APP='reig' or seid>0)then $ $ @ 29105; 70.5; Optimization/Design Sensitivity - AERO - Wrong Answers for rtype =trim or stabder If a static aeroelastic subcase retains a STABDER or TRIM response types, but no element or grid type responses, the required sensitivity analyses will not be performed and wrong answers will be produced for the sensitivities. Avoidance: Include element and/or grid response in the subcase and force its retention through appropriate DSCREEN values. @ 29126; ll running ; Installation - Uninstall Problems on Windows NT, MSCEXE file sti

If MSC/NASTRAN or MSC/PATRAN on Windows NT are unistalled and the system may not reboot and may issue the message "MSCEXE file still running". Avoidance: When performing an uninstall of MSC/NASTRAN or MSC/PATRAN, the user must kill the FLEXlm service daemon. This may be done through the following steps; type ctrl+alt+del, select task manager; find the mscexe process, and then kill it. The system may then be rebooted. @ 29134;

; External Superelements - SEBULK, EXTRN - SFM 3007 (SEP1X) If the external superelement capability is used with the enhanced superelement method (begin super=seid), and the EXTRN entry is missing in the assembly run, then the job will fail with the following unclear messages: *** SYSTEM FATAL MESSAGE 3007 (SEP1X ) ILLEGAL INPUT TO SUBROUTINE SEPEXT FATAL ERROR Avoidance: Provide the proper EXTRN entry.

@ 29139;

70.5; Documentation - QRG, CONVM - Incorrect Documentation

The description of the CONVM bulk data entry in Quick Reference Guide incorrectly states that the EID may be that of a CHBDYE or CHBDYP element. This entry may only reference the CHBDYP entry and only when the FTUBE option is specified. @ 29140; 70.5; Optimization/Design Sensitivity - SIGSEGV Segmentation Violation - Problem Size The message "SIGSEGV - Segmentation violation" can appear in the log file of short word machines for large optimization design tasks. The cause is an integer overflow calculation. A general limit on problem size is that the 2*ndv + ncon SQRT< 2**31-1 = 46,341 (where ndv is the number of design variables and ncon is the number of constraints, including property limits but not side constraints. Avoidance: Reduce the problem size to fit within the above limit or run on a long word machine where the comparable limit is 3 billion and therefore not meaningful. @ 29144; 70.5; Optimization/Design Sensitivity - DSAD, PCOMP, CTRIA3 - UFM 5626 (DSAD13) The above message can occur in error when a DRESP1 entry asks for a CFAILURE response and the deck contains only CTRIA3's or CQUAD8's and no CQUAD4's. Avoidance: Add a dummy, disjoint QUAD4 that references the same PCOMP entry and constrain its response. @ 29150; 70.5; Cyclic Symmetry, Buckling - SOL 16 (or 77), EIGRL - UFM 4418 If a buckling analysis is performed using a cyclic symmetry solution sequence (SOL 16), and the Lanczos eigenvalue solver is requested, then the job will fail with the following error messages: 0*** USER 0 FATAL MESSAGE 4418, NO EIGENVALUE EXTRACTION DATA IS AVAILABLE.

Avoidance: Insert the following DMAP alter: compile sol16 alter 'matmod','' MATMOD CASEds,DYNAMICS,,,,/,/23/S,N,NP/S,N,LANCZOS $ endalter @ 29165; 70.5.1; P-Elements - Shell Elements - Wrong Answers The bending energy is calculated incorrectly for the error estimator for p-shell elements. Thus, p-shell element models subjected to bending may produce poor or wrong answers. Avoidance: There is no avoidance using p-elements. @ 29178; formed 70.5; Superelements - ASETi - ASETi Entries Ignored, Reduction Not Per

The aseti entries are ignored if they are placed in the main bulk data section using the enhanced superelement capabilities (begin super=seid). Avoidance: There are two possible avoidances: 1. Perform a component mode reduction instead of a Guyan reduction on the residual structure. or 2. Place only aset grids in the residual structure and put the rest of the grid points in a separate superelement. @ 29199; 70.5; Cyclic Symmetry - LOADCYT, CYCLIC1 - SFE 4276, EC 7777 Cyclic symmetry problems on the Cray T90/IEEE Version 70 release of MSC/NASTRAN that contain LOADCYT entries may fail with the following traceback in the CYCLIC1 Module. Beginning of Traceback: Called from line 42 (address 15262370c) in routine 'TRBK'. Called from line 39 (address 202316c) in routine 'RECOVR'. Called from line 112 (address 1101772a) in routine 'sysrcv'. Called from address 14765131c in routine '__handlr'. Called from address 1717073a in routine 'CY1CYT'. Called from line 119 (address 1714533c) in routine 'CY1CYD'. Called from line 36 (address 1711300c) in routine 'CY1CCL'. Called from line 267 (address 1747550d) in routine 'CYCL1'. Called from line 116 (address 10753001b) in routine 'XSEM12'. Called from line 342 (address 2117a) in routine 'MAINAL'. Called from line 306 (address 4110d) in routine '$START$'. Avoidance: Set the "METHOD" field to 1 on the LOADCYT to avoid the error. @ 29200;

; Superelements - PARAM,POST,-1 (-2) - Erroneous Output In a superelement analysis, if param,post,-1 or param,post,-2 is used, and is placed inside a subcase, then some of the output for the output2 file will be sent to the log file. A symptom of this error is the erroneous output printed in the log file similar to the following: Process Id = 20562 X 0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ$/+-()'=.*"#%&,;:<>?@![]^_\ 020080 61^GGPL F^4^66^810/81BK33333^71^60^3^61^GGPL F^72^60^3^81BK^813.813*813"813#813 Avoidance: There are two possible avoidances: 1. Place the param,post,-1 or param,post,-2 in the bulk data section or above the first subcase. or 2. Add the following dmap alter and place "param,ounit1,12" inside the same subcase as "param,post,-x".

compile out2geom $ alter 'type.*parm.*ounit1','' $ type parm,,i,y,otape2,ounit1 $

@ 29202; 70.5; Optimization/Design Sensitivity - ADJOINT - Wrong Answers for Mu ltiple B.C.s The release guide for V70 states that the adjoint sensitivity method is not available when there are multiple static boundary conditions. In fact, the adjoint method may still be selected by the code but will yield wrong answers. Avoidance: The following alter avoids the problem: compile descon alter 'IF ( DOSTATX'(,-1) mbcflg=false $ alter 'bcdr '(,-1) mbcflg=true $ @ 29203; 70.5; Optimization/Design Sensitivity - SUPORT1, FBS - SFM 2198 The message "SYSTEM FATAL MESSAGE 2198, INPUT DATA BLOCK 101 HAS BEEN PURGED" may be produced when a design sensitivity task contains static and normal modes subcases and the normal modes subcase contains a SUPORT1 entry that the static subcase does not. Avoidance: Use the following DMAP alter compile desopt alter 'fbs '(,-1) CALL SETQ CASESx// SEID/PEID/S,MTEMP/S,K2GG/S,M2GG/S,B2GG/S,MPC/ S,SPC/S,LOAD/S,DEFORM/S,TEMPLD/S,P2G/S,DYRD/S,METH/ S,MFLUID $ PARAML CASESx//'DTI'/1/255//S,N,SUPORT $ @ 29210; 70.5; Aeroelastic Analysis - AEROS - Incorrect NDDL Description for AE RO Datablock The first record of the AERO datablock was changed for Version 70, but the NDDL description was not updated. Three additional words were added to the datablock to support running static aero and flutter in the same run while using different reference chord lengths. The words are the reference chord length from the third field of the AEROS entry, plus the aerodynamic and reference coordinate system ID's from field 7 and 8. Avoidance: The correct NDDL description for Version 70 is: $ $ $ $ Desc={Aerodynamic Matrix Generation Data } Desc={Contains information for control of } Desc={aerodynamic matrix generation and flutter Desc={analysis } Desc={ }

RECORD=HEADER, NAME(2),CHAR4,{Data block name } EOR, RECORD, ND,I,{Y sym flag SYMXZ } NE,I,{Z sym flag SYMXY } REFCS,RS,{reference chord from AEROS } REFB,RS,{reference span or 0 } REFS,RS,{reference wing area or 0 } REFC,RS,{reference chord from AERO } ACSID,I,{ } RCSID,I,{ } EOR, RECORD, M,RS,{value of m } K,RS,{value of K } EOR, RECORD=TRAILER, WORD1,i,{ = 1 } WORD2,i,{ = 0 } WORD3,i,{ = 0 } WORD4,i,{ = 0 } WORD5,i,{ = 0 } WORD6,i,{ = 0 } EOR, $ Note={ Record 2 is terminated with an integer '-1' } EOF @ 29213; 70.5; Documentation - STRAIN - Clarify Remark in QRG Remark 1 of the Quick Reference Guide description of the STRAIN case control command claims that 'Strains for nonlinear elements are requested by the STRESS command and appear in the nonlinear stress output'. This remark should state that 'Nonlinear strains for nonlinear elements are requested by the STRESS command and appear in the nonlinear stress output'. @ 29310; 70.5; Optimization/Design Sensitivity - DRESP2, DVPREL2 - UFM 316 The message "USER FATAL MESSAGE 316 ILLEGAL DATA ON BULK DATA ENTRY DVPREL2" will be produced in error if the input deck contains DRESP2 entries and the highest numbered DRESP2 entry references DVPREL1 data. Avoidance: Remove the DVPREL1 references from the highest numbered DRESP2 entry. In almost all cases, this is a simple replacement to a DESVAR reference. @ 29311; ; Iterative Solver - SITDR2, SOLVIT, ITER - UFM 6724 (SB1RM) The iterative solver may fail with the message: USER INFORMATION MESSAGE 6447 (SITDR2) ITERATIVE SOLVER DIAGNOSTIC OUTPUT EPS : 1.00000E-06 BICWELL PRECONDITIONING USER FATAL MESSAGE 6724 (SB1RM) TOO MANY RESTARTS.

USER ACTION: INCREASE IPAD VALUE ON BULK DATA CARD ITER. 0FATAL ERROR Avoidance: This message may be caused by a negative term on the factor diagonal.. A potential cause for getting a negative diag. term is that the matrix is ill-conditioned. The error may be avoided by increasing the IPAD parameter on the ITER entry. Another possibility is- to use BICILL instead of BICWELL. @ 29315; 70.5; Accounting - MSCACT, acct, acid, acdata - Missing Documentation & Examples To add accounting of the account ID (acid) and site-dependent account data (acdata) to the acct and MSCACT utilities, the following steps are required: 1) Modify the nastXX.lcl, where XX is "69", "691", "70", to pass the acid and acdata values to acct. For example, in the nast70.lcl file, change %acct=y:% %a.archdir%/acct $acctf $startdate %user% %out% MSC/%appl% %version% to %acct=y:% %a.archdir%/acct $acctf $startdate %user% %out% MSC/%appl% %version% "%acid%" "%acdata%" This change adds the two values, in quotes, to the acct command line. 2) Compile the modified mscact.c source as described in the COG, chapter 6. # cd <ins-dir>/msc70/util # msc70 util mscact This modified program adds "summary=acid" and "summary="acdata". 3) Install the new mscact executable as mscact *and* acct in the architecture directory, for example, # cp mscact <ins-dir>/msc70/<arch>/mscact # cp mscact <ins-dir>/msc70/<arch>/acct 4) Use the "acct=y", "acid", and "acdata" keywords as required. 5) Generate reports using mscact, for example, msc70 mscact summary=acid <ins-dir>/acct/msc97*.acc @ 29316; 70.5; Optimization/Design Sensitivity - DVPREL1 referenced on a DRESP2 - SIGTRAP FP The message SIGTRAP TRACE TRAP FP OVERFLOW may appear in the log file when a DVPREL1 entry is referenced by a DRESP2 entry and sensitivity results are requested. Avoidance: Replace the DVPREL1 reference with a reference to the associated DESVAR's or remove the request for sensitivity output. @ 29324; error 70; SSSALTERs - rflagba.v691, TRD1 -Receiving signal number 10 : Bus

When running a large problem in Version 69.1 on the SGI, if insufficient memory is available, then the job may fail in the TRD1 module with no error messages in the f06 file. A symptom of this error is the presence of the following error message in the log file: Receiving signal number 10 : Bus error This error has been observed in a large transient dynamic run when using the rflagba.v691 sssalter. Avoidance: Increase the memory. @ 29329; ; Virtual Mass - MPRES - Missing Element Lables from Output, Missi ng Output Element labels for complex "PRESSURES FOR ELEMENTS ON FLUID STRUCTURE BOUNDARY" may be missing if SORT2 output is requested. Avoidance: Request SORT1 output. @ 29338; ng Output ; Postprocessing/Nonlinear Analysis - PCOMP, PARAM,POST,-2 - Missi

In solution 106 with lgdisp,1 and composite elements, ply stresses are printed in the f06 file but not in the param,post,-2 output2 file. Also homogeneous stresses are put in the output2 file although they are not requested. Avoidance: Use the following alters: V68: $ $ the following alter is used to work-around a problem when $ using PCOMP in SOL106 with INTOUT=YES on NLPARM (V68.2.3) $ originally the NLSTATIC DMAP writes OESCP multiple times $ with PARAM,POST,-1 which causes problems for PATRAN... $ the fix is to APPEND OESCP within the loop and write it $ afterwards. $ modified 10-9 to add I-DEAS output (POST=-2) $ $ output2 blocks: OESCP laminar stresses $ OESTRCP laminar strains $ COMPILE NLSTATIC SOUIN=MSCSOU NOLIST NOREF $ v68.2 $ ALTER 30 $ allow blocks to be overwritten FILE OESCP=APPEND/OESTRCP=APPEND/OEFIT=APPEND $ ALTER 433,434$ change intermediate output block names $ and then append to original block names $ SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT/ OESCP1,OEFIT1,/0 $ OFP OESCP1,OEFIT1//S,N,CARDNO $ APPEND OESCP1,/OESCP/2 $ APPEND OEFIT1,/OEFIT/2 $ $

ALTER 444,445$ change intermediate output block names $ and then append to original block names $ SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT/ OESTRCP1,OEFITX,/1 $ OFP OESTRCP1//S,N,CARDNO APPEND OESTRCP1,/OESTRCP/2 $ $ ALTER 447,447 $ move output2 outside the loop! ALTER 464 $ IF ( (POST=-1 AND PATVER>=3.) OR POST=-2 ) OUTPUT2, OESCP,OESTRCP//0/OUNIT2//OMAXR/'OESCP'/'OESCP' $ $ $FILE**** cpalter $ $ updated: replaced line numbers with string based alters $ for V69.0 21-mar-97 L.Proctor/MSC. Should help in subsequent $ versions. $ $ the following alter is used to work-around a problem when $ using PCOMP in SOL106 with INTOUT=YES on NLPARM (V68.2.3) $ originally the NLSTATIC DMAP writes OESCP multiple times $ with PARAM,POST,-1 which causes problems for PATRAN... $ the fix is to APPEND OESCP within the loop and write it $ afterwards. $ $ modified 10-9 to add I-DEAS output (POST=-2) $ $ output2 blocks: OES1C laminar stresses $ OESTR1C laminar strains $ OEFIT failure indices $ $ COMPILE NLSTATIC NOLIST NOREF $ $ alter 'FILE.*SAVE' $ allow blocks to be overwritten FILE OES1C=APPEND/OSTR1C=APPEND/OEFIT=APPEND $ alter 'SDRCOMP'(1,0),''(1,1) $ change intermediate output block names $ and then append to original block names $ SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT,bgpdth/ OESCP1,OEFIT1,/0//loadstep $ OFP OESCP1,OEFIT1//S,N,CARDNO $ APPEND OESCP1,/OES1C/2 $ APPEND OEFIT1,/OEFIT/2 $ $ alter 'SDRCOMP'(2,0),''(2,1) $ change intermediate output block names $ and then append to original block names $ SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT,bgpdth/ OESTRCP1,OEFITX,/1//loadstep $ OFP OESTRCP1//S,N,CARDNO APPEND OESTRCP1,/OSTR1C/2 $ $ alter 'OUTPUT2.*OESCP','' $ move output2 outside the loop! alter 'PARAML.*UGNR' $ this is just outside the loop IF ( POST=-2 ) THEN $ OUTPUT2, OES1C,OSTR1C//0/OUNIT2//OMAXR/'OES1C'/'OSTR1C' $

OUTPUT2 OEFIT//OTAPE2/OUNIT2//OMAXR $ ENDIF $ V69: $ $ updated: replaced line numbers with string based alters $ for V69.0 21-mar-97 L.Proctor/MSC. Should help in subsequent $ versions. $ $ the following alter is used to work-around a problem when $ using PCOMP in SOL106 with INTOUT=YES on NLPARM (V68.2.3) $ originally the NLSTATIC DMAP writes OESCP multiple times $ with PARAM,POST,-1 which causes problems for PATRAN... $ the fix is to APPEND OESCP within the loop and write it $ afterwards. $ $ modified 10-9 to add I-DEAS output (POST=-2) $ $ output2 blocks: OES1C laminar stresses $ OESTR1C laminar strains $ OEFIT failure indices $ $ COMPILE NLSTATIC NOLIST NOREF $ $ alter 'FILE.*SAVE' $ allow blocks to be overwritten FILE OES1C=APPEND/OSTR1C=APPEND/OEFIT=APPEND $ alter 'SDRCOMP'(1,0),''(1,1) $ change intermediate output block names $ and then append to original block names $ SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT,bgpdth/ OESCP1,OEFIT1,/0//loadstep $ OFP OESCP1,OEFIT1//S,N,CARDNO $ APPEND OESCP1,/OES1C/2 $ APPEND OEFIT1,/OEFIT/2 $ $ alter 'SDRCOMP'(2,0),''(2,1) $ change intermediate output block names $ and then append to original block names $ SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT,bgpdth/ OESTRCP1,OEFITX,/1//loadstep $ OFP OESTRCP1//S,N,CARDNO APPEND OESTRCP1,/OSTR1C/2 $ $ alter 'OUTPUT2.*OESCP','' $ move output2 outside the loop! alter 'PARAML.*UGNR' $ this is just outside the loop IF ( (POST=-1 AND PATVER>=3.) OR POST=-2 ) THEN $ OUTPUT2, OES1C,OSTR1C//0/OUNIT2//OMAXR/'OES1C'/'OSTR1C' $ OUTPUT2 OEFIT//OTAPE2/OUNIT2//OMAXR $ ENDIF $ $

V70: $ $ V70.

$ $ the following alter is used to work-around a problem when $ using PCOMP in SOL106 with INTOUT=YES on NLPARM (V68.2.3) $ originally the NLSTATIC DMAP writes OESCP multiple times $ with PARAM,POST,-1 which causes problems for PATRAN... $ the fix is to APPEND OESCP within the loop and write it $ afterwards. $ $ modified 10-9 to add I-DEAS output (POST=-2) $ $ output2 blocks: OES1C laminar stresses $ OESTR1C laminar strains $ OEFIT failure indices $ $ COMPILE NLSTATIC NOLIST NOREF $ $ alter 'FILE.*SAVE' $ allow blocks to be overwritten FILE OES1C=APPEND/OSTR1C=APPEND/OEFIT=APPEND $ alter 'SDRCOMP'(1,0),''(1,1) $ change intermediate output block names $ and then append to original block names $ SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT,bgpdth/ OESCP1,OEFIT1,/0//loadstep $ OFP OESCP1,OEFIT1//S,N,CARDNO $ APPEND OESCP1,/OES1C/2 $ APPEND OEFIT1,/OEFIT/2 $ $ alter 'SDRCOMP'(2,0),''(2,1) $ change intermediate output block names $ and then append to original block names $ SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT,bgpdth/ OESTRCP1,OEFITX,/1//loadstep $ OFP OESTRCP1//S,N,CARDNO APPEND OESTRCP1,/OSTR1C/2 $ $ alter 'OUTPUT2.*OESCP','' $ move output2 outside the loop! alter 'PARAML.*UGNR' $ this is just outside the loop IF ( (POST=-1 AND PATVER>=3.) OR POST=-2 ) THEN $ OUTPUT2, OES1C,OSTR1C//0/OUNIT2//OMAXR/'OES1C'/'OSTR1C' $ OUTPUT2 OEFIT//OTAPE2/OUNIT2//OMAXR $ ENDIF $ $ @ 29339; ; Elements - Fluid Points, CELASi, CDAMPi, CMASSi - UFM 2012 If a scalar element is attatched to a fluid point, the following error message may occur: ** USER FATAL MESSAGE 2012 (GP1GSM) IDENTIFICATION 1 SAME BETWEEN GRID, SCALAR OR POINT IN SUPERELEMENT -1 Avoidance: Use DMIG entries for these coefficients or attach the scalar elements to the 1 dof instead of the 0 dof of the fluid point. @ 29340; 70.5; Heat Transfer Analysis - MATT4 - Wrong Answers Transient Heat transfer - CBAR element - Version 69 and version 70 only

If MATT4 references TABLEM1 or TABLEM2 entries that define constant values, then wrong answers result. Avoidance: Do not use the MATT4 if the material properties are constant. Use the MAT4 to specify these properties. Or if you are using the CBAR element,changed to crod element @ 29341; 70.5; Static Analysis - GPFORCE, GPFDR - SFE 4276 FILPS2 EC 1044 If grid point force output is requested for only some of the subcases, and the set does not contain all the grid points, then the job may fail with one of the following error messages below. Similarly in superelements, if there is a gpforce request above the subcase level that references a set and the set does not contain a grid point in each of the superelements then the job will fail. SYSTEM FATAL ERROR 4276.FILPS2 ERROR CODE 1044 PID= 0 GPFDR

*** USER WARNING MESSAGE 2354 (GPFDR) MODULE IS UNABLE TO CONTINUE AND HAS BEEN TERMINATED DUE TO ERROR MESSAGE PRINTED ABOVE OR BELOW THIS MESSAGE. THIS ERROR OCCURRED IN CODE WHERE THE VARIABLE -NERROR- WAS SET = 43 *** SYSTEM WARNING MESSAGE 3002 (GPFDR ) EOF ENCOUNTERED WHILE READING DATA BLOCK SCRATCH (FILE 301) IN SUBROUTINE GPFDFR 0FATAL ERROR Avoidances: There are 3 possible avoidances: 1. Put the gpforce output request above the first subcase. 2. Change the gpforce output to all. 3. Put in a gpforce(plot)=x in the subcase that gpforce printout is not desired. The value of x must reference the same set as the other subcases. @ 29342; n SE0 GL; Superelements - External SE with CMR, CSET - CMR DOF AUTOSPC'd i If a component mode DOF from an external SE is not being attached into the residual structure as anticipated it is an indication that the mode is either a linear combination of the static transformation vectors or the first mode calculated. MSC/NASTRAN will eliminate this mode via the autospc and INREL DMAP mechanism by design. Two avoidances exist. 1) remove the cset entries from the external SE run. or 2) ask for more modes to be calculated and add a second spoint putting it into the qset. @ 29344; nswers

70.5; Elements - CTRIAR, MCID - Error in Material Orientation, Wrong A

Specifying a MCID (integer, material coordinate system) in field 7 of the CTRIAR entry does not work. The resulting THETA (real, orientation angle) will be a large number ( 512.*(MCID+1) ), yielding incorrect stiffness for anisotropic materials. It may also lead to wrong element mass matrix. Avoidance: Do not use an integer in field seven of the CTRIAR entry, use the CTRIA3 element instead. DIAG 38 may be used to print the material angles.

@ 29348; ut

70.5; Acoustic Analysis - PARAM,STRUCTMP, PARAM,FLUIDMP - Missing Outp Requests for structural modal participations and fluid modal participation factors do not work. Avoidances: Structural modal participations can be obtained for Version 69 and later systems with the following dmap alter: compile mpf alter ',ah,' type parm,,char8,y,acsym='yes' $ if ( acsym='yes' ) then delete /ah,,,, $ PARTN bHH,VPHFS,/,mAH,, $ add5 mah,,,,/ah/-1. $ endif $ alter 'z2x' if ( acsym='yes' ) then $ add5 z2x,,,,/z2xx/-1. $ equivx z2xx/z2x/-1 $ endif $ and fluid modal participation factors can be obtained with the following alter: alter 'ENDIF \$ FLUIDMP\<NOFHSET'(,-1) else $ equivx PHDFH1/PHDFH2/-1 $

@ 29356;

70.5; Optimization/Design Sensitivity - DESVAR, mem - SFM 3007 If an optimization job does not have enough memory allocated to it, it may fail with the message " *** SYSTEM FATAL MESSAGE 3007 (DOM9 ) ILLEGAL INPUT TO SUBROUTINE DOM9P4". Avoidance: Allocate more memory (using mem keyword on NASTRAN command line). If more memory is not available, and the model contains only DVPREL1 entries with one associated DESVAR, then you can reduce the memory requirement by setting DELP > DELX on the DOPTPRM entry.

@ 29367; 70.7; Nonlinear Transient Analysis - SADD5 - UFM 4551, UFM 5423, Miss ing Output When a nonlinear transient model has difficulty converging, NASTRAN issues

some variation of the message "USER FATAL MESSAGE 4551". The job stops at that point and results which have been obtained up to that point are printed out. However, if the variation "USER FATAL MESSAGE 4551, STOPPED PROBLEM DUE TO TWO DIVERGENCES IN THE SAME TIME STEP." is issued, then no results are printed, and it is followed by the fatal message: 0*** USER FATAL MESSAGE 5423---ATTEMPT TO ADD INCOMPATIBLE MATRICES, (SADD5) COLS ROWS FORM TYPE NZWD DENS TRAILER FOR PST1 2 875 TRAILER FOR IFS 19 9740 Avoidance: Correct the model so that it will not diverge. @ 29398; 70.7; Nonlinear Analysis - CGAP - Incorrect SPC Force Output When the CGAP element orientation is defined by the orientation vector in fields 6 through 8, the moment due to frictional shear force is not calculated correctly. The resulting moment ignores the length of the element and is incorrectly calculated as the (shear force)/2. As a consequence, the moment component in SPC forces is incorrect. Avoidance: Use CID on CGAP to define the orientation of the Gap element. @ 29411; 70.5; Optimization/Design Sensitivity - ANALYSIS=BUCK - Missing Eigenv alue Table If buckling analysis is used in an optimization run, then only the initial eigenvalue table will be printed. Avoidance: Insert the following dmap alter: compile moders $ alter 1 $ type parm,,logical,n,eigprt $ alter 'if.*nasout.*ofp.*lama'(,-1) $ eigprt=nasout $ nasout=true alter 'if.*nasout.*ofp.*lama' $ nasout = eigprt $ endalter $ $ @ 29412; 70.01; Security - FLEXlM - Missing License Server on Windows NT The initial delivery of the MSC/NASTRAN Version 70 CD-ROM does not contain the FLEXlm network licensing servers for Windows NT. If you want to use your Intel Windows NT system as the license server and do not already have this software, then you can get the network software from MSC by calling our toll-free Order Processing hotline at: 1-800-400-4672. A revised Version 70 CD is available, the Version 70.5 CD will also have the Intel Windows NT license server. IS 28 22 2 1 IS 27 22 2 1

For Digital Alpha Windows NT systems, we do not have the network license server available. We will update the Version 70 Windows NT CD-ROM to contain both licensing servers in the near future. @ 29421; ; Miscellaneous - after keyword - Maximum File Size 2 Giga Bytes On an SGI IRIX 6.4, use of the NASTRAN keyword "after" will lower the maximum filesize from unlimited to 2Gb. This is error is caused by a bug in the SGI operating system. Avoidance: There is no avoidance except not to use the "after" keyword. @ 29425; 70.5; Elements - CTRIA6, CQUAD8, ELFORCE - SFM 4276, EC 7777, Trap Enc ountered If element force data recovery is requested for the CTRIA6 elements, then the job may fail on the IBM RS6000, HP, and possibly other machines. Depending on the machine, the job may fail with the following error message in the f06 file, ***** RECEIVED SIGNAL 8, ABORTING ... = BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX (FILEX) 0 STATUS = 0

or with the following error messages in the log file without any messages in the f06 file. Trap encountered: SIGTRAP - Trace trap FP overflow Traceback: Offset Offset Offset Offset Offset Offset Offset 0x00000024 0x00005044 0x0000075c 0x00000c54 0x00000260 0x00000194 0x00000970 in in in in in in in procedure procedure procedure procedure procedure procedure procedure gmmatd_ str61d_ s2bbrd_ sdr2b_ sdr2_ xsem13_ mainal

Avoidance: If strains and mpcforces are not needed then the following dmap alter will avoid the error: COMPILE SEDRCVR $ AVOID CSR 29425 ALTER 'OQMG1,,OSTR1'(,-1) $ JUMP NOSTR1 $ ALTER 'OQMG1,,OSTR1' $ LABEL NOSTR1 $ @ 29442; 70.5; Post Processing - GPFORCE - Can Not Post Process GPFORCE Results An error in the access key for grid point force results in the xdb

prevents these results from being post processed in transient and modes runs. MSC/PATRAN Version 7.5 supports post processing of these results but MSC/NASTRAN Version 70.5 will be required to produce the corrected xdb. @ 29478;

69.1; Documentation - Release Notes/jet101.dat - UFM 6794 (SEPSEB) The file jet101.dat on the V69 delivery CD, in the msc69/nast/misc/doc/relnotes directory, contains an error which will result in the following message: USER FATAL MESSAGE 6794 (SEPSEB) DUPLICATE BOUNDARY POINTS DETECTED CONNECTING TO SUPERELEMENT GRID 26 FROM DOWNSTREAM 0. LIST FOLLOWS: 4101 4001 4 AT

This error is caused by grid points 4001 and 4002 which are not needed. Avoidance: Remove the definition for grid 4001 and 4002. The correct file is on the delivery CD for Version 69.1. @ 29480; R 1044 2001; Nonlinear Transient Analysis - PBUSHT, KN, TABLED1 - BIOMSG ERRO Nonlinear transient analysis with CBUSH elements that refer to a PBUSHTentry that defines stress dependent properties may fail in the NLTRD2 module with SYSTEM FATAL MESSAGE 4276 BIOMSG ERROR 1044. Avoidance: There is no avoidance. @ 29489; 70.5; P-elements - GMSURF with zero derivatives - Poor Answers If a GMSURF has a point with zero derivatives, such as the center of a cylinder or sphere, linear element edges with grids at the singularity may have skewed parameterizations. This may result in poor answers. Avoidance: An avoidance is to specify an FEEDGE for any such element edge. No GMCURV or POINTs need to be specified for a linear edge. @ 29508; Fault ; Nonlinear Transient Heat Transfer - THERMAL, DBX - Segmentation If a nonlinear transient heat transfer analysis, sol 159, specifies multiple THERMAL case control requests the job may fail with the following message in the log file. Process Id = 4393 Receiving signal number 11: Segmentation fault Job aborting... dbx: Not found Unable to call dbx for trace: Error 0 Avoidance: The data deck that encountered this error was attempting to request printed output for a specific set and punched output for all. The correct way to make this request is to use the OTIME entry to limit the printed request. @

29526;

70.5; Random Response Analysis - TABRND1 - UFM 316 (IFPDRV) The SKIP feature on the TABRND1 entry does not work when LOG-LOG option is used. The following messages are issued. *** USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA ENTRY TABRND1 SORTED ENTRY COUNT = ** *****

Avoidance: Do not use the SKIP feature when LOG-LOG is used. @ 29531; 70.5; Superelements - SEBULK, SECONCT, SEP1X - SIGSEG, Segmentation v iolation When using the enhanced superelement capabilities (begin super=x), only one grid point may occupy the same location for a particular superelement at a particular boundary location. This is a documented limitation. Logically a user may believe that one way to get around this limitation is to use the SECONCT command; however, the use of the command for this purpose will cause the job to terminate with no message in the f06 file and the traceback shown below in the log file. The proper way to get around this limitation is to use the SEEXCLD command. Trap encountered: SIGSEGV - Segmentation violation Traceback: Offset Offset Offset Offset Offset Offset 0x00000a90 0x000003f8 0x00000180 0x00000754 0x00000458 0x00000970 in in in in in in procedure procedure procedure procedure procedure procedure sepmc1_ sepman_ sepsrh_ sep1x_ xsem08_ mainal

Avoidance: Use the SEEXCLD command to exclude all extra boundary points. Note that the SEEXCLD command only works with partitioned bulk superelement. @ 29532; 70.5; Superelements - SEEXCLD, SEMA - SFM 4192, 3062 If the enhanced superelement capabilities (begin super=x) are used, and there is more than one boundary point that occupies the same location, the SEEXCLD can be used to exclude the extra boundary points. However, if the model also contains the old superelement type (seset), then the job may fail with the following misleading error messages: 0*** SYSTEM FATAL message 4192 (sema) GRID POINT 10 IN SUPERELEMENT 0 IS NOT DEFINED IN BGPDT FOR DOWNSTREAM SUPERELEMENT 0*** SYSTEM FATAL MESSAGE 3062 MODULE = SEMA FAILURE OCCURRED IN UTILITY MAPBTI CALLED IN SUBROUTINE SEMA . REQUESTED EXTERNAL ID(S) NOT FOUND. 0FATAL ERROR

Avoidance: Reordering the grid sequencing for the superelements may avoid this error. Reorder the grids such that the lowest superelement number has the lowest ids, the next superelement has the next group of higher ids, and so on. Note that the SEXCLD entry only works with the partitioned bulk superelement. @ 29547; 70.5; Aeroelastic Analysis - MKAERO - UFM 1126 The message: *** USER FATAL MESSAGE 1126 (GNFIST) DMAP MODULE ADD IS ATTEMPTING TO OUTPUT DATABLOCK 201 NAME = AJJ0 , WHICH ALREADY EXISTS. USER ACTION: DELETE THE DATABLOCK, OR USE FILE AJJ0 =OVRWRT is produce when the input deck has m,k (Mach number and reduced frequency) pairs that are identical. Avoidance: Remove duplicate mk pairs. @ 29548; 70; SSSALTER - seaspa.v70 - Possible Wrong Answers An error in the seaspa.v70 alter (wrong value of app to some apd calls) results in un-needed data being checked improperly. It is unlikely that the error in the alter will cause errors in answers; however, the alter has been fixed to remove this error and it also gives better diagnostics if the auxmodel input file is given a wrong value. The revised alter may be down loaded from the web and has a revised date of Oct 23, 1997. The appropriate sample input files have also been updated and posted on the web at http://www.macsch.com. @ 29577; 70.5.2; P-Elements - Singularity in Stresses When Interior Angle of Elem ent is Small P-Elements may encounter a data recovery error when the interior angle of the element is small. Avoidance: Other than modifying the mesh, there is no avoidance. In Version 71 an new system cell 269 is introduced to allow the user to specify a threshold ratio for the Jacobian determinant at a VU-grid with respect to it p-element center. If the ratio is smaller than the threshold specified, then the Jacobian determinant at the VU-grid is replaced with its element-center value when computing the stress and strain. @ 29580; lar 70; Documentation - V69/70 COG - Sec 4.8.3 does not include HP Exemp In the COG, Section 4.8.3 describes a method to flush the F04 and F06 files to disk during a long running job. The method is stated to work on Convex, Cray, and SGI systems. The method also works on HP Exemplar systems. The documentation should also explicitly state that after the buffers are flushed the job is terminated. Edition 2 of the Version 70 Configuration and Operations Guide has been clarified. @ 29581; 70.5; DMAP - MATMOD OPTION 13, COPY - SFM 1269 (SLCDRV)

Copying the lower triangular factor from DECOMP or DCMP module may lead to the fatal message: *** SYSTEM FATAL MESSAGE 1269 (SLCDRV) THE ASSOCIATED FILE FNUMERIC OF DATABLOCK USER ACTION: CONTACT MSC CLIENT SUPPORT. 0FATAL ERROR Avoidance: The error occurs only in Version 70, Version 70.5 corrects the error and prior versions also work. @ 29613; 70.5; Documentation/P-Elements - PVAL - Recommended Minimum p-order. In general, an analysis started at P=1 will not produce the same results as starting at other p-levels. Although it is not suggested in the PVAL documentation where the p-order is set by either default or recommendation, a minimum starting p-level of 3 is recommended for accuracy and efficiency. @ 29615; 70.5; Documentation - DMIG, UACCEL - Incorrect Example in QRG The format and example for the UACCEL entry in the Quick Reference Guide implies that enforced accelerations are input as one column. The correct input format is one column for each enforced acceleration case. A better example with more typical values follows: $ 2 dmig, dmig, dmig, dmig, 3 uaccel, uaccel, uaccel, uaccel, 4 0, 2, 3, 4, 5 9, , , , 6 1, , , , 7 , 2, 2, 2, 8 , 3, 4, 6, 9 , 386.4 3.0 1.0 101 IS ALREADY OPEN.

Remark 2 implies that the SUPER case control command must be present. This is not true. A better remark would be: "2. The load sequence number interpretation depends on the value of NCOL field on dmig, uaccel, 0 entry. The recommended method is to set it equal to the number of loading conditions, 4 in this example. The load sequence number L is then the sequence number of the subcase to which the applied acceleration will be applied. For this example, the following subcase structure might be used: subcase 11 label = Conventional load = 11 subcase 12 label = L value of 2 subcase 13 label = L value of 3 subcase 14 label = L value of 4 loads, no enforced accelerations used here used here used here

If NCOL is left blank, the number of loading conditions is the number of columns input on the dmig, uaccel entry. In this example only 3 subcases would be allowed. Subcase 11 would have to be deleted and the column for L=2 (the first column input) would apply to subcase 12 ( the first subcase

remaining). @ 29619; 70.5; Elements - PBARL, PBEAML - Wrong Answers for K1, K2, Cw, J The BEAM library capability in MSC/NASTRAN Version 70 calculates the K1, K2 (the shear stiffness factor and CW (the warping coefficient) for the the BOX section incorrectly. Avoidance: Calculate the correct cross sectional properties and enter use the pbar and pbeam entries or create a beam server library with the correct cross sectional properties. @ 29622; 70.5; Cyclic Symmetry - SOL 15, EIGRL - SFM 3007 If the EIGRL Bulk Data entry is selected and the CYSUP entry is also specified in SOL 15 then the run may fail with SFM 3007 in REIGL. Avoidance: Run SOL 115, use EIGR entry or insert the following DMAP alter: compile sol15 alter 'reigl ','' REIGL Kxx,Mxx,DYNAMICS,CASEcc,,MR,DMx,vxcompr,,,vaxw/ LAMA,phixc,MI,EIGVMAT,OUTVEC/READAPP/S,N,NEIGV/RECNO/ ///////////nreigl $ @ 29625; 70.5; Nonlinear Analysis - PARAM,POST - DBC Segmentation Violation Nonlinear jobs that fail due to insufficient time or failure to converge that have param,post,0 may continue and issue a traceback from the DBC module: Traceback: Offset Offset Offset Offset Offset 0x000006c4 0x00000db8 0x00001180 0x000002d4 0x00000984 in in in in in procedure procedure procedure procedure procedure dbcaol_ convrt_ dbc_ xsem14_ mainal

--- End of call chain ---

@ 29639; 70.5; Documentation - V70 Release Notes - Missing DMAP Non-upward Comp atibilities The MSC/NASTRAN Version 70 Release Guide did not contain the description of the DMAP module non-upward compatibilities. Below is the information that should have been included. The information is also soon to be included on the MSC external web site. The location will be included in this report when it becomes available. This section summarizes DMAP module changes from Version 69 and 69.1 to Version 70. This information is intended to help you convert your Version 69 and 69.1

DMAP alters or sequences to run in Version 70. A structured solution sequence excerpt shows how the module is actually invoked in the structured solution sequences and the additions or deletions are shown in lowercase in the Version 69 or 70 excerpt, respectively. The following modules have changed in Version 70. (SEP1X changed in V69.1). DSAD DSAH DSAL DSTAP2 DSVG1P GPSP MDCASE SEP1X

_______________ DSAD Module Add 1 data block, BGPDTS, to the end of the input list. Add 4 data blocks; CASDSN, DRDUG, DRDUTB, AND CASADJ, to the end of the output data block list. These data blocks are needed for adjoint sensitivity analysis. Version 69 in SubDMAP DESCON DSAD RSP1CT, R1TAB, RESP12, OBJTAB, CONTAB, BLAMAX,LAMASX, DIVTAB,AUXTAB,STBTAB,FLUTAB,OUGVS1,OESS1,OSTRS1, OEF1AA, EFITS,ES1CS,EA1CS,OQGS1, DSCREN, XINIT, COORDN, OLX, FRQRSP, CASEDS, CASERS, UGX, OPTPRM, PROPI / R1VAL,R2VAL,RSP2R,R2VALR,CVAL,CVALR,OBJTBR,CNTABR, R1TABR,R1VALR,DRSTBL,FRQRPR,UG1,R1MAPR,R2MAPR/ WGTS/VOLS/S,N,OBJVAL/S,N,NR1OFFST/S,N,NR2OFFST/ S,N,NCNOFFST/APP/DMRESD/SEID/DESITER/EIGNFREQ $

Version 70 in SubDMAP DESCON DSAD RSP1CT, R1TAB, RESP12, OBJTAB, CONTAB, BLAMAX,LAMASX, DIVTAB,AUXTAB,STBTAB,FLUTAB,OUGVS1,OESS1,OSTRS1, OEF1AA, EFITS,ES1CS,EA1CS,OQGS1, DSCREN, XINIT, COORDN, OLX, FRQRSP, CASEDS, CASERS, UGX, OPTPRM, PROPI,bgpdts/ R1VAL,R2VAL,RSP2R,R2VALR,CVAL,CVALR,OBJTBR,CNTABR, R1TABR,R1VALR,DRSTBL,FRQRPR,UG1,R1MAPR,R2MAPR, casdsn,drdug,drdutb,casadj/ WGTS/VOLS/S,N,OBJVAL/S,N,NR1OFFST/S,N,NR2OFFST/ S,N,NCNOFFST/APP/DMRESD/SEID/DESITER/EIGNFREQ/ s,n,adjflg/pexist/mbcflg $

DSAH Module Remove data blocks UGSN amd QGSN, and their subsequent commas, from the input list. Remove data blocks UGDSN amd QGDSN, and their subsequent commas, from the output list. Add input parameters ADJFLG and SEID to end of the parameter list.

Structured solution sequence excerpts: Version 69 in SubDMAP RESPSEN DSAH DRSTBL,R1TABR,CASEDS,TABECN,BLAMAX,LAMASX, ugsn,qgsn,OLB,DIVTAB,FRQRPR,VIEWTBDS,CASERS, BUGX,PHGX/ BUDG,PHDG,CASEDSF,LBTAB,BDIAG,LFTAB, COGRID,COELEM,ugdsn,qgdsn,DSEDV,OINTDSF,PELSDSF/ APP/DMRESD/NDVTOT $

Version 70 SubDMAP RESPSEN DSAH DRSTBL,R1TABR,CASDSN,TABECN,BLAMAX,LAMASX, OLB,DIVTAB,FRQRPR,VIEWTBDS,CASERS,csnmb, BUGX,PHGX/ BUDG,PHDG,CASEDSF,LBTAB,BDIAG,LFTAB, COGRID,COELEM, DSEDV,OINTDSF,PELSDSF/ APP/DMRESD/NDVTOT/adjflg/seid $

DSAL Module Add two data blocks, DRDUTB and ADELU, to the end of the input list. Add input parameter ADJFLG to the end of the parameter list. These data blocks and parameter are needed for adjoint sensitivity analysis. Structured solution sequence excerpts: Version 69 in SubDMAP RESPSEN DSAL DRSTBL,DELWS,DELVS,DELB1,DELF1,COGRID,COELEM, OUGVDSN,OESDSN,OSTRDS1,OEFDSN,EFITDS,ES1CDS,EA1CDS, R1VALR,TABDEQ,OLB,DSDIV,DELX,DELS,DELFL,FRQRPR,DELBSH/ DSCM/NDVTOT/DELTAB/EIGNFREQ $

Version 70 in SubDMAP RESPSEN DSAL DRSTBL,DELWS,DELVS,DELB1,DELF1,COGRID,COELEM, OUGVDSN,OESDSN,OSTRDS1,OEFDSN,EFITDS,ES1CDS,EA1CDS, R1VALR,TABDEQ,OLB,DSDIV,DELX,DELS,DELFL,FRQRPR,DELBSH, drdutb,adelu/ DSCM/NDVTOT/DELTAB/EIGNFREQ/adjflg $

DSTAP2 Module Remove the last 4 data blocks, DRSTBLG, FOL1V, TOL2V, and FRQRPRG from the end of the input list. Structured solution sequence excerpts: Version 69 in SubDMAP EXITOPT DSTAP2 R1TABRG,RSP2RG,drstblg,fol1v,tol2v,frqrprg/DSCMCOL $

Version 70 SubDMAP EXITOPT DSTAP2 R1TABRG,RSP2RG /DSCMCOL $

DSVG1P Module Add two data blocks, GPSNTS and ESTDV2B, to the end of the input list. These data blocks are required for optimization with p-elements. Structured solution sequence excerpts: Version 69 in SubDMAP PSLGDV DSVG1P EST,ESTDV2F,BGPDVP,CSTMS,MPTS,DIT,DEQATN,DEQIND,UGX,,DSPT1/ EGKP,/ COUPMASS/K6ROT/ALTSHAPE/WTMASS/NOPSLG/1 $

Version 70 SubDMAP PSLGDV DSVG1P EST,ESTDV2F,BGPDVP,CSTMS,MPTS,DIT,DEQATN,DEQIND,UGX,,DSPT1, gpsnts,estdv2b/ EGKP,/ COUPMASS/K6ROT/ALTSHAPE/WTMASS/NOPSLG/1 $

GPSP Module For correct processing of MPC's and rigid elements during PARAM AUTOSPC operations specify KMM in the second position of the input data block list and the string 'KMM' in the tenth parameter position. Structured solution sequence excerpts: Version 69 in SubDMAP SEKR GPSP KNN, ,USET0,SILS,GPLS,,GEOM4S,EQEXINS/USET,/S,N,NOSSET/ AUTOSPC/PRGPSTX/SPCGEN/EPZERO/ACON/ S,N,SING/EPPRT/S,N,NOSET/S,N,NGERR $

Version 70 SubDMAP SEKR0 IF ( NOMSET>=0 ) UPARTN USET0,KGG/KMM,,,/'G'/'M'/'N' $ GPSP KNN,kmm,USET0,SILS,GPLS,,GEOM4S,EQEXINS/USET,/S,N,NOSSET/ AUTOSPC/PRGPSTX/SPCGEN/EPZERO/ACON/ S,N,SING/EPPRT/S,N,NOSET/S,N,NGERR/'kmm' $

MDCASE Module Add 1 data block, EDOM, to the end of the input data block list and 1 output

parameter, WVFLG to the end of the parameter list. This data block and parameter are needed for adjoint sensitivity analysis. Structured solution sequence excerpts: Version 69 in SubDMAP DESOPT MDCASE CASEXX/ CASESTAT,CASEMODE,CASEBUCK,CASEFREQ,CASEMTRN,CASESAER, CASEDVRG,CASEFLUT,,,,,CASESADV,CASESNMB,CASEXN/ S,N,STATCC/S,N,MODECC/S,N,BUCKCC/S,N,DFRQCC/S,N,MFRQCC/ S,N,MTRNCC/S,N,SAERCC/S,N,DVRGCC/S,N,FLUTCC/ ////S,N,DESOBJ/S,N,DESGLB/S,N,OBJSID/TRUE $

Version 70 SubDMAP DESOPT MDCASE CASEXX,edom/ CASESTAT,CASEMODE,CASEBUCK,CASEFREQ,CASEMTRN,CASESAER, CASEDVRG,CASEFLUT,,,,,CASESADV,CASESNMB,CASEXN/ S,N,STATCC/S,N,MODECC/S,N,BUCKCC/S,N,DFRQCC/S,N,MFRQCC/ S,N,MTRNCC/S,N,SAERCC/S,N,DVRGCC/S,N,FLUTCC/ ////S,N,DESOBJ/S,N,DESGLB/S,N,OBJSID/TRUE/s,n,wvflg $

SEP1X Module In Version 69, SEP1X was executed only if partitioned superelements were present. In Versions 69.1 and 70, SEP1X is now executed for all models, regardless of the presence of superelements. SEP1X was enhanced to refine the grid point connections of RSSCON elements. Three parameters, SEP1XOVR, SEBULK, and TOLRSC, are added to control the operation of the module. SEP1XOVR and TOLRSC are described in the MSC/NASTRAN Version 70 Quick Reference Guide. SEBULK is a logical parameter which is TRUE if partitioned superelements are present. Structured solution sequence excerpts: Version 69 in SubDMAP PHASE0 IF ( SEBULK ) SEP1X SELIST,GEOM1F,GEOM2F,GEOM4F,SETREE,SGPDTF/ EMAP,SGPDT,SCSTM/ S,N,NOSE/CONFAC/'PEID'//S,N,PARTRS/NQSET/'XEID' $

Version 70 SubDMAP PHASE0 SEP1X SELIST,GEOM1F,GEOM2F,GEOM4F,SETREE,SGPDTF/ EMAP,SGPDT,SCSTM/ S,N,NOSE/CONFAC/'PEID'//S,N,PARTRS/NQSET/'XEID'/ sep1xovr//sebulk/tolrsc $

@ 29643;

70.5; Documentation - COMPILER - Missing Documentation The design intent of the COMPILER LIST called by a solution sequence. If you you must input a specific COMPILE LIST desired. Remark 4 should be clarified statement is to list all subDMAPS wish to list only specific subdmaps statement for each subdmap listing to discuss this design intent.

@ 29645;

70.5; Static Analysis - GPFORCE, SUBCOM, SET - Wrong Answers If GPFORCE force is requested for a set of grid points and SUBCOM is used, then the the gpforce output from the SUBCOM subcase will be incorrect. Avoidance: There are two possible avoidances: 1. Request GPFORCE=all or

2. Use SUBCASE instead of SUBCOM @ 29646; 70.5; Optimization/Design Sensitivity, Differential Stiffness - STATSU B - UFM 1126 When performing a buckling analysis in Optimization (SOL 200) with no mass on any unconstrained degree of freedom in the model and with scr=yes on the nastran command line, then the message "USER FATAL MESSAGE 1126 (GNFIST) DMAP MODULE UPARTN IS ATTEMPTING TO OUTPUT DATABLOCK 202 NAME = KSF, WHICH ALREADY EXISTS. USER ACTION: DELETE THE DATABLOCK, OR USE FILE KSF=OVRWRT", may be issued. The error may also occur in SOLs 103 and 107 through 112 if differential stiffness is requested with the STATSUB command. Avoidance: There are three avoidances: 1. Insert a dummy mass element on any unconstrained degree of freedom. 2. Insert the following DMAP avoidance in V70. compile sekr nolist, noref $ $ alter 10 $ $ $ check for KFF $ if KFF exists, then so does KFS, KSF, KSS $ if it doesn't exist generate KFF, KFS, KSF, KSS $ dbstatus kff//s,n,nokff $ if (nokff < 1) then $ $ $ if in this if statement, then either first time to $ generate KFF, or KFF was scratch datablock $ either way KFF needs to be generated $ alter 13, 13 $

if (nosset >= 0) then $ $ $ delete upartn output in case this is scratch $ run and KFS, KSF, KSS are on database $ delete /KSF,KFS,KSS,, $ UPARTN USET,KNN/KFF,KSF,KFS,KSS/'N'/'F'/'S' $ $ endif $ $ alter 15 $ endif $ 3. Specify scr=no on the command line. @ 29648;

70.5; Eigenvalue Analysis - EIGRL - SFM 4276, EC 7777 MSC/NASTRAN Version 70 on Dec Alpha OSF may fail in the EIGRL module with the message SYSTEM FATAL MESSAGE 4276, ERROR CODE 7777. The following traceback will be issued in the log file: handler(0x11fffdff8,0x11fffdf98,0x8,0x0,0x11fffdff8) L. 28 <<?? pc=0x12180e0b4>>(0x0,0x0,0x0,0x8,0xc00000007) L. 28 dot3rd_(0x140602658,0x1208397bc,0x1409cbddc,0x12164ccc8,0x1409eb378) 93 lncrdd_(0x1404c1678,0x165000b38,0x1404c1678,0x1409eb378,0x165001708) 266 lncr3d_(0x165001c30,0x1407ac4f0,0x165001c50,0x165001c70,0x165000b38) 357 lncr2d_(0x1650023c8,0x165002550,0x1650026d8,0x1650008f8,0x140645210) 429 lncr1d_(0x165002550,0x165001cc8,0x1650026d8,0x165000250,0x1650002f8) 457 lncs1d_(0x1405d2348,0x14064522c,0x140645230,0x140645228,0x1406450c8) 893

L. L. L. L. L. L.

Avoidance: In the single case that this error has been encountered the error was avoided by changing the EIGRL entry from specifying ND only to specifying a specific frequency range. @ 29655; 70.6; SSSALTERs - propa.v70, propf.for - Wrong Answers The sssalter propa.v70 and accompanying fortran program propf.for enables the flutter analysis of propellers. Two errors have been discovered in this package. The first is that the propeller blade aspect ratio is based on the propeller diameter when it should be based on the radius. The second is that the example case based on NASA TN 1807, uses a reference semichord that is twice that specified in the NASA TN. Avoidance: Either input the aspect ratio into the preprocessing routine directly or modify the supplied FORTRAN code to use radius in the aspect ratio calculation. For the test case, divide the input chord lengths by two. @ 29663; 70.5; Elements - CRAC3D, TEMP(LOAD), GP3D - SFM 3007 If TEMP is applied to a CRAC3D element that has 32 grids or more the run

will fail with the message: *** SYSTEM FATAL MESSAGE 3007 (GP3 ) ILLEGAL INPUT TO SUBROUTINE GP3D Avoidance: If possible use TEMPD to apply the temperature load. @ 29672; 70.5; PCOMP, STRAIN - Missing Output In post Version 68 systems, the case control command STRAIN=n is supposed to write ply strains to the .f06 file. This command only works if stress is also requested. Thus, there is no way to get ply strains only. Avoidance: To get only strain output, specify the STRESS=m command where SET m specifies an undefined element. @ 29691; 70.5; Optimization/Design Sensitivity - ANALYSIS=FLUTTER, K2GG, M2GG B 2GG - SFM 6296 If flutter optimization is attempted with CASE CONTROL commands K2GG, M2GG or B2GG then the run will fail and the message SYSTEM FATAL MESSAGE 6296 Unable to locate a member of datablock QHHL" will be issued. Avoidance: Don't use the K2GG, M2GG or B2GG commands in an ANALYSIS=FLUTTER subcase in SOL 200. A complex workaround involving TABEDIT's is available if this avoidance is unacceptable. @ 29699; 70.5; Postprocessing - PARAM,OUNIT1 PARAM,OUNIT2 - Misplaced Output The parameters ounit1 and ounit2 are ignored in Version 70 when used with param,post,-2. Avoidance: If the ounit1 and ounit2 are needed, the following DMAP alter may used: $ $ for v70.0 $ compile super3 nolist, noref $ alter 335,335 $ type parm,nddl,i,y, ounit1 $ type parm,,i,n, ounit1x $ message //' ounit2 before out2geom call= '/ounit2 $ IF ( POST<0 ) THEN $ IF ( OUNIT1>0 ) THEN $ OUNIT1X=OUNIT1 $ ELSE $ OUNIT1X=OUNIT2 $ ENDIF $ ENDIF $ IF ( ( ((DROPT=0 OR DROPT=4) AND APP1<>'BKL1') OR APP='BKL0') AND (OGEOM='YES' OR OGEOM=' ') AND (POST=-2 OR POST=-4) ) CALL OUT2GEOM, CSTM,GPL,GPDT,EPT,MPT,GEOM2,GEOM3,GEOM4, GEOM1,DIT,DYNAMICS,EQEXIN,BGPDT,CASECC// OUNIT1x/S,NP $

@ 29704;

70.5; Performance - MPYAD - Large MPYAD has Excessive Run Time It has been observed on a very large test problem (1.5M DOF) that the matrix operation PhiT*M required more than 1000 passes, where M is symmetric and nearly diagonal, and Phi has the same number of rows of M and 70 columns. This type of operation is unlikely to occur in a standard solution, but may occur in a user-dmap. This operation is, for example, used in the checka.v70 sssalter. The MPYAD method selected was method 1T, which unpacks the (mostly null) rows of M in memory. This is on a model for which eigenvectors had already been computed, and the time spent up to the start of the mpyad was less than the estimate for the multi-pass solution. Avoidance: Instead of solving for x=PhiT*M in one step, break the operation into two steps as follows: xt = M*Phi x = trnsp(xt) This way the sparse methods are used in the first multiply, which is done in less than 10 passes.

@ 29705;

70.5; Postprocessing - PARAM,POST,-2, SCR=MINI - Missing OUTPUT2 Data When an input file is submitted with the scr=mini option and param post,-2 is used, no op2 output will be generated. Avoidance: Do not use the scr=mini option. If the scr=mini option and param,post,-2 are needed, then the following dmap alter may be used. COMPILE SUPER3 SOUIN=MSCSOU NOLIST NOREF ALTER 335,335 IF ( (OGEOM='YES' OR OGEOM=' ') AND (POST=-2 OR POST=-4) ) CALL OUT2GEOM, CSTM,GPL,GPDT,EPT,MPT,GEOM2,GEOM3,GEOM4, GEOM1,DIT,DYNAMICS,EQEXIN,BGPDT,CASECC// OUNIT2/S,NP $

@ 29710;

70.7; Miscellaneous - PBARL,PBEAML, ECHO=NONE - Unrequested punch file If both the beam library and echo=none are used in the same run, then a punch file containing the equivalent beam properties will be created. Avoidance: If this additional punch file is not desired, then remove the echo=none case control command.

@ 29723; FF)

70.7; Buckling & Differential Stiffness - PARAM,FOLLOWK - SFM 3002 (TA When performing buckling analysis or stress stiffened analysis (STATSUB) with follower force effects (PARAM, FOLLOWK, YES), and a LOAD bulk data entry references more than one PLOAD4 entry, then the following fatal message will occur: *** SYSTEM FATAL MESSAGE 3002 (TAFF ) EOF ENCOUNTERED WHILE READING DATA BLOCK (NONE) (FILE SUBROUTINE TAFMEG 0) IN

Avoidance: 1) Set PARAM, FOLLOWK, NO -- thus follower force effects will be ignored. 2) Do not reference more than one PLOAD4 card on the LOAD bulk data entry. @ 29724; 70.5; Miscellaneous - Incorrect OS Version on Fujitsu systems The Operating System version reported by the nastran command and in the MSC/NASTRAN F06 title block is incorrect. The correct version is obtained with the 'uname -v' command. The result "1" indicates "V10L10"; "2" indicates "V10L20"; and "3" is expected to indicate "V20L10". The correct version may be specified with the "s.osv" keyword in any RC file, for example s.osv=V10L10 @ 29736; 70.5; Optimization/Design Sensitivity - ANALYSIS=MFREQ - SFM 4276 EC 3 001 module DSAD If FREQ3, FREQ4 or FREQ5 Bulk Data entries are specified in SOL 200 run with ANALYSIS=MFREQ and a METHOD entry, the following message is erroneously received: USER WARNING MESSAGE 6645 (FRSRTD) FREQ3,FREQ4, AND FREQ5 ENTRIES ARE PRESENT IN A DIRECT FREQUENCY RESPONSE ANALYSIS. and the following message is issued in the DSAD module: BIOMSG: ERROR 3001 HAS OCCURRED IN ROUTINE GREAD 0 STATUS = 0 0*** SYSTEM FATAL ERROR 4276.GREAD ERROR CODE , FILE INDEX (FILEX) =

3001 PID=

Avoidance: Insert the following DMAP alter: compile presens alter 'dopr3 ' type parm,nddl,logi,n,zuzrl1 $ if ( not zuzrl1 ) then $ delete /CONTAB,R1TAB,RESP12,RSP1CT, $ delete /frqrsp,CASEDS,ointds,pelsetds, $ endif $ not zuzrl1 alter 'enddo'(,-1) if ( not zuzrl1 ) delete /objtab,,,, $ alter 'return'(,-1) if ( not zuzrl1 ) delete /objtbg,,,, $ compile desopt alter 'frlgen ','modacc ' alter 'CALL PRESENS,'(2,-1) IF ( DESCYCLE=1 ) then $ ISOLAPP=6 $ FRLG CASErS,,,FRL,,,,/ ,,,fOL,/ 'modal'//'freqresp' $ PARAML FOL//'TRAILER'/1/S,N,NOFREQ $

matgen ,/pjunk/7/1/nofreq $ modacc cases,fol,,pjunk,,/oli,,xxx,,/'freqresp' $ type parm,nddl,logi,n,zuzrl1 $ type parm,,logi,n,npl $ zuzrl1=true $ indicates 2nd call to presens npl=gmbs $ save original value of gmbs gmbs=true $ force 2nd call to presens endIF $ DESCYCLE=1 alter 'CALL PRESENS,'(2) gmbs=npl $ restore original value of gmbs $ added to try to avoid FATAL error... alter 188,188 @ 29759; 70.5; Eigenvalue Analysis - EIGR entry with METH=SINV - Inefficient Pr ocessing Auto-omit processing is done unnecessarily with METHOD=SINV on the EIGR entry. Auto-omit requires a decomposition to condense out the massless dofs which produces a denser stiffness matrix. Avoidance: Insert the following DMAP alter: compile moders alter 'noared'(2) noared=-1 $ @ 29772; 70.5; Documentation - QRG, DRESP1 - Wrong units The units for eigenvalues is incorrect as described in remark 18 of the Quick Reference Guide's description of the DRESP1 entry. The correct units are (radian/time)**2 instead of (radian/time). @ 29778; 70.5; Plotting - PLTSET - Poor Performance The PLTSET module for Version 70 has a severe performance degradation as compared to Version 68.1. For one probelm, the two PLTSET calls in Version 68.1 used less than 11 CPU seconds while Vhile 70 consumes 470 CPU seconds. This problem has been corrected in Version 70.5 @ 29779; ; P-Elements - h-p Interface, FEEDGE - Incorrect Results When a model consists of both h- and p-elements, the edges of the p-elements which are on the h-p boundary will not remain at p-level 1 if these edges are defined with curved geometry. These edges will adapt to higher p-level, giving incorrect results. Avoidance: Do not define curved FEEDGE geometry on the h-p element boundary. @ 29780; 70.5; P-Elements - Singular Stiffnes Matrix Singular stiffness matrices or negative eigenvalues may occur in QUAD p-elements for lower p-values when the p-distribution is nonuniform. This is due to an error in bubble functions which are used for edges with p=1 and p=2. The error is fixed in V70.5 Avoidance: Turn bubble functions off with system cell 225, specify NASTRAN SYSTEM(225)=1 in the input file.

@ 29781; ed

70.5; Optimization/Design Sensitivity - PBARL, LINKER - Trap Encounter In an optimization run, if the PID (field 4) on the DVPREL1 entry references a PBARL entry, and the FID (field 5) references an invalid field on the PBARL entry, then the job will terminate in the f06 file without any messages. In this case, the job terminates after generating the equivalent PBAR entries. This error will result in the following error messages in the log file. Trap encountered: SIGSEGV - Segmentation violation Traceback: Offset Offset Offset Offset Offset Offset 0x000006f8 0x000005dc 0x0000028c 0x000010c0 0x00000284 0x00000934 in in in in in in procedure procedure procedure procedure procedure procedure domcb_ domc_ dopr1f_ dopr1_ xsem03_ mainal

Avoidance: Reference a valid PBARL FID. @ 29788; 70.5; DMAP - MATMOD(12) - Illegal Trailer, BIOMSG: ERROR 211 MATMOD(12) may output matrices with an incorrect trailer of the form of 7, when it should be a form of 2. Some modules may accept this vector, while others may crash in unpredictable manners. The ADD module, for example, exits with the message below: BIOMSG: ERROR 211 HAS OCCURRED IN ROUTINE BLDPK, FILE INDEX (FILEX) = 0 Avoidance: Use the MODTRL module to change the form to 2. @ 29797; 007 ; Optimization/Design Sensitivity - DSAD, ANALYSIS=FLUTTER - SFM 3 Flutter optimization cannot be performed with superelements unless the design model is in the residual. If attempted the run will fail and issue the following messages: ^^^ FLUTTER ANALYSIS COMPLETED. DESIGN CYCLE NUMBER= *** SYSTEM FATAL MESSAGE 3007 (DSAD ) ILLEGAL INPUT TO SUBROUTINE DSAD14 0FATAL ERROR 1

@ 29798; 70.5; Optimization/Design Sensitivity - DRESP1, FRSPCF, TSPCF, DSADJ UFM 5423, 3007 If the DRESP1 entries requests spc forces (FRSPCF or TSPCF) in a frequency response or transient response optimization task and all the constraints are inactive then an incompatible matrix multiply may result leading to the following message: *** SYSTEM FATAL MESSAGE 3007 (DSADJ ) ILLEGAL INPUT TO SUBROUTINE DSDJ2D 0FATAL ERROR

or *** USER FATAL MESSAGE 5423---ATTEMPT TO ADD INCOMPATIBLE MATRICES, (SADD5) Avoidance: Use the DSCREEN entry to force a constraint in each subcase to be active.

@ 29800; N

70.5; Virtual Mass - SOLs 3 and 7 through 12, MFLUID - SFM 3001 in MGE In SOLs 3 and 7 through 12, if there is an ELIST1 entry and no free surface nor plane of antisymmetry then SFM 3001 is issued from the MGEN module. Avoidance: Use SOLs 103 and 107 through 112.

@ 29802;

70.5; Documentation - QRG, SPLINE3 - Incorrect Equation The equation written in Remark 2 for the SPLINE3 Bulk Data Entry is incorrect. The sign on the right-hand side of the equation should have a minus sign. That is the processing of the SPLINE3 entry is similar to the MPC entry and the equation should be written as: u + sum(A * u ) = 0 d i i

@ 29817; 70.0.5; Materials - MATT1, MATT2 - Temperature Dependent Materials May F ail, SFM 3007 In Version 70, material table lookups can fail if there are points outside the range of the supplied tables. This is true for both linear and nonlinear solution sequences. Symptoms of this error are the following messages: *** USER WARNING MESSAGE 4698. STATISTICS FOR DECOMPOSITION OF MATRIX KLLRH . THE FOLLOWING DEGREES OF FREEDOM HAVE FACTOR DIAGONAL RATIOS GREATER THAN 1.00000E+07 OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL. *** USER WARNING MESSAGE 3057 (NLITER) MATRIX IS NOT POSITIVE DEFINITE. *** SYSTEM FATAL MESSAGE 3007 (NLITER) ILLEGAL INPUT TO SUBROUTINE NLINIT Avoidance: Define the temperature dependent material tables over a wider range. @ 29825; 70.5; Superelements - BEGIN BULK SUPER, GENEL - UFM 316, Partitioned S uperelements If GENEL elements are used in a partitioned superelement run, then "USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA ENTRY GENEL xxx SORTED ENTRY COUNT = yyy" will be incorrectly issued if any GENEL ids in

the second and subsequent bulk data sections are greater than or equal to any GENEL ids in a previous bulk data section. Avoidance: 1) Do not use duplicate GENEL ids, and 2) Number your GENELs and order the bulk data sections such that the GENEL ids of the second and subsequent bulk data sections are less than those of the preceding bulk data sections. @ 29829; 70.5; Static Analysis - STRSORT - Stress Sorting may be Inefficient In very large models, if stress sorting is requested, then the time spent in the stress sorting module STRSORT may dominate the total runtime. This is most apparent in less efficient I/O machines. Avoidance: There is no known avoidance. @ 29837; 70.5; Differential Stiffness - STATSUB, SOLs 107, 108, 109, scr=no - U FM 9058 If scr=no and the Case Control command STATSUB are specified in SOLs 107, 108, or 109 to select the static subcase for differential stiffness, then the run will fail with the message: USER FATAL MESSAGE 9058 (STATICS2) THE SOLUTION FOR THE RESIDUAL STRUCTURE AND THE CURRENT SUBCASE DOES NOT EXIST. Avoidance: Submit the job with scr=yes or insert the following DMAP alter: compile phase1dr alter 'not(statics) or kdiff','' IF ( true ) then $

@ 29851;

70.5; Installation - Not all the files could be Decompressed. If the "GZIP" environment variable is set, mscsetup may fail with the error message Not all the files could be decompressed. <req> MB free space are required. <free> MB were available in the installation directory /some/dir/ <free> MB were available in the work directory /some/otherdir/ Please make more space available and redo this installation using the command /some/dir/mscsetup.pbk -t <work_directory> where <work_directory> is an alternate directory with at least <req> MB free space that can be used for scratch purposes during the installation process.

Prior to this error message, you may also see lines like: 7 22:18 36976 120864 69.4% ./msc70/irix64/demo1 11 Jul 7.gz: No such file or directory 22:18.gz: No such file or directory 36976.gz: No such file or directory 120864.gz: No such file or directory 69.4%.gz: No such file or directory ./msc70/irix64/demo1.gz: 69.4% -- replaced with ./msc70/irix64/demo1 7 22:17 28057 74876 62.5% ./msc70/irix/demo2 7 Jul 7.gz: No such file or directory 22:17.gz: No such file or directory 28057.gz: No such file or directory 74876.gz: No such file or directory 62.5%.gz: No such file or directory ./msc70/irix/demo2.gz: 62.5% -- replaced with ./msc70/irix/demo2 Avoidance: Unset the GZIP environment variable: unset GZIP unsetenv GZIP # Bourne and Korn shell # C shell (ugh)

and rerun the installation. @ 29852; 70.0.5; Grid Point Force - GPFORCE - UWM 2354 (GPFDR) Large TETRA models, about 35000 grids, with a GPFORCE request may fail with the message: . *** USER WARNING MESSAGE 2354 (GPFDR) MODULE IS UNABLE TO CONTINUE AND HAS BEEN TERMINATED DUE TO ERROR MESSAGE PRINTED ABOVE OR BELOW THIS MESSAGE. THIS ERROR OCCURRED IN CODE WHERE THE VARIABLE -NERROR- WAS SET = 37 *** SYSTEM WARNING MESSAGE 3003 (GPFDR ) ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK SCRATCH (FILE 302) IN SUBROUTINE GPFDFB 0FATAL ERROR Avoidance: Increase memory. @ 29859; 6137 70.7; Complex Eigenvalue - CEAD, EIGC, HESS - Traceback in hess1, UWM Complex eigenvalue analysis that selects the Hessenberg method may issue a traceback in the hess1 routine with no other message whenever the mass matrix is singular. Other runs may issue a warning message that the input matrix is rank deficient or may explicitly state that the mass matrix is singular and proceed and give unreliable answers. Examples of the traceback and warning message are given below: Offset Offset Offset Offset Offset 0x00000030 0x00000ec4 0x00000c60 0x0000006c 0x00000994 in in in in in procedure procedure procedure procedure procedure pack_ hess1_ cead_ xsem11_ mainal

*** USER WARNING MESSAGE 6137 (DFMN)

INPUT MATRIX IS RANK DEFICIENT, RANK = 159 USER ACTION: CHECK MODEL *** USER INFORMATION MESSAGE 4158 (DFMSA) ----STATISTICS FOR SPARSE DECOMPOSITION OF DATA BLOCK CMDXX FOLLOW NUMBER OF NEGATIVE TERMS ON FACTOR DIAGONAL = 6 MAXIMUM RATIO OF MATRIX DIAGONAL TO FACTOR DIAGONAL = 3.5E+16 AT ROW NUMBER 209 Avoidance: When using the Hessenberg method of complex eigenvalue extraction, carefully inspect the results for all warning messages and use aset/omit entries to remove massless degrees of freedom from the dynamic set. Otherwise select the complex lanczos method on the eigc entry which is not sensitive to massless degrees of freedom. @ 29890; 70.5; Frequency Response - DLOAD, RLOADi, NEC - Possible Wrong Answers On the NEC, if the DLOAD entry references more than 12 RLOADi entries in a frequency response analysis, only the last 12 RLOADi entries will be included in the applied load. The other loads are ignored. Avoidance: If more than 12 combined loadings are required, use a single RLOADi entry which in turn references multiple DAREA entries with the same ID. @ 29898; 70.5; Modal Augmentation - PARAM,RESVEC, SUPORTi - UIM (RESVEC) No Sta tic Shapes When a structure is free-free and a SUPORTi entry exists, there will be no static shapes to augment the mode shapes when there are no constraints in the model (i.e., no s-set, m-set or o-set). This condition leads to the following message: ^^^ USER INFORMATION MESSAGE ( RESVEC ) ^^^ NO STATIC SHAPES WERE FOUND TO AUGMENT MODE SHAPES Avoidance: Add a disjoint grid point, with no elements attached to it. Autospc will cause an s-set to exist. @ 29901; 70.5; Superelements - CBEND - UFM 2016 If a superelement contains a CBEND element that uses a MAT1 that is not referenced by any elements type, the MAT may not get copied internally into the bulk data associated with the superelement. This will results in: UFM 2016 MODULE = EMG NO MATERIAL PROPERTIES EXIST Avoidance: Make sure that each MAT1 that is used by a CBEND elements is also used by an element type other than the CBEND. This may require a dummy element such as CBAR be included in the appropriate superelements. This dummy element does not need to be part of the structure and may be fully constrained. @ 29909; 2001; Frequency Response - FREQ - SFE 4276, EC 32, Missing Output In a frequency response analysis, if too many continuation entries are used for a FREQ entry, then the job will either fail with the following

error messages: SYSTEM FATAL ERROR 4276.GETSTR ERROR CODE 32 PID= 0

or the output will not be generated for all the frequencies. Avoidance: Use a combination of the FREQ3, FREQ4, and FREQ5 in a modal frequency response analysis. @ 29919; 70.5; Documentation - SEEXCLD - Incorrect description The description for field 3 (SEIDB) of the SEEXCLD entry should be changed from SEIDB Superelement identification. (Integer > 0 or Character = "ALL") to SEIDB Superelement identification. (Integer >= 0 or Character = "ALL") @ 29922;

; Rigid Elements - RSSCON, SECNSD - UFM 6692 When a highly skewed plate element is attached to a solid element with an RSSCON entry, the run may fail with the message: *** USER FATAL MESSAGE 6692 (SECNSD ) RSSCON NUMBER *** CANNOT CONNECT AN EDGE OF QUAD ELEMENT *** TO HEXA ELEMENT *** Avoidance: The only avoidance is to reduce the skewness of the plate element or to use the GRID option on the RSSCON entry.

@ 29932;

69; Installation - MSC/NASTRAN will not run on SGI IRIX 6.4 MSC/NASTRAN V68.2.3 (r2) for the SGI R8K may not run on IRIX 6.4 or later systems. The failure will occur at program start, before any processing has begun, with the following message: MSC/NASTRAN started Mon Nov 17 08:45:14 PST 1997 Process Id = 27828 idalloc: address of dynamic common is too low current brk value: 0x12700000 ( 309329920) current address of XNSTRN: 0x121f0000 ( 304021504) ***memory allocation failed, program stop*** real 0m0.04s user 0m0.02s sys 0m0.02s MSC/NASTRAN finished Mon Nov 17 08:45:14 PST 1997 The specific reason is that the amount of code and data put into the user's address space by the runtime libraries and OS has become too large and interferes with MSC/NASTRAN's open core common block /XNSTRN/. Avoidances: 1) Do not run IRIX 6.4 or later or run a more current version of MSC/NASTRAN.

2) It appears that at the one site that could run on IRIX 6.4 they had removed patch SG0002422. Thus, this may be an avoidance. However, be aware that another site that did not have this patch installed still could not run. @ 29944; 70.5; Static Aeroelastic Analysis - QSETi, SEQSETi - UWM 5229, SFMs 21 98, 2199, 3007 If QSETi run will AESTATRS messages entries are present in SOL 144, Static Aeroelastic Analysis, the fail in subDMAP AESTATRS. The message issued from subDMAP varies. If the a-set contains only q-set dofs, the following are issued:

*** USER WARNING MESSAGE 5229 (DDCOMP) INPUT FILE TO MODULE DECOMP PURGED. MODULE RETURNS WITH PURGED OUTPUT. *** SYSTEM FATAL MESSAGE 2198 MODULE = FBS INPUT DATA BLOCK 101 HAS BEEN PURGED. *** SYSTEM FATAL MESSAGE 2199 MODULE = FBS SUMMARY, ONE OR MORE OF THE ABOVE FATAL ERRORS WAS ENCOUNTERED IN SUBROUTINE DFBS . If the a-set contains q-set and t-set dofs the the following message are issued: *** SYSTEM FATAL MESSAGE 3007 (UMERGE) ILLEGAL INPUT TO SUBROUTINE MERGE QSETi and SEQSETi entries are not permitted in SOL 144. However this is undocumented. Avoidance: Remove the QSETi and SEQSETi Bulk Data entries. @ 29965; 70.5; DBC - SWM 6065 - Undocumented Limit in DBC, SWM 6065 (DBC) The DBC module always has had a limit of 9999 allowable output positions available for any process list, ie time, frequency, modes, etc. The enforcement of this limitation has not always been consistent until V70. This limit may lead to the message: ^^^ SYSTEM WARNING MESSAGE 6065 (DBC) ^^^ DBCMSG ERROR MESSAGE 43 FROM SUBROUTINE DBCAOL OUTPUT LIST SUPPLIED EXCEEDS INTERNAL LIMIT, SUPPLIED INTERNAL LIMIT 10000.

20013

Version 70.5 uses a new dynamic access key technique which effectively increases the limit for many problems. MSC/XL and other programs may no be compatible with this new technique. SYSTEM(274)=1 (a.k.a. DBFACT=1) is provided to access the old technique. This system cell may be set in the system or user rc file as well as the individual test deck. @ 29976; ; Online Documentation - MNE - Irregular Fonts The characters on the menus of the MSC/NASTRAN Online Encyclopedia (MNE) appear as irregular fonts when displayed on a Digital UNIX V4.X.

Avoidance: Change the Iview*fontList resource value from -*-bold-r-*-*-14-* to -*-bold-r-*-*-12-* @ 29998; 70.5.2; Matrix Methods - MPYAD METHOD 4T - UFM 4276 EC -31 Certain problems cause MPYAD METHOD 4T to be selected when other methods would solve the problem much more efficiently. The problem will take longer to run than it should and may run out of disk space. The symptom of the bad method selection is that MPYAD METHOD 4 T is selected with a large number of passes. This can be seen in the f04 file, for instance, as follows: 20:36:57 267:07 53987.9 MPYAD END 20:36:58 267:08 53998.1 MPYAD BEGN CPU= 81.7, I/O= 20:36:58 267:08 20:36:58 267:08 20:36:59 267:09 20:36:59 267:09 20:36:59 267:09 20:36:59 267:09 20:36:59 267:09 626.3# 6843.0 0.5 6843.1 188.9# 0.0 FDRMGEN FREQRS 15

METHOD 4 T , NBR PASSES=****, EST. 40.0, TOTAL= 121.8 53998.1 0.0 6843.1 0.0 MPYAD BGN P=2930 53998.9 0.8 6843.2 0.1 MPYAD PASS= 1 53999.1 0.3 6843.3 0.1 MPYAD PASS= 2 53999.6 0.5 6843.4 0.1 MPYAD PASS= 3 53999.9 0.3 6843.5 0.1 MPYAD PASS= 4 54000.1 0.3 6843.6 0.1 MPYAD PASS= 5 54000.6 0.5 6843.8 0.1 MPYAD PASS= 6

Avoidance: The avoidance it to force another method by deselection of method 4T at the MPYAD where the fatal message occurred. Deselecting various methods is documented in the MSC/NASTRAN Version 68 Numerical Methods User's Guide. For instance, for the run above, use the following alter: compile subdmap=freqrs alter 14 putsys(128,66) alter 15 putsys(0,66) endalter @ 30005;

70.5; External Superelement - ASETi, SE - SEP1X, SFM 3007 In an external with the ASETi superelements, matrices. The Avoidance: superelement generation run, if the DMIGPCH option is used entries, and the external superelement contains upstream then the punch file will not contain the physical boundary job may also fail with SFM 3007.

There are two possible avoidances:

1. Use BSETi instead of ASETi entries to represent the boundary points. 2. Use the DMIGDB, DMIGOP2, or MATRIXDB option instead of the DMIGPCH option. @ 30007; Answers ; Documentation/Nonlinear Analysis - Multiple Subcases, Different

A transient nonlinear analysis may yield slightly different results when the problem is defined with multiple subcases rather than a single

subcase. The difference occurs because loads are averaged over three timesteps. The loads at the end of a subcase and the beginning of the subsequent subcase may then be slightly different. @ 30009; 70.5; Installation - SIGNAL 11 (SIGSEGV) CODE 5, or CODE 6 The submittal script on MSC/NASTRAN Version 70 on Convex computers has an error which may cause jobs that request more than 10 million words of memory to fail with the message: ***** RECEIVED SIGNAL 11 - Segmentation fault (SIGSEGV); CODE 5 - Invalid page table reference (SEG_INVPTP_TRAP) JOB ABORTING... Avoidance: Modify the nast70.lcl file in the msc70/bin directory as follows: %MSC_ARCH=convex:% rather than %MSC_ARCH=convex:% @ 30010; $tcmd %a.solver% $asgf %mem% $tcmd %a.solver% %a.rel=68:% $asgf %mem%

; Superelements - BEGIN BULK SUPER, RESTART - SFM 3001 A restart of a superelement problem will result in the fatal message shown below, if the cold start contains superelements defined in the main bulk data section as well as other superelements defined in the partitioned bulk data sections. *** SYSTEM FATAL MESSAGE 3001 (SEP1X ) THE EXECUTIVE FILE 0 DOES NOT EXIST. PROGRAMMER INFORMATION: SUBROUTINE SEPMAB Avoidance: Define all superelements in either the main bulk data section or the partitioned bulk data section, but not both.

@ 30011; 70.5; Optimization/Design Sensitivity - MFREQ, OFREQ - SFM 4276.GREAD EC 3001 If a multidisciplanary optimization run uses MFREQ as one of the disciplines, contains an OFREQ Case Control requests and has a value of P2>=8 on the DOPTPRM entry to indicate that the responses are to be printed, then the job will fail with the following fatal message in the DOM10 Module: SYSTEM FATAL ERROR 4276.GREAD ERROR CODE 3001 PID= 0 USER INFORMATION MESSAGE: TO OBTAIN A NASTRAN DUMP, RESUBMIT JOB WITH DIAG 44 INSERTED IN THE EXECUTIVE CONTROL SECTION. Avoidance: Remove the OFREQ request or lower the value of P2 on the DOPTPRM entry. @ 30015; 70.5; ESTIMATE - Incorrect estimates of DOF on TRIAX6 elements. The ESTIMATE utility may incorrectly count the degrees of freedom associated with CTRIAX6 elements. If the singular degrees of freedom are explicitly constrained, either on spc entries, grid entries or grdset entries, then the ESTIMATE will assume that a constraint in the 2 degree of freedom is removing one of physical degrees of freedom. This is not correct as a CTRIAX element is define in the x-z plane not the x-y plane.

As a result, ESTIMATE will undercount the DOF by one per CTRIAX6 element. Avoidance: Do not explicitly constrain the singular degrees of freedom, let autospc perform this function. @ 30020; 70.7; Matrix Methods - MPYAD 2 or 3 - Wrong Answers When very low memory is requested such that MPYAD method 2 or 3 are selected, and multiple passes are required, then the job will produce wrong answers. Insufficient memory causes the sparse method to be deselected and MPYAD METHODs 2 and 3 to spill in multiple passes. This information is contained in the f04 file with DIAG 19. Avoidance: Deselect method 2 and 3 by adding the following system cell to the top of the input file or system rc file: nastran system(66)=60 This may cause some jobs that require methods 2 or 3 to fail with the message: *** SYSTEM FATAL MESSAGE 3001 (*** ) THE SCRATCH FILE 399 DOES NOT EXIST. PROGRAMMER INFORMATION: SUBROUTINE GOPEN 0FATAL ERROR The most likely jobs to encounter this message are optimization jobs in DOPR1. If the fatal message is encountered, set system(66) back to the default of 0, turn on diag 19, and give the job plenty of memory. Then, scan the Execution Summary (f04 file) for MPYAD and ensure that multiple PASSES were not required for METHODs 2 or 3. @ 30023; 70.5.1; Miscellaneous - Ishell does not work on HP Systems The ISHELL module will not work on the HP Version 70 system because the allocated size of the stack and child thread was too small for HP machine. The original cause of this error was fixed in Version 70.5.0 but another error effectively renders the error not fixed on HP-UX 10.2 systems. The error has been fixed in Version 70.5.1. Avoidance: There is no avoidance. @ 30027; 70.5; Optimization/Design Sensitivity - SPCD, scr=no, auxiliary models - Wrong Answers In a shape optimization run with SPCD loads on auxiliary models and scr=no the answers are wrong after the first cycle. Avoidance: Use scr=yes or insert the following DMAP avoidance: compile update alter 'delete \/yg1,' DBVIEW ygp=yg (WHERE DESITER=DESP) $ RESTART YGp,YG,//TRUE $ DETECT ENFORCED DISP CHANGES endalter @ 30030; 70.5; Aeroelastic Analysis - SSG2 - Poor Efficiency The ssg2 module reduces the spline matrix gtkg to several other

components. The r-set component is not used anywhere, and this operation takes a considerable amount of cpu time. Avoidance: An alter which eliminates this computation is to remove this component from the dmap call: COMPILE AESTAT NOLIST NOREF $ V70 ALTER 'SSG2','' $ REMOVE GTKR FOR EFFICIENCY. NOT USED DOWNSTREAM. $SSG2 USET,GM,,,GOA,DM,GTKG/GTKR,GTKO,GTKS,GTKA,GTKL $ SSG2 USET,GM,,,GOA,DM,GTKG/,GTKO,GTKS,GTKA,GTKL @ 30037; 70.5; Optimization/Design Sensitivity - DESOPT, TFL - UFM 3055 If the TFL command is specified and ANALYSIS=MFREQ or DFREQ in a SOL 200 run that performs adjoint sensitivity then the run will fail with UFM 3055 at statement 404 in subDMAP DESOPT. Avoidance: Insert the following DMAP alter: compile desopt alter 'PARAML CASERS' PARAML CASERS//'DTI'/1/ 15// S,N,TFL $ @ 30041; 70; Installation/Security - No read access to authorize.dat or licen se.dat Files If the user running mscsetup has their umask set to a value that disables group or other access, e.g., 026 or 027, mscsetup may create files that are not readable by the user community. The specific files of concern are: <install-dir>/flexlm/licenses/license.dat and/or <install-dir>/conf/authorize.dat. Note, having a restrictive umask value is a necessary, but not sufficient, condition for the error to occur. This error has been corrected on the MSC/NASTRAN V70 (r4) CD. Avoidance: Either of the following actions below will avoid the error. 1) Set the umask to 022 before running mscsetup. or 2) Change the permissions of the affected files after installation, e.g., chmod a+r <install-dir>/flexlm/licenses/license.dat chmod a+r <install-dir>/conf/authorize.dat chmod a+r <install-dir>/conf/nast70rc.dat @ 30045; tput 70.5; Frequency Response Analysis - CDAMP1, CBUSH - Wrong SPCFORCE Ou If frequency dependent damping and no stiffness is specified on a CBUSH-element then the SPC forces will be wrong.

Avoidance: Insert the following DMAP alter: compile freqrs alter 'equivx bggf\/xggf','' ADD bggf,/xGGf/iomega $ endalter

@ 30054;

70.7; Eigenvalue Analysis - STATSUB, PARAM,INREL - UFM 9058 (SUPER3) If a PARAM,INREL,-1 is erroneously placed in a SOL 103 run along with a STATSUB command for differential stiffness, then the following error message will be received: ^^^ USER FATAL MESSAGE 9058 (SUPER3) - THE SOLUTION FOR THE RESIDUAL STRUCTURE DOES NOT EXIST. This error needs to be trapped in order help the user correct the problem. Avoidance: Remove the PARAM,INREL,-1 since it is only valid for inertial relief in SOL 101.

@ 30058; 70.5.2; Documentation - Incorrect Specification of FLEXlm License Server s on Windows NT The MSC/NASTRAN Version 70.5 NT "Configuration and Operations Guide" indicates the format for a list of alternate FLEXlm licensing files or license server nodes is specified as value:value:... The correct format for NT machines is value;value;... Note, ANY list of values, such as a PATH, or the FLEXlm license spec, use semicolons on Windows NT. @ 30073; 70.6; Nonlinear Analysis - APPEND - SFM 4276, EC 1039 In a nonlinear run, it is possible that the run may terminate in APPEND with SFM 4276, EC 1039. Avoidance: Set SYSTEM(158)=1000 and SYSTEM(159)=500 @ 30083; 70.5; Transient Response Analysis - TLOAD, TRL ,- time: command termin ated abnormally If a transient response analysis has a DLOAD that references a nonexistent TLOAD id, the job aborts without any messages except: *** TERMINATING /pw/devl/svr4/structures/t477/nastran/msc70/solaris/analysis.ult ra

*** Received signal 11 (SIGSEGV) time: command terminated abnormally in the log file. Avoidance: Remove the reference to the nonexistent TLOAD id. @ 30101;

70.5; Superelements - Component Mode Synthesis - UFM 3007 (UMERGE) If an insufficient number of q-set degrees of freedom are allocated through the SEQSETi entry to hold all of the internally generated residual flexibility vectors, the run will fail in the UMERGE module with the message "USER FATAL MESSAGE 3007" indicating illegal input to a subroutine. Avoidance: Supply a larger number of SEQSETi degrees of freedom.

@ 30104;

70.5; Documentation, QRG - TIME - Incorrect Default Value The default value for the executive control section entry "TIME" changed in V69. The value is now 1.892160e9 (seconds), not 60 minutes as documented. Avoidance: Specify a user defined value for TIME.

@ 30109;

70.5; DMAP - NORM - SFM 4276 EC 7777 The norm module sets an output parameter to the size of the maximum magnitude number in a matrix. On short word machines these numbers can be as large as 1.E60+. The number is stored in an output parameter of single precision whose maximum size is 1.E30+. Therefore, matrices which are legal for other modules cause overflow in norm. Avoidance: Short of not generating matrices with large numbers in them, there is no avoidance.

@ 30124; 70.0.3; Miscellaneous - GPWG, PARAM,GRDPNT,0 - time: command terminated abnormally Jobs with PARAM,GRDPNT on SUN SOLARIS Version 70.0.0 systems may fail in the GPWG Module with the message: *** TERMINATING /local/msc/msc70/solaris/analysis *** Received signal 11 SIGSEGV time: command terminated abnormally. Avoidance: Remove the PARAM,GRDPNT from the data deck. The error could not be repeated on Version 70.0.3 or 70.5. @ 30136; 70.5; Installation - "The following items are missing:" If a user or site has replaced the standard tar(1) command with another "equivalent" utility, e.g., pax(1), mscsetup may fail with the following error: The following items are missing: msc70/access

where msc70/access is any file to be loaded from the CD-ROM. Avoidance: Make sure the tar command invokes the standard tar(1) executable. @ 30147;

; Nonlinear Analysis - NLPCI, CRIS - Negative Load Factors If the Crisfield arc length method (TYPE=CRIS) is used in a post buckling analysis, then the job may not converge. A symptom of this problem is the printing of large negative load factors as shown below. SUBID 2 LOOPID FACTOR-7.38852477 SUBID 2 LOOPID FACTOR-8.89144518 SUBID 2 LOOPID FACTOR*********** 29 30 31 LOAD STEP LOAD STEP -6.389 -7.891 LOAD LOAD LOAD

LOAD STEP -10.059

Avoidance: Use the RIKS method. @ 30200; 70.5.2; Rigid Elements - RSSCON - UFM 6694, Incorrectly Worded Fatal Mes sage When two solid elements are connected to one shell element using the ELEM option on the RSSCON entry, the following error message is issued: USER FATAL MESSAGE 6694 (GP4RS4) SHELL GRID POINT xxx IS CONNECTED TO MORE THAN TWO SHELL ELEMENTS BY RSSCON BULK DATA ENTRIES. THIS IS NOT ALLOWED IF THE SOLID ELEMENT FACES CONNECTED ARE TRIANGLES. Although the user made an illegal connection, the error message is incorrectly worded. The problem is not that the shell is connected to a shell but that it is connected to two solids. The RSSCON entry can not connect more than one solid element to one shell element when the ELEM option is used. If the shell to solid connection spans across more than one solid element, the GRID option must be used. @ 30201; 70.7; Loads - FORCE - Unexpected Answers A negative sign in the CID field of the FORCE entry may not be trapped and the analysis may continue as if the field was left blank indicating the loads are to be applied in the basic coordinate system. Avoidance: Other than checking the oload output and entering proper data in the field, there is no avoidance. @ 30204; ; Sensitivities/Optimization - MSGMESH - MSGMESH FE 9104 Running an input file with an auxiliary model generated with MSGMESH fails with the fatal error: 1 CGEN QUAD4 2001 2000 1 2054 INPUT ** *** MSGMESH FATAL ERROR 9104 *** N1=1 N4=0 ERROR ** 3 N2=5 2001 N3=0

for any MSGMESH input entries. Avoidance: Do not use MSGMESH entries in the auxiliary subcase. @ 30205; 70.5; Dynamics - PARAM,RESVEC - Wrong Answers with MAX Normalization The Modal Augmentation (Residual Vector) Method introduced in Version 70 gives wrong answers if the modes are normalized to MAX. Avoidance: Use MASS normalization of modes. @ 30212;

70.5; Superelements - SUPORTi, CSETi, GDR - SFM 3007 (UMERGE) When performing component mode synthesis using generalized dynamic reduction (GDR) and CSETi and SUPORTi entries are present the run may fail in the GDR module with the following message: ^^^ USER INFORMATION MESSAGE ( RESVEC ) ^^^ NO STATIC SHAPES WERE FOUND TO AUGMENT MODE SHAPES *** SYSTEM FATAL MESSAGE 3007 (UMERGE) ILLEGAL INPUT TO SUBROUTINE MERGE 0FATAL ERROR Avoidance: Apply the following DMAP alter to avoid the error: compile semr3 alter 'call resvec','' call resvec pg,uset,gm,kfs,ys,bgpdts,scstm,cstms,medge,mff, cmpho0,loo,koo,cmlama0,moo,got,koo,cmphrc/ cmpho,cmlama/ nozset/norc/seid/'o ' $ alter 'equivx cmpho0','' CALL SEGOA uset,GOT,/GOAt/NOOSET/NOQSET/NOTSET $ EQUIVX CMPHO0/PHIVZ/NORC $ IF ( NORC>-1 ) Umerge1 USET,CMPHO0,CMPHRC,,/ PHIVZ/'V'/'O'/'C'/1 $ INREL USET,PHIVZ,,,,GOAT,/GOQ/0/EPSRC $ dotrip=true $ alter '/goq/' $ paraml goq//'trailer'/1/s,n,noqset1 $ if ( noqset1>noqset ) then $ Truncate GOQ to the true q-set size nqq1=noqset1-noqset $ matgen ,/vq1qx/6/noqset1/noqset/nqq1 $ partn goq,vq1qx,/phx3,,,/1 $ equivx phx3/goq/-1 $ dotrip=true $ endif $ compile resvec alter 'if ( nobset','' if ( nobset>-1 ) upartn uset,got/,,gorc,/'t'/'b'/'c'/2 $

@ 30229;

70.5; Installation - SGI R10000 misidentified during V70 installation On multi-processor SGI system running IRIX 6.2 or later, mscsetup may report a syntax error such as the following:

./mscsetup[24]: 10000^J10000^J10000^J10000^J10000^J10000^J10000^J10000: syntax error Avoidance: Set the MSC_ARCH environment variable to "irix64" for any R8 or R10 processor running IRIX64, or "irix" for any other combination of processor and OS. Note, specifically "O2" systems use the "irix" architecture. For example, on an Origin2000 system, set export MSC_ARCH=irix64 setenv MSC_ARCH irix64 @ 30235; # Korn shell $ C-Shell

70.5; Superelements - SUPORT1, PARAM,SESEF - UFM 3055 (MPYAD) The use of SUPORT1 entries with PARAM,SESEF,>=0 and superelements will result in *** USER FATAL MESSAGE 3055 (MPYAD ) AN ATTEMPT TO MULTIPLY OR MULTIPLY AND ADD NON-CONFORMABLE MATRICES TOGETHER WAS MADE IN MODULE MPYAD. Avoidance: $V70 compile postreig nolist noref alter 4 $ type parm,nddl,i,n,suport $ alter 25 $ after call setq paraml casedr//'dti'/1/255//s,n,suport $ ENDALTER

@ 30250; 70.5.2; Optimization/Design Sensitivity - REPCASE - SFM 3002 (DSADJ ) The message: SYSTEM FATAL MESSAGE 3002 (DSADJ) may occur when the adjoint method is invoked and there are REPCASE case control commands. Avoidance: The best avoidance is to remove the REPCASE commands. A more complex and less robust avoidance is to force the direct sensitivity method. This latter technique entails forcing the retention of an element response. @ 30251; 70.5; Grid Point Force - GPFORCE, SUBCOM - GPFDR - SFM 4276, EC 7777 If a run contains SUBCOMs and a GPFORCE request, the run will fail on Digital Alpha UNIX systems in the GPFDR module with the following message and traceback: SYSTEM FATAL ERROR 4276.tbk ERROR CODE 7777 PID= 0 USER INFORMATION MESSAGE: TO OBTAIN A NASTRAN DUMP, RESUBMIT JOB WITH DIAG 44 INSERTED IN THE EXECUTIVE CONTROL SECTION. Process Id = 9467 handler(0x11ffff138,0x0,0x120974d00,0x150fb59a4,0x1405fd6c0) L. 24 <<?? pc=0x12105c844>>(0x120974d00,0x150fb59a4,0x1405fd6c0,0x8,0xc00000007) L. 24 gpfdr_(0x1409ac7b0,0x11ffff494,0x14057e9b0,0x14057e9b4,0x1409ac7b4) L.

1057 xsem13_(0x12005727c,0x11ffff915,0x11ffff710,0x0,0x1) L. 94 mainal_(0x11ffff710,0x0,0x1,0x0,0x3ffc0082620) L. 349 <<?? pc=0x12005727c>>(0x140a40188,0x0,0x100000000,0x100000001,0x100000000) L. 349 <<?? pc=0x120056fa0>>(0x1,0x11ffff918,0x0,0x11ffff939,0x11ffff94e) L. 349 10:57:44 NSEXIT: EXIT(0) Avoidance: The only avoidance is to remove either the SUBCOM or GPFORCE requests. Equivalent SUBCASES can be defined to replace the SUBCOMs. @ 30252; File 70.5; Heat Transfer - RADMTX - Duplicate Continuation Entries in Punch If a heat transfer analysis contains multiple cavities and each cavity exceeds 100,000 elements the duplicate continuation entries for RADMTX may be output to the punch file. This problem only occurs when each cavity exceeds 100,000 elements. If each cavity has more than 100,000 elements,then you will get the duplicate continuation entries on the RADMTX bulk data entry. This error will lead to: *** USER FATAL MESSAGE 4392. CONTINUATION CARD ERRORS. EXPLANATIONS FOLLOW LIST OF CARDS IN ERROR when the punch file is used. Avoidance: Manually edit the punch file. The continuation fields may be left blank as long they are in order. @ 30254;

70.5; Complex Eigenvalue Analysis - CEAD - SFM 4276, EC 7777 If a complex eigenvalue analysis is missing a cmethod entry in the Case Control the job may end with a traceback in the CEAD module. Offset Offset Offset Offset Offset Offset Offset Offset Offset Offset 0x00000020 0x00004110 0x00000134 0x00000240 0x0000019c 0x00004048 0x000002c0 0x000010d8 0x00000094 0x000009b4 in in in in in in in in in in procedure procedure procedure procedure procedure procedure procedure procedure procedure procedure raise bioxxx_routine bioxxx_ nsexit_ pexit_ gperr_ mesage_ cead_ xsem11_ mainal

--- End of call chain --Avoidance: Add a cmethod entry in the Case Control. @ 30259; 70.5; Heat Transfer - TETRA, PENTA, CHBDYG - Poor Answers The ten noded TETRA element and the 15 noded PENTA element produce poor answers when the convection coefficient is large relative to the thermal conductivity of the material. For example , when h*delta(x)/k is larger than 10, where h is the convection coefficient, delta(x) is the element

length and, the k is the thermal conductivity of the material poor answers result. Avoidance: Use other element types such as the 4 noded TETRA which uses CHBDYG,AREA3 for convection. @ 30273;

; DMAP - CALL statement - UFM 72, Poor message If a value or variable is not specified for all parameters on a CALL statement then UFM 72 is issued with misleading information. For example, 83 CALL FRRD2X KDDZ,BDDZ,MDDZ,,PDFZ,FOL,CASES, EDT,SILD,USETD,/ UDVF1,FOLI1/ENFORCE///////FRRDSET//FIRSTBD0 $ 0*** USER FATAL MESSAGE 72 (XGPIDG ) PARAMETER SYNTAX ERROR IN ABOVE STATEMENT. PARAMETER IS ENFORCE --ERROR CODE = 2 Avoidance: Remove the unspecified parameters or specify dummy variables or values for the parameters.

@ 30293;

2001; Optimization/Design Sensitivity - DEQATN - Undocumented Limit The intrinsic functions MIN and MAX for the DEQATN entry have an open ended list of arguments. There is an undocumented limitation that the number of arguments for these functions has to be less than 100. This problem is considered fixed in Version 2001 by raising the limit to 1000. Avoidance: If it is desired to take the max (or min) of more than 100 arguments, several intermediate functions can be defined that find the max of a number of arguments less than 100 and then the max of the intermediate functions will result in the overall maximum.

@ 30299;

70.5; Optimization/Design Sensitivity - BCDR, Superelements - SFM 3002 If all the design variables and responses are specified in upstream superelements then the following messages may be issued in subDMAP PSLGDV and module BCDR: *** SYSTEM FATAL MESSAGE 3002 (BCDR ) EOF ENCOUNTERED WHILE READING DATA BLOCK (NONE) (FILE 101) IN SUBROUTINE BCDR. Avoidance: There are two possible avoidances: 1. Specify at least one design variable or response in the residual structure. 2. Insert the following DMAP alter: compile desopt alter 'call descon' seid=0 $ peid=0 $ equivx cases/casdsn/-1 $

endalter @ 30322; 70.0.5; Heat Transfer Analysis - TETRA - Wrong Answers In MSC/NASTRAN Version 70, a temperature look-up for a temperature dependent material property for the TETRA element only occurs at temperature=0.0. This causes wrong answers. Avoidance: Use other element types such as HEXA or PENTA elements when there are temperature dependent material properties in the problem. @ 30325; ; DMAP - Reserved Names If a parameter or datablock on a type statement uses a name that has "cond" as a prefix, for example, condnosq, then the results using these parameters or datablocks in subsequent operations are unpredictable. For example, in the following dmap operation, type parm,,rs,n,condnosq $ condnosq = 1.0 $ message //condnosq $ does not produce a value of 1.0 for the condnosq parameter. The reason is that "cond" is a dmap exec statement for branching operations. Avoidance: Do not define a datablock, parameter, or label using "cond" as a prefix for the name. @ 30336; r Message

; DMAP - Building a New Solution Sequence - UIM 1179 (RESDBS), Poo

An attempt to LINK a "called SUBDMAP" (one with input and output arguments on the SUBDMAP line) will result in the following error message, which does not explain the problem: *** USER INFORMATION MESSAGE 1179 (RESDBS) A DB-DIRECTORY IMPROPERLY CLOSED IN THE PREVIOUS RUN IS BEING RESTORED BY THIS RUN. MAIN: "Access violation" (C0000005) exception encountered. MAIN: For reference, A(MAINAL) = 408540, A(/SYSTEM/) = 14ef640 MAIN: Exception occurred at address 89050D. MAIN: Program attempted to access data at location 19970738. MAIN: Context Flags 0001001F Avoidance: Do not LINK a "called SUBDMAP" only MAIN SUBDMAPs with no arguments in the SUBDMAP line are to be linked. @ 30367; 70.5; Dynamic Analysis - TEMP(LOAD) - Incorrect Stresses It is an undocumented limitation that thermal loads cause incorrect stresses in dynamic analysis. This limitation will be documented under TEMP(LOAD) case control entry in the Version 70.5 Quick Reference Guide.

@ 30414;

70.7; Miscellaneous - SET - SFM 3008

Long SET case control entries may cause NASTRAN fail with the message SYSTEM FATAL MESSAGE 3008. The following SET entry resulted in this fatal: SET 1 = 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21 Avoidance: The above SET entry may be broken up as follows: SET 1 = 1,2,3,4,5,6,7,8,9, 10,11,12,13,14,15,16,17,18,19, 20,21 @ 30419; 70.5; Documentation - QRG - Description of Large Field Format Continua tion Entries In the discussion in the QRG on the "Format of Bulk Data Entries" there is no mention made of how to continue a large field format Bulk Data Entry. There is an example of a large field format entry, which shows the correct use of the "*" in column 1 that marks field 1B as a continuation mnemonic of the parent entry begun in field 1A. However, in the "Continuations" discussion, there is no mention made of how to continue with child entries that are also of the large field format. These types of continuations require an "*" in column 1 of field 1B along with the continuation mnemonic of the first half of the entry. Additionally, it is stated that continuation fields can be auto-generated (the recommended practice) within MSC/NASTRAN, provided that those continuations immediately follow the parent and fields 1 and 10 of the continuation lines are left blank. Under this usage, blank fields mean small field format and errors can result from the first half of the entry being in large field format and the second half being in small field format. In order to use auto-generated continuation fields on large field format entries, field 6 of the first line, as indicated in the large field format description must be blank and field 1B of the second line must contain an "*" in column 1. All additional continuations using large field format data must also conform to the above or they will be incorrectly classified as small field data and processed as such.

@ 30420;

70.5; Nonlinear - EIGB - SFM 6707, SFM 3007, Poor Message In a nonlinear buckling analysis, if the method on the EIGB entry is SINV or INV and no eigenvalues are found in the range, then the following misleading messages will be printed in the .F06: *** SYSTEM WARNING MESSAGE 3030 (OFP) THE OFP MODULE CANNOT PRINT THE DATABLOCK LAMA USER ACTION: USE THE MATPRN OR TABPT MODULE. PROGRAMMER INFORMATION: OFP LOC CODE = 40 FILE = 102 *** SYSTEM FATAL MESSAGE 6707 (PARAML) THE PARAML MODULE WITH P1=DTI DETECTS A STRING RECORD. *** SYSTEM FATAL MESSAGE 3007 (PARAML) ILLEGAL INPUT TO SUBROUTINE PARAML or

0*** SYSTEM FATAL MESSAGE 4540, INPUT ESTNL DATA BLOCK IS NOT NONLINEAR FORMAT. *** SYSTEM FATAL MESSAGE 3007 (NLITER) ILLEGAL INPUT TO SUBROUTINE NLINIT Avoidance: Choose a more appropriate range on the EIGB entry. @ 30425; N ; Miscellaneous - (PUNCH) - Segmentation Fault in OFP Source OFPPU Large punch jobs on the HP and SUN may fail with the following message: Segmentation fault in OFP source OFPPUN SIGSEGV in ofppun.f and in the log file: 11:21:41 Beginning Analysis Process Id = 23091 Process 23091 received signal 8 Process traceback ... ( 0) 0x0009700c sysrcv + 0x320 [/apps/msc70/msc70/hpux/analysis.pa2_0] ( 1) 0x00f83310 _sigreturn [/apps/msc70/msc70/hpux/analysis.pa2_0] ( 2) 0x00f48644 __convert_and_assign_data + 0x1464 [/apps/msc70/msc70/hpux/analysis.pa2_0] ( 3) 0x00f4a368 __cnv_fc + 0xb0 [/apps/msc70/msc70/hpux/analysis.pa2_0] ( 4) 0x00f335bc __a_c_t + 0xed4 [/apps/msc70/msc70/hpux/analysis.pa2_0] ( 5) 0x00f33ecc __a_c_t + 0x17e4 [/apps/msc70/msc70/hpux/analysis.pa2_0] ( 6) 0x00f342d8 __io_tn1 + 0x118 [/apps/msc70/msc70/hpux/analysis.pa2_0] ( 7) 0x00f3cb74 __F90_DO_IO_ITEM + 0xfc [/apps/msc70/msc70/hpux/analysis.pa2_0] ( 8) 0x00f2fb58 FTN_DO_R4IOA + 0x30 [/apps/msc70/msc70/hpux/analysis.pa2_0] ( 9) 0x00418098 ofppun_ + 0x4a8 [/apps/msc70/msc70/hpux/analysis.pa2_0] (10) 0x0016c4b4 ofp_ + 0x2f38 [/apps/msc70/msc70/hpux/analysis.pa2_0] (11) 0x00093b88 xsem14_ + 0x294 [/apps/msc70/msc70/hpux/analysis.pa2_0] (12) 0x0008ec44 mainal_ + 0x99c [/apps/msc70/msc70/hpux/analysis.pa2_0] Avoidance: Other than not requesting punch output there is no avoidance. @ 30453; 2001; Documentation - ALTER - String Based Alter, UFM 802 If a string-based DMAP alter uses the ! in the expression (find occurrence at the beginning of a line), it is possible that MSC/NASTRAN will fail with USER FATAL MESSAGE 802 stating that no match was found, even if there should be a match. Avoidance: If you encounter UFM 802, check to see that the desired line is not preceded by a comment in the DMAP. The search logic using the "!" assumes the line begins on the preceding comment. In order to avoid this, remove the ! from the text string (be careful to verify that the DMAP alter is inserted at the correct location, as this will then search for

any occurrence of the text on a line, rather than only searching for lines which begin with the text). If necessary a more unique search string will have to be selected. @ 30459; 70.5; Estimate - Leading 0's in Integers - Poor Results If an integer has leading 0's (e.g., 05 instead of 5) on a NASTRAN bulk data entry, MSC/ESTIMATE may not interpret the data correctly. Avoidance: Remove any leading 0's from the bulk data of a file to be processed by ESTIMATE. @ 30462; 70.5; Estimate - Excessive Errors - Run Aborted If 'memory=estimate' or 'buffsize=estimate' is specified on the nastran command line or in an RC file, messages generated by the ESTIMATE utility may overflow a buffer within the nastran command. When this overflow occurs, the following message will appear: *** SYSTEM WARNING MESSAGE (nastran.estimate_job_requirements) a.estimate=<install-dir>/msc705/<arch>/estimate (program default) ESTIMATE executable filename. This keyword may only be specified on the command line or in the command initialization file (see <0.ini>). It appears that ESTIMATE has failed. Ensure that the program is properly installed. Assuming default value for buffsize. Avoidance: Either of the following avoidances will work: 1. Fix the errors in the input file and rerun with estimate. 2. Run estimate by itself. Use the suggested parameters to run MSC/NASTRAN. @ 30468; 70.5; Miscellaneous - VECPLOT - SFM 3062 If a very large memory is requested in a run, then the job may fail in vecplot with the following error messages: 0*** SYSTEM FATAL MESSAGE 3062 MODULE = VECPLOT FAILURE OCCURRED IN UTILITY GETGPT CALLED IN SUBROUTINE VECPL1 . PREXYZ FLAG ARGUMENT WAS SET WITH AN INCORRECT VALUE TO GET THE REQUESTED DATA RETRIEVAL. In this case, a memory of 400mw was requested. Avoidance: Reduce the amount of memory request. @ 30496; ; Restarts - Delete Entry - Various User Fatal Messages When a restart delete entry, using the small field format, is incorrectly shifted a column to the left, the beginning and end values are misread without a fatal message. For instance the following delete bulk data entry: / 65266 65267

is read as: / @ 30507; 5266 5267

Avoidance: Use free field, comma separated input for this entry. 70.5; Optimization/Design Sensitivity - SDR2 - SFM 3062 A shape optimization task may fail with a system fail message, 0*** SYSTEM FATAL MESSAGE 3062 MODULE = SDR2 FAILURE OCCURRED IN UTILITY MAPXTI CALLED IN SUBROUTINE SHLNRM . REQUESTED EXTERNAL ID(S) NOT FOUND. The failure may occur when the following conditions are met: 1. A shape optimization job; 2. Shell normal is invoked; 3. The identification id for grid points is larger than the total number of grids in the model. Say the model has 10 grids but the ten grids are identified by 100001 to 100010. The possible avodance: 1. Always start grid ids from 1; 2. Set SNORM=0.0. Notice that if shell normal can not be ignored, the avoidance # 1 should be used. @ 30515; 70.5; Restarts - EPOINT - Inefficient Restart If a restart is made from a solution 103 normal models run and epoints are added, the modes will be recalculated even though they are unaffected by the epoints. Avoidance: Use the following alter, it removes the check on DYNAMIC(s) and the restart skips the READ or REIGL module. THIS AVOIDANCE PRODUCES INCORRECT RESULTS! compile ifpl alter 158,158 compile phase0 alter 590,590 endalter @ 30522; 70.5; Optimization/Design Sensitivity - PCOMP - SFM 5628 (DSAD12) In design optimization, if there are composite responses (DRESP1 = CSTRESS, CSTRAIN or CFAILURE) and force responses (DRESP1=FORCE), a USER FATAL MESSAGE 5628 as shown below will be encountered unless the highest element ID has a composite response. 0*** USER FATAL MESSAGE 5628 (DSAD12) DRESP1 ID = 5363003 IS FOR RTYPE = FORCE AND ELEMENT ID = 5682. THE RESPONSE COULD NOT BE FOUND, EITHER BECAUSE THE ELEMENT DOES NOT EXIST OR BECAUSE THE RESPONSE TYPE IS NOT AVAILABLE FOR THE ASSOCIATED ELEMENT

TYPE. 0*** USER FATAL MESSAGE 5628 (DSAD12) DRESP1 ID = 5363004 IS FOR RTYPE = FORCE AND ELEMENT ID = 5682. THE RESPONSE COULD NOT BE FOUND, EITHER BECAUSE THE ELEMENT DOES NOT EXIST OR BECAUSE THE RESPONSE TYPE IS NOT AVAILABLE FOR THE ASSOCIATED ELEMENT TYPE. 0FATAL ERROR AVOIDANCE: It is not sufficient to make sure that the highest constrained element id is a composite element since the design sensitivity analysis spawns element ID's that are out of the user's control and will produce a gread error in DSAL. Instead, the only known avoidance is to remove the TYPE=FORCE dresp1 entries. They could be replaced with equivalent STRAIN or STRESS dresp1 entries. @ 30542;

70.5; Response Spectra Analysis - SUBCASE - Missing Output When using Response Spectrum ananlysis in SOL 103, if you have multiple SUBCASEs (solutions) and there are diffenrent output requests in various SUBCASEs, only the requests which match the first SUBCASE will be honored. Avoidance: Insert the following DMAP alter: compile postreig alter 'sdr2 ','' SDR2 CASEDR,CSTMS,MPTS,DIT,EQEXINS,,,,BGPDTS,OLB,PHQG, PHIGH,EST,XYCDBDR,,,,GPSNTS,,,/ ,IQG1,IPHIG1,IES1,IEF1,/ 'mmREIG' $ endalter

@ 30550;

70.5; Documentation - QRG, INIT,SCRATCH - Incorrect Default The default value for the SCRATCH dbset is stated as 175,000 gino blocks in the Quick Reference Guide. The correct default value is 250,000 gino blocks.

@ 30560;

70.5; Restarts - scr=mini, SUPORT command - UWM 4410, Missing output If SCR=MINI is specified on the command line and the SUPORT command is specified in the Case Control, then subsequent data recovery restarts will terminate with UWM 4410 and no data recovery output. Avoidance: Insert the following DMAP alter: compile sestatic alter 'call setq' PARAML CASES//'DTI'/1/255//S,N,SUPORT $ endalter

@ 30563;

70; Transient Analysis - OTIME, op2, f06 - wrong answers In Version 69.1, the OTIME command is ignored for element-type output (e.g., forces, stresses). Furthermore, the time labels are in the correct

sequencial order, but the label values are incorrect. The OTIME command functions properly for grid-type output (e.g., disp, velo, acc). This problem has been corrected in Version 70. Avoidance: Either remove the OTIME command or upgrade to Version 70. @ 30575; GL; Optimization/Design Sensitivity - Limit on the Number of Design Variables There is a limit imposed by the optimization algorithm used for design optimization that the number of design variables (NDV). squared must be less than the number of machine precision words available. For example, if a job is run with 40 million single precision words, the number of design variables must be less than the sqrt(40,000,000/2) = 4472 on a short word machine and 6324 on a long word machine. @ 30578; ; Nonlinear Analysis - PARAM,LGDISP, PLOADX1, LOAD - Wrong Answers If PLOADX1 (pressure loads for axisymmetric elements) is referenced by a bulk data LOAD record, follower forces (in nonlinear solution sequences) do not work. Pressure loads are computed based upon the initial geometry. Avoidance: An avoidance is to write PLOADX1 records for every load magnitude needed, and not use the LOAD records to combine loads. @ 30585; N 70.5; Installation/Security - Unable to Obtain Licenses for MSC/NASTRA There is an error in the FLEXlm libraries provided by Globertrotter that result in the libraries ALWAYS examining the LM_LICENSE_FILE environment variable even though a valid license file or server has been explicitly named in the "authorize" keyword. As a result of this, the following limitation is set: If the LM_LICENSE_FILE environment variable is set, it shall point to a valid license file. This error is most easily seen by setting the "authinfo" value to 9 and examining the LOG file output. If a license file not named by the "authorize" keyword is present, it was obtained from the LM_LICENSE_FILE environment variable. Avoidance: Delete the LM_LICENSE_FILE environment variable if it does not point to a valid FLEXlm license file. @ 30590; 70.5; Nonlinear Static Analysis - Superelements, Restarts - SFM 4540 When downstream superelements are processed in a SOL 106 coldstart by using SEALL=n, where n is the superelement number, NASTRAN incorrectly issues the following fatal message: SYSTEM FATAL MESSAGE 4540, INPUT ESTNL DATA BLOCK IS NOT NONLINEAR FORMAT. Avoidance: If you are using partitioned superelements and you are going to restart to process the residual, ignore the fatal message and restart anyway. The fatal message is benign and restarts are possible. @ 30617; 70.5; Documentation - QRG, Force(corner), strain(corner) - Incorrect d

efault The Quick Reference Guide incorrectly lists cubic as being the same as the corner output for the force and strain output. The correct default corner option for the force and strain corner output is BILIN, similar to the stress output. @ 30631; d ; Documentation - Reference Manual, GPSTRESS - Strain Not Supporte In the second paragraph of Section 10.1 of the MSC/NASTRAN Reference Manual it states "The calculation of strains may be sustituted for stresses using the STRAIN command in place of the STRESS command in the Case Control Section" inside the same paragraph that describes the GPSTRESS features. This may lead users to incorrectly conclude that grid point strain can be obtained by substituting the STRESS Case Control Command with the STRAIN command. Grid point strain output is not supported. @ 30638; 70.5; Grid Point Force Balance - GPFORCE, PLOAD - Wrong Answers Grid point force balance output of element force contributions are not correct when pressure loads are applied. For example, PLOAD1 on BAR, BEAM and BEND elements will produce different element forces with the ELFORCE data recovery than from that printed in the grid point force balance output. The same is true of the plate elements under pressure loads. Avoidance: There is no avoidance, this is currently a limitation of the GPFDR module. The GPFDR module has as input the discrete load vector to perform the data recovery operations. The discrete load vector does not contain any of the distributed load effects and will calculate element force contributions based on the discrete loads not the distributed loads. The limitation has not been removed but is now documented under the GPFORCE case control command. @ 30639; 70.5; Optimization/Design Sensitivity - PLOAD1/BEAM - Wrong Answers A design optimization job may report incorrect element response value (Stress/Strain/Force) for CBAR, CBEAM and CBEND elements subjected to pressure loads because SOL200 does not account for distributed forces of the PLOAD1 entry while recovering those element response quantities. Avoidance: There is no avoidance. @ 30645; 70.5; DMAP - User Supplied subdmaps may Discard datablocks on type Sta tements If a user subdmap is of the form: subdmap mine in1/out1 type db sils mod1 sils/out1 etc. etc. . . . return end $

and sils is a datablock in the nddl, which means it is stored in the database, then assuming qualifiers are set correctly, module mod1 will find sils and use it appropriately. However, at the end of the execution of the subdmap the LTU logic will automatically delete the datablock. This will not be discovered until the datablock is needed by a downstream subdmap, when it will no longer be in the database. Consequences include wrong answers or mysterious fatal error crashes such as "SILS NOT AVAILABLE" for modules that allow purged inputs. These errors occur even though SILS was in the database in prior subdmaps. Avoidance: Do not write the type statement, but instead input the datablock in a calling argument: subdmap mine in1,sils,out1 $ etc. @ 30703; GL; Postprocessing - XDB Limitation - Maximum Number of Grids There is a limit on the number of GRID points which can be processed by the DBC module for the ".xdb" file. The limit is dependent on many factors, one of the most significant factors is the solution sequence. For dynamic and nonlinear solution sequences the limit is 214748 GRID points. If more GRIDs than this exist in the model, some of the desired data may not be processed and placed in the ".xdb" file. Although there is still a limit on the total amount of data DBC can process, Version 70.5 contains changes that take into account the actual number of timestep or load steps actually solved. Thus, the program can more accurately predict the amount of data that it must process. The maximum number of grids is now determined dynamically and this change effectively greatly increases the maximum number of grids allowed in most dynamic and nonlinear models. Avoidance: If your model has more than 214748 GRID points in divide it into superelements, each with less than 214748 GRID it. If data recovery is only required for a few grid points, numbering these points first and referencing a SET with these also avoid the error. @ 30704; 70.5; Nonlinear Heat Transfer - CHBDYG - UFM 2157 The backside on an ambient plate element cannot be used for thermal radiation. If attempted the follow message results: USER FATAL MESSAGE 2157, MATERIAL ID = 0 FOR ELEMENT ID = DOES NOT APPEAR ON ANY RADM RADIATION MATERIAL DATA CARD FATAL ERROR Avoidance: Add a front face RADM entry. 1 it, then points in then points will

@ 30712;

70.5; Optimization/Design Sensitivity - PARAM,INREL, ADD - UFM 5423 The USER FATAL MESSAGE below may occur when performing an inertial relief sensitivity analysis with multiple subcases and one or more subcases does not have any retained responses. The message is: 0*** USER FATAL MESSAGE 5423---ATTEMPT TO ADD INCOMPATIBLE MATRICES,

(SADD5) TRAILER FOR PL TRAILER FOR PLR6Y TRAILER FOR PLIPA FATAL ERROR IS IS IS COLS ROWS FORM TYPE NZWD DENS 6 57 2 2 50 3596 12 57 2 2 64 5614 12 57 2 2 34 2281

Avoidance: Insert the following DMAP alter: compile desopt alter 'call descon' solapp='stnmbk' $ call dbstore caseds,casdsn,,,//0/0/' '/0 $ compile inersen alter 'plbase','qrbase' call dbfetch /caseds,casdsn,,,/0/0/0/0/0 $ type parm,,i,n,icase0=1,icase1=1 $ type parm,,logi,n,loopflg=true,sync=true $ matgen ,/one/1/1 $ 1x1 identity matrix matgen ,/zero/7/1/1 $ 1x1 null matrix file vc=append $ do while ( icase0>0 ) $ paraml caseds//'dti'/s,n,icase0/1//s,n,sid0 $ if ( icase0>0 ) then $ if ( sync and icase1>0 ) then $ paraml casdsn//'dti'/icase1/1//s,n,sid1 $ if ( icase1>0 ) icase1=icase1+1 $ endif $ if ( sid0=sid1 ) then $ append zero,/vc/2 $ sync=true $ else append one,/vc/2 $ sync=false $ endif $ icase0=icase0+1 $ endif $ enddo $ trnsp vc/vr $ append vc,/vc6///6 $ trnsp vc6/vr6 $ $ DBVIEW CRXbase0 = CRX WHERE (DELTA=FALSE) $ partn crxbase0,vr,vr6/crxbase,,,/1 $ $ DBVIEW QRbase0 = QR WHERE (DELTA=FALSE) $ partn qrbase0,vr6,/qrbase,,,/1 $ $ DBVIEW PLBASE0 = PL WHERE (DELTA=FALSE) $ GET FOR LAST 6 COLUMNS partn plbase0,vr6,/plbase,,,/1 $ endalter @ 30721; Order 70.5; Optimization/Design Sensitivity - MODTRK - Eigenvectors in Wrong Mode tracking in design optimization seeks to place modes back into their original order if any mode switching has occurred as a result of changes made to the design. The eigenvalue summary, however, will not be correctly correlated with the eigenvector output if modes have switched,

and the parameter, NASPRT is set to a positive integer to request data recovery for the intermediate design cycles. The values are correct but the order does not agree with the eigenvalue summary table. Avoidance: The following is a DMAP alter for Version 70 that correctly reorders the eigenvector summary output as a result of mode tracking. This alter has been built into Version 70.5 of MSC/NASTRAN along with some additional module MODTRK enhancements. $ $ v70 avoidance for csr 30721: $ compile fea souin=mscsou list noref alter 267,268 if (feaopt='analysis') then $ message //' dmap information message 9051 (fea)' $ message //' normal modes analysis initiated for analysis ='/ analysis/' and design cycle number ='/descycle $ endif $ $ alter 284 before computing modes, determine if mode tracking is on: PARAML CASEM//'DTI'/1/275//S,N,MODETRAK $ alter 285,285 IF ( NOT NOPH2 ) CALL MODEFSRS, MR,USET,DM,CASEM,DYNAMIC,MAA,KAA,GPLS,SILS,EED, EQEXINS,VAFS,PG,GM,GOT,KFS,YS,LLL,KLL/ PHA,LAMA,PHFA,LAMAF,PHSA,LAMAS,PHSA0,LAMAS0/ ' '/NORSET/NOLSET/ASING/NOASET/FS/NOFASET/ NOSASET/METH/METHF/NASOUT/NOQSET/NSUBM/modetrak $ alter 286,286 $ remove misplaced paraml $ $ alter 297 if modes cannot be tracked output latest lama table, even though unordered: message //' modes could not be tracked for design cycle '/ descycle $ message //' though the eigenvalue summary table for the '/ 'latest design follows, be aware that the mode '/ 'order corresponds to the untracked state.' $ ofp lamas,/ $

alter 338 $ for mode tracking, output reordered lama table on every design cycle: if ( modetrak>0 ) ofp lama,/ $ alter 405,405 CALL MODERS ,,USET,DM,CASEM,DYNAMIC,KDAA,KAA, GPLS,SILS,EED,EQEXINS,,/ BPHA,BLAMA/ NORSET/NOLSET/'BUCKLING'/ASING/ FALSE/FALSE/NASOUT/NOQSET/FALSE/NSUBB/modetrak $ endalter

compile modefsrs souin=mscsou list noref alter 1,1 SUBDMAP MODEFSRS MR,USET,DM,CASES,DYNAMICS,MAA,KAA,GPLS,SILS,EED, EQEXINS,VAFS,PG,GM,GOT,KFS,YS,LLL,KLL/ PHA,LAMA,PHFA,LAMAF,PHSA,LAMAS,PHSA0,LAMAS0/ APP/NORSET/NOLSET/ASING/NOASET/FS/ NOFASET/NOSASET/METH/METHF/NASOUT/NOQSET/NSUB/modetrak $ alter 6 type parm,nddl,i,n, modetrak $ alter 37,37 CALL MODERS MR,USET,DM,CASES,DYNAMICS,MSAA,KSAA,GPLS,SILS,EED, EQEXINS,,VAFS/ PHSA0,LAMAS0/ NOSRSET/NOLSET/'MODES'/ASING/FALSE/SECND/ NASOUT/NOQSET/FS/NSUB/modetrak $ alter 59,59 CALL MODERS MR,USET,DM,CASES,DYNAMICS,MFAA,KFAA,GPLS,SILS,EED, EQEXINS,,VAFX/ PHFA,LAMAF/ NOFRSET/NOLSET/'MODES'/ASING/FALSE/TRUE/ NASOUT/NOQSET/FS/NSUB/modetrak $ endalter compile moders souin=mscsou list noref alter 1,1 SUBDMAP MODERS MR,USET,DM,CASES,DYNAMICS,MMAA,MKAA,GPLS,SILS,EED, EQEXINS,VACOMPR,VAFS/ PHA,LAMA/ NORSET/NOLSET/READAPP/ASING/ CYCLIC/SECND/NASOUT/NOQSET/FS/NSUB/modetrak $ alter 7 type parm,nddl,i,n, modetrak $ $ $ alter 126,126 output eigenvalues only if mode tracking is not active. otherwise, reserve output until after modes have been sorted. if ( nasout and modetrak=0 ) ofp lama,oeigs/ $ endalter @ 30738; 2001; Frequency Response - QUAD4, sort2, corner, - SFM 3001, Possible wrong answers In a frequency response analysis, if SORT2 output is requested, then the following error message may be issued in data recovery after printing the element force output. *** SYSTEM FATAL MESSAGE 3001 (DPD ) THE OUTPUT DATA BLOCK IN POSITION 11 DOES NOT EXIST.

Furthermore, on the SGI (and possibly other machines), a sort2, corner output request for quad4 elements may produce incorrect results that are evident by the printing of very large element forces (e.g. E+20) or "Inf". Avoidance: Use a sort1 output request with frequency response. If corner output is not being requested, and if random response is not being performed, then the following alter may avoid the problem in Version 70. compile sedrcvr souin=mscsou nolist noref $ V70 alter 'dpd'(1,-2),'dpd'(1,7) $ Delete the offending DPD module and associated modules. endalter @ 30740; e AMPC 70.5; Aeroelastic Analysis - AMP, CAERO1, CAERO2 - UFM 3005, Subroutin The message: USER FATAL MESSAGE 3005 (AMP ) ATTEMPT TO OPERATE ON THE SINGULAR MATRIX SCRATCH IN SUBROUTINE AMPC FATAL ERROR will occur in the AMP module if multiple group ids are used in defining the aerodynamic model and the model contains both lifting surface entries (CAERO1) and body entries (CAER02). Avoidance: Place all the aerodynamic panels and bodies in the same interference group. @ 30766; 70.5; MNE - Missing Collection on V70 MNE If you choose an installation option that does not install the PRIMARY (V70) documentation set onto the hard disk, you must do the following after executing the installation script: 1. Change the current directory to <install_dir>/mne70, where <install_dir> is the directory in which you installed the MNE: cd <install_dir>/mne70 2. Remove the existing erroneous symbolic link to v70qrg: rm v70qrg 3. Add the correct link for v70qrg: ln -s /<CDROM>/doc/mne70/v70qrg where <CDROM> is the CD-ROM mount point. @ 30775; 70.5; Loads - PLOADi - Incorrect Follower Forces There is an error in the follower force stiffness matrix (see PARAM FOLLOWK) due to PLOAD, PLOAD2 or PLOAD4, if the global displacement system is not in the basic coordinate system. This error occurs if the first grid point listed on the PLOADi has a non-basic coordinate system reference in field seven of the corresponding GRID entry. @ 30776; GL; Element Strain Energy - ESE - Incorrect for Buckling The element strain energy output for modes found in SOL 103 using STATSUB is incorrect. It includes only the linear portion of the ESE; the differential stiffness contribution is ignored.

Avoidance: The following alter will calculate the strain energy contribution of the differential stiffness in addition to the linear portion: $ the following dmap is for V70 SOL 103 with STATSUB $ after the solution completes, there are 2 ESE outputs $ 1st: unstiffened ESE due to KELM $ 2nd: stiffened ESE due to KDELM $ combine these to get the total modal ESE $ this has not been tested with superelements. COMPILE SEKDR,NOLIST,NOREF ALTER 'EMG.*KDELM' $ CALL DBSTORE KDELM,KDDICT,,,//-99/SEID/'DBALL'/0 $ COMPILE SEDRCVR,NOLIST,NOREF ALTER 'OFP.*ONRGY1' $ after unstiffened gpfdr MESSAGE //'DIFFERENTIAL ESE CONTRIBUTION FOLLOWS:' $ TYPE PARM,,I,N,OKKDGET $ CALL DBFETCH /KDELM,KDDICT,,,/-99/SEID/0/0/S,OKKDGET $ $MATPRN KELMX,KDICTX,KDELM,KDDICT// $ GPFDR CASEDR,UG,KDELM,KDDICT,ECTS,EQEXINS,GPECT,PG1,QG, BGPDTS,SILS,CSTMS,VELEM,PTELEM/ ONRGY1D,OGPFB1D/APP2/TINY//CYCFLG $ OFP ONRGY1D//S,N,CARDNO $ @ 30782; ; Superelements/Buckling - Wrong Answers. Superelements give wrong answers for buckling. Avoidance: Do not use superelements.

@ 30797;

70.5; Utilities - ESTIMATE - Poor Diagnostics The ESTIMATE program is currently limited to estimating resources for models composed of GRID and connectivity entries. If no GRIDS are present, ESTIMATE will return a USER FATAL MESSAGE or may indicate NaNQ for memory or disk estimates. If no recognized connectivity bulk data entries are present, ESTIMATE will return DOF=-2**(32-1). Avoidance: Do not use the ESTIMATE program for models as described above. Do not use either the "memory=estimate" or "buffsize=estimate" keywords on a nastran command for models as described above.

@ 30805; ; Nonlinear Analysis - Multiple Subcases - Error in Bisection logi c, Infinite Loop In a nonlinear analysis, in some instances, the load factor tolerences may be in error. When this occurs the value of the load factor exceeds 1.0 and the solution may enter an infinite loop or produce some other error later on. The problem is identified when one of the load factors > 1.0 Avoidance: On the NLPARM entry increase the ninc field by 1 and restart into the offending subcase. This will change the tolerance

@ 30810;

2001; Optimization/Design Sensitivity - SEP1A - Infinite CPU Loop When the analytic boundary shape option is used for a shape optimization task and a local coordinate system is defined for the auxiliary boundary model, the job may fall in an infinite loop in the SEP1A module. Avoidance: Do not define local coordinate systems in the auxiliary boundary model.

@ 30852;

GL; Postprocessing - PARAM,POST,-4 - UIM 1207, UIM 1206 If PARAM,POST,-4 is used and FORMATTED output is requested in the OUTPUT2 file, error messages 1207 and 1206 will occur indicating that several datablocks have not been written to the file. An example of these message are: USER INFORMATION MESSAGE 1207 (OUTPN2) THE DATABLOCK KDICT /KDICT DEFINED AS NDDL TYPE KDIC IS NOT SUPPORTED BY NEUTRAL FILE TRANSLATION. THIS DATABLOCK WILL NOT BE TRANSLATED TO NEUTRAL FORMAT 0*** USER INFORMATION MESSAGE 1206 (OUTPN2) THE DATABLOCK OPHIG /OPHIG IN AN OUTPUT2 REQUEST IS NOT DEFINED IN THE MSC/NASTRAN DATA DEFINITION LANGUAGE (ND This is due to the fact that these datablocks are not described in the NDDL. Avoidance: Use binary OUTPUT2 if PARAM,POST is set to -4.

@ 30903;

GL; Superelements - PARAM,AUTOSPC - Incorrect Constraint If two or more superelements which have no internal constraints applied, meet at a single GRID point, and no elements which are not in those superelements connect to that GRID point, AUTOSPC may incorrectly constrain that point. The reason for this is that the reduced stiffness matrix at a single point for a determinate structure with no constraints is a null matrix. This limitation is unlikely to affect real world models since it is unlikely that the boundary between superelements will be comprised of a single point. Avoidance: 1. Put the DOFs data entries. 2. Put in small not large enough practical method you want as dynamic analysis variables on spcoff bulk springs to ground, large enough to turn off autospc but to change dynamic analysis results. This is the most for superelement analysis.

@ 30947;

70.5; Static Analysis - MATTi - SDR2, biomsg error 7777 For a large job on the hp (hpux), if temperature-dependent material property exists in the model, then the job may fail in SDR2 with following error messages: ***** RECEIVED SIGNAL 8, ABORTING ... BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX (FILEX) =

0 Avoidance: Remove the MATTi entry and define the desired properties at the appropriate temperature on the MATi entries since for a linear analysis, the temperature-dependent lookup is only performed once in the beginning of the run. @ 30959; ; Nonlinear Analysis - CGAP, PGAP - Possible Wrong Answers when U0 is Negative Use of a gap element (tmax>=0.) with a negative value of U0 when GA and GB are NOT coincident may lead to wrong answers. Consider two grids on the x-axis with GA to the left of GB and the CID field blank and U0 specified to a negative value. Then the gap will always try to push GA to the left and GB to the right regardless of whether GA comes first or GB comes first in field 4 on the CGAP entry. Thus, a press fit with the inner radius of the outer disk at GA and the outer radius of the inner disk at GB would always pull in the wrong direction when UO is negative. Avoidance: To make the gap sensitive to the order of GA and GB with U0 negative, the CID entry of field 9 must be specified. @ 30969; 02 70.7; GPFORCE - SET, GPFORCE - Poor Message, SFM 4276, EC 3061, SWM 30 If a GPFORCE request is made for a SET that only contains grid points attached to rigid elements then the deck will fail in the GPFDR module. The exact message encountered may vary. The following messages have been encountered: SYSTEM FATAL MESSAGE 4276, ERROR CODE 3061 or *** USER WARNING MESSAGE 2354 (GPFDR) MODULE IS UNABLE TO CONTINUE AND HAS BEEN TERMINATED DUE TO ERROR MESSAGE PRINTED ABOVE OR BELOW THIS MESSAGE. THIS ERROR OCCURRED IN CODE WHERE THE VARIABLE -NERROR- WAS SET = 33 *** SYSTEM WARNING MESSAGE 3002 (GPFDR) EOF ENCOUNTERED WHILE READING DATA BLOCK SCRATCH (FILE 303) IN SUBROUTINE GPFDFA Avoidance: GPFDR does not support rigid elements, request MPCFORCE output instead. @ 30976; 70.5.1; Online Documentation - Installation The UNIX installation aborts if you choose an installation option that installs the V69.X documentation set on the hard disk. The old V69 documentation is duplicated on this CD only for those customers still running MSC/NASTRAN V69 or V69.2. If you use MSC/NASTRAN V70, we recommend not installing V69.X documentation onto your hard disk, since all the documents applicable to V70 are already contained in the Standard installation. If you occasionally need to refer to V69 documentation, it is best to access it directly from the CD. If you frequently need V69-specific documentation and you prefer to access it from the hard disk instead of the CD, please follow the Installation

Instructions on the pink sheet included with the delivery package. @ 30981; 70.5; Online Encyclopedia - INSTALL.txt on V70 Online Encyclopedia is Incorrect The INSTALL.txt file on the MSC/NASTRAN Version 70 Encyclopedia (MNE) is incorrect. It is the file for MSC/NASTRAN Version 70 not the MNE. Avoidance: Either view the INSTALL.htm file using a web browser or request a printed copy of the file from MSC. @ 30987; GL; Eigenvalue Analysis - EIGRL, GMINTC, GMINTS - Possible Missing R oots, SFM 5407 The Lanczos method of eigenvalue extraction does not work well when interface elements (GMINTC or GMINTS) are present, roots may be missing in the range. This may lead to the message: SYSTEM FATAL MESSAGE 5407 (LNCS1D) INERTIA (STURM SEQUENCE) COUNT DISAGREES WITH THE NUMBER OF MODES ACTUALLY COMPUTED IN A (PARTIAL) INTERVAL Avoidance: If interface elements are present, the SINV method is recommended. @ 30995; 70.5; Acoustic Analysis - GE on MAT10, PARAM,ACSYM,YES - Wrong Answers If GE is specified on the MAT10 entry and PARAM,ACSYM is YES (default) then the answers will be wrong. Avoidance: Insert the following DMAP alter: compile gma alter 'if ( app=\'tranresp\''(2),'' alter 'if ( app=\'tranresp\''(3,-1) if ( app='tranresp' or soltyp='direct' or modacc>=0 ) then $ alter 'k4hh','' if ( nok4dd>0 ) then $ if ( acsyml ) then $ add5 ifdd,,,,/mifdd/-1. $ MERGE isdd,,,mifdd,VAFSE,/idd $ mpyad k4dd,idd,/k4ddx $ SMPYAD PHDH,K4DDx,PHDH,,,/K4HH/3////1 $ else $ SMPYAD PHDH,K4DD,PHDH,,,/K4HH/3////1 $ endif $ endif $ nok4dd>0 endalter @ 31000; GL; Eigenvalue Analysis - READ - Initial Vectors Ignored The READ module has provisions for initial vectors as input for the Lanczos method. These vectors are ignored in the READ module. Avoidance: Use the REIGL module. @ 31003; 70.5; DMAP - MATMOD Option 12 - Wrong Answers

The form of the column vector output in MATMOD option 12 is 7 (labeled "ROW VECTOR") and should be 2 (labeled "ROW RECTANG."). Avoidance: Run MODTRL on the output; for example, MATMOD I1,I2,I3,,,/O1,O2/12 $ MODTRL O2// / /2 $ @ 31027; 70.5; Miscellaneous - PARAM,CHECKOUT and PARAM,PRTPG, SSG1 - SFM 3003 If PARAM,CHECKOUT,YES and PARAM,PRTPG,YES are specified, and thermal loads are also applied, then SYSTEM FATAL MESSAGE 3003 will terminate the run in the SSG1 module in the PHASE0 subDMAP. Avoidance: Insert the following DMAP alter: compile phase0 alter 'ssg1 ','' SSG1 SLT,BGPDTS,CSTMS,SILS,EST,MPTS,ETT,EDT,MJJX,CASESX, DIT,,DEQATN,DEQIND,GPSnts,cstm0,scstm/ PJX,PTELEM,NEWSLT/ LUSETS/1//'ALL '/ALTSHAPE/TABS/seid $ endalter @ 31037; 70.5; Superelements - SEMA, SEBNDRY, SECONCT, SEEXCLD - UFM 4192, UFM 7341, UFM 6787 Two errors relating to checking the input on the SEBNDRY, SECONCT and SEEXCLD entries have been discovered. 1. The SEIDA field is not verified to be restricted to partitioned superelements. The failure to check this may cause UFM 4192 to occur in SEMA. This message may be followed by a SFM 3062 concerning the MAPBTI utility. The true message is UFM 7341. 2. Conflicts in grid usage between the SECONCT and SEEXCLD is not properly verified and should have result in UFM 6787. Avoidance: Follow the rules for the entries delineated in the Quick Reference Guide. @ 31051; 70.5.2; P-elements - GMSURF, GMCURV, GP0 - Segmentation Violation If you use an EQUATION or Aries BSP for geometry evaluation, in other words, if "MSCGRP1", or "MSCABS" is specified for the 'group name' field on the GMCURV or GMSURF entries the job may fail with the following traceback in GP0: Trap encountered: SIGSEGV - Segmentation violation Traceback: Offset Offset Offset Offset Offset 0x000003d4 0x000000e8 0x000000a8 0x00000b80 0x00000680 in in in in in procedure procedure procedure procedure procedure get_edge_info abspnt ngpont_ gp0dx_ gp0dc_

Offset Offset Offset Offset

0x000017c8 0x00000cf8 0x000004b0 0x000009a8

in in in in

procedure procedure procedure procedure

gp0da_ gp0_ xsem02_ mainal

--- End of call chain --Avoidance: There is no avoidance. @ 31054; 2001; Documentation - SEBULK, SECONCT, SEEXCLD, SELABLE, SELOC, SENQSE T - UWM 9013 In the interest of upward compatibility, several of the bulk data entries implemented in Version 69 will only work if PART superelements (BEGIN SUPER) exist. These entries are: SEBULK, SECONCT, SEEXCLD, SELABEL, SELOC, SEMPLN, SENQSET, SETREE. Avoidance: Do not use these entries unless you have defined PART superelements. @ 31067; 70.5.1; Miscellaneous - SFM 4276 EC 7777, signal 10 It is possible, but not probable, that a file run on the HP9000 using the analysis.pa2_0 executable may fail with SFM 4276, EC 777, signal 10. This is due to a kernel error on the PA2 machines. Avoidance: There are several possible avoidances: 1) run using a different BUFFSIZE or memory 2) try using a different SMEM 3) run using the older 'analysis' executable (this will avoid the error, but is slower) 4) change bpool @ 31069; 42 ; Optimization/Design Sensitivity - DVPREL1, DVPREL2, EMG - UFM 20 In an optimization analysis, if a negative FID is specified on a DVPREL1/2 entry when PTYPE is not PBEAM, the job may result in the following error message: USER FATAL MESSAGE 2042 (EMG ) MISSING MATERIAL TABLE ******** FOR ELEMENT IIIIIIII

Avoidance: Always use a positive FID on a DVPREL1/2 entry when PTYPE is not PBEAM. @ 31070; 70.5; Nonlinear Analysis - File indexing error A file indexing error may cause a nonlinear analysis to fail in the NLTRD2 module. Avoidance: Increase the bufferpool or smem. @ 31071;

70.5; DMAP - PARAML DTI option - Wrong Answers

If the PARAML module DTI option is used to extract a character string then the variable contains both the value from the selected word in the record and also the value in the next word; e.g., PARAML TABLE//'DTI'/1/45////S,N,STRING $ the variable STRING will contain the character value from the 45th and 46th word. Avoidance: Add the following statement after the PARAML module: STRING=SUBSTRIN(1,4,STRING) $ @ 31073;

70.6; DMAP - MATMOD OPTION(17) - UFM 6968 MATMOD OPTION(17) is designed to read the PARTN case control entry. When PARTN is present, it sets p4 to 0, properly. When it is not present it sets p4 to -1, also properly, but also outputs *** USER FATAL MESSAGE 6968 (MTMD17) NO GRID OR SCALAR POINTS IN SELECTED SET. PARTITIONING VECTOR N ULL. However, execution continues. Avoidance: Ignore the message, or test for the presence of a PARTN command, and branch over the module if none is present. SSSALTER evds103.f705 demonstrates this avoidance.

@ 31076; ce

70.5; Nonlinear Transient Heat Transfer - PENTA(15) - Failed Convergen If there are temperature dependent material properties in the model and you are running a transient thermal analysis using the PENTA(15) element the run will fail to converge. Avoidance: Use lower order PENTA elements.

@ 31083; 70.7.2; Superelement - SET, THRU - Extra Output In a superelement analysis, if output is requested with a SET command using multiple "thru" ranges, then extra output that is not specified in the range may appear in the output in addition to the requested output. An example of the offending SET command is as follows: SET 100 = 1 THRU 10, 80 THRU 90 DISP = 100 The displacements for grid points that are outside of the range may appear in addition to output for the above grid points. The answers are, however, correct. Avoidance: If the extra output is not desired, then specify the SET explicitly without using the THRU range. The above example can be modified as shown below: SET 100 = 1,2,3,4,5,6,7,8,9,10,

80,81,82,83,84,85,86,87,88,89,90 disp = 100 @ 31096; 70.5; P-Elements - GP4 - BIOMSG: ERROR 1044 OCCURRED IN ROUTINE FILPS2 A P-element analysis with coordinate systems may fail in the module GP4 with the message: BIOMSG: ERROR 1044 HAS OCCURRED IN ROUTINE FILPS2 , FILE INDEX (FILEX) = 0 STATUS = 0 Avoidance: Increase the memory request. @ 31101; ; P-Elements - PLOT, SDR2 - SFE 4276 EC 211 The nastran plotter does not support the plotting of p-element results. If the nastran plotter is requested with p-elements, then the job will fail with the following unclear messages. = BIOMSG: ERROR 0 STATUS = 211 HAS OCCURRED IN ROUTINE BLPK68 , FILE INDEX (FILEX) 0

0*** SYSTEM FATAL ERROR 4276.BLPK68 ERROR CODE 211 PID= 0 0*** USER INFORMATION MESSAGE: TO OBTAIN A NASTRAN DUMP, RESUBMIT JOB WITH DIAG 44 INSERTED IN THE EXECUTIVE CONTROL SECTION. Avoidance: If plotting of p-element results is desired, use MSC/PATRAN. @ 31104; 70.5; Matrix Methods - MPYAD - SFM 4276, EC 7777 It has been observed on a very large problem, if the matrices are very dense, then the job may fail in MPYAD. On some machines, the following message may appear at the bottom of the f06 file. ***** RECEIVED SIGNAL 8, ABORTING ... BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX (FILEX) = 0 STATUS = 0

On other machines, there is no messages in the f06 file, but the following messages may appear in the log file: Trap encountered: SIGTRAP - Trace trap Traceback: Offset Offset Offset Offset Offset Offset 0x00000024 0x000007bc 0x00000708 0x000004f8 0x0000029c 0x000008f4 in in in in in in procedure procedure procedure procedure procedure procedure gino_ mpytim_ mpyad_ dmpyad_ xsem07_ mainal

This is due to an overflow in the timing estimate routine. Avoidance: The only avoidance is to break up the model into

superelements. @ 31105; 4 70.5; Superelements - Duplicate Point ID's on Boundary Lead to UFM 679 When a model with list superelements has part superelements added to it, duplicate grid points at the boundary of the list superelements are not allowed, although they were allowed prior to Version 69. This condition leads to following error message: USER FATAL MESSAGE 6794 (SEPSEB) DUPLICATE BOUNDARY POINTS DETECTED IN UPSTREAM SUPERELEMENT xxx WHILE SEARCHING 0 AT GRID yy. LIST FOLLOWS: Avoidance: Remove the part superelements, or move one of the offending grid points. @ 31114; 70.5; Database - DBUNLOAD - Signal 11 Writing LLL If the NEUTRAL option is used in DBUNLOAD, the run will fail. The f06 will likely contain no message, the log file will contain a message of the form: ***** RECEIVED SIGNAL 11, ABORTING ... BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX (FILEX) 0 STATUS = 0

Avoidance: Use the BINARY DBUNLOAD. @ 31118; 70.0.5; GPFORCE - GPFDR, TEMP(LOAD) - SFM 6218 If a model has thermal loadings and GPFORCE is requested, the run may terminate with SFM 6218 in the GPFDR module. Avoidance: Remove the request for GPFORCE or the thermal loads. Defining SETs (which omit the GRID points connected to the element mentioned in the message) for GPFORCE may avoid the problem. @ 31139; 70.5; DMAP - SAMEAS Keyword in NDDL - Poor Message and Wrong Answers. If SAMEAS,recnam is specified in the NDDL and RECORD=recnam is specified below SAMEAS,recnam then the following message is issued but the run is not terminated. *** SAMEAS CTETRA NOT FOUND

RECORD=recnam must appear above SAMEAS,recnam. Avoidance: There is no avoidance. @ 31141; ers ; Eigenvalue Analysis - EIGRL, very large v2 - Possible Wrong Answ If the specified V2 value on the eigrl entry is substantially larger than the last mode found and the absolute value of V2 is large (e.g., 99999.),

then the eigenvectors may be incorrect. This can lead to incorrect responses in a subsequent modal response analysis. This error is most likely to occur if there are massless degrees of freedom and/or large mass in the model and the above conditions exist. If the model is stable, a possible symptom of this error is large rotational displacements in the eigenvectors. This error does not occur with the new Lanczos method which is the default method in Version 70.5. Avoidance: There are two possible avoidances: 1. In v70.5, use the new Lanczos (the default method). 2. Based on the last eigenvalue calculated and the sturm sequence output, reduce v2 on the eigrl entry accordingly. @ 31145; 70.5; Nonlinear Analysis - Restarts into Dynamic Analysis for Base Exc itation A method to perform automatic restarts from static nonlinear analysis (SOL 106) to linear dynamic analysis (SOLs 103-112) has been avaialble for several versions. The resart runs have limitations on the type of data that can be changed. A Technical Application Note (TAN) describing this capability can be obtained from your MSC Technical Represenative. @ 31150; 70.7; Optimization/Design Sensitivity - PARAM,OPTEXIT - Incorrect Term ination Message When optexit is set to -4 (output design sensitivity information into an external file) for a design optimization input file, the task terminates with an incorrect message: RUN TERMINATED DUE TO MAXIMUM NUMBER OF DESIGN CYCLES = IIIII.

The correct message should be RUN TERMINATED DUE TO PARAMETER OPTEXIT = -4. @ 31154; 70.5; Superelements - Duplicate Element ID's not Detected It is a rule of superelement usage that element ids must be unique across all element types, including structural, rigid, and plot elements. When part and grid list superelements are in one model, duplicated element ids in the list superelements are not detected. The results are difficult to predict. Some models have computed correct results, but it is possible that others would either give wrong answers or take confusing fatal exits. Avoidance: An avoidance is to perform a data check run where all of the part superelements are removed. This will result in a different path through the code which will identify all illegal duplicate element ids. @ 31155; 6174 ; Buckling, Nonlinear Analysis - CBEAM, CQUAD4, ZOFF, OFFSET - UFM Offsets should not be used for the beam and plate elements in a buckling or nonlinear analysis. This is a documented limitation. However, the action taken for the beam and plate element with offsets is inconsistent. In the case of a beam element, the job terminates with UFM 6174; however, in the case of the plate element, the job continues without any messages. @

31189; 70.5.2; Nonlinear Analysis - NLITER, CREEP - SFM 4276, EC 7777 Nonlinear analysis with creep and temperature loads may fail in the NLITER module with the message: 0*** SYSTEM FATAL ERROR 4276.tbk ERROR CODE 7777 PID= 0

The traceback will indicate the routine nmcrd. Other machines may produce NaN's in the output. Avoidance: The only avoidance is to use Version 70.5.2 where the error has been fixed. @ 31209; GL; Eigenvalue Analysis - READ, REIGL - SFM 5401 If f1 is set on an eigrl entry to avoid computing rigid body modes, and f1 is much closer to the origin than the first flexible mode's frequency, than SYSTEM FATAL MESSAGE 5401 may be encountered. Avoidance: An avoidance is to first run with f1 blank. It should be possible to compute the flexible modes, as well as the rigid body modyes. On a second run, set f1 10% below the first flexible mode. @ 31218; 70.5; GPFORCE - SUBCOM, TEMP(LOAD) - Missing Forces of SPC GPFORCE output for SUBCOMs with thermal loads will be missing the contribution of the Forces of Single Point Constraint. Standard SUBCASE processing will be correct. @ 31225; 70.5; Elements - PBAR/PBARL, PBEAM/PBEAML - UFM 4382, Wrong Answers A job that includes PBARs and PBARLs with the same PID will issue USER FATAL MESSAGE 4382 and then continue to run. The same behavior is seen with PBEAM/PBEAML. *** USER FATAL MESSAGE 4382 (IFP9 ) PBARL AND PBAR IDENTIFICATION @ 31232; xxx WAS DUPLICATED

; Transient Analysis - LOADSET - UWM 4582 If a static pre-load is used in a modal solution, then LOADSET entry may not be used to define the dynamic loading. This combination will lead to either UWM 4582 show below (if the number of LSEQ entries is not equal to the number of static subcases) or wrong answers (if the number of LSEQ entries is equal to the number of static subcases). USER WARNING STATIC LOADS USER WARNING STATIC LOADS MESSAGE 4582---, MATRIX COLUMN MESSAGE 4582---, MATRIX COLUMN LSEQ CARD SID= 2 ( NCOLS= 1 LSEQ CARD SID= 3 ( NCOLS= 1 100 REFERS TO A NON-EXISTENT ) 100 REFERS TO A NON-EXISTENT )

Avoidance: Do not use the LOADSET command when using static pre-loads. @ 31237; ; Online Encyclopedia - Installation If MSC/NASTRAN Online Encyclopedia (MNE) is installed on a directory path

name that is too long or contains a name with embedded space characters (e.g. "Program Files"), the documentation browser will not execute properly. Avoidance: Use the suggested default directory <Disk>:\msc @ 31249; 70.5; Heat Transfer Analysis - SFM 3002 (TAHT) If a control node is used in a very large heat transfer analysis model, then the following fatal message may be encountered: *** SYSTEM FATAL MESSAGE 3002 (TAHT ) EOF ENCOUNTERED WHILE READING DATA BLOCK NEWSLT SUBROUTINE TAHA Avoidance: Run the job with more memory. @ 31256; 3021 70.5; Optimization/Design Sensitivity - SDR2, SUBCOM - SWM 2092, ERROR If a design optimization design task contains subcoms and one or more of the subcom related subcases has no retained responses, the following fatal message may occur: SYSTEM WARNING MESSAGE 2092 MODULE = SDR2 SDR2 FINDS A SYMMETRY SEQUENCE LENGTH = xx AND AN INSUFFICIENT NUMBER OF VECTORS AVAILABLE = yy WHILE ATTEMPTING TO COMPUTE STRESSES AND FORCES. ALL FURTHER STRESS AND FORCE COMPUTATION TERMINATED. BIOMSG: ERROR 3021 HAS OCCURRED IN ROUTINE GREAD , FILE INDEX (FILEX) = 0 STATUS = 0 In a related error, if a subcom contains subseq coefficients that are 0.0 and the subcase associated with this zero coefficient has no retained responses, the above fatal message may occur. Avoidance: Either remove the subcase that has no retained responses or force the retention of at least one response in the subcase. @ 31262; Zero. 2001; Documentation - PBEAML - Warping Coefficients are Always Set to The documentation for the PBEAML does not state that the equivalent PBEAM created will not contain warping coefficients. Avoidance: Provide the terms using a PBEAM entry. @ 31263; 70.5; Documentation - QRG - PARAM,ACSYM The description for the parameter "acsym" in the Quick Reference Guide states that the parameter only works for coupled fluid-structure analysis in direct frequency response. This parameter also works for coupled fluid-structure analysis in modal frequency response. @ 31266; 70.5; Documentation - QRG, DRESP1 (FILE 101) IN

The Quick Reference Guide states that the PKNL method can be used in the evaluation of a DRESP1 with RTYPE=FLUTTER if the word `PKNL' is placed in the ATTB field of the DRESP1 entry. This is an error, it should be placed in the PTYPE field of the DRESP1 entry. Avoidance: Move PKNL from the ATTB to the PTYPE field. @ 31271; 2001; Documentation - DECOMP, PARAM,AUTOSPC,NO - Unexpected Answers If the automatic grid point singularity check is turned off (param,autospc,no), and there are singularities in the model, then the run may continue and produce large deflections. The next release of the Quick Reference Guide will warn the user that if param,autspc,no is set, then the user should take extra caution analyzing the results in the GRID POINT SINGULARITY TABLE, EPSILON output as well as all warning messages. @ 31278;

70.5; Superelements - DBLOCATE, SERST,SEMI - UFM 4188 A superelement job that contains PARAM,SERST,SEMI and attempts to add the bulk data for a new superelement fails with the message: USER FATAL MESSAGE 4188 QUAD4 ** REFERENCES UNDEFINED GRID POINT **. Avoidance: Apply the following alter: COMPILE IFPS1 nolist noref alter 60 $ putsys(1,109) endalter

@ 31304; swers

; Transient Analysis - PARAM,ADSTAT,YES with TEMP(LOAD) - Wrong An In a transient analysis, if "param,adstat,yes" is used to include a preload effect in the transient analysis data recovery, and the preload is due to thermal loads, then the transient results will be incorrect. Avoidance: Use solution 129.

@ 31308; t

2001; Element Strain Energy - ESE(PUNCH), CELASi - Missing Punch Outpu In MSC/NASTRAN Version 70, if ESE(PUNCH) is requested and there are CELAS1 elements in the model, then there will be no output in the punch file. Avoidance: Insert NASTRAN SYSTEM(210)=1 at the top of the MSC/NASTRAN input file.

@ 31312;

70.5; Aeroelastic Analysis - CAERO2 - Possible Wrong Answers If the aerodynamic coordinate system is not aligned with the basic coordinate system, the geometry of CAERO2 slender bodies IN THE BASIC COORDINATE SYSTEM (BGPA or BGPDT where MODLTYPE=aerostruc) may be incorrect. These data are used in Spline calculations, so that, if the geometry is affected in such a way as to shift the location of the body relative to the spline computational coordinate system, erroneous results

may be obtained. The geometry in the aerodynamic coordinate system is not affected, so the "rigid aerodynamic" data are not affected--only the spline matrix MAY BE affected. It has been observed, however, that certain erroneous transformations in combination with TYPEs of CAERO2 bodies (Z, Y or ZY) can yield identical answers even if the transformation error has moved the body relative to the spline computational coordinate system. This results because the relevant degrees of freedom of the body are transverse to the translation caused by this error. Avoidance: Aerodynamic models with CAERO2 bodies in which the aerodynamic coordinate system is rotated relative to the basic coordinate system will need to be modified and checked. @ 31318; 70.5; Composite Analysis - SEDRCVR - SFM 4276, EC 7777 Large composite models that request large amounts of data recovery may fail on CRAY systems with SYSTEM FATAL MESSAGE 4276, ERROR CODE 7777 in the SEDRCVR module. @ 31322; 70.5; Heat Transfer - RSSCON and PARAM,HEATSTAT - SWM 3001 in SEP1X, W rong answers If PARAM,HEATSTAT,YES is specified in SOL 101, then SWM 3001 will be issued by the SEP1X module. Theis warnings may safely be ignored unless there are RSSCON elements present, in which case, wrong answers may result. Avoidance: Insert the following DMAP alter: compile sestatic alter 'hsflag\=false' peid=-1 $ seid=-1 $ endalter @ 31323; ; Miscellaneous - SFM 3063, Memory Allocation Error If a run is submitted using the MEM= option and the memory allocated is above or near the limit of what is available, it is possible that the run may fail with SFM 3063 indicating that there is no more dynamic memory available. Avoidance: Always leave at lease 8mW of memory available for the system when using the MEM keyword. For example, if you are using a queue which has a 128mW limit, use a MEM value of 120mW or less. @ 31327; 2001; Documentation - CSUPER - Should Reference EXTDR, EXTDROUT, EXTOU T and EXTUNIT The Version 70.5 Quick Reference Guide documentation for the CSUPER entry should be modified to reference parameters EXTDR, EXTDROUT, EXTOUT and EXTUNIT in section 6 of the QRG. @ 31334; 2001; Documentation/Optimization - DVPREL1, DVPREL2 - QRG Not Updated for V70.5

Version 70.5 has simplified the interface for the design of constant section beams. The revised user interface requires changes in the Quick Reference Guide's description of the DVPREL1 and DVPREL2 entries. These changes were not completed in time before the Version 70.5 QRG was sent to press. Avoidance: The revised descriptions can be found in Appendix B of the Version 70.5 Release Notes. @ 31351; 70.5; Optimization/Design Sensitivity - SUBCOM - SFM 4276.GETSTR OR CODE 31 The message: SYSTEM FATAL MESSAGE 4276, GETSTR ERROR CODE 31 may occur if the user employs SUBCOMs and the adjoint sensitivity method is selected. Avoidance: p-elements, whether the the model. Since the adjoint method is not selected when there are an easy avoidance is to tell the module that determines adjoint method can be selected that there are p-elements in This is done with the following dmap alter: ERR

compile descon $ alter 'dsad'(1,-1) $ type parm,,logical,n, hold $ hold=pexist $ pexist=true $ alter 'dsad' $ pexist=hold $ @ 31364; 70.5; Aeroelastic Analysis - AEROS, ZY Bodies - Wrong Answers IF there are zy bodies in the model and the aerodynamic coordinate system is rotated about the x-axis relative to the aerodynamic reference system, wrong answers can occur in the stability derivative prints, and hence, the trim solution. Avoidance: If ZY bodies are included in the analysis, make sure the aerodynamic coordinate system and the reference coordinate systems appearing on the AEROS entry are not rotated relative to each other. @ 31374; 70.5; Elements - PBEAM, IFP - SFM 469 If a PBEAM entry contains an integer in a real field, then the job may fail in IFP without a meaningful message. Avoidance: The error was only experienced on CRAY systems. If you have access to another system then run the job on that system to identify the invalid entries and correct them. @ 31381; ; Miscellaneous - INPUTT2 - UFM 1019, SFM 6551 If the .dat file that is used to read the INPUTT2 file has the same prefix as the referenced INPUTT2 file--for example, test.dat is the .dat file and test.op2 is the referenced INPUTT2 file--then depending on the value of STATUS=x, the job may fail with one of the following error messages:

ASSIGN INPUTT2='test.op2',UNIT=12 USER FATAL MESSAGE 1019 (OPNPFL) FORTRAN UNIT OPEN ERROR IOSTAT = 6, No such file or directory(errno=2) LOGICAL = INPUTT2 FILE = test.op2 SYSTEM FATAL MESSAGE 6551 (INPTX2) AN END-OF-FILE WAS DETECTED ON UNIT 12 DURING BINARY INPUTT2 PROCESSING PROGRAMMER INFORMATION: 1. A KEY VALUE WAS BEING READ 2. THE LAST KEY VALUE READ: 0 FATAL ERROR 0 NO. FORTRAN RECORDS READ = 0 NO. FORTRAN WORDS READ = 0 Avoidance: Modify the name of either the .dat file or the INPUTT2 file. @ 31392; ltiply 70.5; SSSALTERS - GPFORCE, trncompa - SFM 5423, Incompatible Matrix Mu If GPFORCE output is requested when using the trncompa alter, then the job may fail with the following error messages: *** SYSTEM FATAL MESSAGE 5423 (DMPYAD) ATTEMPT TO MULTIPLY INCOMPATIBLE MATRICES COLS ROWS FORM TYPE NZWD DENS TRAILER FOR QG2 2 2 0 12 77 TRAILER FOR UHDISP 18 7792 TRAILER FOR (NONE) IS 0 0 TRANSPOSE FLAG IS 0 0FATAL ERROR There are two possible avoidances: 1. Add "param,ddrmm,-1" to the input file or 2. Replace the following dmap statement mpyad qg,uhdisp,/qgfull/ $ with equivx qg/qgfull/always $ @ 31399; 2001; Documentation - PARAM,USETPRT, PARAM,USETSEL - Incorrect Decimal Equivalent the Version 70 and 70.5 Quick Reference guide parameter section indicates that the decimal equivalent of set U6 (to be specified on PARAM,USETSEL for PARAM,USETPRT) is 76108864, it should be 67108864. 0 0 0 0 IS 102 21 11 2 IS 101 21 739 2

@ 31422;

2001; Documentation/QRG - XYPLOT, PARAM,CURVPLOT - UWM 979 If XYPLOT commands are specified for plotting in transient or frequency response analysis with grid points on the abscissa, as defined under PARAM,CURVPLOT, then USER WARNING MESSAGE 979 is issued and no plots are output. PARAM,CURVPLOT is applicable only in static and normal modes analysis; i.e., SOLs 1, 3, 5, 14, 15, 16, 101, 103, 105, 106, 114, 115, 144, and 200 even though Table 1 of Section 6 in the Quick Reference Guide indicates that it is applicable in all the structured solution sequences. Table 2 is documented correctly for CURVPLOT. Section 13.3.6 of the Reference Manual is also documented incorrectly with regards to solution sequence applicability.

@ 31423;

; Interface Elements - GMBNDC - GMCURV Option Results in UFM 6403 The GMCURV option on the GMBNDC entry may incorrectly result in the following fatal message: *** USER FATAL MESSAGE 6403 (CIFCRV) THE GMBNDC ID = 2 REFERENCED BY GMINTC CONTINUOUS AT GRID ID= 1. ID = 5001 IS NOT

Avoidance: The exact cause of the problem is not yet known. The FEEDGE and GRID options may be used instead. @ 31427; d 70.5; Miscellaneous - Access violation (C0000005) exception encountere When there is an illegal bulk data entry "0" as shown below, then the job may fail in various ways. NT systems have failed with the message "Access violation (C0000005) exception encountered" 0 0 4602 4606 4603 4607 4604 4605

Avoidance: Correct the offending bulk data entries. @ 31430; 2001; Documentation/Elements - CMASS1, CMASS2, GPWG - QRG Unclear The CMASSi elements have provisions for defining a pair of DOFs, although they are often used for point masses. To input a point mass, define only a single DOF. When two DOFs are used a coupled mass matrix is produced. If not input correctly it may cause internal constraints. No checks are made for internal constraints in such elements. The following remarks should be added to the CMASS1 and CMASS2 Quick Reference Guide Descriptions: 7. Scalar elements input coupled mass matrices when the second pair of fields is entered. When uncoupled point masses are desired input only the first pair of fields. When a coupled mass matrix is requested the submatrix added has M on the diagonal, and -M on the off-diagonal. The element is not checked for internal constraints, which is the user's responsibility if desired. There are instances where elements with internal constraints are desired, although not frequently. To identify the presence of internal constraints caused by coupled mass, inspect GPWG

output, OLOAD output due to GRAV loads, and rigid body modes of free structures. Some forms of coupled mass will cause coupling of rigid body translational mass terms in GPWG output, and poor rigid body modes in modal analysis. @ 31436;

2001; Documentation - PBARL, PBEAML - Missing Documentation There are undocumented limits in the use of the beam library. PBEAML: For the

The beam can have no more than 14 dimensions per station. The total number of dimensions at all stations must be less than 200. The transfer of data with the beam server is limited to 4000 words. None of these limits are exceeded with the MSC beam library, but a user defined beam library could. The data transfer limit would most likely be exceeded when the number of dimensions for a PBARL exceeds 235 (4000 / 17 properties per bar). There is a further limit that the NSM values input on the PBARL and PBEAML must be >= 0. No other property types have this limit. @ 31447; 70.5; Aeroelastic Analysis - EPOINT, SOL 145, Multiple Eigenvalue Subc ases - UFM 1126 In aeroelastic flutter analysis (SOL 145 only) UFM 1126 as shown below is issued when there are multiple normal mode boundary conditions and the model contains epoints. UMERGE1 IS ATTEMPTING TO OUTPUT DATABLOCK NAME = GMD which already exists Avoidance: Insert the following DMAP alter: compile sefluttr alter 1 file gmd=save,ovrwrt/god=save,ovrwrt/k4dd=save,ovrwrt $ endalter @ 31452; ; P-Elements - GMINTC and GMINTS - Non-Zero Rigid Body Stresses Models that connect curved surfaces with the GMINTC and GMINTS interface elements may experience non-zero stresses due to rigid body motion. These non-zero stresses are caused by the differences in cubic geometry on the dissimilar mesh. Straight interfaces have no such problem, since the geometry is identical. Avoidance: The tolerance TOL on the PINTC and PINTS property entries will generate a warning message for differing geometry, and may be set lower to be more conservative. @ 31461; 2001; Documentation - PARAM,NQSET - Missing QRG Documentation For part superelements, generalized coordinates can be defined automatically by senqset entries, as documented in the qrg, or by the

param, nqset entry. The param method was introduced in Version 69 but has yet to be documented. The documentation follows: NQSET Default=0

If NQSET>0, and the PARAM entry is in case control, all part superelements that do not contain qset entries, or are not referenced by senqset entries in the main bulk data deck, have NQSET generalized coordinates assigned to them. These qset variables are functionally equivalent to those generated by senqset entries. @ 31463; 70.5; Optimization/Design Sensitivity - ANAL=STATICS - UFM 9058 In solution 200 with ANALYSIS=STATICS, if the static loads are null then UFM 9058 will be issued during the sensitivity analysis: THE SOLUTION FOR THE RESIDUAL STRUCTURE DOES NOT EXIST Avoidance: Insert the following DMAP alter: compile pslgdv alter 'call sela1'(,-1) dbstatus egxp//s,n,noegxp $ if ( noegxp<=0 ) matgen ,/egxp/7/lusets/nopgs $ endalter @ 31469; GL; Superelements - SECONCT - UFM 4156 If PART superelements are used in a model and manual connections are specified through the SECONCT entry and there are coincident exterior points in the PARTs, the run will fail when attempting to connect the reduced matrices with UFM 4156 indicating an incorrect matrix trailer as shown below: USER FATAL ERROR 4156--- DIMENSIONS OR TYPE OF DATA BLOCK XAAV , SUPERELEMENT xx TRAILER=( 1278 1278 6 2 58 52) ARE INCONSISTENT WITH SUPERELEMENT MAP. FATAL ERROR Avoidance: Coincident points are not supported when connecting PART superelements. It is possible that the error could be avoided by only connecting one of the points to the other PARTs in the model and using an RBAR or CBUSH to connect the coincident points in the PART. If the automatic connection logic were used, the program would fail with a fatal message indicating that there were coincident points in the PART. @ 31472; 2001; P-Elements - PBEAM - UFM 3005 (SNGERR) BEAM p-elements fail in an adaptive loop if J(A)>0 is not specified on the PBEAM entry. The following fatal message is issued. *** USER FATAL MESSAGE 3005 (SNGERR) ATTEMPT TO OPERATE ON THE SINGULAR MATRIX D IN SUBROUTINE ADPF1D 0FATAL ERROR Avoidance: Specify J(A)>0 in the 8th field of the PBEAM entry. @

31473;

2001; Documentation - CVISC - Stress Plot Codes not Applicable The Appendix A in the Quick Reference Guide documents stress plot codes for the CVISC element. The CVISC element only supports force output.

@ 31475; 70.6; Optimization/Design Sensitivity Analysis - Multi-Level Superelme nts - UFM 9058 "USER FATAL MESSAGE 9058" may appear in a superelement sensitivity analysis if the model contains multilevel superelements and an upstream superelement contains a design model but more than one of its downstream superelements does not. Avoidance: Place dummy design information in downstream superelements. @ 31479; hines 70.5; P-Elements - TETRA, GMLOAD, GMCORD - UFM 6497 on SGI and SUN Mac TETRA p-elements with GMLOAD and GMCORD may lead to the following fatal message: *** USER FATAL MESSAGE 6497 (TRNJFD) THE POINT AT: X= 0.346875D+03 Y= 0.500000D+03 Z=-0.203698D+03, IN THE BASIC COORDINATE SYSTEM COULD NOT BE LOCATED ON FEFACE ID= 1. *** USER FATAL MESSAGE 2025 MODULE = SSG1 UNDEFINED COORDINATE SYSTEM 1 0FATAL ERROR Avoidance: Either: - mesh with HEXA elements. - use a system other than GMCORD with GMLOAD - use PLOAD4 instead of GMLOAD @ 31488; 70.5; ESTIMATE - May Ignore Entry After TITLE or SUBTITLE Under certain circumstances, ESTIMATE will ignore the entry following a TITLE, SUBTITLE, or LABEL entry. This may cause ESTIMATE to ignore a METHOD, MPC, SPC, STRAIN, or STRESS CC entry if it immediately follows a TITLE, SUBTITLE, or LABEL that contains a comma "," character. Avoidance: One of the following avoidances may be used: 1) Enter a comment following the TITLE, SUBTITLE, or LABEL. or 2) Specify the missing METHOD, MPC, or SPC entry on the command line. @ 31498; ; Restart - DBALL offline - UFM 1127, SFM 6357, plus others If scr=no is specified in a cold start, but the data base path and name is not assigned by the dbs keyword or the ASSIGN statement, then a restart run in a different directory with only an ASSIGN statement for the MASTER dbset, and not DBALL, will fail with UFM 1127. This problem does not occur if the dbs keyword

specifies the location of the database on the command line. possible error messages are shown below: *** USER FATAL MESSAGE 1127 (GNFIST) DMAP MODULE PLTSET IS ATTEMPTING TO ACCESS DATA BLOCK = GEOM2 WHICH IS ON AN UNAVAILABLE DBSET 0FATAL ERROR *** SYSTEM WARNING MESSAGE 3001 (SEP1X ) THE INPUT DATA BLOCK NAMED GEOM1QS AT POSITION EXIST. *** SYSTEM FATAL MESSAGE 6357 (SEP1A) NO GRID POINTS FOUND. Avoidance: There are three avoidances:

Some of the

103 NAME

2 DOES NOT

1. Submit the restart from the same directory as the cold start 2. Specify ASSIGN statements for all the DBSETs 3. Use the dbs keyword with the full pathname

@ 31537;

2001; Documentation - QRG, PARAM,AUTOSPC - Incorrect Default Starting with Version 69, the QRG states that the default for PARAM,AUTOSPC is NO in SOL's 1 thru 12, when in fact it is YES for all of these solutions except SOL 4.

@ 31540;

69.1; Postprocessing - Incorrect Approach Code in XDB Solution 1 in Version 69 will incorrectly set the approach code to 4 in the xdb file. The correct value is 1. Avoidance: Either insert the following DMAP alter or use solution 101: compile sol1 alter 'dbc '(2),'' DBC CASECC,,,,,,,,,,,,,,,,,,,// 'CASECC'//////////////////// -1/DBCPATH/S,N,CP/soltype//V,Y,GEOMU/V,Y,LOADU/V,Y,POSTU/ V,Y,DBCDIAG/V,Y,DBCCONV/V,Y,DBCOVWRT $ endalter

@ 31568;

70.7; Restarts - DBLOC - UFM 723 (LCDEFW) The WHERE clause on the DBLOC command does not work consistantly with the VERSION and PROJECT qualifaiers with the DBDICT entry. The following statement: DBLOC DATABLK=(QKKS) WHERE (VERSION=* AND PROJECT=*) LOGICAL=QTODB5

caused the diagnostic message: 0*** USER FATAL MESSAGE 723 (LCDEFW) PROJECT ID EXPRESSION HAS AN UNBALANCED NUMBER OF SINGLE QUOTES Other variations of the WHERE clause with these qualifiers may cause data to not be found. @ 31578; e Loop ; Nonlinear Analysis - Load Factor Greater than 1.0 causes Infinit In a nonlinear analysis, on some instances, the load factor tolerances may be in error. When this occurs the value of the load factor exceeds 1.0 and the solution may enter an infinite loop or produce some other fatal message later on. The problem is identified when one of the load factors > 1.0 Avoidance: On the NLPARM entry increase the NINC field by 1 and restart into the offending subcase. This will change the tolerance calculation. @ 31582; 2001; Miscellaneous - PARAM,DESPCH1 - Punch output with small field fo rmat not allwed Bulk Data paramter DESPCH1 can be used to request the format type for the punched output. A negative DESPCH1 requests the small field format while the positive DESPCH1 requests the large field format. However, when an optimization deck specifies the negative DESPCH1 and also specifies the second parameter DESPCH>0, the small field format is ignored. Avoidance: Apply the dmap alter below: compile nddl=nddl include 'nddl.dat' $ where the nddl.dat is the modified nddl list that includes the following two lines: param,despch =0, type=I, PATH=GEOM,LOCATION=master param,despch1 =6, type=I, PATH=GEOM,LOCATION=master compile desopt alter 'despch' type parm,,i,y,despch1 compile exitopt alter 'despch1','despch1' type PARM,,I,Y,cnvflg,despch,omaxr,despch1,xyunit @ 31583;

; Acoustic Analysis - METHOD(STRUCTURE), SDFREQ - UFM 4276 EC 1305 In modal frequency response analysis, solution 111, with fluids, if the METHOD(STRUCTURE)=n case control command is missing, the following error will occur: USER INFORMATION MESSAGE 5222 (SQFREQ) COUPLED SOLUTION ALGORITHM USED. BIOMSG: ERROR 1305 HAS OCCURRED IN ROUTINE ENDPUT , FILE INDEX (FILEX) = 0 STATUS = 0 ********* NASTRAN FILE TABLE ********* INDEX LOGNAME 2 OBJSCR STREAM_ID 00000002 NAME /work/30314_42.OBJSCR

3 DBA 00000003 /work/fnd1.dba 4 MAST 00000004 /work/fnd2.mast 6 SCRA 00000006 /home/no4/emc/emc3fnd/fnd5.scra 7 SCRU 00000007 /tmp/fnd6.scru 8 MASTER 00000008 /work/30314_42.MASTER 9 SC2A 00000009 /work/30314_42.SC2A 10 MASTERA FFFFFFF6 /msc/msc70/aix/SSS.MASTERA 11 MSCOBJ FFFFFFF5 /msc/msc70/aix/SSS.MSCOBJ 0*** SYSTEM FATAL ERROR 4276.ENDPUT ERROR CODE 1305 PID= 0 0*** USER INFORMATION MESSAGE: TO OBTAIN A NASTRAN DUMP, RESUBMIT JOB WITH DIAG 44 INSERTED IN THE EXECUTIVE CONTROL SECTION. 1 * * * END OF JOB * * * This user input error need to have a valid error message given. Avoidance: Put a METHOD(STRUCTURE)=n case control command in the input deck. @ 31590; ; Nonlinear Analysis - Missing Intermediate output on a nonlinear restart run If additional subcases are added in a restart of a solution 106 nonlinear job then there may be missing intermediate output. Avoidance: The problems is under investigation. Other than performing the job as a cold start there is no avoidance. @ 31591; 70.5.2; Postprocessing - DBC, PARAM,POST PARAM,GEOMU,0 - NASTRAN Stops i n DBC Module If PARAM,GEOMU,0 is used to suppress the writing of geometry into XDB (i.e. PARAM,0) is in the deck, then the job will stop without message in the DBC module. The log file may contain a message of the form: 1525-001 The READ statement on the file ./s31591.xdb cannot be completed because the end of the file was reached. The program will stop. @ 31608; ; File Management Section - ASSIGN, DELETE - DBALL is not Deleted The fms statement ASSIGN includes the DELETE qualifier, which specifies that an existing data base at the start of a run should be deleted, and replaced with a new version. If this qualifier is applied against a MASTER database, it properly deletes the MASTER database. However, MASTERs point to one or more data base members, such as DBALL. The index for these databases are stored on MASTER, so they are useless after MASTER is deleted and they too should be deleted but are not. @ 31622; 70.5.1; Heat Transfer Analysis - CHBDYE, TA1A - Segmentation Error Heat Transfer models with CHBDYE elements may fail with a segmentation error. Avoidance: There is no avoidance. @ 31636; issued

; Optimization/Design Sensitivity - DESSUB - UFM 6293 incorrectly

If an attempt to place limits on a global response (such as weight) is applied using a DESSUB command, the program will incorrectly issue UFM 6293. Avoidance: Global responses (such as WEIGHT) may only be constrained using the DESGLB command. They may not be constrained by the DESSUB command. @ 31655; 70.5.2; Optimization/Design Sensitivity - DOM9 - SFM 4276, EC 3002, EC 1 409, EC 31 The message: SYSTEM FATAL MESSAGE 4276, ERROR CODE 1409 can occur in the DOM6 or DOM9 module if there are DRESP2 responses and two or more of the DRESP1 id's invoked by a single DRESP2 are duplicates. Avoidance: Remove the duplicate DRESP1 id's from the DRESP2 entry. It is most likely not the user intent to have duplicated the id's, but it will be permitted once this error is corrected. @ 31662; 70.7; Miscellaneous - DBVIEW - Infinite CPU Loop If the view name and the datablock name are identical on the DBVIEW statement, and the datablock is local scratch, then an infinite CPU loop will occur in GNPREP when the view name is used as input to a module. The following input deck demonstrates the problem: sol AERBG diag 8,24$ compile AERBG $ SUBDMAP AERBG dbview uset = uset ( where peid=1) tabprt uset END $ $ CEND BEGIN BULK ENDDATA @ 31664; 70.7; Rigid Elements - RBE3, GP4 - EXIT from GP4, No Fatal Message The um fields of the RBE3 element contain grid point and component data. A GID-component pair may appear only once in the um fields. When a pair appears more than once, the gp4 modlule takes an exit without producing any diagnostics. The most likely clue that this has happened is to inspect the .f04 file. If the last several modules to run are gp4, append, bcdr,prstum, and exit, a bad rbe3 may be the cause. Avoidance: An avoidance is to check the um fields of all rbe3 entries. @ 31665; ; P-Elements - MATGPR - GRIDN Labels Incorrect

When using the MATGPR module to print a matrix for p-elements, the components of the higher-order degrees of freedom represented by the GRIDNs are incorrect. They are labeled T1,T2,T3,R1,R2,R3 followed by garbage, since the number of components may be much higher than 6. They should simply be labeled with integers. @ 31668; 70.5.1; Optimization/Design Sensitivity - PBEAM - Wrong Answers The Version 70.5 Release Notes state that it is now possible to design properties for the constant cross section PBEAM by referencing the END A fields only on the PBEAM bulk data entry. (Previously, one had to reference the EPT and explicitly design END A and END B). The implementation of this new capability was incomplete in that it does not change the DVPREL2 invoked designed properties at END B. As a consequence, sensitivities of END B responses are most likely incorrect. If an optimization task in undertaken with END B responses, the redesigned structure will provide incorrect analysis results at END B and lead to invalid designs. Avoidance: Either do not use the new feature or do not ask for data recovery at END B in the design task. @ 31671; 2001; Documentation - MODES - Inefficient Processing, Excessive Output The MODES case control command may be used in all solution sequences as a subcase replicator. The most common and obvious use is in eigenvalue solutions and is demonstrated under the Quick Reference Guide's description of this entry. This command may be of no use in other solution sequences such as transient response and may cause the data recovery processing to be repeated and printed multiple times. @ 31674; 2001; Documentation - PARAM, CB1, CB2, CK1, CK2, CK3, CP1, CP2, CM1, CM2 - UFM 6011 Parameters CB1, CB2, CK1, CK2, CK3, CP1, CP2, CM1, and CM2 are incorrectly indicated to be real parameters with a default of 1.0 when they are complex parameters with a default of (1.0,0.). Avoidance: If you use these parameters, define both the real and imaginary terms. @ 31686; 70.7; Elements - CQUAD4, CQUAD8, CTRIA3, CTRIA6 - UFM 316 The MCID field on the CQUAD4, CQUAD8, CTRIA3 and CTRIA6 entries cannot be greater than 999999 or else UFM 316 will be generated. However, on the PSOLID and CGAP, coordinate system IDs > 999999 are accepted. @ 31689; ; Nonlinear Static Analysis - PGAP - Unnecessary Restriction for Non-Adaptive Gap The adaptive gap element requires that MU2 <= MU1; however, this restriction is unnecessarily imposed on the non-adaptive gap element. Avoidance: This limitation can be avoided by proper choice of the orientation vector.

@ 31694;

70.7; Dynamics - SOLs 103-112, PARAM - Wrong Answers In SOLs 103, 105, and 107 through 112, if the Case Control section contains multiple dynamic subcases or a static subcase for differential stiffness, then PARAM commands specified in the dynamic subcase(s) are ignored. Avoidance: Specify PARAMs in the Bulk Data section or insert the following DMAP alter: compile sekdr alter 'pvt ','' pvt pvts,casebuck/pvtxxx $ endalter

@ 31703; 70.5.1; Aeroelastic Analysis - AELIST, ADG - UFM 6944 If an AELIST set is referenced by more than one AESURF entry, then the job will fail with the message: USER FATAL MESSAGE 6944 (ADG) AELIST ENTRY ID = xxx THAT IS REFERENCED ON AESURF ENTRY ID =yyy DOES NOT EXIST. The id's specified in the message are also scrambled in that aesurf id's are given in the aelist position and vice versa. Avoidance: Create another AELIST entry with a unique ID but with the same set of aero boxes. @ 31717; 70.7; DMAP - SOLVE - Inefficient Processing The internal resequencing option for decomposition in the SOLVE module is not performed. This may lead to inefficient processing. Avoidance: Use the DCMP and FBS modules. @ 31721; 70.5.1; Miscellaneous - Missing Release Guide Test Problems The MSC/NASTRAN Version 70.5 Release Guide as well as the README file in the Version 70.5 relnotes directory refer to the problems ar29be01.dat, ar29src1.dat, tpsnp.dat, and tpsp.dat. These problems are missing on the UNIX R0 cd. Avoidance: The input decks are listed below separated by $fname. $ ar29be01.dat $ $********************************************* $ Test of nonlinear shock absorber $ under large deformation $******************************************** TIME 1000 SOL 129 $******************************************** CEND $

TITLE = Shock absorber and spring loaded with sine pulse LABEL = CBUSH1D with SHOCKA option and SPRING option ECHO = SORT SET 1= 1,2 DISP(SORT2,PLOT) = 1 SET 2 = 1, 100 STRESS(SORT2,PLOT) = 2 OLOAD(SORT2,PLOT) = 1 $ SPC = 1 $ SUBCASE 10 DLOAD = 100 TSTEPNL = 10 $ SUBCASE 20 DLOAD = 100 TSTEPNL = 20 $ OUTPUT(XYPLOT) XAXIS = YES YAXIS = YES XGRID LINES = YES YGRID LINES = YES XTITLE = TIME $ YTITLE = Applied load TCURVE = Applied load at GRID 2 in vertical direction, T3, z XYPLOT OLOAD 10,20/2(t3) $ YTITLE = DISPLACEMENT TCURVE = Vertical displacement at beam end, GRID 2, T3, z XYPLOT DISP 10,20/2(T3) $ $************************************************************************ * BEGIN BULK $ PARAM, POST, -1 PARAM,AUTOSPC,YES PARAM,MAXRATIO,1.E+8 PARAM LGDISP 1 PARAM LANGLE 2 $ $ CBUSH1D element (SHOCKA option with TABLE for S(u) $ and TABLE for SPRING constant) $ CBUSH1D, 100, 101, 8, 19 $ $ CVT = 0.2 damping for tension $ CVC = 0.4 damping for compression $ S(u) = 1.0 constant scale factor $ EVT = 1.0 exponent of velocity for tension $ EVC = 0.5 exponent of velocity for compression $ PBUSH1D, 101, 10.0, 0.4, , , , , ,+PB1 +PB1, SHOCKA, TABLE, 0.2, 0.4, 1.0, 0.5, 999, , +PB2 +PB2, SPRING, TABLE, 998 $ TABLED1, 999, , , , , , , , +TB999

+TB999, -1000., 1.0, 0.0, 1.0, 1000., 1.0, ENDT $ TABLED1, 998, , , , , , , , +TB998 +TB998, -1000., -10000., 0.0, 0.0, 1000., 10000., ENDT $ PLOTEL, 9999, 8, 19 $ $ Loading - Unit amplitude half sine $ $TLOAD2 ID DAREA DELAY TYPE T1 T2 F P TLOAD2 100 10 0 .0010 .5010 1.0 -90.0 $ $ Point of application, direction and magnitude of load $ $DAREA ID P1 C1 A1 DAREA 10 2 3 900.0 $ $ Time step integration. Set MAXBIS to a negative number, $ terminate if convergence not reached. $ $TSTEPNL, ID, NT, DT , NO, METHOD, KSTEP, MAXITER, CONV TSTEPNL, 10, 1200, 0.0005, , ADAPT , , , ,+T10a +T10a, , , , , , , , ,+T10b +T10b, -9 $ TSTEPNL, 20, 800, 0.0010, , ADAPT , , , ,+T20a +T20a, , , , , , , , ,+T20b +T20b, -9 $ $ Geometry $ GRID 1 0.0000000.0000000.000000 GRID 2 1200.0000.0000000.000000 GRID 3 100.00000.0000000.000000 GRID 4 200.00000.0000000.000000 GRID 5 300.00000.0000000.000000 GRID 6 400.00000.0000000.000000 GRID 7 500.00000.0000000.000000 GRID 8 600.00000.0000000.000000 GRID 9 700.00000.0000000.000000 GRID 10 800.00000.0000000.000000 GRID 11 900.00000.0000000.000000 GRID 12 1000.0000.0000000.000000 GRID 13 1100.0000.0000000.000000 GRID 19 -92.82040.000000399.9991 $ $ Connectivity $ CBEAM 1 1 7 80.000000-1.000000.000000 +E61 +E61 0 00.0000000.0000000.0000000.0000000.0000000.000000 CBEAM 2 1 1 30.000000-1.000000.000000 +E62 +E62 0 00.0000000.0000000.0000000.0000000.0000000.000000 CBEAM 3 1 3 40.000000-1.000000.000000 +E63 +E63 0 00.0000000.0000000.0000000.0000000.0000000.000000 CBEAM 4 1 4 50.000000-1.000000.000000 +E64 +E64 0 00.0000000.0000000.0000000.0000000.0000000.000000

CBEAM 5 1 5 60.000000-1.000000.000000 +E65 +E65 0 00.0000000.0000000.0000000.0000000.0000000.000000 CBEAM 6 1 6 70.000000-1.000000.000000 +E66 +E66 0 00.0000000.0000000.0000000.0000000.0000000.000000 CBEAM 7 1 8 90.000000-1.000000.000000 +E67 +E67 0 00.0000000.0000000.0000000.0000000.0000000.000000 CBEAM 8 1 9 100.000000-1.000000.000000 +E68 +E68 0 00.0000000.0000000.0000000.0000000.0000000.000000 CBEAM 9 1 10 110.000000-1.000000.000000 +E69 +E69 0 00.0000000.0000000.0000000.0000000.0000000.000000 CBEAM 10 1 11 120.000000-1.000000.000000 +E6A +E6A 0 00.0000000.0000000.0000000.0000000.0000000.000000 CBEAM 11 1 12 130.000000-1.000000.000000 +E6B +E6B 0 00.0000000.0000000.0000000.0000000.0000000.000000 CBEAM 12 1 13 20.000000-1.000000.000000 +E6C +E6C 0 00.0000000.0000000.0000000.0000000.0000000.000000 $ $ Beam property and material $ PBEAM 1 2600.0000180000.05000.0000.0000000.0000000.000000 $ MAT1 22.00E+6 7.69E+050.3000007.80E-09 $ $ Boundary conditions $ SPC 1 1 1234 0.000000 SPC 1 2 246 0.000000 SPC 1 3 246 0.000000 SPC 1 4 246 0.000000 SPC 1 5 246 0.000000 SPC 1 6 246 0.000000 SPC 1 7 246 0.000000 SPC 1 8 246 0.000000 SPC 1 9 246 0.000000 SPC 1 10 246 0.000000 SPC 1 11 246 0.000000 SPC 1 12 246 0.000000 SPC 1 13 246 0.000000 SPC 1 19 1234 0.000000 $ ENDDATA $ $ $ $ $ $ $ $ $ $ ar29scr1.dat Name: ar29src1 A swinging pendulum is modeled. The pendulum is loaded with gravity load. The analysis goes through one and a half cycles, starting in horizontal position. Large deformation is turned on. The pendulum is modeled with two seperate systems. 1. CROD and CVISC, CROD rotates, CVISC stays in horizontal

$ orientation. $ 2. CBUSH1D with SPRING and DAMPER, spring and damper rotate. $ TIME 10 $ SOL 129 $ CEND $ TITLE= ar29src1 swinging elastic pendulum SUBTITLE= demonstrate rotating damping ECHO=BOTH SET 1 = 22, 42 SET 2 = 104, 108 DISP=1 STRESS=2 spc= 4 $ SUBCASE 1 DLOAD=100 TSTEPNL=100 $ OUTPUT(XYPLOT) XTITLE = TIME IN SECS XGRID LINES = YES YGRID LINES = YES YTITLE = X DISPLACEMENT TCURVE = DISPL of ROD and VISC compared to CBUSH1D XYPLOT DISP RESP/22(T1), 42(T1) YTITLE = Z DISPLACEMENT TCURVE = DISPL of ROD and VISC compared to CBUSH1D XYPLOT DISP RESP/22(T3), 42(T3) YTITLE = STRESS/FORCE TCURVE = STRESS/FORCE of ROD compared to CBUSH1D XYPLOT STRESS RESP /104(2), 108(2) $ BEGIN BULK $ param, lgdisp, 1 $ $ GEOMETRY $ grid,21, , 0., 2., 0. grid,22, , 1., 2., 0. grid,41, , 0., 4., 0. grid,42, , 1., 4., 0. $ $ Boundary conditions $ spc1, 4, 12356, 21, 41 spc1, 4, 246, 22, 42 $ $ mass $ conm2, 53, 22, , 1000. conm2, 55, 42, , 1000. $ $ rod with viscous damping $ crod, 104, 201, 21, 22 $ prod, 201, 222, 1.

mat1, 222, 1.e+7 mats1, 222, , plastic, 0., 1, 1, 1.e+12 $ cvisc, 114, 310, 21, 22 pvisc, 310, 1000. $ $ CBUSH1D spring and rotating viscous damper $ cbush1d, 108, 208, 41, 42 pbush1d, 208, 1.e+7, 1000. $ $ $ $ $ $ $ $ $ SOLUTION STRATEGY $ TSTEPNL, 100, 400, .01, , adapt, , , , +TSP100 +TSP100, , , , , , , , , +TSP100 +TSP100, , 0 $ $ LOADING $ DAREA 100 22 3 10000. DAREA 100 42 3 10000. $ TLOAD1 100 100 0 100 TABLED1 100 +ABC +ABC 0. 1. 10. 1. ENDT $ ENDDATA

$ tpsnp.dat $ DEC/CMS REPLACEMENT HISTORY, Element HA144E.DAT $ *2 6-JUL-1994 14:07:00 A_BOYADJIAN "68 PLUS/G/ CHANGE DBSDIR: TO TPLDIR: FOR INCLUDE CARDS" $ *1 5-JUL-1994 16:51:13 A_BOYADJIAN "68 PLUS/G/ NEW FOR V68 AERO_SS BOOK" $ DEC/CMS REPLACEMENT HISTORY, Element HA144E.DAT ID MSC, HA144E $ E_JOHNSON V68 5-JUL-1994 $ID MSC, HA144E $$$$$$$$ HANDBOOK FOR AEROELASTIC ANALYSIS EXAMPLE HA144E $$$$$$$$ $ $ $ MODEL DESCRIPTION FULL SPAN 30 DEG FWD SWEPT WING $ $ WITH AILERON, CANARD AND AFT SWEPT $ $ VERTICAL FIN AND RUDDER. $ $ BAR MODEL WITH DUMBBELL MASSES. $ $ $ $ SOLUTION QUASI-STEADY AEROELASTIC ANALYSIS $ $ USING DOUBLET-LATTICE METHOD $ $ AERODYNAMICS AT MACH NO. 0.9. $ $ $ $ OUTPUT PLOTS OF THE STICK MODEL AND AERO $ $ GRID, LISTS OF RESTRAINED AND $ $ UNRESTRAINED SYMMETRIC AND ANTI$ $ SYMMETRIC STABILITY DERIVATIVES, $ $ AERODYNAMIC FORCES AND PRESSURES $ $ PLUS STRESSES AND DEFLECTIONS FOR $ $ LEVEL FLIGHT AND SEVERAL UNSYM$ $ METRICAL MANUEVERS. $ $ $ $ $

$$$$$$$$ $$$$$$$$ TIME 5 $ CPU TIME IN MINUTES SOL 144 $ STATIC AERO CEND TITLE = EXAMPLE HA144E: 30 DEG FWD SWEPT WING WITH 3 CONTROLS HA144E SUBTI = UNSYMMETRIC FLIGHT CONDITIONS, DOUBLET-LATTICE AERO LABEL = FULL-SPAN MODEL ECHO = BOTH SPC = 1 $ SYMMETRIC CONSTRAINTS SET 1 = 1 THRU 999999 DISP = 1 $ PRINT ALL DISPLACEMENTS STRESS = ALL $ PRINT ALL STRESSES FORCE = ALL $ PRINT ALL FORCES AEROF = ALL $ PRINT ALL AERODYNAMIC FORCES APRES = ALL $ PRINT ALL AERODYNAMIC PRESSURES SUBCASE 1 TRIM = 1 $ HIGH SPEED LEVEL FLIGHT SUBCASE 2 TRIM = 2 $ HIGH SPEED ROLLING PULLOUT SUBCASE 3 TRIM = 3 $ HIGH SPEED PULLOUT WITH ABRUPT ROLL SUBCASE 4 TRIM = 4 $ HIGH SPEED SNAP-ROLL ENTRY SUBCASE 5 TRIM = 5 $ HIGH SPEED CLIMBING TURN OUTPUT(PLOT) PLOTTER = NASTRAN SET 1 = ALL FIND SCALE, ORIGIN 1,SET 1 PLOT SET 1 PLOT STATIC DEFORMATION 0, ORIGIN 1, SET 1, OUTLINE BEGIN BULK $ $ DEC/CMS REPLACEMENT HISTORY, Element FSWTWO.DAT $ *1 5-JUL-1994 16:46:14 A_BOYADJIAN "68 PLUS/G/ NEW FOR V68 AERO_SS BOOK" $ DEC/CMS REPLACEMENT HISTORY, Element FSWTWO.DAT $*** * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * ***$ $ $ $ * * * STRUCTURAL DATA * * * $ $ $ $ (LB-FT-SEC SYSTEM) $ $ $ $ * * GRID GEOMETRY * * $ $ $ $ GRID 90 - 100 (T3) FUSELAGE POINTS $ $ GRID 110 - 122 (T3) WING POINTS $ $ GRID 310 - 312 (t3) FIN POINTS $ $ $ $ * FUSELAGE GRID * $ $ $ $ THE GRID ENTRY DEFINES THE LOCATION OF A STRUCTURAL GRID $ $ POINT. LISTED ARE ITS COORDINATE SYSTEM ID, ITS LOCATION, $ $ THE ID OF THE COORDINATE SYSTEM IN WHICH ITS DISPLACEMENTS $ $ ARE DEFINED, ITS PERMANENT SINGLE-POINT CONSTRAINTS AND $ $ ITS ASSOCIATED SUPERELEMENT ID. $ $ $ $ ID CP X1 X2 X3 CD PS SEID GRID 90 15. 0. 0. GRID 97 0. 0. 0.

GRID GRID GRID $ $ $ $ GRID GRID GRID GRID GRID GRID $ $ $ $ GRID GRID GRID GRID GRID GRID $ $ $ GRID GRID GRID $ $ $ $ $ $ $ $ $ $ $ $ CBAR CBAR CBAR CBAR $ $ $ $ $ $ $ $ $ $ $ $ $ PBAR +PB1F

98 99 100

10. 20. 30.

0. 0. 0.

0. 0. 0. $ $ $ CD PS SEID

* RIGHT WING GRID * ID 111 110 112 121 120 122 CP X1 X2 24.61325 +5. 27.11325 +5. 29.61325 +5. 18.83975+15. 21.33975+15. 23.83975+15. X3 -2. 0. 2. 2. 0. -2.

* LEFT WING GRID * ID 211 210 212 221 220 222 CP X1 X2 24.61325 -5. 27.11325 -5. 29.61325 -5. 18.83975-15. 21.33975-15. 23.83975-15. X3 -2. 0. 2. 2. 0. -2. CD PS SEID

$ $ $

* FIN GRID * 310 311 312 32.88675+0. 30.38675+0. 35.38675+0. 5. 5. 5.

$ $ $

* * STRUCTURAL STIFFNESS PROPERTIES * * * FUSELAGE STRUCTURE * THE CBAR ENTRY DEFINES A SIMPLE BEAM ELEMENT. LISTED ARE ITS PROPERTY ENTRY ID, THE TWO GRID POINTS JOINED BY THE BEAM AND COMPONENTS OF A VECTOR FROM THE FIRST POINT. THIS VECTOR DEFINES THE DIRECTION OF THE STRUCTURAL DEFLECTION OF THE POINT AND ITS POSITIVE SENSE. EID 101 102 103 104 PID 100 100 100 100 GA 97 98 90 99 GB 98 90 99 100 X1,GO 0. 0. 0. 0. X2 0. 0. 0. 0. X3 1. 1. 1. 1.

$ $ $ $ $ $ $ $ $ $ $

THE PBAR ENTRY DEFINES GEOMETRIC PROPERTIES OF THE BEAM. LISTED ARE ITS ASSOCIATED MATERIAL ENTRY ID, ITS CROSS SECTIONAL AREA, AREA MOMENTS OF INERTIA, TORSIONAL MOMENT OF INERTIA AND NON-STRUCTURAL MASS PER UNIT AREA. THE OPTIONAL CONTINUATION ENTRY CONTAINS STRESS RECOVERY COEFFICIENTS, I.E., Y,Z COORDINATES WHERE STRESSES ARE TO BE COMPUTED. K1 AND K2 ARE AREA FACTORS FOR SHEAR STIFFNESS (DEFAULT IS BLANK; THEN SHEAR STIFFNESS IS INFINITE, I.E., SHEAR FLEXIBILITY IS ZERO. I12 IS THE AREA PRODUCT OF INERTIA. PID 100 MID 1 A 4.0 I1 I2 .347222 .30 J 1.0 NSM

$ $ $ $ $ $ $ $ $ $ $ $

$ +PB1F +PB2F $ +PB2F $ $ $ $ $ $ $MAT1 +MT MAT1 $ $ $ $ $ $ $ $ $ $ $ $ CONM2 CONM2 CONM2 CONM2 $ $ $ $ $ $ $ $ $ PARAM $ $ $ $ $ $ $ PARAM $ $ $ $ PARAM $ $ $ $ $ $ $

C1 1.0 K1

C2 1.0 K2

D1 1.0 I12 0.0

D2 -1.0

E1 -1.0

E2 1.0

F1 -1.0

F2 -1.0

THE MAT1 ENTRY DEFINES THE MATERIAL PROPERTIES. LISTED ARE ITS ID, ITS ELASTIC MODULUS, SHEAR MODULUS, POISSONS RATIO, MASS DENSITY, TEMPERATURE EXPANSION COEFFICIENT, REFERENCE TEMPERATURE AND A STRUCTURAL DAMPING COEFFICIENT. MID 1 E G NU RHO 0.0 A TREF GE

$ $ $ $ $ $

1.44+9 5.40+8

* * MASS AND INERTIA PROPERTIES * * * FUSELAGE MASSES * THE CONM2 ENTRY DEFINES A CONCENTRATED MASS. LISTED ARE ITS ID, GRID LOCATION, COORDINATE SYSTEM TO LOCATE THE CENTER OF GRAVITY, THE MASS VALUE AND THE LOCATION OF THE CENTER OF GRAVITY RELATIVE TO THE GRID LOCATION. EID 97 98 99 100 G 97 98 99 100 CID 0 0 0 0 M X1 3000.0 3000.0 3000.0 3000.0 X2 X3

$ $ $ $ $ $ $ $ $

* * STRUCTURAL PARAMETERS * * THE PARAM,GRDPNT,XX ENTRY CAUSES THE GRID POINT WEIGHT GENERATOR TO BE EXECUTED USING GRID POINT XX AS THE REFERENCE POINT. THEN THE INERTIA MATRIX, THE TRANSFER MATRIX FROM BASIC TO PRINCIPAL AXES AND OTHER PERTINENT INERTIA DATA ARE PRINTED. GRDPNT 90 THE PARAM,WTMASS,GINV CAUSES ALL THE STRUCTURAL MASSES AND MASS DENSITIES TO BE MULTIPLIED BY GINV (I.E., BY ONE OVER THE ACCELERATION OF GRAVITY). THE DYNAMIC PRESSURE SUPPLIED FOR AERODYNAMIC FORCE CALCULATIONS WILL NOT BE MULTIPLIED BY GINV. WTMASS .031081 THE PARAM,AUNITS,GINV PERMITS THE ACCELERATIONS ON THE TRIM ENTRY TO BE SPECIFIED IN UNITS OF LOAD FACTOR (I.E., IN G'S) AUNITS .031081 * * STRUCTURAL CONSTRAINTS * * THE SPC1 ENTRY CONSTRAINS THE LISTED GRID POINTS IN THE SPECIFIED DOF COMPONENTS. SID C G1 G2 G3 G4

$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $

$ $ $ $ $ $

SPC1 $ $ $ $ $ $ $ $ $ $ SUPORT $ $ $ $ $ OMIT1 $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ AEROS $ $ $ $ $ $ $ $ CORD2R +CRD1 $ +CRD1 $ $ $ $ $ $ CORD2R +CRD100 $ +CRD100 $ $ $ $

90 $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $

THE SUPORT ENTRY IDENTIFIES A GRID POINT OR A SCALAR POINT AND SPECIFIES THE DOF COMPONENTS IN WHICH THE USER DESIRES REACTIONS TO BE APPLIED TO PREVENT RIGID BODY MOTION. IT THUS INVOKES THE SOLUTION OF THE BALANCE EQUATIONS TO DETERMINE THE REACTIONS. IN THE STATIC AEROELASTIC SOLUTION THE DOF COMPONENTS MUST BE CONSISTENT WITH THE UNDEFINED VARIABLES ON THE TRIM ENTRIES. ID 90 C 23456

THE OMIT1 ENTRY IDENTIFIES GRID POINTS TO BE OMITTED FROM THE REMAINDER OF THE ANALYSIS. ID 4 G 110 G 120 G 210 G 220 310

* * * AERODYNAMIC DATA * * * (LB-FT-SEC SYSTEM) * * ELEMENT GEOMETRY * * THE AEROS ENTRY IS UNIQUE TO THE STATIC AEROELASTICITY SOLUTION, SOL144. ACSID IDENTIFIES THE AERO COORDINATE SYSTEM, RCSID IDENTIFIES THE REFERENCE COORDINATE SYSTEM FOR RIGID BODY MOTION. REFC IS THE REFERENCE CHORD. REFB IS THE REFERENCE SPAN. REFS IS THE REFERENCE WING AREA. SYMXZ AND SYMXY ARE SYMMETRY KEYS. ACSID 1 RCSID 100 REFC 10.0 REFB 40.0 REFS 400.0 SYMXZ SYMXY

THIS CORD2R ENTRY DEFINES THE AERO COORDINATE SYSTEM FLAGGED BY THE AEROS ENTRY. THE ORIGIN IS AT THE CANARD QUARTER CHORD. LISTED ARE THE ORIGIN, A POINT ALONG THE Z AXIS AND A POINT IN THE X-Z PLANE, ALL IN THE RID COORDINATE SYSTEM. CID 1 C1 20. RID 0 C2 0. A1 12.5 C3 0. A2 0. A3 0. B1 12.5 B2 0. B3 -10.

THIS CORD2R ENTRY DEFINES THE NACA COORDINATE SYSTEM TO WHICH ALL THE STABILITY DERIVATIVES AND TRIM CONDITIONS WILL BE REFERENCED. CID 100 C1 0.0 RID 0 C2 0.0 A1 15.0 C3 0.0 * WING AERODYNAMIC MODEL * A2 0.0 A3 0.0 B1 15.0 B2 0.0 B3 -10.0

$ $ $ $ $

$ $ $ $

$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ CAERO1 +CARW $ $ +CARW $ $ $ $ $ PAERO1 $ $ CAERO1 +CALW +CALW $ $ $ $ CAERO1 +CARC +CARC $ $ CAERO1 +CALC +CALC $ $ $ CAERO1 +CA1FI +CA1FI $ $ $ $ $ $ $ $ $ $

THE CAERO1 ENTRY IS USED FOR DOUBLET-LATTICE AERODYNAMICS. LISTED ARE ITS PAERO ENTRY ID AND THE COORDINATE SYSTEM FOR LOCATING THE INBOARD AND OUTBOARD LEADING EDGE POINTS (1 AND 4). NSPAN AND NCHORD, OR LSPAN AND LCHORD, ARE USED TO PARTITION THE WING INTO AERODYNAMIC PANELS, THE FORMER FOR UNIFORMLY SPACED PANELS AND THE LATTER FOR NON-UNIFORMLY SPACED PANELS. IGID IS THE ID OF ITS ASSOCIATED INTERFERENCE GROUP. THE CONTINUATION ENTRY DEFINES POINTS 1 AND 4, THE ROOT CHORD AND TIP CHORD. THE BOXES FORMED BY THE GRID LINES WILL BE NUMBERED BEGINNING WITH EID SO CHOOSE A NUMBER THAT IS UNIQUE, AND IS GREATER THAN ALL STRUCTURAL GRID, SCALAR AND EXTRA POINT IDS. EID 1100 PID 1000 CP * RIGHT WING * NSPAN NCHORD LSPAN 8 4 LCHORD IGID 1 CHORD X14 10.

$ $ $ $ $ $ $ $ $ $ $ $ $ $

( FWD LEFT POINT ) X1 Y1 Z1 25. 0. 0.

CHORD ( FWD RIGHT POINT ) X12 X4 Y4 Z4 10. 13.45299+20. 0.

$ THE PAERO1 ENTRY IS REQUIRED EVEN THOUGH IT IS NON-FUNCTIONAL $ (BECAUSE THERE ARE NO ASSOCIATED BODIES IN THIS EXAMPLE). $ $ PID B1 B2 B3 B4 B5 B6 1000 $ * LEFT WING * 2100 1000 8 4 1 13.45299-20. 0. 10. 25. 0. 0. 10. $ $ $ 1 5. 0. 10. $ 2000 10. 1000 -5. 0. * LEFT SIDE * 2 4 10. 10. 0. 0. 1 10. $ $ $ 1 0. 0. 10. $ $ $ $ $ $ $ $ $ $

* CANARD AERODYNAMIC MODEL * 1000 10. 1000 0. 0. * RIGHT SIDE * 2 4 10. 10.

* FIN AERODYNAMIC MODEL * 3100 1000 10. 4 10. 4 25.

30.7735 0.

* * SPLINE FIT ON THE LIFTING SURFACES * * * BEAM SPLINE FIT ON THE WING * THE SPLINE2 ENTRY SPECIFIES A BEAM SPLINE FOR INTERPOLATION OVER THE REGION OF THE CAERO ENTRY (ID1 AND ID2 ARE THE FIRST AND LAST BOXES IN THIS REGION). SETG REFERS TO A SET1 ENTRY WHERE THE STRUCTURAL GRID POINTS ARE DEFINED. DZ AND DTOR ARE SMOOTHING CONSTANTS FOR LINEAR

$ $ $ $ $ $ SPLINE1 SET1 122 $ $ SPLINE1 SET1 222 $ $ $ $ SPLINE2 +SPRC +SPRC $ SET1 $ $ SPLINE2 +SPLC +SPLC $ $ $ SPLINE2 +SP2FI +SP2FI $ SET1 $ $ $ $ $ $ $ $ $ CORD2R +CRD2RW $ +CRD2RW $ $ $ $ CORD2R +CRD2LW $ +CRD2LW $ $ $

ATTACHMENT AND TORSIONAL FLEXIBILITIES. CID IDENTIFIES THE CORD2R ENTRY THAT DEFINES THE SPLINE AXIS. DTHX AND DTHY ARE ROTATIONAL ATTACHMENT FLEXIBILITIES (-1. SPECIFIES NO ATTACHMENT). 1601 1100 1100 99 1100 100 * RIGHT WING * 1131 1100 111 110 0. 112 TPS 121

$ $ $ $ $ $

120 $ $

2601 2100

2100 99

* LEFT WING * 2100 2131 2100 100 211 210

0. 212

TPS 221

220 $ $ $ $

* BEAM SPLINE FIT ON THE CANARD * 1501 1. 1000 2501 1. 1000 -1. * RIGHT SIDE * 1000 1007 1000 0. 1. 1

$ 98 2000 -1. * BEAM SPLINE FIT ON THE FIN * 3100 -1. 3100 3100 -1. $ 99 100 311 312 $ $ $ $ $ $ $ $ THE CORD2R ENTRY DEFINES THE COORDINATE SYSTEM IN WHICH THE BEAM SPLINE EXTENDS ALONG THE WING Y-AXIS. IT LISTS THE ORIGIN, A POINT ALONG THE Z-AXIS AND A POINT IN THE X-Z PLANE. * RIGHT WING SPLINE AXIS * CID 2 CS 0 A1 30. C3 0. * LEFT WING SPLINE AXIS * CID 20 CS 0 A1 30. C3 0. * FIN SPLINE AXIS * $ $ $ A2 0. A3 0. B1 30. B2 0. B3 10. $ $ $ A2 0. A3 0. B1 30. B2 0. B3 10. 3100 3115 3100 0. 1. 300 $ $ $ 99 * LEFT SIDE * 2000 2007 1000 $ $ 0. 1. 1

C1 C2 38.66025+5.0

C1 C2 38.66025-5.0

CORD2R +CRD2FI +CRD2FI $ $ $ $ $ $ $ $ $ AESURF AESURF AESURF $ $ $ $ $ AELIST AELIST AELIST AELIST AELIST $ @ 31724;

300 20.0

0 0.0

30.0 5.7735

0.

0.

30.0

10.0

0. $ $ $ $ $ $ $ $

* CONTROL SURFACE DEFINITION * THE AESURF ENTRY DEFINES AN AERODYNAMIC CONTROL SURFACE. LISTED ARE THE ALPHANUMERIC NAME OF THE SURFACE, THE ID OF A COORDINATE SYSTEM THAT DEFINES THE HINGE LINE AND THE ID OF AN AELIST ENTRY. ID 505 517 518 LABEL ELEV AILERON RUDDER CID1 1 110 301 ALID1 1000 1100 3000 CID2 1 210 ALID2 2000 2100

THE AELIST ENTRY LISTS AERODYNAMIC BOXES THAT LIE ON THE CONTROL SURFACE. SID 1000 2000 1100 2100 3000 E1 1000 2000 1119 2103 3103 E2 THRU THRU 1123 2107 3107 E3 1007 2007 1127 2111 3111 ETC 1131 2115 3115

$ $ $ $

; Optimization/Design Sensitivity - OFREQ - Wrong Answers If sensitivity is requested in a frequency response analysis and there is an OFREQ request to limit output, the sensitivity results will be in error. Avoidance: Remove the OFREQ request from case control.

@ 31738; p

70.7; Optimization/Design Sensitivity - DRESP1 - Possible Infinite Loo In optimization analysis, if none of the connectivity cards reference existing property entries and a dresp1 selects an existing property card for an element response, it is possible to enter into an infinite loop in subroutine DOMLOC of module DOPR3. Avoidance: Correct the missing property entries or the incorrect connectivity references.

@ 31751;

2001; Nonlinear Analysis - CBUSH - Wrong Answers This error points out two problems: 1. A code problem where the CBUSH has a secant logic for load deflection which can not properly track the curve. 2. A user problem with use of the field 9 CID. The CID sets the element axes in cement. The element then becomes sensitive to the order of the grids when computing an internal force. The relation is Ke * (Ub - Ua) If Ua > Ub a compressive force will result. This is unlike the G0 option which flips the element axies if the grids are interchanged.

31757; 70.5.1; Transient Analysis - TRLG - UFM 2025 The following fatal message may be incorrectly encountered in the TRLG module: *** USER FATAL MESSAGE 2025 (TRLG) UNDEFINED COORDINATE SYSTEM ID = 1065353216 0FATAL ERROR Avoidance: The error exists in Version 69.1, 70. and 70.5. The best avoidance is to request MSC/NASTRAN Version 70.5.1. The correction of this error has been integrated in to Version 70.0.5 for NASTRAN for Windows. The fixed version remains at 70.5.1 since this error was never an issue for the prior release of NASTRAN for Windows which was based on Version 68.2. @ 31759; 2001; Documentation - RELEASE - UFM 6755 The documentation for the RELEASE entry does not state that it is not supported if the model has PART superelements. If a model containing PART superelements contains a RELEASE entry, the program will issue UFM 6755 and terminate the run. @ 31763; 70.5.1; Eigenvalue Analysis - EIGRL, READ - SFM 7340 (LNNHERR), LNNDRVD In the rare case that Lanczos shifts directly on a mode, the new lanczos shift logic in MSC/NASTRAN Version 70.5 may encounter an error which may cause the job to fail with messages similar to those below instead of performing a new shift. *** USER WARNING MESSAGE 4698 (DCMPD) STATISTICS FOR DECOMPOSITION OF MATRIX (NONE) . THE FOLLOWING DEGREES OF FREEDOM HAVE FACTOR DIAGONAL RATIOS GREATER THAN 4.85302E+11 OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL. *** SYSTEM FATAL MESSAGE 7340 (LNNHERR) process error reported by subroutine LNNP2CD see LNNP2CD abstract (ier = -729) *** SYSTEM FATAL MESSAGE 7340 (LNNHERR) warning reported by subroutine LNNDRVD see LNNDRVD abstract (ier = 729) Avoidance: The simplest, most guaranteed avoidance is set system(273)=1 to select the Version 70 Lanczos shift logic. @ 31769;

70.5; Online Encyclopedia - RANDPS - RANDPS is Missing from Index The RANDPS index is missing from the Table of Content section of the online documentation. The actual RANDPS entry is in the online CD. Avoidance: Click on the RANDT1 button and scroll one page up to find the RANDPS entry.

@ 31775; 70.5.1; Utilities - PLOTPS - Fail with UNRECOVERABLE library error On all Cray platforms, PLOTPS will fail with the following error message

lib-1171 plotps: UNRECOVERABLE library error An output list item is incompatible with its data edit-descriptor. Encountered during a sequential formatted WRITE to an internal file (character variable) Error initiated at line 647 in routine '_wrfmt'. Abort Beginning of Traceback: Interrupt at address 74547a in routine '_lwp_killm'. Called from line 16 (address 75571a) in routine 'raise'. Called from line 92 (address 15347c) in routine 'abort'. Called from line 59 (address 352112a) in routine '_ferr'. Called from line 647 (address 424651c) in routine '_wrfmt'. Called from line 232 (address 343053c) in routine '_xfer_iolist'. Called from line 396 (address 423231c) in routine '_FWF'. Called from line 1483 (address 12245c) in routine 'SYMBOL'. Called from line 620 (address 4641b) in routine 'DRWBCD'. Called from line 1225 (address 10656a) in routine 'PLTFIL'. Called from line 358 (address 3532c) in routine 'PLOTPS'. Called from line 316 (address 36334b) in routine '$START$'. End of Traceback. Abort(coredump) Avoidance: Change line 1529 in <install-dir>/msc705/util/plotps.F from 50 to 50 format(f6.0,3i6,' s',i2.2) @ 31778; ; Nonlinear Analysis - DECOMP - UWM 4698 Does not Identify Degrees of Freedom When running sparse decomposition in solution 106, the message: USER WARNING MESSAGE 4698, STATISTICS FOR DECOMPOSITION OF MATRIX ***. THE FOLLOWING DEGREES OF FREEDOM HAVE FACTOR DIAGONAL RATIOS GREATER THAN *** OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL" does not list the grids/dofs corresponding to the negative diagonal terms. Avoidance: Run the job with SPARSE=1 to identify the negative diagonal terms. @ 31784; 70.5.2; Nonlinear Analysis - CBUSH, PBUSHT - Missing CBUSH STRESS/FORCE Output In a SOL 106 run, if a CBUSH element is present using the nonlinear force-deflection feature (KN on the PBUSHT entry) and there is no other nonlinear material present, then the STRESS/FORCE output for the CBUSH element may be missing. This has been observed on some, but not all problems. Avoidance: Add param,lgdisp,1 to the input file. A safer avoidance would format(4i6,' s',i2.2)

be to define a dummy rod or quad4 element with an associated MATS1 entry @ 31787; 70.5.1; Heat transfer Analysis - CHBDYG, CHBDYP - Wrong Answers If the user has nonunique CHBDYP element ID that are the same as a CHBDYG element ID, the NASTRAN program does not know which CHBDYx element ID should be used in the convection calculation. The fix will allow the program to flag the non-unique CHBDYP and CHBDYG id's in the gp2 module. Avoidance: Change the conflicting CHBDYP and CHBDYG element id's so that they are unique. @ 31796; 70.5.1; P-Elements - ADAPT, EPSTOL - Wrong Answers When a shell or beam p-element is omitted from the error analysis because of the EPSTOL parameter on the ADAPT bulk data entry, all subsequent shell and beam p-elements will also be omitted. Avoidance: An avoidance is to set the EPSTOL parameter to zero on the ADAPT entry so that no elements are omitted. @ 31801; 70.5.1; P-Elements - ADAPT, EPSTOL - Poor Answers If a shell be omitted ADAPT bulk have a low or beam p-element has a low stress at one grid, the element may from the error analysis because of the EPSTOL parameter on the data entry. The element should not be omitted unless all grids stress.

Avoidance: An avoidance is to set the EPSTOL parameter to zero on the ADAPT entry so that no elements are omitted. @ 31804; 70.5.2; Thermal Loads - PSHELL - Wrong Stresses If the T1 and T2 fields are specified on the TEMPP1 entry and the MID1 field on the PSHELL record is left blank, then the calculated element stress will be wrong. Avoidance: Specify the MID1 field on the PSHELL entry. @ 31814; 70.5.1; Matrix Methods - MPYAD Method 1 NT/T Storage 1 - Memory Allocati on Error Runs on aix platforms with large multiply-adds, in which Method 1 NT/T Storage 1 is selected, may unexpectedly run out of virtual memory. Encountering the error depends on such variables as the size of the problem and the amount of system paging space available on the machine. When the error occurs, the run terminates with the following error message: SYSTEM ERROR MESSAGE (gotsig33): SYSTEM RAN OUT OF PAGING SPACE. Avoidance: Note that when this error is encountered, increasing the memory allocation with the NASTRAN mem parameter will not generally avoid the problem. The avoidance is to increase the system paging space. @

31817; 70.5.2; Dynamics - PARAM,VMOPT,1 - Inefficient Processing If PARAM,VMOPT,1 is specified then the MGEN module is executed twice. Also, if component modes are not requested then the fluid mass should not be computed at the g-set level; in other words, PARAM,VMOPT,1 should be ignored. Avoidance: Insert the following DMAP alter: compile semrm alter 'mgen '(,-1) if ( vmopt=0 or not((DYNRED=-1 or METHOD=-1) and noqset>-1) ) then $ alter 'vmopt'(2),'' alter 'call mgen1'(,1) endif $ NOMGEN>-1 alter 'PURGEX \/mvaa','' dbstatus mej//s,n,nomej $ if ( nomej>0 ) then $ alter 'mpyad meo'(,-1) CALL SEGOA uset,GOT,GOQ/GOA/NOOSET/NOQSET/NOTSET $ alter 'mpyad meo'(,1) else $ purgex /mea,,,,/-1 $ endif $ nomej>0 alter 'ENDIF \$ NOMGEN','' compile semg alter 'vmopt'(2),'' type db uset0 $ PARAML CASES//'DTI'/-1/150//S,N,DYNRED $ PARAML CASES//'DTI'/-1/5 //S,N,method $ PARAML USET0//'USET'//////'q'/S,N,NOqSET $ IF ( vmopt=1 and (DYNRED=-1 or METHOD=-1) and noqset>-1 ) THEN $ endalter @ 31832; ; Miscellaneous - PARAM,USETPRT - Wrong Sort If there is a user defined degree of freedom set (USET, USET1), and row sorting is selected (PARAM,USETPRT,0 or 10), then the uset print tables will not be sorted in ascending order, but will be in a random order. Avoidance: Select 'Column'(val=1 or 11) or 'Row and Column'(val=2 or 12) sort on the PARAM,USETPRT entry. @ 31842; 70.7; Miscellaneous - INCLUDE, SOL - UFM 443 If the SOL executive statement points to a user-written dmap, and the dmap is inserted in front of the SOL statement using the include command, then the job may fail with the following messages on some machine. *** USER FATAL MESSAGE 443 (DBALOC) THE SUBDMAP MAIN REFERENCED BY THE SOL, LINK OR COMPILE STATEMENT WAS NOT FOUND ON THE MSCOBJ DBSET. Avoidance: Either move the SOL statement before the include command or insert the dmap into the actual file instead of using the include command. @ 31844; 70.7; Aeroelastic Anaslysis - AEKX, AESMETH=ITER - Slow Convergence When using AESMETH=ITER or AESMETH=AUTO in SOL 144 and the displacement and force splines are not the transpose of one another (e.g., if the USAGE

field is not BOTH on all SPLINEi entries), the iterative aeroelastic solution may converge slowly or even fail to converge. This is caused by an error in the AEKX subdmap in which it was mistakenly assumed that the splines were transposes of one another (as they had been required to be prior to Version 70.5). If the force and displacement splines are transposes of one another, this problem does not occur. Avoidance: An avoidance is to use identical FORCE and DISP splines (which was a requirement prior to Version 70.5). Another avoidance is to apply the following alter to Version 70.5: compile aekx list noref $ $ fix errors due to mqg' and g being different matrices. $ COMPUTE P = KS*U + MQG*Q*G*U = KS*U + KA*U, $ AVOIDING COMPUTING OF Ka FOR REASONS OF EFFICIENCY $ IF TRANSP=1, ka' IS USED IN PLACE OF ka $ alter '/ksu','/p ' $ if (transp=0) then $ MPYAD G,U,/UK $ MPYAD Q,UK,/PK $ Q IS K BY K AND DENSE MPYAD MQG,PK,/kau $ else $ mpyad mqg,u,/uk/1 $ ka'*u = g'*q'*mqg'*u mpyad q,uk,/pk/1 $ mpyad g,pk,/kau/1 $ endif $ MPYAD KS,U,kau/p $ end of aekx.v705 alter endalter @ 31849; 70.5.1; Submittal - nastran.ini on Windows NT - Application Error If the last line of the nastran.ini file (found in the $MSC_BASE\msc705\<arch>\ directory) is not terminated, then the NASTRAN command will fail with the following message generated by Windows: An application error has occured and an application error log is being generated nastran.exe Exception: access violation (0xc0000005), Address: 0x0041c608 This error is easily created with Notepad. Note that the ini file is now only used to set a very small number of keywords, e.g., msc_base, and version. In most other cases the rcf files are now used. If other keywords are set in the nastran.ini file they are ignored. Avoidance: Put an end-of-line at the end of the last line in the nastran.ini file. @ 31855; ; Documentation - CGAP, PARAM,G - Unreasonably Large Damping With penalty CGAP (default), W3 damping can become excessive because of large CGAP penalty stiffness. This is a limitation that should be documented under the CGAP entry.

Avoidance: When gap elements are used in nonlinear transient analysis, param,g damping should be avoided. Damping should be specified on the MATi entries and PARAM,W4 set. @ 31884;

70.7; Optimization/Design Sensitivity - ADD5, PARAM, DSNOKD - UFM 5423 The message "USER FATAL MESSAGE 5423" indicating an incompatible matrix multiple is output from the ADD5 module in SOL 200 when PARAM DSNOKD is non-zero and there are multiple subcases that have active buckling design constraints. Avoidance: Set DSNOKD to 0.0 or restrict the buckling design task to a single subcase.

@ 31885;

70.7; Rigid Elements - RBE3 - UWM 7343 Diagnostic USER WARNING MESSAGE 7343 for the RBE3 element appears, properly, when an RBE3 element has inadequate connectivity and is approaching singularity. It may also appear for a stable element when there is a large distance between the grid points to which it attaches. This is because the test used to measure singularity is not independent of the units of measure of the model. A test model of an rbe3 connecting 4 grid points in a square pattern to a point at their center does not cause the message when the points are 100 units apart, but does cause the message when they are 10,000 units apart on certain computer types. Avoidance: The criterion used to print this message is that the computational zero ("error value") associated with its use is greater than 1.E-10. This conservative value was chosen because a related fatal message (new in Version 70.5) and client feedback had not been obtained with its use. If the error value is less than 1.E-6 it may be safely ignored. The warning message can generally be avoided if the grid points to which the element is attached are moved, or one is added, which allows them to approach a description of a plane, rather than a condition approaching a line. If the grid points are widely separated the only method known to avoid the message is to use a different unit of measure for grid point locations.

@ 31890; 70.5.2; Rigid Elemets - RBE3 - UFM 2038, Only one illegal RBE3 found per run All bad rigid elements should be found and identified in a single execution. However, the RBE3 is an exception. It causes the program to stop after the first bad element is found. This is a particular problem with Version 70.5 because improved logic for detecting singular RBE3 elements was introduced in this version, and the program may fail RBE3 elements that earlier versions passed. Avoidance: The error has been corrected in Version 70.5.2. Another avoidance is to run the model in Version 70 with param, checkout yes present. This makes tests on the model including equibrium tests on the rigid element constraint matrix. Elements which will pass tests in Version 70 but fail with singularity messages in Version 70.5 are likely to show poor equilibrium properties in the param, checkout tests. Look for output of the EMH matrices. It will identify dependent DOFs that do

not meet equilibrium checks, which may be tied back to RBE3 elements. All elements which fail will be identified in Version 70. This technique will not work in Version 70.5.0 because the program will take a fatal exit before the checkout operations begin. @ 31891; 2001; Documentation - GPSTRESS - No Printout for Midside Nodes MSC/NASTRAN does not printout GPSTRESS for midside nodes. This limitation is documented in the Application Manual but since that manual is no longer current a remark should be added to the Quick Reference Guide. @ 31898; 70.5; Frequency Response - FREQ - UWM 6498 In a frequency response analysis run (SOL 108 and 111), if the FREQ Case Control command does not exist, then the job will terminate with the following unclear error messages on some machines. *** USER WARNING MESSAGE 6498 (MsgGetStr) Message number is not found PROGRAMMER INFORMATION: Check call to MSGPOP from FRLG with the message number 3031 Severity is USER WARNING Parameter 1 is 0 Parameter 2 is FRL Parameter 3 is FRRD USER ACTION: Please contact MSC Client support 0FATAL ERROR On other machines, the job may terminate with a signal 11 error in the log file. Avoidance: Add the appropriate FREQ case control command. @ 31906; 70.5.2; Nonlinear Analysis - DBC - UFM 4551 If a nonlinear analysis fails to converge and PARAM,POST,0 is set, the job will issue fatal messages indicating that the analysis has not converged but then will continue and will fail in the DBC module with a machine dependent message. An example of the messages are: *** USER INFORMATION MESSAGE 6194, *** STOPPED ITERATIONS DUE TO REACHING MAXIMUM ITERATION LIMIT WITHOUT CONVERGENCE *** *** USER INFORMATION MESSAGE 6193, *** MAXIMUM NUMBER OF BISECTIONS OR MINIMUM LOAD STEP HAS BEEN REACHED. *** USER FATAL MESSAGE 4551, *** STOPPED PROBLEM DUE TO FAILED CONVERGENCE ^^^ USER INFORMATION MESSAGE 9005 (NLSTATIC) ^^^ THE SOLUTION FOR THE LAST CONVERGED LOOPID= 85 IS SAVED FOR RESTART ***** RECEIVED SIGNAL 8, ABORTING ... BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX (FILEX) = 0 STATUS = 0 *** SYSTEM FATAL ERROR 4276.RECOVR ERROR CODE 7777 PID= 0 *** USER INFORMATION MESSAGE: TO OBTAIN A NASTRAN DUMP, RESUBMIT JOB WITH

DIAG 44 INSERTED IN THE EXECUTIVE CONTROL SECTION. On other computers the f06 file ends abruptly after UFM 4551, the f04 file ends abruptly after a DBC call and a system message occurs in the log file such as the following from a SUN computer: *** TERMINATING /app/msc/msc70/solaris/analysis *** Received signal 11 (SIGSEGV) time: command terminated abnormally. Avoidance: The failure in dbc does not affect the rest of the run. The analysis needs to be modified to achieve converge. @ 31919; 70.7; Installation - Sun/Solaris - fatal : libsunmath.so.1: can't open file errno = 2 For Sun SuperSPARC V70.5 and v70.6 and Sun UltraSPARC V70.0.3 Clients only, If you do not have a Fortran compiler on your system, you may receive the following error message in the LOG file when you attempt to run MSC/NASTRAN: ld.so.1: <install-dir>/msc705/solaris/analysis: fatal : libsunmath.so.1: can't open file errno = 2 This error occurs when the system cannot find the "libsunmath" dynamic library. This error is corrected on the MSC/NASTRAN V70.5.1 (r2) CD-ROM. You can also obtain the missing libraries from your MSC representative or directly from MSC's Web site at the URL: ftp://ftp.macsch.com/pub/msc-products/NASTRAN/V70.5/solaris/libsunmath.so 1 Once you have obtained the file, please copy it to the solaris architecture directory, i.e., <install-dir>/msc705/solaris where <install-dir> is your MSC/NASTRAN V70.5 installation base directory. Avoidance: Download the libsunmath.so.1 file from the ftp site. @ 31931; 70.6; Acoustics - ACMODL, CTETRA, CTRIA3, GP5 - SFM 6169 Some models with TRIA3 structural elements and TETRA fluid elements at the fluid/structure interface using IDENT method on the ACMODL bulk data entry fail in the GP5 module with a message similar to: *** USER FATAL MESSAGE 6169, FOUND TWO OR MORE STRUCTURAL GRID POINTS WITH IDENTICAL COORDINATES WHICH LIE ON THE FLUID/STRUCTURE BOUNDARY. USER ACTION: CHECK GRIDS AND/OR SET1 CARDS WHICH BELONG TO PANEL OR ACMODL CARDS.

GRID ID = 12 GRID ID = 101979 ^^^ USER WARNING MESSAGE 9031 (ERRPH1) SEMG

NOGO ENCOUNTERED IN SUBDMAP

This message is in error since one of the referenced structural grids is really a fluid grid. Abitrary changes to the model can make it run or not run. Avoidance: Use the DIFF method on the ACMODL bulk data entry. @ 31933; ; Documentation - MATGEN Option 5 - DMAP Module Dictionary The documentation for MATGEN option 5 states that the parameter p5 may be used to select pseudo-random or non-repeating random numbers, but the equation used is incorrect. It states that if P5<0, a matrix of random numbers that do not repeat from call to call is produced. By implication, values outside this range produce pseudo random numbers. The usual definition of this term is that while the numbers produced should have reasonable distributions for a set of random numbers, the matrix produced on each call should be the same, both on one run, and on different runs. The correct formula is "if p5 is less than or equal to 0" the non-repeating ("seeded") type of random numbers are used. Note that this changes the action for the default value of p5, which is zero.

@ 31934;

2001; Documentation - RFORCE - Clarification of the CD Field The content description of the CID and R1, R2, R3 fields for the RFORCE entry currently states CID must be a Caresian system. It should state that if a non-rectangular system is specified, RFORCE will use the CID as if it were a rectangular system and unexpected answers may result. The desrcription of CID, R1, R2, R3 should read as follows and remark 14 should be added.

CID otation vector. R1, R2, R3 ctor

Coordinate system used to define the components of the r (integer >= 0, Default = 0) Rectangular components of the rotation vector R. The ve

will pass through G, please see remark 14. (R1**2 + R2**2 + R3**2 > 0.0) Remark 14. If CID is not a rectangular coordinate system, RFORCE will treat it as if it were a rectangular system when defining the components of the rotation vector. @ 31958; ; Memory Management - REIGL, GPFDR - SFM 4275 and SFM 4276 On a large model, it has been observed that the job may fail in REIGL or GPFDR with the following unclear messages:

*** SYSTEM FATAL MESSAGE 4275. DURING A READ OPERATION, UNIT = 26, DATA BLOCK=NUMERIC , A BLOCK NUMBER CHECK CONDITION OCCURRED. ( IONAST CODE= 304) 0*IONAST* err 304 pid,name,b+1,z4(b+1),nxtblk,pbknbr= 0 118 62402415 0 3377 5912 BIOMSG: ERROR 304 HAS OCCURRED IN ROUTINE IONAST , FILE INDEX (FILEX) = 0 STATUS = 0 The exact cause of this error is currently unknown. MSC has been unable to obtain the input file that shows this error due to the confidentiality of the input file. Avoidance: Increasing memory cured this particular problem. @ 31966; ZREAD 70.7; Nonlinear Analysis - CBUSH, CELAS1, CELAS2, MATHP - UWM 6286 in If nonlinear stress output is requested for CBUSH, CELAS1, and CELAS3 elements which reference PBUSHT and PELAST in nonlinear analysis then UWM 6286 will be issued. This also occurs for hyperelastic element grid point output. Avoidance: Insert and/or replace the following changes into nddl.dat under DATABLK OES: OR,40,{ Bush1d } EITHER,BACK(FCODE)=0,{ Real } FE,RS,{Element Force} UE,RS,{Axial Displacement} VE,RS,{Axial Velocity* } S,RS,{Axial Stress*} E,RS,{Axial Strain*} EP,RS,{Plastic Strain*} FAIL,I,{Failed Element Flag} OR,1,{ Real / Imaginary } FER,RS,{Element Force} UER,RS,{Axial Displacement} SR,RS,{Axial Stress*} ER,RS,{Axial Strain*} FEI,RS,{Element Force} UEI,RS,{Axial Displacement} SI,RS,{Axial Stress*} EI,RS,{Axial Strain*} OR,216,{ Nonlinear form of hyperelastic 4 NODE TETRA } TYPE,CHAR4,{ } ENTRY, ID,I,{ } SX,RS,{ } SY,RS,{ } SZ,RS,{ } SXY,RS,{ } SYZ,RS,{ } SZX,RS,{ } PRESSURE,RS,{ } VOLSTR,RS,{ } EX,RS,{ }

EY,RS,{ } EZ,RS,{ } EXY,RS,{ } EYZ,RS,{ } EZX,RS,{ } ENDENTRY,COUNT=004, OR,217,{ Nonlinear form of hyperelastic 3 NODE Strain } TYPE,CHAR4,{ } ENTRY, ID,I,{ } SX,RS,{ } SY,RS,{ } SZ,RS,{ } SXY,RS,{ } PRESSURE,RS,{ } VOLSTR,RS,{ } EX,RS,{ } EY,RS,{ } EZ,RS,{ } EXY,RS,{ } ENDENTRY,COUNT=003, OR,218,{ Nonlinear form of hyperelastic 20 NODE } TYPE,CHAR4,{ } ENTRY, ID,I,{ } SX,RS,{ } SY,RS,{ } SZ,RS,{ } SXY,RS,{ } SYZ,RS,{ } SZX,RS,{ } PRESSURE,RS,{ } VOLSTR,RS,{ } EX,RS,{ } EY,RS,{ } EZ,RS,{ } EXY,RS,{ } EYZ,RS,{ } EZX,RS,{ } ENDENTRY,COUNT=008, OR,219,{ Nonlinear form of hyperelastic 9 NODE Strain } TYPE,CHAR4,{ } ENTRY, ID,I,{ } SX,RS,{ } SY,RS,{ } SZ,RS,{ } SXY,RS,{ } PRESSURE,RS,{ } VOLSTR,RS,{ } EX,RS,{ } EY,RS,{ } EZ,RS,{ } EXY,RS,{ } ENDENTRY,COUNT=004, OR,220,{ Nonlinear form of hyperelastic 15 NODE } TYPE,CHAR4,{ } ENTRY, ID,I,{ } SX,RS,{ }

SY,RS,{ } SZ,RS,{ } SXY,RS,{ } SYZ,RS,{ } SZX,RS,{ } PRESSURE,RS,{ } VOLSTR,RS,{ } EX,RS,{ } EY,RS,{ } EZ,RS,{ } EXY,RS,{ } EYZ,RS,{ } EZX,RS,{ } ENDENTRY,COUNT=006, OR,221,{ Nonlinear form of hyperelastic 10 NODE } TYPE,CHAR4,{ } ENTRY, ID,I,{ } SX,RS,{ } SY,RS,{ } SZ,RS,{ } SXY,RS,{ } SYZ,RS,{ } SZX,RS,{ } PRESSURE,RS,{ } VOLSTR,RS,{ } EX,RS,{ } EY,RS,{ } EZ,RS,{ } EXY,RS,{ } EYZ,RS,{ } EZX,RS,{ } ENDENTRY,COUNT=004, OR,222,{Nonlinear form of hyperelastic 3 NODE Axisym } TYPE,CHAR4,{ } ENTRY, ID,I,{ } SX,RS,{ } SY,RS,{ } SZ,RS,{ } SXY,RS,{ } PRESSURE,RS,{ } VOLSTR,RS,{ } EX,RS,{ } EY,RS,{ } EZ,RS,{ } EXY,RS,{ } ENDENTRY,COUNT=003, OR,223,{ Nonlinear form of hyperelastic 9 node Axisym } TYPE,CHAR4,{ } ENTRY, ID,I,{ } SX,RS,{ } SY,RS,{ } SZ,RS,{ } SXY,RS,{ } PRESSURE,RS,{ } VOLSTR,RS,{ } EX,RS,{ } EY,RS,{ }

EZ,RS,{ } EXY,RS,{ } ENDENTRY,COUNT=004, OR,224,{ Nonlinear ELAS1 } F,RS,{ Force } S,RS,{ Stress } OR,225,{ Nonlinear ELAS3 } F,RS,{ Force } S,RS,{ Stress } OR,226,{ Nonlinear BUSH } FX,RS,{ Force X } FY,RS,{ Force Y } FZ,RS,{ Force Z } STX,RS,{ Stress Translational X } STY,RS,{ Stress Translational Y } STZ,RS,{ Stress Translational Z } ETX,RS,{ Strain Rotational X } ETY,RS,{ Strain Rotational Y } ETZ,RS,{ Strain Rotational Z } MX,RS,{ Moment X } MY,RS,{ Moment Y } MZ,RS,{ Moment Z } SRX,RS,{ Stress Rotational X } SRY,RS,{ Stress Rotational Y } SRZ,RS,{ Stress Rotational Z } ERX,RS,{ Strain Rotational X } ERY,RS,{ Strain Rotational Y } ERZ,RS,{ Strain Rotational Z } @ 31970; lement ; Superelements - CSUPEXT Ignored if no Elements Exist in a Supere If there are no elements inside a superelement, but there are interior points, and CSUPEXT entries are used to define the exterior points, the CSUPEXT entries will be ignored. Avoidance: Add PLOTEL elements between interior and exterior points. @ 31980; 70.5.1; Nonlinear Analysis - NLCOMB - Traceback If a problem has BEAM, BAR, or BEND elements and thermal loads in a nonlinear solution sequence (106 and 129), the job may fail in module NLCOMB. This error first occurred in Version 70.5. This has been observed only on the Cray, but might occur on other systems. On the CRAY the following traceback was encountered: Beginning of Traceback: Called from line 42 (address 15517150b) in routine 'TRBK'. Called from line 39 (address 132053d) in routine 'RECOVR'. Called from line 112 (address 140623d) in routine 'sysrcv'. Called from address 15237006a in routine '__handlr'. Called from address 5616165c in routine 'NLCWRK'. Called from line 126 (address 5607145c) in routine 'NLCOMB'. Called from line 90 (address 10632226d) in routine 'XSEM15'. Called from line 363 (address 1734c) in routine 'MAINAL'. Called from line 306 (address 3647a) in routine '$START$'. End of Traceback.

Avoidance: There is no known avoidance. @ 31982; 70.6; SSSALTERs - checka.v70 - UFM 13 In the checka.v70 and checka.v705 alters, there are two lines which exceed 80 columns. If these alters are pasted into the data file, the line will wrap in many editors, and cause the following fatal message: 0*** USER FATAL MESSAGE 13 (XGPIDG ) ERROR IN DMAP INSTRUCTION Q-SET INSTRUCTION NO. 50 DMAP INSTRUCTION NOT IN MODULE LIBRARY. Avoidance: Shorten the comments following the DMAP instructions "type db,aagds,ptga $", and "upartn uset0,aagds/ptga,,,/'G'/'A'/'S'/1 $". Another option is to avoid pasting the file in, but instead use the include statement.

@ 31986; 70.5.2; Superelements - DBDN DBset Too Large The DBDN DBset is larger in V70 and V70.5 than in earlier versions of MSC/NASTRAN. The increased size is due to several data blocks which are generated during dynamic reduction being incorrectly stored on DBDN. The only effect of this error is an increase in the DBDN Dbset disk usage. Avoidance: Insert the following DMAP alter: compile semr3 alter 'call cmpmode' type parm,nddl,char8,n,dbupx,dbdnr1 $ type parm,,char8,n,dbtmp $ dbtmp=dbdnr1 $ dbdnr1=dbupx $ alter 'call sesum' dbdnr1=dbtmp $ compile sekr alter 1 type parm,nddl,char8,n,dbupx,dbdnr1 $ type parm,,char8,n,dbtmp $ dbtmp=dbdnr1 $ dbdnr1=dbupx $ alter 'end \$'(,-2) dbdnr1=dbtmp $ compile semrm alter 'mvaa'(2,-1) type parm,nddl,char8,n,dbupx,dbdnr1 $ type parm,,char8,n,dbtmp $ dbtmp=dbdnr1 $ dbdnr1=dbupx $ alter 'end \$'(,-2) dbdnr1=dbtmp $ endalter @ 31990; 2001; Documentation - PBUSH1D The MSC/NASTRAN Quick Reference Guide for Version 70.5 incorrectly states that SHOCKA, CVT, IDTS, IDETS, IDETSD, TYPE, IDT, IDTTU, IDTDV are required on the PBUSH1D entry. These fields may be required based on

values in other fields and options selected, but are not mandatory. @ 32011; 2001; Documentation - PBUSH - Element Property Direction Not Described The documentation for PBUSH does not say which coordinate system the "directions 1 through 6" refer to. These directions refer to the element coordinate system.

@ 32014; 70.5.2; Transient Analysis - NLRGAP, TRD1 - Poor Answers, Traceback Input files using the NLRGAP in SOL 109 may fail with a floating point exception on NT platforms or may issue a traceback in the TRD1 module, routine tabld on SGI R8000/R10000 systems. The error is caused by the table look up for the coefficient of friction for the NLRGAP. All 32 bit, short-word, machines will return an incorrect value for the coefficient of friction; however, this is not likely to have a significant effect on the results of most problems. Avoidance: Run the problem in solution 129. A different error with similar symptoms may also be encountered in the same area of the code, please see error number 32685. @ 32015; 112. 2001; Documentation - Release Guide - NLRGAP Does Not Work in Solution The MSC/NASTRAN Release Guide for Version 70.5 incorrectly states that the NLRGAP element works in SOL 112. The NLRGAP does not work in this solution, but it does work in SOL 9. @ 32016; 2001; GAP Constraints - PARAM,CDITER - UWM 9096, Wrong Answers If the iterative solution for GAP constraints in SOL 101 does not converge, a warning message is issued (9096), indicating that the solution has not converged; however, the results for displacements, etc. are still output. ^^^ USER WARNING MESSAGE 9096 (CSTRDISP) ^^^ CONSTRAINED DISPLACEMENT ITERATIONS FAILED, ITERATION LIMIT LOAD CASE = 1 Avoidance: If warning message 9096 is issued, do not use the results. Increasing the value of CDITER may help the problem to converge. @ 32022; 70.5.1; Miscellaneous - NLRGAP - Incorrect TPL Problem The tpl test deck nlrgap.dat referred is not the problem described in the Version 70.5 Release Guide. The tpl deck nlrgap.dat has been replaced with the correct input deck on the Version 70.5.1 r2 cd. A listing of the correct input follows: $ $ $ $ $ $ shaft1x.dat radial gap, .002 clearance, contact stiffness = 1.e6 constant example: rotating load on cantilevared shaft 40" shaft, circular x-section, 2" diameter shaft 10" diameter rotor 15 hz. imbalance

sol 129 time 80 cend title= 15 hz. rotating imbalance, radial gap set 1 = 9,10 set 2 = 1,8 set 3 = 1001,1002 $sdisp=3 $ epoint displacements (relative displacements) disp(plot) = 1$ NLRGAP grid displacements force(plot) = 2 nlload=1 $ NLRGAP load output nonlinear = 200 $ NLRGAP set ID tfl=900 $ transfer function ID for relative displacement check dload = 1 tstepnl = 21 $ 0 - .4 seconds output(xyplot) xgrid= yes ygrid= yes xtitle = Time (seconds) ytitle = x relative displacement epoint 1001 xyplot sdisp /1001(t1) ytitle = y relative displacement epoint 1002 xyplot sdisp /1002(t1) ytitle = X Nonlinear loads Grid 9 xyplot nonlinear /9(t1) ytitle = Y Nonlinear loads Grid 9 xyplot nonlinear /9(t2) $ytitle = RZ Nonlinear loads Grid 9 $xyplot nonlinear /9(r3) $ begin bulk param coupmass1 param grdpnt 0 param,post,-1 param,w4,419.0 $ $ dynamic loads simulating 15 hz. 3 lb. centripetal imbalance dload 1 1. 1. 10 1. 20 $ x direction darea 40 9 1 3.0 darea 41 9 2 3.0 $ cosine imbalance tload2 10 40 0.0 100.0 15.0 $ sine imbalance tload2 20 41 0.0 100.0 15.0 $ 0-.03 s tstepnl 21 300 0.0001 1 adapt + + + $MAT1 mat1 $PBAR $pbar + $+ pbar + + 1. $CBAR 0. EID 0. PID 1. GA GB -1. 0. X1,G0 0. X2 X3 -1. .5 1 0. 100 0. 3.14 .5 .8 -.5 .8 0. 1.6 0. 0 MID 100 PID 1

-90.0

+ E 3.0e7 MID 100 G .3 A 1.0 I1 .0491 NU .7-4 I2 .0491 J .0982 -.5 RHO A TREF 1.0 NSM GE

cbar 1 1 1 cbar 2 1 2 cbar 3 1 3 cbar 4 1 4 cbar 5 1 5 cbar 6 1 6 cbar 7 1 7 cbar 8 1 8 $ grid 1 0.0 grid 2 0.0 grid 3 0.0 grid 4 0.0 grid 5 0.0 grid 6 0.0 grid 7 0.0 grid 8 0.0 $ shaft grid and rotor mass grid 9 0.0 $ 10" diameter, 1" thick disk conm2 109 9 .036 .036 $ Housing grid and housing mass grid 10 0.0 conm2 110 10

2 3 4 5 6 7 8 9 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 .0057

1. 1. 1. 1. 1. 1. 1. 1. 0.0 5.0 10. 15. 20. 25. 30. 35. 40. .071 123456

0.0 1.

40.

345

1. $ ground point for housing grid 11 0.0 0.0 40. 123456 $CELAS2 EID K G1 C1 G2 C2 GE S celas2 10 1.0+6 10 1 11 1 100. celas2 11 1.0+6 10 2 11 2 100. celas2 12 1.0+6 10 6 11 6 100. $ $***************************************************************** $NLRGAP SID GA GB PLANE TABK TABG TABU RADIUS nlrgap 200 9 10 XY 961 965 967 5.0 $ $ Constant contact stiffness vs time tabled1 961 0. 1.+6 10. 1.+6 endt $ $ Constant gap clearance vs time tabled1 965 0. 0.002 10.0 0.002 endt $ $ gap friction vs time, positive value assumes rotation $ about dof 3 from axis 1 toward axis 3 tabled1 967 0. 0.1 10. 0.1 endt $***************************************************************** $ gap dof relative displacements for verification epoint,1001 epoint,1002 $ tf 900 1001 1.0 10 1 1.0 9 1 -1.0 tf 900 1002 1.0 10 2 1.0

9 enddata @ 32047; mping

-1.0

70.7; Optimization/Design Sensitivity - Wrong Answers with Material Da Material damping defined on a material bulk data entry is ignored in an adjoint sensitivity analysis. Avoidance; If damping is a constant across all materials, param,g may be used. Otherwise the adjoint method needs to be avoided by forcing the direct sensitivity method using the following alter: COMPILE DESCON alter 'dsad'(,-1) $ type parm, ,logi,n,hold $ hold = pexist $ pexist = true $ alter 'dsad' pexist = hold $

@ 32070; 70.5.2; Differential Stiffness - SOLs 103, 107-112 - Wrong answers If QSETi, SEQSETi, or SENQSET entries are used to compute component modes in SOLs 103 and 107-112 and differential stiffness is requested with the STATSUB Case Control command, then the differential stiffness will not include the effects of the q-set degrees-of-freedom. Avoidance: Insert the following DMAP alter which is valid for Version 70.5 only: compile sekdr $ this alter is for Version 70.5 only alter 'call segoa','' type db goq $ CALL SEGOA USET,GOT,goq/GOAT/NOOSET/NOQSET/NOTSET $ endalter @ 32073; 70.5.2; Superelements - Partitioned SE - Wrong Answers or Fatal Messages If scalar points are defined in a partitioned superelement and the scalar points are also defined as exterior to the superelement (e.g.; QSETi) then the run may produce wrong answers and/or issue a fatal message. This error can also occur using the SENQSET entries. Avoidance: Insert the following DMAP alter: compile phase0 alter ',SILS,VGFS' delete /vgfs $ alter ',SILS,VGFS'(2) delete /vgfs $ endalter @ 32074; 70.5.2; Superelements - Data Recovery for Externals in SOLs 110-112 - UW M 9098

Performing data recovery using external superelements in SOLs 110, 111, and 112 with SEBULK type external superelements will result in user warning message 9098 indicating that the data necessary to perform data recovery is not available. The run will then fail in an unexpected manner. Avoidance : Include the following DMAP alter in the run which performs the residual structure solution compile super3 alter 'call extdrout'(4),'' if ( extdrout<>' ' ) then $ if ( app1='mmreig' ) then $ call extdrout uls,ol1m,eqexins,maps/ dum/app1/' '/true $ call extdrout uh2,ol1r,eqexins,maps/ dum/app/soltyp/true $ else $ call extdrout uls,ol1m,eqexins,maps/ dum/app1/soltyp/true $ endif $ endif $ extdrout<>' ' endalter $ @ 32080;

70.7; Complex Eigenvalue Analysis - CEAD - SFM 4276, EC 1605 On the SGI, complex Lanczos eigenvalue analysis may produce the unexpected system fatal message 4276 error code 1605. This appears to be caused by insufficient memory for the CEAD module. Avoidance: In most if not all cases experienced to date, increasing the memory has allowed the runs to run successfully.

@ 32087;

70.7; Frequency Response - TLOADi - Wrong answers TLOADi entries are ignored in direct frequency response (SOLs 8, 11, and 108). Avoidance: Insert the following DMAP alter in SOL 108: compile sedfreq alter 'CASE CASErs' type parm,,logi,n,fourier $ paraml dynamics//'bulk'//////'tload1'/s,n,tload1/ 'tload2'/s,n,tload2 $ PARAML CASEF//'DTI'/-1/38//S,N,TSTEP $ fourier=(tstep=-1 and (tload1=-1 or tload2=-1)) $ alter 'call freqrs','' CALL freqrs CASEF ,USETD ,DLT ,FRL ,trl , rpd ,DIT ,EST ,cstms ,mpts , bgpdts ,gpsnts ,GPECT ,SILS ,edts , ckdd ,cmdd ,cbdd ,k4dd , , , , ,sild ,ditid , vafs ,rgdfd ,rsgg ,rmgg ,ppf , fol , ,vgfd ,vgfs / UDF ,qmgf ,qsgf ,pdf / soltyp /freqdep /noa /fourier /false / 1 /false /0. /dompcffd/dospcffd/ fs /noue /0 $

alter 'call vdr1'(,-1) IF ( Fourier ) then $ Fourier transform delete /ppf,psf,,, $ app='tranresp' $ endif $ compile freqrs alter 'ift '(,-1),'' IF ( not(delta) and Fourier ) THEN $ endalter Insert the following DMAP alter for SOL 8 or 11, making sure to specify "SOL8" or "SOL11", respectively, on the COMPILE statement: COMPILE SOL8 or SOL11 alter 'lblgo'(,-1) paraml dynamics//'bulk'//////'tload1'/s,n,tload1/ 'tload2'/s,n,tload2 $ PARAML CASExx//'DTI'/-1/38//S,N,TSTEP $ IF ( tstep=-1 and (tload1=-1 or tload2=-1) ) THEN $ type parm,,i,y,iftm=0 $ IFT Uhv,CASExx,TRL,olst/ UHT1,TOL1/IFTM $ EQUIVX TOL1/olst/-1 $ EQUIVX UHT1/Uhv/-1 $ delete /pp,ps,pd,, $ soltype='tranresp' $ ENDIF $ endalter @ 32091; 70.7; Nonlinear Analysis - PCOMP, STRESS, STRAIN, PARAM,POST,0 - Missi ng XDB Output If PARAM,LGDISP,>0 and PARAM,POST,0 is specified in SOL 106, then laminar stress and strain output is missing from the .xdb file. Avoidance: Insert the following DMAP alter: compile nlstatic alter 'oescp,oestrcp' file oescpr=append/oestrcpr=append $ append oescp,/oescpr/2 $ append oestrcp,/oestrcpr/2 $ alter 'OUTPUT2 oesnlxr' IF ( POST=0 ) then $ DBC OEScpr,,,,,,,,,,,,,,,,,,,// 'OESC'/'CASECC'/'OL'////////////////// -1/DBCPATH/S,N,CP/app//GEOMU/LOADU/POSTU/ DBCDIAG/DBCCONV/DBCOVWRT $ DBC OEStrcpr,,,,,,,,,,,,,,,,,,,// 'OESC'/'CASECC'/'OL'////////////////// -1/DBCPATH/S,N,CP/app//GEOMU/LOADU/POSTU/ DBCDIAG/DBCCONV/DBCOVWRT $ endif $ post=0 endalter @ 32093; 70.5.1; Aeroelastic Analysis - CSSCHD - UFM 311 (IFPDRV) If a subcase invokes multiple CSSCHD bulk data entries, the user fatal message shown below incorrectly states that the CSSCHD ID's must be

unique. *** USER FATAL MESSAGE 311 (IFPDRV) NON-UNIQUE FIELD 2 ON BULK DATA ENTRY CSSCHD SORTED ENTRY COUNT = 153 10

Avoidance: There is no avoidance other than to limit the scheduled control surface functionality to a single control surface for each subcase. @ 32094; 70.7.2; Optimization/Design Sensitivity - Nonlinear Trim Analyis - BIOMS G: ERROR 100 Version 70.5 added a basic capability to perform nonlinear static aeroelastic trim analysis. This capability is not available in SOL 200 for sensitivity analysis and its use results in a nondescriptive fatal message such as BIOMSG: ERROR 100 HAS OCCURRED IN ROUTINE PACK

Avoidance: Either linearize the static aeroelastic trim analysis or do not attempt to perform this type of sensitivity analysis. Version 70.7.2 has been corrected to provide a meaningful message. @ 32098; 2001; Documentation, QRG - PSHELL, remark # 14 - typographical error There is a typographic error for Remark 14 of the PSHELL entry in the Version 70 and 70.5 Quick Reference Guide. The term (Ts)10**8(Ts) should simply be 10**8 instead. @ 32099; TPUT2 File ; Postprocessing - PARAM,POST,-2 - Multiple Geometry Written to OU

If multiple boundary conditions are specified in a normal modes analysis and param,post,-2 is set, the geometry data blocks are written to the output2 file in each subcase rather than just once. This unnecessarily increases the size of the output2 file. Avoidance: Specify PARAM,OGEOM,NO in all but the last subcase or insert the following DMAP alter: compile super3 alter 'call out2geom'(,-1) type parm,nddl,i,n,ncase $ if ( ncase<=1 ) then $ alter 'call out2geom' endif $ endalter @ 32101; 70.7; Plotting - XYPLOT of Grid Point Stress in SOL 101 - UWM 979 If an XYPLOT of grid point stress (param,curvplot,1) in solution 101 will lead to the following message: *** USER WARNING MESSAGE 979 (XYTRAN) AN XY-OUTPUT REQUEST FOR POINT OR ELEMENT ID xxx -STRE- CURVE IS BEING PASSED OVER. THE ID COULD NOT BE FOUND IN DATA BLOCK Avoidance: Insert the following alter:

COMPILE SEDRCVR $ALTER 'SORT2\=','' $ for V70 ALTER 'SORT2\='(2),'' $ for V70.5 SORT2=NOT(ANDL(NOSORT2,NOSTR2) OR CURVPLOT>-1) $ ENDALTER @ 32112; 70.7.2; Miscellaneous - MSCACT Missing Summary Output The MSCACT utility delivered with MSC/NASTRAN Version 70.5 on the UNIX and Windows NT systems has an error that disables the "standard" output summary. If one of the alternate summaries is selected via the "summary" keyword, the correct detail lines will be generated, but the total line will be missing. This error can be corrected by changing the following lines in the <install-dir>/msc705/util/mscact.c source file and rebuilding the MSCACT utility as described in Section 6 of the "Configuration and Operations Guide" for you operating system. NOTE: This error does not affect the accounting logger functions. The changes are: Change line 1339 from if( total == -1 && !strcmp( p, TOTAL_LABEL ) ) to if( total == EOL && !strcmp( p, TOTAL_LABEL ) ) Change line 1341 from else if( null == -1 && !strcmp( p, NULL_LABEL ) ) to else if( null == EOL && !strcmp( p, NULL_LABEL ) ) Change line 1386 from if( null > -1 ) to if( null != EOL ) Change line 1390 from if( total > -1 ) to if( total != EOL ) @ 32113; 2001; Miscellaneous - DMIG, DMI - Poor Messages The messages encountered when there is an error on a DMIG or DMI entry do not point to the location of the error. This can make it almost impossible to locate the error in a large matrix input. @ 32116; 70.5.1; Miscellaneous - ISHELL - Program Called by ISHELL Not Found If the program called by the ISHELL module has a name with fewer than 8 characters without including the extension, the module will not search properly for the extension. For example, the file named "EFM.EXE" will be found if the name "EFM.EXE" is entered, but will not be found if the name "EFM" is entered.

Avoidance: Use 8 characters in the filename if you do not include the extension. A related error that may occur on Windows NT if a 16-bit executable is to be run. If ISHELL processing fails with an "UNABLE TO CREATE A NEW PROCESS" error, specifying the standard NT command processor as the processor to be used may solve the problem. That is, setting the MSC_ISHELLEXT environment variable to exe#cmd, either in the System/Environment Control Panel tab or by a command line set option ("set MSC_ISEHHEXT=exe#cmd") will usually correct the problem, especially if the program can be run from a command prompt. Note: The MSC_ISHELLEXT capability will only be available starting with MSC/NASTRAN V70.5.1 (R2 CD). The short name problem has been corrected in MSC/NASTRAN V70.5, Round 11 @ 32125; 70.5.1; Submittal - DBINIT - Long filenames on SGI If a very long filename is used, either via the job datafile pathname, the dbs or out keywords, or an ASSIGN statement, the job may fail without any messages in the F06, F04, or LOG files. A clue to this error may be the termination of the job in the DBINIT module (as indicated in the F04 file) or incomplete reports for the "SUMMARY OF FILE ASSIGNMENTS" or "SUMMARY OF PHYSICAL FILE INFORMATION" or "SUMMARY OF PHYSICAL FILE I/O ACTIVITY" tables. Avoidance: Reduce the lengths of the longer pathname components. A pathname component is either a directory name (i.e., the text between slashes) or the actual filename part. @ 32127; 70.5.1; Aeroelastic Analysis - AELINK, CSSCHD - Wrong Answers If AELINK and CSSCHD entries appear is the same subcase in a static aeroelastic analysis, wrong answers are produced. Avoidance: There is no avoidance. @ 32134; 70.5.1; Installation - SGI IRIX64 Only When trying to run Version 70.5 on SGI IRIX 64 systems the following error message may appear in the command window: 3708:/msc/msc705/irix64/analysis: rld: fatal error: cannot successfuly map soname 'libcomplib.sgimath.so' under any of the filenames < then a list of library files follows >. Avoidance: If the client does not have a complete set of sgi math libraries they will need to get a /usr/lib64/mips4/libcomplib.sgimath.so from MSC. This can be found at the following locations: On the web at: ftp.macsch.com/pub/msc-products/NASTRAN/V70.5/irix64 filename: libcomplib.sgimath.so or the following computer file on MSC's net:

/net/oberon/oberon/disk1/msc705/unix/msc705/irix64/libcomplib.sgimath.so @ 32135; 70.6; Complex Eigenvalue Analysis - CEAD - Missing Left Hand Vectors The new CEAD module option for the CLAN method in Version 70.5 will produce both left and right hand eigenvectors. However, if there are less than three DOFs in the problem only the right hand vectors are output. Avoidance: An avoidance is to add a third DOF, uncoupled, if necessary. @ 32136; 2001; Documentation/QRG - PARAM,CURVE - Incorrect Equation The documentation for PARAM,CURV in the Quick Reference Guide indicates that if: n dot lm >= .4, then the projection of the y-axis on the surface of the element is taken as the reference axis. The left hand side of the inequality should be squared. ( n dot lm)**2 >= .4. @ 32143; ; P-Elements - TETRA - Incorrect Mass Matrix The mass for TETRA p-elements at p-level 1 is incorrectly calculated. Avoidance: Run the analysis at a p-level of 2 or higher when using TETRA p-elements are present. @ 32145; 70.5.2; Transient Analysis - TRD1 - Received Signal 10 Transient analysis may fail in the TRD1 module with no message. The log file may contain a machine dependent message of the form: *** Received signal 10 (SIGBUS) time: command terminated abnormally Avoidance: The error is machine dependent and is caused by a boundary alignment problem. The error will only occur on HP, SUN and SGI systems (Digital will print warning messages but will complete successfully). The error should be avoided by increasing memory by exactly 2 words on HP and SUN system and by 6 words on SGI systems. @ 32157; ment ; Online Documentation - Installation - Not a Valid WorldView Docu The MSC/NASTRAN Online Encyclopedia issues a "Not a valid WorldView document:" error message when any hotlink button on a book selected from the main books menu is picked. Iview configuration file entries for fulltextPath and welcomeDoc are valid and the documents could be selected manually from the Open/File menu picks. Resolution/Avoidance: This problem could occur on any system where:

1) the mne collections are moved from one location to another 2) the mne installation directory is nfs mounted or linked to a differently named location As part of the installation process, several *.cfg files are created on the fly and locations are hardcoded with pathname of the installation directory. If the mne collection directory is changed, it is not sufficient to update the Iview*fulltextPath list in the Iview configuration file - these *.cfg files must be edited also. Example 1: Moving mne collection location -MNE initially installed to /disk1/msc/mne705. -The entire collection is later copied/moved over to /disk2/msc/mne705 and /disk1/msc/mne705 is removed. -The Iview*fulltextPath list in the Iview configuration file is updated to reflect the disk1 -> disk2 change. Avoidance: edit the *.cfg files to accomdate the new location (disk1 -> disk2) Example 2: Installation dir not same as nfs mount point -MNE installed on server1:/disk1/msc/mne705 -That directory is nfs mounted on client2 as /msc/mne705. -The Iview*fulltextPath list in the Iview configuration file on client2 is updated to use the "/msc/mne705" path. Avoidance: Use a path that will be common to ALL workstations that will nfs mount the collection (e.g. /msc/mne705) AND IS ALSO ON THE SERVER NODE (/msc/mne705 should also exist there). Edit the *.cfg files to update to the common location. The *.cfg files that must be updated are: docmenu/docmenu.cfg 70all.cfg v69all/v69all.cfg v70prim.cfg v69prim/v69prim.cfg v705all/v705all.cfg v705prim/v705prim.cfg recomend/recomend.cfg v70all/v v70prim/

@ 32168; 70.5.2; Matrix Methods - DCMP - SFM 4276 EC 7777 It is possible that a job that should fail in DCMP with a max ratio error, identifying the problem degrees of freedom, instead fails with a traceback in sparse decomposition. The following is an example of the traceback from an hpux system. Process traceback ... ( 0) 0x0009700c sysrcv + 0x320 [/cae_k_apps/msc/msc70/hpux/analysis.pa2_0] ( 1) 0x00f83310 _sigreturn [/cae_k_apps/msc/msc70/hpux/analysis.pa2_0] ( 2) 0x005efba0 dfmsj_ + 0x130 [/cae_k_apps/msc/msc70/hpux/analysis.pa2_0] ( 3) 0x0022b7b4 dfmsa_ + 0x2b20 [/cae_k_apps/msc/msc70/hpux/analysis.pa2_0]

( 4) 0x004c355c sdcomp_ + 0xb8 [/cae_k_apps/msc/msc70/hpux/analysis.pa2_0] ( 5) 0x00220eb8 dcmpa_ + 0x2b8 [/cae_k_apps/msc/msc70/hpux/analysis.pa2_0] ( 6) 0x000c0fe8 dcmp_ + 0x130 [/cae_k_apps/msc/msc70/hpux/analysis.pa2_0] ( 7) 0x0009232c xsem07_ + 0x5c4 [/cae_k_apps/msc/msc70/hpux/analysis.pa2_0] ( 8) 0x0008ecb4 mainal_ + 0xa0c [/cae_k_apps/msc/msc70/hpux/analysis.pa2_0] Avoidance: If it is possible run the analysis on another machine to identify the problem degrees of freedom and then correct the modeling problems. Running the deck with sparse decomposition tuned off by setting SPARSE=1 on the NASTRAN command may also allow the problem degrees of freedom to be identified. @ 32171; 70.7; Composites - PARAM,POST,0 - Missing output Laminar stress and strain output is not written to the xdb file in SOL 103 and 105. Avoidance: Insert the following DMAP alter: compile sedrcvr alter 'DBC OSTR1' endIF $ SORT2 alter 'endIF $ SORT2'(3),'' endalter @ 32174; 70.5.1; Installation - xnastran not delivered The XNASTRAN utility is missing from the Version 70.5 (R1) CD. The executable file, <install-dir>/msc705/<arch>/Xnastran, can be copied from any earlier version. @ 32176; 2001; Documentation - EIGR - Point Normalization In modal analysis (sol 103), when POINT normalization is selected on the EIGR bulk data entry, and the defined component is zero, NASTRAN does not revert to MAX as specified in the Quick Reference guide for the EIGR bulk data entry: NORM or "POINT"; Default+"MASS") . . POINT Normalize to a positive or negative unit value of the component defined in fields 3 and 4 (defaults to "MAX" if defined component is zero). Documentation of NORM POINT should read: . . POINT Normalize to a positive or negative unit value of the component defined in fields 3 and 4 Method for normalizing eigenvectors. (Character: "MASS," "MAX,"

(defaults to "MASS" if defined component is zero). @ 32177; 2001; Documentation - QRG, INIT FMS - Incorrect Default values Under the description of the INIT command in the Quick Reference Guide, the default values for DBALL and SCR300 are stated incorrectly as 25000 and 175000 blocks. The correct default value is 250000 blocks for each of the DBSETs. @ 32193; 70.6; Nonlinear Analysis - CELAS1, CELAS3, PELAST, NLITER - SFM 2085 If the CELAS1 or CELAS3 element references the PELAST entry in a nonlinear analysis where the TKNID field is used to define the nonlinear force vs. displacement relationship, the job will fail with the following message: *** SYSTEM FATAL MESSAGE 2085 (NLITER) ELEMENT TYPE 11 NOT SUPPORTED FOR NONLINEAR ANALYSIS. *** SYSTEM FATAL MESSAGE 3007 (NLITER) ILLEGAL INPUT TO SUBROUTINE NLEMG Avoidance: Use the CBUSH entry and define the nonlinear force vs. displacement relationship through the KN field on the PBUSHT entry. @ 32218; 2001; Optimization/Design Sensitivity - CBUSH, DVGRID - SFM 3007 In shape optimization, if the CBUSH element is present, (even when not referenced by a DVGRID), the job fails with the following fatal message, "element type 102 is not supported *** SYSTEM FATAL MESSAGE 3007 (DOPR6) ILLEGAL INPUT TO SUBROUTINE Dopr6a". Avoidance: Don't use the CBUSH element in a shape optimization run. @ 32220; ; Transient Analysis - TSTEP - UFM 4562 In a modal transient analysis, the following error messages may be erroneously issued after the completion of the job. 0*** USER FATAL MESSAGE 4562, TSTEP (TIME STEP) DATA IS MISSING

Avoidance: This message can be safely ignored if all the requested output is obtained. @ 32225; 2001; Documentation - QRG, System Cell 206 - Missing Option Option 9 of system cell 206 which selects the better of METIS and MMD, is not documented in the Version 70.5 Quick Reference Guide. @ 32259; 70.7; Aeroelastic Analysis - Incorrect Answers for Unrestrained Hinge Moment Coeff. Two errors have been detected that relate to the printing of unrestrained hinge moment coefficients in SOL 144 (static aeroelasticity): 1. The unrestrained hinge moment coefficient values printed for the "INTERCEPT" coefficients are in error if they are nonzero.

2. If there are multiple subcases with the same Mach and q (dynamic pressure) conditions, the unrestrained hinge moment coefficient values printed for the second and subsequent subcases are in error. The coefficients for the first subcase are deemed correct with the exception of the INTERCEPT value mentioned previously. Avoidance: Make the following DMAP change: compile aestatrs list alter 'totu','kbrzx' dbstatus ipzf//s,n,noipz $ if ( noipz=-1 ) matgen ,/ipzf/7/norset/1 $ add ipzf,/mipzf/-1.0 append mipzf,kr1zx/totu $ alter 'kbr1zx','urddi' solve mirr,totu,,,/urddi//-1 $ DBSTATUS M4rr//S,N,NOM4rr $ IF( NOM4RR=-1 ) THEN $ MPYAD M1RL,AMLR,/M3RR//-1 $ FBS L2RR,U2RR,M3RR/M4RR $ endif $ nom4rr=-1 @ 32264; 69; Parallel Processing - EMA - Wrong Answers Due to an error in parallel EMA, any system running parallel EMA (even if they are only using one processor), could get wrong answers. Avoidance: Use the serial version. @ 32282; 70.7.2; Utilities - XNASTRAN - xmonast Started with the Wrong File Names The XNASTRAN utility does not properly handle an alternate input data file type, e.g., "example.bdf", when directly starting XMONAST to monitor a job. It will attempt to directly start XMONAST on the files "example.bdf.f04", "example.bdf.f06", and "example.bdf.log". Instead, it should be setting the nastran command's "xmon" keyword and relying upon the nastran command to correctly monitor the job. Avoidance: Set "xmon=yes" in an RC file or as an "advanced argument" and do not use XNASTRAN's ability to start XMONAST. @ 32286; 70.6; Superelement - STATSUB, multiple subcases - SFM 6710 (TAFF) When more than one static subcase exists in the bulk data for forming the differential stiffness matrix for a normal modes analysis (STATSUB), and the model contains superelements, then the job will fail with the following unclear messages: *** USER INFORMATION MESSAGE 7325 (SEP2DR) PROCESSING OF SUPERELEMENT 10 IS NOW INITIATED. ^^^ USER INFORMATION MESSAGE 9049 (SEKDR) ^^^ SUPERELEMENT DIFFERENTIAL STIFFNESS GENERATION, ASSEMBLY, AND REDUCTION. *** SYSTEM FATAL MESSAGE 6710 (TAFF) THE THIRD INPUT PARAMETER TO MODULE TAFF REFERENCES AN UNDEFINED LOAD SET ID. 0FATAL ERROR This job will fail even if there is only one static subcase callout, but

there are multiple loadings in the bulk data section that are not used. Avoidance: There are two possible avoidances: 1. Do not use superelements or 2. Only keep a single static and a single normal modes subcase and remove the extra loading entries from the bulk data section. @ 32294; 70.5; Miscellaneous - ISHELL - Insufficient Memory on CRAY Use of the ISHELL utility on a CRAY computer may cause the job to fail due to insufficient memory. Avoidance: Specify twice the memory that the job would normally take. @ 32296; D)

; Differential Stiffness - STATSUB, Superelements - UFM 5423 (MPYA In SOL 105, or if differential stiffness is requested (STATSUB Command) in SOLs 103 and 107 through 112, when superelements are present and no SUPER command, or SUPER=ALL, is specified, then static loads will be generated for the second or dynamic subcase and the job will fail with the following error messages. *** USER FATAL MESSAGE 5423 (MPYAD) ATTEMPT TO MULTIPLY INCOMPATIBLE MATRICES COLS ROWS FORM TYPE NZWD DENS TRAILER FOR GOAT 2 2 2 60 4333 TRAILER FOR UAS 8 3333 TRAILER FOR UOOX 40 2000 TRANSPOSE FLAG IS 0 Avoidance: Specify a separate subcase for the upstream superelements. IS 104 2 50 2 IS 103 1 12 2 IS 106 12 50 2

@ 32303; t

; Nonlinear Transient Heat Transfer - SUBCASE, SET - Missing Outpu In nonlinear transient heat transfer (SOL 159), if different subcases request different outputs, only the request for the first subcase will be honored and no output will be obtained for subsequent subcases. In the following example thermal results for grid 1 and 2 will be output for subcase 1 but no thermal output will be obtained for subcase 2. SUBCASE 1 TSTEPNL = 100 DLOAD = 200 SET 1 = 1,2 THERMAL =1 SUBCASE 2

TSTEPNL = 100 DLOAD = 200 THERMAL = ALL Avoidance: The only avoidance is to request the same output for all subcases above the first subcase. In the above example, request THERMAL=ALL above the first subcase. @ 32320; 70.7; Inertia Relief - PARAM,ALTRED,YES - Wrong Answers If PARAM,INREL,-1 and PARAM,ALTRED,YES are specified then wrong answers are produced. Avoidance: Insert the following DMAP alter: compile phase0 alter 'LFIXEDB OR INREL\=-1','' IF ( LFIXEDB ) ALTRED='NO' $ alter 'SEID\=0 and (LFIXEDB','' IF ( SEID=0 and (LFIXEDB or cditer>0) ) INREL0=false $ compile sedisp alter 'MPYAD LAO,UA,UOX','IF ( FIXEDB\>-1 ) CALL SEGOA,' if ( inrel0 ) then $ IF ( SEID=0 ) THEN $ IF ( inrel=-1 ) then $ MPYAD UOx,CRX,/UOXx $ else $ equivx uox/uoxx/-1 $ endif $ else $ MPYAD UOx,CRj0,/UOXx $ endif $ else $ equivx uox/uoxx/-1 $ endif $ MPYAD LAO,UA,UOXx/UOY/1 $ FBS LOO,,UOY/UOO///-1 $ ENDIF $ NOOSET>=0 endif $ IF ( APP<>'CYC' ) THEN $ alter 'MPYAD UOO,CRX,\/UOOX','' if ( altred<>'yes' ) then MPYAD UOO,CRX,/UOOX $ else $ EQUIVX UOO/UOOX/ALWAYS $ endif $ alter 'MPYAD UOO,CRJ0,\/UOOX','' if ( altred<>'yes' ) then MPYAD UOO,CRJ0,/UOOX $ else $ EQUIVX UOO/UOOX/ALWAYS $ endif $ alter 'sdr1'(3,-1) IF ( altred<>'yes' and FIXEDB>-1 ) CALL SEGOA, uset,GOt,/GOAt/NOOSET/NOQSET/NOTSET $ endalter @ 32321; 70.7; Static Analysis - PARAM,ALTRED,YES and MPCFORCE - Missing output

MPCForce output is not computed if PARAM,ALTRED,YES is specified. Avoidance: Insert the following alter: compile sedisp alter 'sdr1'(3,1) ENDIF $ ALTRED alter 'ENDIF \$ ALTRED','' endalter @ 32361; ot Work ; Online Documentation, MNE - Links in DMAP Module Dictionary Do N

In the Version 70 MSC/NASTRAN Online Encyclopedia, the chapter contents links in the first page of each chapter of the DMAP Module Dictionary do not work. Avoidance: Use the main TOC in the quick access bar located at the top of every page. @ 32362; 2001; Documentation, Interface Elements - GMBDC, GMBNDS - DMAP Fatal M essage 9050 The Quick Reference Guide bulk data definitions for the GMBNDC and GMBNDS entries needs the following remark added: The interface elements may generate high or negative matrix/factor diagonal ratios. If there are no other modelling errors, these messages may be ignored and PARAM,BAILOUT,-1 may be used to continue the run. @ 32364; 70.5.2; Postprocessing - TABPRT, OUTPUT2 - Wrong Answers, SWM 1204 Incorrect data may be output by a TABPRT and/or an OUTPUT2 request using the neutral option. Many occurences of SYSTEM WARNING MESSAGE 1204 may be an indication of a problem. Avoidance: For the OUTPUT2 neutral option, increase mem such that the value is at least f2 x size of largest record + 2 x nasbuff + compiled length of the NDDL of that datablock For TABPRT NDDL-output, increase mem such that the value is at least 4 x size of largest record + 1 x nasbuff + compiled length of the NDDL of that datablock @ 32365; ction 2001; Documentation/QRG - SESUP - Problems with free-free Dynamic Redu Models with free-free component modal superelement reduction give poor answers (or may fail). This limitation should be documented under the SESUP entry by adding the following remark: 5. SESUP bulk data records are not allowed for Part (partitioned bulk data) Superelements. Use the SUPORT bulk data records to identify component rigid body modes.

@ 32375; 70.5.2; P-Elements - Transverse Shear Strain Output Mis-Labeled The p-shell element's two transverse shear strain outputs for the default curvature option are switched. Otherwise the values are correct. Avoidance: There is no avoidance. @ 32376; 2001; Documentation/P-Elements - OUTRCV - Missing P-Shell Output When the PROJ=F option is used on the OUTRCV bulk data entry for p-shell elements, no output is produced in the .f06 file. The option should be used if a postprocessor requires results in terms of 3D vectors or tensors, for example, stress tensors with 6 stress components. The option does not produce output in the print file *.f06. @ 32388; ; Nonlinear Transient Analysis - TSTEPNL, ADD5 - UFM 4276 GETSTR E RROR CODE 31 If a nonlinear transient analysis has trouble converging and output is not requested for every timestep, the job mail fail with one of the following type messages in ADD5. 0*** SYSTEM FATAL ERROR 4276.NXTCOL CELAS2 0*** SYSTEM FATAL ERROR 4276.GETSTR CBUSH1D ERROR CODE ERROR CODE -31 PID= 31 PID= 0 0 -> ->

Avoidance: Request output for every timestep. @ 32390; 70.7; Cyclic Symmetry - READ, REIGL - UIM 6137, SFM 3001 The READ module may issue the message: UIM 6137 INPUT MATRIX TO DECOMPOSITION HAS xx SINGULARITIES, followed by *** SYSTEM FATAL MESSAGE 3001 (READ). This is due to null columns in the matrices. For cyclic symmetry AUTOSPC does not remove unused degrees of freedom if they are for grid points listed on CYJOIN records. Avoidance: For solid element models, add a bulk data record GRDSET,,,,,,456 to remove unused rotation degrees of freedom. Setting NASTRAN SYSTEM(273)=1 (Use earlier version Lanczos method) has been shown to work as well in NAST705. The following ALTER may be applied to version 70.5: COMPILE MODERS ALTER 'MATMOD *CASES' SETVAL //S,N,METHTYP/0 $ MESSAGE //'DO A-SET REDUCTION FOR MASSLESS DEGREES OF FREEDOM' @ 32394; 70.5.1; Complex Eigenvalue Analysis - EIGC - SFM 3008 (UTMCOR) The new block complex lanczos method in Version 70.5 (selected by setting system(108)=4) does not accurately predict the memory requirement. For large problems this may lead to the following fatal message:

*** SYSTEM FATAL MESSAGE 3008 (UTMCOR) MODULE = CEAD INSUFFICIENT CORE AVAILABLE FOR SUBROUTINE CLDRV USER ACTION: INCREASE OPEN CORE BY AT LEAST 1130847 WORDS PROGRAMMER INFORMATION: LOCATION KEY = TSLSPILL Avoidance: The fatal message may not be able to be avoided by increasing memory. The analysis may have to be performed with another method. @ 32395; 70.5.2; Eigenvalue Analsysis - EIGRL, READ - SFM 7340 (LNNHERR) The new lanczos in MSC/NASTRAN Version 70.5 may fail with some variation of the following message: *** SYSTEM FATAL MESSAGE 7340 (LNNHERR) process error reported by subroutine LNNDRVS see LNNDRVS abstract (ier = -1) The message that should have been output is: FACTORIZATION ERROR DURING FINITE INTERVAL ANALYSIS AT FLEFT. FLEFT = -3.9478D-05 *** SYSTEM FATAL MESSAGE 5299 (LNCS1D) FACTORIZATION ERROR AT A BOUNDARY SHIFT. POTENTIAL FIXES: IF LOWER BOUND IS ZERO USE A SMALL NEGATIVE VALUE. IF LOWER BOUND IS A SMALL NEGATIVE VALUE DECREASE IT. ELIMINATE MASSLESS MECHANISMS THE MODEL MAY CONTAIN. Avoidance: Follow the direction of SFM 5299 above. @ 32411; 70.7; Aeroelastic Analysis - APD, RESTART LOGICAL - UFM 1245 If a read-only restart (RESTART LOGICAL FMS statement) is performed in SOLs 144, 145, or 146, then the run will fail in the APD module with the following message: 0*** USER FATAL MESSAGE 1245, (BIO) AN ALLOCATION ERROR HAS OCCURRED FOR A DATABASE FILE, FILX = WRITE ATTEMPTED ON FILX WHICH WAS OPENED FOR READONLY. Avoidance: Use RESTART with the COPY option. @ 32426; 70.7; P-Elements - GP0 - Distorted Element Edges May Lead to Traceback If a p-element edge is too distorted, the program may fail with the traceback that follows in GP0 instead of exiting with a meaningful message. SIGTRAP - Trace trap FP division by zero Traceback: Offset Offset Offset Offset 0x00001380 0x00000c54 0x00001710 0x00000cd4 in in in in procedure procedure procedure procedure evlpnd_ gp0dc_ gp0da_ gp0_ 10.

Offset 0x000004b0 in procedure xsem02_ Offset 0x000009dc in procedure mainal --- End of call chain --Avoidance: Other than correcting the distorted geometry, there is no avoidance. @ 32438; 70.5.2; Submittal - Wrong executable may be selected on HP-UX 11 Apparently HP will make both 32- and 64-bit versions of HP-UX 11.0 available. As a result of this, if the client is running the 32-bit version of HP-UX 11.0 on a PA-RISC 2.0 system, the nastran command will incorrectly choose the V-class executable. This error can only occur with MSC/NASTRAN V70 and MSC/NASTRAN V70.5 (r1). The bit size of the OS is determined with the following command: $ getconf KERNEL_BITS The command will print 64 on a 64-bit kernel, e.g., on a V-Class system; and 32 on a 32-bit kernel. Avoidance: Set the following keyword in the either system (e.g., <install-dir>/conf/nast705rc) or node (e.g., <install-dir>/conf/net/<node>/nast705rc) RC files. On a PA-RISC 1.1 system: proc= On a PA-RISC 2.0 system: proc=pa2_0 The processor type can be determined using the nastran command, e.g., msc705 nastran system @ 32457; 70.5.2; P-Elements - Beam Error Estimator - Poor Answers, Inefficient Pr ocessing The error estimators for the beam p-element are incorrect by a factor based on the length of the beam. This error may lead to over adapting or under adapting. That is the error may lead to poor answers or it may lead to an unnecessarily high p-level. Avoidance: There is no avoidance. @ 32458; 70.6; ESTIMATE - P-Elements - Insufficient Memory MSC Estimate underestimates memory for input files with P-elements. This limitation will be removed with version 71 for P-element runs without interface elements (GMINTC/GMINTS). The limitation still exists for runs with interface elements. Avoidance: Do not use estimate to determine memory and disk requirements of runs with p-elements prior to Version 71, and do not use estimate to determine memory and disk requirements of any run with interface elements. @

32467; 70.5.5; Parallel Processing - Incorrect CPU Times When parallel processing is selected, the f04 per processor cpu times may show negative cpu times on Version 70.5 NT systems. This problem has been corrected in the Version 70.5.5 release. @ 32476; ; Utilities - RCOUT2, BGPDT - rcout2 utility does not transfer dat a correctly The rcout2 utility will not transfer the BGPDT correctly on short-word UNIX systems. @ 32477;

70.6; Estimate - SPC/SESET THRU 9999999 - performance If an input file has an SPC entry specifying 1 THRU 9999999, then estimate will take a lot of CPU time. This performance error will be removed in version 71. Avoidance: Specify the last grid id instead of 9999999.

@ 32482;

2001; DMAP - DRMS1 - SFE 4276 EC 1305 in ENDPUT. The DRMS1 module may fail with SFE 4276 EC 1305 in ENDPUT. Avoidance: Reduce the number of subcases (or modes) in the input data blocks to the DRMS1 module.

@ 32489;

; Aeroelastic Analysis - AEROF - Incorrect Aerodynamic Box ID's In SOL 146 for Harmonic Gust (Frequency Response), the AEROF=<setid ALL> option is supported. However, if the ALL option is used in Version 70.5, the f06 output listing incorrectly lists the Aerodynamic Box ID's as 0 (zero) for all boxes. Whereas, if a specific SET is used, the correct results are obtained. The complex forces printed in both cases, are correct.

@ 32490;

70.7; Inertia Relief - UPARTN, PARAM,INREL,-1 - Poor Message, UFM 3007 If PARAM,INREL,-1 is specified without any SUPORTi entries then the run will fail in the UPARTN module with the message: *** SYSTEM FATAL MESSAGE 3007 (UPARTN) ILLEGAL INPUT TO SUBROUTINE PARTN Avoidance: Remove PARAM,INREL,-1 or specify SUPORTi entries.

@ 32495; tput

70.6; Nonlinear Statics - PARAM,NMLOOP,-1 and POST,0 - Missing .xdb Ou If PARAM,NMLOOP,-1 and PARAM,POST,0 are specified along with multiple subcases in SOL 106, then only the eigensolution results for the last subcase are written to the .xdb file. Avoidance: Insert the following DMAP alter:

compile nlstatic alter 'call super3'(2),'' CALL SUPER3 CASECC ,casei , , ,PCDB EMAP ,MAPS SLT ,ETT , USET ,SILS GM , CMPHO ,CMLAMA DYNAMICS,BGPDTS EST ,GEOM2S INDTA ,KELM ,ditid ,GEOM4S ,EPT GEOM3 ,GEOM4 casei , ,gpsnts , MMG ,MSF , , geom1s , / PHG ,PHQG , , APP /APP FALSE /0 0 /FALSE FALSE /false ' '/false endalter

,PHA ,olb , ,XYCDB ,EQEXINS ,GOT , , ,KFS ,CMPHA ,CSTMS ,GEOM3S ,KDICT ,pvtbc , ,MPT ,PTELEM , ,pvt , , , , , ,phqmg , , /RSONLY /FALSE /0 /false /false

,OLB , , ,POSTCDB ,PVTS ,GOQ , ,EDTs ,KSS ,MAR ,MPTS ,ECTS ,GPECT ,scstm , ,GEOM1 ,MELM , ,DEQATN ,KSF , , , , ,LAMA1 , , /FALSE /S,PFILE /FALSE /0 $

, , , ,FORCE ,CASES , ,GPLS , ,QR ,MEA ,DIT ,EPTS ,VELEM , , ,GEOM2 ,MDICT , ,DEQIND ,KMG , , , ,ugni

, , , , , , , , , , , , , , , , , , , , , , , ,

, , , , / /FALSE / /S,CARDNO/ /0 / /false /

@ 32496; ; Documentation - PSDF, ATOC, RALL - Random Requests not yet Suppo rted in PATRAN The documentation for the new random analysis data recovery request options PSDF, ATOC and RALL for DISP, VELO, ACC, FORCE, and STRESS documented in the Version 70.5 Quick Reference Guide is misleading. These options are not yet supported in PATRAN and thus are of no use and should not be selected. Presently if they are selected, the random output is written to output2 file and no random output is written to the f06 file. Currently, nothing is written to the xdb graphics data base as might be implied by the Quick Reference Guide remarks. Furthermore, if PARAM,POST,-1, is set, and the OUTPUT2 file is FORMATTED and the new format of the output requests for RANDOM results is used, such as DISP (PSDF) = all then the run will issue UWM 1206 indicating that the datablocks created by the RANDOM module are not written to the OUTPUT2 file. Avoidance: Currently the avoidance is simply do not use these output requests since they are not supported in PATRAN.

@ 32502; Message

; Optimization/Design Sensitivity - Superelements - UFM 9068, Poor

If superelements are used in design optimization and a DRESP2 entry references DRESP1 entries which select responses in different superelements, the run may fail with UFM 9068 incorrectly stating that there are no DRESP1 or DRESP2 entries for the DESOBJ selected. The program should issue a message indicating that the DRESP1 entries incorrectly spans superelements. Avoidance: Remark 2 in the Quick Reference Guide description of the DRESP2 entry states that DRESP1 entries referenced on a DRESP2 may not span superelements. However, the error message issued is misleading. Simply place all of the referenced items into a single superelement. @ 32504; ; Restarts - TEMPD - Inefficient Processing If a bulk data entry used to define a temperature loadings (TEMP,TEMPD,etc) is changed on a restart, the restart logic will delete the stiffness matrix and create a new one. This results in inefficient restarts. Avoidance: Either:

1) Place all TEMPi entries in the initial input file, or 2) Insert the following DMAP in the restart run; $ DMAP alter to remove checking for changed in temperature $ BE CAREFUL WHEN USING THIS, AS IT REMOVES SOME OF THE RESTART LOGIC compile ifpl alter 'restart.*geom3,geom3r','' compile phase0 alter 'restart.*geom3s,geom3sr','' @ 32514; 2001; Documentation - LOADSET - Only One LOADSET Command is Allowed pe r Superelement The Quick Reference Guide description of the LOADSET command does not explicitly state that only one LOADSET command may be used for each superelement (or, in the case of no superelements, only one per run). This command should occur in the first SUBCASE for each superelement (or above the SUBCASEs). Any LOADSET commands in subsequent SUBCASEs will be ignored. @ 32522; 70.5.2; External Superelements - CSUPER, SPOINTS - Wrong Answers If an external superelement is brought in using the CSUPER entry and SPOINTs are used for some of the attached dof, it is possible that the matrices will be attached incorrectly. In the problem demonstrating this, if all GRID points were listed before the SPOINTs on the CSUPER, the matrices were attached correctly. If the SPOINTs were listed, then the GRID points, then the matrices were incorrectly attached. (NOTE - the matrices are attached in the order listed on the CSUPER and care must be exercised to verify that the correct order is listed - do not simply change the order on the CSUPER to correct this problem) Avoidance: If possible, use only GRID points on the CSUPER, or be careful to create the matrices for the external superelement in such a way that the terms for the GRID points occur first. Another possible avoidance is

to use all SPOINTs on the CSUPER and then to use MPCs to connect to the GRID points. Bringing in the external superelement as a PART using the EXTRN entry also works. @ 32526; 70.5.2; Matrix Methods - MPYAD - SFM 4276, EC 3012 or Wrong Results On the SUN computer, if a MYPAD operation uses the DOT3RD kernel, then the answers may be incorrect or the job may terminate with one of the following error messages: BIOMSG: ERROR 3012 HAS OCCURRED IN ROUTINE QOPEN process error reported by subroutine LNNP2CD see LNNP2CD abstract (ier = -728) *** SYSTEM FATAL MESSAGE 7340 (LNNHERR) Avoidance: We have been unable to describe in modeling terms when the erroneous method is chosen. To date the error has only been observed in two similar models using superelements while calculating the system modes at the residual. To prevent the incorrect method from ever being selected it is recommended to add the following system cell to the top of the input file or to the system rc file for all runs on Version 70.5 SUN Ultra SPARC systems. nastran system(252)=0 Note that the SUN Solaris system rc file on the Version 70.5.1 R3 cd which added the Fujitsu UXP/V system was modified to include this avoidance. @ 32550; ; Static Analysis - SUBCASE, SDR2 - Missing Output, SFM 3002 If a subcase is repeated in a model with many subcases and multiple boundary conditions, the job may issue the message: *** SYSTEM FATAL MESSAGE 3002 (SDR2) EOF ENCOUNTERED WHILE READING DATA BLOCK UG2 SUBROUTINE SDR2 or may be missing output for some subcases. Avoidance: Remove the duplicated subcases. @ 32551; 2001; Optimization/Design Sensitivity - DESVAR, IFP - SFM 4276, Error Code 7777 If integer values are used for XINIT, XLB, or XUP on DESVAR entries, MSC/NASTRAN may fail in the IFP module with either SIGNAL 8, error 7777, or SFE 4276, EC 7777. Avoidance: Correct the format on the DESVAR entries. @ 32557; LG 70.7; Differential Stiffness - STATSUB, Superelements - UFM 4562 in TR When performing transient analysis (SOL 109) with superelements and differential stiffness, with STATSUB, then UFM 4562 may (FILE 112) IN , FILE INDEX (FILEX) =

be issued. Avoidance: Insert the following DMAP alter: compile super3 alter 'equivx CASERS' else if ( nocasrs1>-1 ) then $ equivx CASERS1/CASE0/-1 $ for non-SOL 200 with dynamic looping endalter @ 32561; 70.5.2; Frequency Response/Random Response - RANDPS - Time Command Termi nated Abnormally An underflow error may occur on some machines when the random module squares stress values that are numeric zeroes. If this error is encountered the job will stop in the RANDOM module, the f06 file will not contain a message, and the log file may contain and overflow/underflow message. The DEC ALPHA OSF/1 system issued the message: Time command terminated abnormally. Avoidance: Do not request random xy output for very small output quantities. If the quantity is near 1.e-19, when it is squared for random, it can cause an underflow to occur. Most machines will catch this and output a zero in the results. The ALpha stops with no meaningful messages @ 32572; 70.6; Superelements - SUBCASE,SUPER - UFM 5423 in SDR1 or UFM 7324, m issing output If there exists more subcases for a superelement than the residual structure, then the run may fail with UFM 5423. If the load sequence number is specified on the SUPER commands for the superelement subcases, then UFM 7324 is issued. However, the run continues and output is missing for the extra superelement subcases. Avoidance: Remove the extra superelement subcases or add residual structure subcases. @ 32608; 70.6; Inertia relief - DMIG, UACCEL - UFM 6143 If less than 6 degree of freedom are specified on a single grid point on the DMIG,UACCEL entry, then the job will fail with the following messages: *** USER FATAL MESSAGE 6143, (MERGE1) INPUT MATRICES AND VECTORS ARE INCOMPATIBLE Avoidance: Insert the following dmap alter: $ $ for v70, v70.5 $ compile statrs alter 'UPARTN USET,UACCEL','' IF ( NOUACCEL>=0 ) UPARTN USET,UACCEL/URAx,,,/'G'/'R'/'S'/1 $ alter 'dbstatus','' dbstatus urax//s,n,noura $ alter 'solve ' endif $

alter 'endif \$ nullqrr','' endalter

@ 32609;

2001; Documentation - PARAM,UNSYMF - Missing Documentation Complex eigenvalue analysis may be performed in nonlinear static analysis, SOL 106, with PARAM,UNSYMF,YES. This feature is documented in the Version 70.5 Release Guide but the PARAM,UNSYMF is not documented in the Version 70.5 Quick Reference Guide. The parameter should be documented and referred to from the cmethod case control command in the QRG.

@ 32611; 70.5.2; P-Elements - Beam Error Estimator If there are beam p-elements in a model with no strain, such as rigid body modes in modal analysis, the error estimator may crash. The EPSTOL parameter does not work with beam p-elements so it will not help avoid this error. Avoidance: Exclude beam p-elements from the error estimate in such models, or turn off the error estimator. @ 32612; ; P-Elements - BEAMP - Wrong Answers at P=2 A BEAMP p-element has bad residual bending flexibility when p=2. The error is more apparent when small values for the modulus "E" are used. Avoidance: Set NASTRAN SYSTEM(226)=1, to turn off the residual bending flexibility. @ 32615; ; Submittal - A "$" Sign in Command Line Argument Causes Argument to be Truncated If a command line argument contains a dollar sign ('$'), the command line argument is truncated at the dollar sign. This probably should only occur for NASTRAN cards encountered in RC file processing. Avoidance: There is no avoidance for Windows NT. The error can be avoided on UNIX by quoting the argument and escaping the dollar sign, however, a dollar sign is a very unusual UNIX filename character. @ 32626; o ; Matrix Methods - DECOMP - UFM 9050 (SEKRRS) Excessive Pivot Rati If negative terms occur on the factor diagonal matrix during matrix decomposition using the SPARSE (default) solver, UFM 9050 will correctly be printed, but the dof containing negative factors on the factor diagonal matrix will not be identified. Terms which have factor diagonal ratios greater than the value of PARAM,MAXRATIO will be properly identified. Avoidance: If the problem is run using NASTRAN SPARSE=1, the degrees of freedom with negative terms in the factor diagonal matrix will be properly identified in the listing. @ 32633; 70.6; ACCESS - smplr - Program to Read .xdb Does not work Properly. A dynamic factor was added for the access key in MSC/NASTRAN Version 70.5.

Thus, the program smplr.f delivered with MSC/NASTRAN Version 70.5 does not work properly. Avoidance: An illustration of the protocode required is: Call READR (file,ia,0,iret) <- when output object is open attempt to locate entry with key=0 if ( iret.eq.0 ) then factor = ia(2) <- obtain dynamic factor for access key else if ( approach=statics ) then <- conventions before V70.5 more or less factor = 1 else factor = 10000 endif key = intid*factor + offset <- intid from EQEXING/E and offset from TOL,FOL, etc. @ 32639;

70.6; Complex Eigenvalues - EIGC, HESS - wrong answers The HESS method may produce one pair of roots that are wrong for diagonal input matrices. The symptoms indicate that the shift may not be removed from the last pair calculated. The solchk alter in the sssalter libray may be used when it is suspected that roots may be wrong. It computes a normalized residual for each eigenvector. Similar code is built into the complex eigensolver for systems after v70.5. It has been confirmed that a new HESS method implemented there has solved the two problems discovered to date that have wrong answers. Avoidance: An avoidance is to use the CLAN method instead. In Version 70.6 and beyond this error occurs only when the old HESS method is selected by system cell 108. An avoidance is to remove the system cell request from the NASTRAN card. The new default method is the QZ HESS method, which does not have this error.

@ 32642;

70.7; Complex Eigenvalue Analysis - CEAD - Floating Point Exception An error in the spill logic on "small endian" computers which include Digital and Windows NT systems may cause a floating point exception to occur in the complex eigenvalue solver. The message obtained is machine dependent, one example is: "The FPU has been reset after floating point exception" Avoidance: Increase the available amount of memory until spill is not necessary. Memory on Intel NT systems can be increased above 25 mw only in Version 70.5.

@ 32647;

70.6; P-Elements - EMA - Wrong Answers A wrong stiffness matrix may occur if both h- and p-elements occur in the same model, the h-element grid points use a non-basic coordinate system for displacements, and multiple passes are required for matrix assembly. Since some columns are output using the Basic coordinate system rather than the local system, the matrix will be unsymmetric. The wrong answers

should be evident by high epsilon values and the spcforce resultants will not balance the applied loads. Avoidance: Either increase memory to avoid multiple passes or avoid the use of displacement coordinate systems. @ 32660; 70.5.2; Installation - time: command terminated abnormally MSC/NASTRAN will fail to run if the hard limit for the system's data segment size, see the "data(kbytes)" line of the "ulimit -H -a" command, has been set to any value other than "unlimited". Avoidance: Reset the hard limit to unlimited. This can only be done by root, it will require the user to exit the affected shell. @ 32664; 70.6; Random Analysis - RANDPS - UFM 5418, UFM 316 The RANDPS bulk data entry has an error that limits the number of RANDPS entries that are processed. This error occurs only for the special case of uncoupled PSDF (see remark 8 in the QRG). The error may results in various fatal messages. Those experienced include USER FATAL MESSAGE 5418 and USER FATAL MESSAGE 316. The text of these messages indicate that there are illegal RANDPS entries and NO GRID POINTS IN THE MODEL. It is speculated that there is a transition area somewhere between 100 and 320 bulk data entries where different types of bulk data entries may be discarded, leading to many different error messages. This error does not apply to other types of random analysis input, where the number of RANDPS entries that are allowed depends only on memory requirements. Avoidance: There is no error for the coupled case. Add a dummy RANDPS entry, where J<K are distinct subcase identifiers, and both X and Y are 0. @ 32670; 70.7; P-Elements - QUADP - Spurious Zero Frequency Modes The QUADP element may have spurious modes (more zero frequency modes than expected) for certain values of p along the edges. Avoidance: An avoidance is to set SYSTEM(225)=1, which turns off the bubble functions. @ 32677; 2001; Documentation - PARAM,EIGD - PARAM,EIDG Does Not Exist In the Version 69, 70, and 70.5 Quick Reference guides the parameter description for EIGD says: "PARAM,EIDG,YES requests ..." But, it should say: "PARAM,EIGD,YES requests ..." (note EIDG vs EIGD) @ 32685; 70.5.2; Transient dynamics - NLRGAP with SOL 109 - TRD1, Signal 10, Bus error When using the NLRGAP element in SOL 109, the job may terminate with no error messages in the TRD1 module on some machines. Depending on the machine, the log file may contain messages similar to the following: Receiving signal number 10: Bus error

The FPU has been reset after floating point exception. Avoidance: Look at the "TOTAL MSC/NASTRAN MEMORY LIMIT" in the Execution Summary (f04 file) and increase the memory by exactly 2 words on SUN systems and 6 words on SGI systems by specifying mem= on the nastran submittal command. See also error 32014 before proceeding. @ 32705; 2001; Documentation - RSPLINE - Maximum of 100 Grids The RSPLINE has an undocumented limit of 100 grid points. A remark will be added to the next release of the Quick Reference Guide. @ 32707; 70.5.2; Acoustic Analysis - CTETRA - UFM 4683 (READ) Acoustic Analysis with the 10 noded CTETRA element fails with the message: *** USER FATAL MESSAGE 4683 (READ) MASS MATRIX NEEDED FOR EIGENVALUE ANALYSIS Avoidance: Other than using the 4 noded TETRA element there is no avoidance. @ 32717; error ; Complex Eigenvalue Analysis - CEAD, EIGC - signal number 10: Bus If the INV complex eigenvalue method is used in version 70+ and the model contains transfer function callout (tfl) and p-elements, then the job may fail with one of the following error messages: Receiving signal number 10: Bus error Trap encountered: SIGSEGV - Segmentation violation Access violation Avoidance: Use the complex lanczos method instead of the inverse method. @ 32725; ; Postprocessing/Component Mode Synthesis - PARAM,FIXEDB - Missing Output If PARAM,POST is set to 0 to generate an MSC/XL database and PARAM,FIXEDB,-1 is used, the ".xdb" file will not contain the component modal results. Avoidance: There is no avoidance. @ 32728; 70.5.2; Cyclic Symmetry - Incorrect UWM 6231 is given is SOL 114 An erroneous message of the following form may appear in SOL 114: *** USER WARNING MESSAGE 6231 (CY1CCT) BOUNDARY CONDITIONS IN SUBCASEKKLL ARE DIFFERENT THAN THOSE IN THE FIRST SUBCASE. SUBCASE 1 SPC= .001 AND MPC= SUBCASE .001 SPC= AND MPC= .001

USER INFORMATION: ONLY ONE BOUNDARY CONDITION IS PERMITTED FOR ALL SUBCASES IN CYCLIC SYMMETRY ANALYSIS. THEREFORE, THE CONDITIONS IN THE FIRST SUBCASE WILL BE USED FOR ALL SUBCASES. Avoidance: Ignore the message. @ 32731; 70.5; SSSALTERs - sebloada.v705 - Possible Wrong Answers Due to an error in the sssalter sebloada.v705, output from the alter in transient solutions may not be printed correctly. Rather than printing a time history of OLOAD (boundary force), it may print only a few points with an incorrect time step. Avoidance: Change the alter: change the line: EQUIVX OL1/OUTPLIST/ALWAYS $ OUTPUT LIST to: EQUIVX OL1r/OUTPLIST/ALWAYS $ OUTPUT LIST @ 32733; ffness ; External Superelements - PARAM,EXTOUT - Ignores differential sti When creating an external superelement using PARAM,EXTOUT, the first residual structure SUBCASE is used. This means that it is not possible to include a static preload. Avoidance: Apply the following DMAP alter when creating the reduced superelement: $ $ alter to allow static pre-load when creating reduced external superelement matrices $ $ For of the case control: $ $ SUBCASE 1 $ LOAD = 99 $ SUBCASE 2 $ STATSUB = 1 $ METHOD = 3 $ compile phase1dr alter 1 $ remove external superelement creation alter 'IF ( EXTOUT.*) THEN','ENDIF $ EXTOUT' $ $ compile semodes, list $ $ insert external superelement creation $ alter 'xadd5.*kaa,kdaa' $ type parm,,i,n,gotit $ type db,gpak,gdak type parm,nddl,char8,y,extout IF ( EXTOUT<>' ' ) THEN $ CALL EXTOUT MKAA,MAA,BAA,K4AA,PA,GPAK,GDAK,EQEXINS,USET/ DUM/LUSETS $

CALL PRTSUM /RSTJUNK $ EXIT $ ENDIF $ EXTOUT<>' $ $ fix for error 32070 $ compile sekdr $ this alter is for Version 70.5 only alter 'call segoa','' type db goq $ CALL SEGOA USET,GOT,goq/GOAT/NOOSET/NOQSET/NOTSET $ endalter @ 32748; 70.5.2; Documentation - Network Mounted Database Files Affect Elapsed Ti me Performance Database files that are remotely mounted will affect elapsed time performance of MSC/NASTRAN jobs. There is no explicit documentation or user warning to this fact. Depending on the I/O characteristics of the job, the relative performance of the processor, and the relative performance of the network, the penalty can be extreme. One user saw a 24:1 ratio of elapsed time to CPU time with NFS files and 2:1 with local files. Higher ratios would not be surprising. @ 32751; 70.5.2; Nonlinear Transient Heat Transfer - EMG, MATT4 - SFM 4276, EC 7 777 Nonlinear transient heat transfer analysis that involves a phase change in MSC/NASTRAN Version 70.5 may fail in the EMG module with the following traceback: Trap encountered: SIGTRAP - Trace trap FP division by zero Traceback: Offset Offset Offset Offset Offset Offset Offset Offset 0x00001250 0x0000059c 0x0000aeb4 0x0000054c 0x00000b68 0x000004a0 0x000000fc 0x00000948 in in in in in in in in procedure procedure procedure procedure procedure procedure procedure procedure tab_ hmat1_ ehexd_ emgbrd_ emgpro_ emg_ xsem09_ mainal

--- End of call chain --Avoidance: User avoidance is to add an additional point at the end of the TABLEMi table. @ 32760; ; Utilities - estimate program, replication entries - UFM (estimat e.generate_) If the input file uses the nastran replicator to generate continuation entries, the ESTIMATE program may fail with the following error messages: *** USER FATAL MESSAGE (estimate.generate_bulk_entry) In file "run1.dat", at line 26:

*1,

*2, *2

Field 10 is blank. Avoidance: Echo the input file into a sorted nastran punch file and run the ESTIMATE program on the new file. @ 32773; 70.5.2; Rigid Elements - RSSCON with ELEM option - UFM 6695 (SEGRDD) When using the ELEM option of the RSSCON entry, the following error may result when the shell element has a significant acute angle causing attachment to the wrong edges of the solid: *** USER FATAL MESSAGE 6695 (SEGRDD) RSSCON 12 CANNOT CONNECT SHELL GRID POINT ACTION, CHECK THE GEOMETRY *** USER FATAL MESSAGE 6695 (SEGRDD) RSSCON 13 CANNOT CONNECT SHELL GRID POINT ACTION, CHECK THE GEOMETRY @ 32783; 44 : USER 43 : USER

70.7; Installation/Security - flexlm, rcf - Incorrect Default On NT systems using flexlm, the authorization code specified on the System environment variable MSC_LICENSE_FILE (under settings/control panel/system/environment) overwrites the rcf file but not the command line. If you want the RC file to override MSC_LICENSE_FILE, you must lock the keyword in the RC file. For example, auth=what_ever_the_filename_is lock=auth Any attempt to set a new auth value will be silently ignored. Unfortunately, an error in prior versions of the nastran command allowed the environment to override a locked keyword. This error has been corrected for v70.7. Avoidance: Either specify the authorization file on the command line or remove the MSC_LICENSE_FILE variable from the system environment.

@ 32784;

2001; Documentation - EIGC - Incorrect OMEGAAJ Value Under the Quick Reference Guide description of the EIGC entry, field 3, the OMEGAAJ-field, of the first continuation entry should be +5.6 instead of -5.6

@ 32793; RLG

70.7; Transient Analysis - LOADSET, LSEQ - Segmentation Violation in T In SOL 129, various computer types may fail with a segmentation fault of a "signal" error if LOADSET-LSEQ is used to define the dynamic loadings and the static load ids on the LSEQ entries are identical to the associated DAREA ids. Avoidance: Always use unique ids for the static loadings and DAREA ids

@ 32828;

; Parts superelements - SECONCT - Wrong Answers External superelements at present require the presence of a seconct entry

in the main bulk data deck, and an extrn entry in the part superelement. If the number of DOFs listed on the seconct entry is less than the number on the extrn entry, then what is called the g-set option is being used. There is an error in the map matrix which resequences the boundary point, leading to incorrect assembly of the external superelement in the assembly run. The error does not exist when the a-set option is used. For this option the number of DOFs listed on the seconct entry is exactly equal to the number on the extern entry. Avoidance: If it is required to release some of the boundary dofs this can be done by coupling them to an extra grid point in the downstream superelement, then connecting this extra point to the actual downstream boundary point with an rbar element which connects only the desired DOFs. @ 32829; 70.5.2; Beam Library - Missing bmsrv.o file for build The main program object file for the Beam Server is missing from the MSC/NASTRAN V70.5 CD's. This file is only needed by clients building a beam server. If a build is attempted messages similar to the following will be encountered: f77 -o beamserv /big_fish/msc/msc705/hpux/bmsrv.o bsmsg.o bsbrc.o bsbrg.o bsbri.o bsbrp.o bsbrt.o bscon.o bsgrq.o mevbr.o msbri.o msbrg.o msbrc.o brtup.o brtui.o brtug.o brtuc.o /big_fish/msc/msc705/hpux/libbmsrv.a -Wl,-a,archive +O3 +U77 +Oparallel_env +Onolimit +DS770 -v -lM f77: Cannot read file /big_fish/msc/msc705/hpux/bmsrv.o f77: Errors detected, no link *** Error exit code 1 Stop. Avoidance: The files will soon be available from the following URL: ftp://ftp.macsch.com/pub/msc-products/NASTRAN/V70.5/ Select the appropriate directory for your system, and download the "bmsrv.o" file. @ 32837; 70.5.2; Submittal - Remote Execution Via node Command will Fail. Remote execution of a nastran job using the "node" keyword will fail. A typical error message that will appear on the terminal or window is: ksh: 128: not found. Avoidance: Edit <install_dir>/bin/nast705.rmt and change all four (4) occurrences of the text "0.name" to "rcmd". @ 32846; ; External Superelements - SECONCT - Wrong answers The g-set option allows connecting less DOFs to the downstream superelement than exist in the boundary matrices. If the number of DOFs listed on the SECONCT entry is less than the number on the EXTRN entry, then what is called the g-set option is being used. The method, used to discard unwanted DOFs, is to discard their rows and columns from the boundary matrices, which is analogous to spcing them. This will make internal constraints in the model, that is, spcs whose affects are visible

implicitly in missing rigid body modes, load and element force unbalance at some of the grid points, etc. but the internal spc forces do not appear in spcforce output. Also be aware of error 32828. Avoidance: An avoidance is to use the a-set option only. See the discussion in error 32828 on how to convert to this option. When using the a-set option, boundary DOFs may be disconnected from the downstream superelement by connecting the unwanted boundary point to an extra grid point in the downstream superelement which is not connected to any other structure. @ 32852; ; Security - FLEXlm - Can't locate Multi-processors When using FLEXlm, if the number of processors specified by the PROC entry in the license file is less than the total number of processors on a machine, MSC/NASTRAN will use only one processor even if more are requested. Avoidance: Obtain a license file that allows the maximum available (or unlimited) number of processors. @ 32854; 2001; Documentation/Superelements - SETREE, SELABEL - Entry Ignored, I ncorrect SE Tree The SETREE entry will only work if PART (BEGIN SUPER) superelements exist in the model. If there are no PARTs in the model, the SETREE entries will be ignored. A remark should be added to the SETREE and SELABEL entries stating this limitation. Avoidance: Either: 1) Change BEGIN BULK to BEGIN SUPER to force new superelement logic or: 2) use DTI, SETREE @ 32858; 2001; Documentation/Aeroelastic Analyis - SPLINE2, GI - Poor Message If an aeroelasticity run, contains a SPLINE2 element that incorrectly references only one grid point, the GI module issues a FATAL ERROR message with no explanation of the cause of the error. Avoidance: Increase the number of grids to the required two or more. @ 32860; 70.6; Differential Stiffness - LOADSET, superelements, STATSUB - Wrong answers or UFM If the LOADSET command is used in SOLs 107 through 112 with superelements and differential stiffness (STATSUB Case Control command) then wrong answers may occur. For example, if LOADSET is specified above the subcase, the static loads in the static subcase are wiped out. If the LOADSET command is specified in the dynamic subcase following the static subcase, the dynamic load via LSEQ disappears. Avoidance: Specify two LSEQ Bulk Data entries with the same SID and two different DAREA IDs; one for the static preload and one for the dynamically applied static load. The LSEQ with the smaller DAREA ID must

correspond to the static preload. The LSEQ with the larger DAREA ID must correspond to the dynamically applied static load. Also, insert the following DMAP alter avoidance: compile statics2 alter 'bcdr ','bcdr '(,2) LCGEN CASES,SLT,ETT/CASESX/S,N,DUMMY $ BCDR CASESx//SEID/SOLAPP/ S,N,NSKIP/S,N,NLOADS/S,N,BCFLAG/S,N,SPC/S,N,MPC/ S,N,SUPORT/S,N,LOAD/S,N,LSEQ///1 $ PARAML CASESx//'DTI'/nskip/7//S,N,tempLD $ PARAML CASESx//'DTI'/nskip/6//S,N,deform $ endalter @ 32861; 70.5; Utilities - DBC.XDB missing from NT delivery The DBC.XDB schema database was not included with the NT Version 70.5 delivery. This will cause the ddlqry and ddlprt utilities to fail or not have access to the Version 70.5 description information. @ 32864; 70.5.2; Superelements - EXTRN, PARAM,EXTOUT, SMPYAD - SFM 3001 If external superelement matrices are used in SOL 101 or 103, then the run may fail with SFM 3001 in SMPYAD. Avoidance: Insert the following alter: compile extin alter 'smpyad xmap,x_k','mpyad xmap,x_p' type parm,,char8,n,name $ trnsp xmap/xmapt $ smpyad xmap,x_k ,xmapt,,,/kxx/3////////6 $ call fname /mxx,ZUZR11/S,NAME if ( name<>' ' ) smpyad xmap,x_m ,xmap,,,/mxx/3////1////6 $ call fname /bxx,ZUZR11/S,NAME if ( name<>' ' ) smpyad xmap,x_b ,xmap,,,/bxx/3////1////6 $ call fname /k4xx,ZUZR11/S,NAME if ( name<>' ' ) smpyad xmap,x_k4,xmap,,,/k4xx/3////1////6 $ mpyad xmap,x_p,/px $ @ 32870; 70.5.2; Nonlinear Transient Heat Transfer - CTRIAX6 with MATT4 - UFM 455 1 Nonlinear transient heat transfer models that use the 6 noded CTRIAX6 axi-xymmetric element will not converge and will fail with the message: *** USER FATAL MESSAGE 4551 (NL2CON) Avoidance: Use the 3 noded axisymmetric element instead. @ 32873; 70.5.2; P-Elements/Plotting - Wrong VUGRID ID for TRIAP elements The VUGRID id for TRIAP elements may be output as zero. This will cause missing points in deformation plots.

Avoidance: There is no avoidance. @ 32884; wers 70.7; Nonlinear Transient Heat Transfer - CBEND with MATT4 - Wrong Ans The CBEND element produces wrong answers in nonlinear transient heat transfer analysis. Avoidance: Use the CBEAM element. @ 32890;

70.6; Aeroelastic Analysis - GI, SPLINE2 - Unnumbered Fatal Message Spline checking improvements made in V70.5 can result in erroneous and unlabeled "FATAL ERRORs" out of the GI module. These errors occur if a SPLINE2 is attached to a single grid point (a situation that can produce a singular spline matrix). Due to the code error, ANY SPLINE2 with a single structural grid point will generate the singularity fatal, whereas those attached to a single grid point for both translation and rotations will produce the correct results. Unfortunately, a secondary error causes the message to be lost; hence, the unlabeled FATAL. Avoidance: In the case of a single grid point accounting for both translation and rotations, in order to avoid the erroneous fatal message, RBAR another GRID to it. The other GRID has no use in the analysis, but just allows the model to avoid the fatal message. Correct results will occur.

@ 32899; 70.6; Eigenvalue Analysis - EIGRL, READ, PARAM,AUTOSPC,NO - SFM 3001, SFM 3002 A model which contains degrees of freedom with no mass or stiffness may cause the Lanczos eigensolver to fail with one of the messages below. The most likely cause for this is submitting a model with PARAM,AUTOSPC,NO. This error is also likely in cyclic problems with solid elements and points on the axis of symmetry since these points are unaffected by PARAM,AUTOSPC,YES and are singular in the rotational degrees of freedom. *** SYSTEM FATAL MESSAGE 3001 (READ) THE INPUT DATA BLOCK IN POSITION 13 DOES NOT EXIST. USER INFORMATION: THIS ERROR IS CAUSED BY ONE OF THE FOLLOWING: 1. THE DATA BLOCK IS UNSPECIFIED ON THE DMAP MODULE. 2. THE DATA BLOCK IS SPECIFIED ON THE DMAP MODULE AND ON THE SUBDMAP STATEMENT BUT NOT ON THE CORRESPONDING CALL STATEMENT. *** SYSTEM FATAL MESSAGE 5299 (LNCS1D) UNRECOVERABLE TERMINATION FROM LANCZOS ITERATION INTERNAL ERROR CODE = -99 USER ERROR RETURN = *** *** SYSTEM FATAL MESSAGE 3002 (READ) EOF ENCOUNTERED WHILE READING DATA BLOCK SCRATCH (FILE 314) IN SUBROUTINE REIGL3 0FATAL ERROR Avoidance: Do not use PARAM,AUTOSPC,NO unless you are sure that there are no singularities in your matrices. The avoidance for solid element cyclic problems is to constrain the boundary grids in dof 456 or to select the

old Lanczos with system(273)=1. @ 32901; 70.5.2; Nonlinear Transient - NLRGAP, AUTO, TSTEP, TABK - Possible Wrong Answers When using the NLRGAP element, if the gap stiffness varies as a function of time, and either the AUTO or TSTEP method is selected on the TSTEPNL entry, then the answers may be incorrect. Avoidance: Use the default ADAPT method. @ 32906; ; Aeroelastic Analysis - SPLINEi Referencing Non-existent GRID - P oor Message If an aeroelastic model is created in which a SPLINEi entry references non-existent aerodynamic grid points (boxes), a misleading error message is issued in which the missing box id's are correctly listed, but the the message associates them to a SET1 or SPLINE3 entry that is, itself, nonexistent. Furthermore, the SET1/SPLINE3 association is only applicable to missing structural grids and, in this case, it is the aerodynamic grids that are missing. These data are on the SPLINEi entry and not on the SETi entry. @ 32909; s ; Nonlinear Transient Heat Transfer - PARAM,AUTOSPC - Wrong Answer The default for PARAM,AUTOSPC is NO for the nonlinear solution sequences, this includes solution 159 for nonlinear transient heat transfer. This is because AUTOSPC only acts on the linear portion of the kgg matrix and thus if it was YES it could lead to wrong answers. To prevent possible wrong answers, Version 70.5 ignores a YES setting for AUTOSPC. The correct action would be for AUTOSPC to act on the sum of the kggl and the kggnl matrices. The correction of this error will be to take this action. Avoidance: Do not use the PARAM,AUTOSPC,YES in transient heat transfer analysis. Identify any isolated/disjoint grids and constrain them.

@ 32917; 2001; Superelements/Plotting - SET, INCLUDE - All Elements Plotted ins tead of Subset When using superelements, any plot request which references a set definition using the INCLUDE option to select elements will plot the entire superelement. Avoidance: Use the EXCEPT feature. Example: SET 1 = ALL EXCEPT QUAD4 @ 32929; $ will plot all elements except QUAD4s.

2001; Documentation - P-Elements, PLOAD2, GP3 - UFM 2105 The MSC/NASTRAN Quick Reference Guide does not state that the PLOAD2 may not be applied to p-elements. If a PLOAD2 references a p-element the message: *** USER FATAL MESSAGE 2105 MODULE = GP3

PLOAD2 CARD FROM LOAD SET xxx REFERENCES MISSING OR NON-2-D ELEMENT OR HYPERELASTIC PLANE ELEMENT yyy will be encountered. receive UFM 2105 stating that the loading references a missing or non 2-d element or hyperelastic element. Avoidance: Use a PLOAD4 to load p-shells. @ 32931; ng Answers ; Restarts - MATi, RESTART - Restarts Ignore Material Damping, Wro

If a restart is performed from solution 106 (nonlinear static analysis) to solution 108 the GE field of the MAT1 entry is ignored. Avoidance: Specify PARAM,G damping instead. @ 32937; 70.6; Optimization/Design Sensitivity - UWM 9058, UFM 4574 If the direct sensitivity method is used in one iteration and then the adjoint method is selected by the program in a subsequent iteration, then the run will fail with the following messages: ^^^ USER WARNING MESSAGE 9058 (FEA) ^^^ THE PSEUDO SOLUTION AND THE PSEUDO LOADS FOR THE RESIDUAL STRUCTURE AND THE CURRENT SUBCASE ARE ZERO. *** USER FATAL MESSAGE 4574, DATA RECOVERY STARTED BUT NO DISPLACEMENT VECTORS AVAILABLE Avoidance: Insert the following DMAP alter: compile pslgdv alter 'isolapp\=1' psload=false $ endalter @ 32938; 70.5.2; P-Elements - TEMP, TEMPF - Wrong Answers If a model with p-elements has both TEMPF and TEMP entries, a solid element without a TEMPF entry will be assigned the default temperature, rather than using the nodal TEMP temperatures. Avoidance: A workaround is to use all TEMPF or all TEMP entries to assign temperatures. @ 32944; 2001; Documentation - PARAM,FLUIDSE - Missing Documentation The parameter fluidse is only documented in the Version 70.5 Release Guide, it is not documented in the Version 70.5 Quick Reference. Prior to Version 70.5 the fluid-structure analysis capability required that all fluid elements and all points (structure and fluid) at the fluid-structure interface be in the residual structure. With this parameter the structure points may now be assigned to upstream

superelements. @ 32945; ; Miscellaneous - BEGIN BULK - Wrong answers The BEGIN BULK command may contain the words SUPER=seid or AUXMODEL=auxmid. If any other word is input, then the BEGIN BULK command is not recognized as a delimiter and no related fatal or warning message is issued. Avoidance: There is no avoidance. @ 32951; ; Nonlinear Analysis - CGAP - Poor Answers when GAPs used with Mid -Side Nodes GAP elements do not work well with elements with midside grid points. Irregular stresses will be produced in elements near the GAPs. This problem is severe when quadrilateral faces are involved (HEXA-20) and moderate when triangular faces are used (TETRA-10). Avoidance: Use elements without midside nodes. @ 32960; 70.5.2; Restarts - GP2 and MCE2 - Poor Performance In the static solution sequences; e.g., SOLs 101, 114, 106, MCE2 may be reexecuted unecessarily during a restart. In SOLs 101-200, GP2 is reexecuted unecessarily. Avoidance: Insert the following alter in all runs including the restart: compile sekmr alter 'ENDIF \$ SEMR'(,-1) else if ( statics ) then $ purgex /goq/-1 $ compile phase0 alter 'gp2 ','' GP2 GEOM2,EQEXIN,EPT,geom2a,epta/ECT,ecta/S,N,ACOUSTIx $ alter 'fscoup\=','' FSCOUP=(ASCOUP='YES' AND ACOUSTIx=2) $ endalter @ 32961; 70.6; Aeroelastic Analysis - AMP, RESTART - SFM 4276 EC 1044 A restart into SOL 146 from SOL 103 may fail with SYSTEM FATAL MESSAGE 4276, ERROR CODE 1044 in the AMP module. Avoidance: If this error occurs, it is because the spline matrices are not being generated on the restart, insert the following DMAP alter. compile phase1a alter 'aspline'(,-1) delete /gpjk,gdjk,,, $ @ 32972; 70.5.2; P-Elements - PARAM,PEDGEP, FEEDGE - Parameter Ignored If PARAM,PEDGEP,1 is specified in the bulk data file, the POINTs on the FEEDGEs are not re-parameterized. This parameter only applies to cubic edges; for quadratic edges, the parameter is ignored.

Avoidance: There is no avoidance; however, it is unlikely that this inconsistency would negatively affect answers. @ 32973; ; SSSALTERS - alter2ga, alter2ha - Potential Inefficient Restarts In some instances, a restart from a run using alter2ga.vxx or alter2ha.vxx may try to "re-process" the superelement brought in using the alter. The cause of this is under investigation. Avoidance: Insert the following DMAP into the run which is failing (not the run with alter2ga or alter2ha). BEWARE - this DMAP will cause the program to ignore any changes made in the model during a run which uses it. compile chkgoaq alter 3,3 $ $ compile phase0 $ alter 581,585 $ V70 alter 694,697 $ V70.5 @ 32979; 2001; Documentation - Static Aeroelasticity - Missing Remark from PAER O2 Entry The following remark needs to be added to the Quick Reference Guide for the PAERO2 bulk data entry as follows: Remark: 8. For half models, the theta arrays LTH1 and LTH2 should encompass a full 360 degree range. @ 32981; 70.5.2; P-Elements - PSHELL, NSM - Wrong Answers If NSM (non-structural mass) is used on shell p-elements wrong answers may result. The NSM is interpreted as mass per unit volume instead of mass per unit area. This leads to wrong answers in the Grid Point Weight Generator, eigenvalues and static response due to gravity loads. Avoidance: The only avoidance is to adjust NSM to take into account the thickness of the element. @ 32983;

70.7; Superelements - SEBULK - SFM, EC 1044 IN FILPS2 When a sebulk entry references a seid that is not on a begin super entry, the following system error occurs: BIOMSG: ERROR 1044 HAS OCCURRED IN ROUTINE FILPS2 A log file message also appears, with a machine-dependent message.

@ 32987; 70.5.2; Nonlinear Analysis - CBUSH1D, SDR2 - SFM 6985 The BUSH1D element when used in Nonlinear Statics (106) will abort in the SDR2 module when STRESS data recovery is requested. The fatal message

is shown below. *** SYSTEM FATAL MESSAGE 6985 (SDR2RD) PREVIOUS ELEMENT-ID = 101 DID NOT REQUEST CORRECT NUMBER OF PHASE-1 OUTPUT WORDS. 0FATAL ERROR Avoidance: An avoidance is to request the BUSH1D data recovery with the NLSTRESS Case Control Command and to EXCEPT the range or BUSH1D elements from the STRESS Case Control Selection. @ 32991; 70.0.5; Nonlinear Analysis - PARAM,FOLLOWK, GRAV, RFORCE - SFM 3002, TAF F SOL 106 Follower forces. EOF ENCOUNTERED WHILE READING DATA BLOCK (NONE) (FILE 0) IN SUBROUTINE TAFMEG In nonlinear analysis, solution 106, when follower forces are used and there are GRAV and/or RFORCE inertial loads, the job may fail in the TAFF module with the message: SYSTEM FATAL MESSAGE 3002 EOF ENCOUNTERED WHILE READING DATA BLOCK (NONE) (FILE 0) IN SUBROUTINE TAFMEG. Avoidance: A DMAP avoidance is under investigation. Currently there is no practical avoidance other than inserting PARAM,FOLLOWK,NO or removing the inertial loads. Although this error exists in Version 70.5.0 (it was fixed in 70.5.2) the error correction was taken to Version 70.0.5 for NASTRAN for Windows. Therefore, the fixed version has been changed to 70.0.5 under the assumption that all 70.5 clients should have upgraded to 70.5.2. @ 33006; 70.5.2; Frequency Response Analysis - MPYAD, SOLs 108, 111, 146, 200 - P oor Performance The MPYAD module in subDMAP FREQRS for frequency response analysis that is used to compute the PDF matrix does not take advantage of the tendency of the dynamic load matrix to be very sparse and this leads to poor performance. Avoidance: Insert the following alter: compile freqrs alter 'mpyad '(,-1) trnsp ppf/ppft $ matmod ppft/vppf/12/s,n,nullppf/1 $ if ( nullppf>-1 ) then $ partn ppf,,vppf/ppf1,,,/1 $ partn rpx,,vppf/rpx1,,,/1 $ mpyad rpx1,ppf1,/pdf/1 $ else $ alter 'mpyad ' endif $ endalter @ 33012; 2001; Superelements - SPCs on Boundary Grids - SFM 6804

If a boundary (exterior) point for a PART (BEGIN SUPER = i) superelement is constrained, then the fatal message below will be issued if different displacement coordinate systems are used for the coincident GRID points. SYSTEM FATAL MESSAGE 6804 (BNDSP3) THE DISPLACEMENT COORDINATE SYSTEM (CD FIELD ON GRID ENTRY) SPECIFIED FOR BOUNDARY GRID ID 4 B IN SUPERELEMENT 1 IS NOT CONGRUENT WITH DISPLACEMENT COORDINATE SYSTEMS IN OTHER CONNECTING SUPERELEMENTS. Avoidance: Either apply the constraints in the downstream superelement (the one which has the coincident point as an interior point) or use the same displacement coordinate system on all of the points which are coincident. @ 33016; 2001; Documentation/QRG - SEEXCLD The Quick Reference Guide description for the SEEXCLD entry states that SEIDA(field 2) must be larger than zero. A value of 0 (zero) is also allowed. Also, the thru range is indicated as 61<62 it should be GIDA1<GIDA2 @ 33023; ; Superelements - SEPLOT - No Contour Plot Output Currently, the SEPLOT module does not support contour plots of superelements. Avoidance: There is no avoidance. @ 33026;

70.7; Nonlinear Buckling Analysis - UFM 3007 (UPARTN) from CALCV A nonlinear buckling analysis will fail in the NLSTATIC module with the message: *** SYSTEM FATAL MESSAGE 3007 (UPARTN) ILLEGAL INPUT TO SUBROUTINE CALCV if the run contains superelements and there are no spc or mpc sets in the residual structure. Avoidance: Apply the following dmap avoidance: compile nlstatic alter 'upartn.*dugni.*dulv'(2),'upartn.*dpgni.*dpl0' paraml usetnl//'uset'//////'G'/s,n,Gsize/'L'/s,n,Lsize $ message //'debug: from UsetNL, G-size is '/Gsize message //'debug: from UsetNL, L-size is '/Lsize if(Gsize = Lsize) then message //' fyi: copying dugni --> dulv, dpgni --> dpl0' equivx dugni/dulv/-1 $ else message //' fyi: doing normal upartn on DUGNI and DPGNI' UPARTN USETNL,DUGNI/DULV,,,/'G'/'L'/'M'/1 $ endif endalter

33027;

2001; Optimization/Design Sensitivity - DEQATN - Incorrect ID On CRAY systems DEQATN id's may be printed incorrectly in error message output.

@ 33045; 2001; Nonlinear Analysis - CGAP, TEMP(LOAD), ASETi - Wrong Answers at Omitted DOF Nonlinear analysis in Version 70 and 70.5 systems with multiple subcases, thermal loads, asets and gaps will give erroneous results at omitted degrees of freedom in the first printed load increment in the second subcase. If intermediate load increment results are not requested, the errors will occur in the final subcase. Avoidances: 1) Use a single subcase, which should work since nonlinear elements are not allowed to be omitted, or, 2) use superelements instead of ASETs, or, 3) printout intermediate load increments and ignore the first increments results of the second subcase, or, 4) use a version prior to 70. @ 33075; 70.7; Superelements - SECONCT - Wrong Answers If coincident grid points exist in a superelement and a SECONCT entry is used to connect one of them to a grid point in another superelement, it is possible that the SECONCT entry will not work properly and incorrect answers will result. If no SECONCT is used, then the program will correctly issue UFM 6794 Avoidance: Review the "BOUNDARY SEQUENCE ASSIGNMENT TABLE" for grids that are attached to multiple downstream GRIDS instead of the usual one to one correspondence. Do not create coincident grid points in a superelement where BEGIN SUPER is used. @ 33076; 70.5.2; Buckling - EIGRL, READ - UFM 4276 7777 signal 10 It is possible that the new Lanczos eigensolver may terminate with a machine-dependent error (signal 10 on an HP) followed by BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR in the READ module. Avoidance: For the model that demonstrated this error, adding V1 and V2 on the EIGRL entry in addition to the ND specification allowed the problem to be solved. Setting SYSTEM(273)=1 to select the old Lanczos method will allow the problem to be completed. @ 33078; 70.7; DMAP - CALL DBSTORE - Fails with Signal Trap If CALL DBSTORE and DIAG 47 are used then the run will fail with a hard crash in the PARAML module. Avoidance: Remove DIAG 47 or insert the following DMAP alter: compile dbstor

alter 't6','' type parm,,i,n,t6 $ endalter @ 33080; 70.7; Nonlinear Transient Analysis - NLTRD2 - BIOMSG: ERROR 1044 If a nonlinear transient analysis has only stiffness, but no mass or damping the job may fail with the message BIOMSG: ERROR 1044 HAS OCCURRED IN ROUTINE QOPEN 0 STATUS = 0 in NLTRD2. Avoidance: Add appropriate damping and mass to the model. @ 33101; 2001; Nonlinear Transient Analysis - TRLG - UWM 6498 Module TRLG issues the following message if a nonlinear transient analysis has a DLOAD command which refers to non-existant TLOADi entries: *** USER WARNING MESSAGE 6498 (MsgGetStr) Message number is not found PROGRAMMER INFORMATION: Check call to MSGPOP from TRLG with the message number 3031 Severity is USER WARNING Parameter 1 is 111 Parameter 2 is DLT Parameter 3 is TRLG USER ACTION: Please contact MSC Client support Avoidance: Remove references to non-existant loadings on the DLOAD entry. @ 33108; 2001; Documentation/Buckling - EIGRL - UFM 4405 The Lanczos buckling shift logic in MSC/NASTRAN Version 70.5 was enahanced. The enhanced code allowed many outstanding difficult problems to be solved without resorting to adjustments of the EIGRL entry. However, a very unusual case has been found that gives the Lanczos new shift logic trouble, wheras the old Lanczos shift logic solved the problem. The problem case occurs when there are on the order of thousands of eigenvalues below 1. and there are on the order of thousands of eigenvalues in the frequency range of interest. The old Lanczos shift logic would shift near V1 first, thus if ND was specified and there were ND roots near V1, the old shift logic would solve the problem. The new shift logic has a tendency to shift near 1. first or V2 if V2 is less than 1. The new logic solves the more common problem of users specifying 0. for V1 even though the first root is not near zero. Thus, if there are thousands of roots between V1 and V2 or below 1, the code finds those nearer to V2 or 1. but is never able to find the lowest ND roots. Avoidance: Several avoidances are possible. First it is unusual to pose a problem that has thousands of buckling modes below 1 so the model and the loading should be reviewed. A smaller eigenvalue range of interest could be specified or the SHFSCL field on EIGRL entry can be specified to force the first shift near the lower end of the V1, V2 range. @ 33113; 70.6; Documentation/Superelements - CSUPER - BIOMSG: ERROR 3021 from , FILE INDEX (FILEX) =

FILPS2 Although the documentation does not specifically state so, a superelement which is a copy (mirror or otherwise) of another superelement may not be copied or mirrored. Version 70.6 addresses this error by issuing a more meaningful fatal message. Avoidance: If additional copies are required, make more copies of the original superelement. @ 33118; 70.5.2; Installation - IRIX 6.5; libblas.so missing MSC/NASTRAN Version 70.5 does not run on IRIX 6.5 with a standard operating system configuration. Error messages such as those shown below will be encountered: MSC/NASTRAN V70.5 (Silicon Graphics IRIX64 6.5) Fri Aug 21 20:11:26 1998 MSC/NASTRAN beginning job test. PID: 4669. jever 5% 4735:/msc/msc705/irix64/analysis: rld: Fatal Error: Cannot Successfully Map soname 'libblas.so' under any of the filenames /msc/msc705/irix64/libblas.so:/usr/lib64/libblas.so:/usr/lib64/internal/l bblas.so:/lib64/libblas.so:/opt/lib64/libblas.so:/msc/msc705/irix64/libbl s.so.1:/usr/lib64/libblas.so.1:/usr/lib64/internal/libblas.so.1:/lib64/li blas.so.1:/opt/lib64/libblas.so.1: Avoidance: If the client does not have a complete set of sgi math libraries they will need to get a /usr/lib64/mips4/libcomplib.sgimath.so from MSC. This can be found at the following locations: On the web at: ftp.macsch.com/pub/msc-products/NASTRAN/V70.5/irix64 filename: libcomplib.sgimath.so filenames < then a list of library files follows >. @ 33119;

70.6; Elements - TRIAX6, TEMP(LOAD) - Wrong Answers The TRIAX6 element gives incorrect stresses if THETA (see CTRIAX6 entry) is nonzero, and there is thermal expansion. The deflections are not affected. The THETA field is most likely used for MAT3 orthotropic materials. Avoidance: The only known avoidance when there is thermal expansion, is to input the material properties in the r-z system, so THETA is zero.

@ 33121; 70.5.2; P-Elements - BEAMP - Element Coordinate System, Wrong Answers If the curvature of a beam p-element exceeds 90 degrees, the element coordinate system may reverse itself. For a beam with a symmetric cross section, this will probably only affect the data recovery points. Avoidance: Use beam p-elements that have a curvature of less than 90 degrees. @ 33131; ; Iterative Solver - P-Elements, Solvit - UFM 4157

The iterative solver may fail without issuing a meaningful message if the model contains adaptive p-elements. USER FATAL MESSAGE 4157 may be issued indicating insufficient memory; however, the message may not indicate the amount of memory required or it may indicate an incorrect amount. Avoidance: The failure may be the result of insufficient memory. Thus, increasing memory may solve the problem. If increasing memory does not solve the problem, then do not use the iterative solver. @ 33171; ; Elements - CBUSH1D, SHOCKA - No Warning Message By design intent NASTRAN allows the unmodified CBUSH1D element to be used in a linear static analysis where the linear initial stiffness is used. However, the SHOCK option has no stiffness and NASTRAN does not issue a warning message. @ 33181; 70.5.2; Performance - SDR2 - Poor Performance on NEC V70.5.1 The SDR2 module in V70.5.1 on the NEC system may be up to 60% slower than in V70.0. Avoidance: The problem has been fixed in Version 70.5.2. @ 33186;

; Nonlinear Analysis - RESTARTS - SFM 4252 Nonlinear analysis restarts may fail with the message: ^^^ PHASE 1 - SUPERELEMENT GENERATION, ASSEMBLY AND REDUCTION. *** SYSTEM FATAL MESSAGE 4252 (SEMA) MATRIX XAAV FOR UPSTREAM SUPERELEMENT 10 DOES NOT EXIST. USER INFORMATION: THE CURRENT SUPERELEMENT REQUIRES ALL OF IT'S UPSTREAM SUPERELEMENT MATRICES FOR ASSEMBLY INTO THE CURRENT SUPERELEMENT. THE PROBABLE CAUSES ARE: 1. THE UPSTREAM SUPERELEMENT'S DBSET OR DATABASE IS NOT ATTACHED. 2. THE UPSTREAM SUPERELEMENT DID NOT COMPLETE REDUCTION IN A PREVIOUS RUN. Avoidance: This problem is still under investigation but is believed to occur when temperature loads are applied in the cold start with TEMP(INIT) but are removed in the restart. Please see general limitation 4609.

@ 33192; NG

2001; Documentation - STATSUB in SOL 200 only supports ANALYSIS=BUCKLI Remark 1 of the MSC/NASTRAN Version 70.5 Quick Reference Guide's description of the STATSUB command indicates that STATSUB is supported in SOLs 103, 105, 107 through 112 and 200. However, the support in solution 200 is limited to ANALYSIS=BUCKLING.

@ 33202; wers

2001; Modal Response Analysis, Complex Eigenvalues. - GKAM - Wrong Ans It has been found that a very low memory request may result in poor complex eigenvalues being produced when using the modal method or poor answers in a modal response. In the case that demonstrated the problem the memory was set to one million words, well below the default of four

million words. The one million word value was barely above the minimum needed for subroutine xspace. This error is not expected to occur if more than 1mw of open core memory is available to nastran, which is the case if the default memory and buffsize are used. The error appears to have occurred in the GKAM module which assembles the modal matrices. Avoidance: Make sure that sufficient open core memory is available. @ 33233; 70.6; Inertia Relief - SUPORT, PARAM,INREL - Wrong Answers If an inertia relief analysis in solution 101 or 200 specifies less than six suport degrees of freedom in Version 70.5 of MSC/NASTRAN, then wrong answers may result. Avoidance: Apply the following dmap alter or use solution 1. COMPILE STATRS ALTER 'upartn ','fbs ' UPARTN USET,DRG/D6T,,,/'G'/'T'/'S'/2 $ TRNSP D6T/DT6 $ MPYAD DT6,DT6,/D6T2/1 $ MODTRL D6T2////6 $ DECOMP D6T2/LD6T2,,////////1.+8/16 $ MATGEN ,/IRR/1/NORSET $ UMERGE USET,DM,IRR/DML/'T'/'L'/'R' $ MPYAD DT6,DML,/DT6TDML/1 $ FBS LD6T2,,DT6T/CRYY $ compile inersen alter 'upartn ','fbs ' UPARTN USET,DRG/D6T,,,/'G'/'T'/'S'/2 $ TRNSP D6T/DT6 $ MPYAD DT6,DT6,/D6T2/1 $ MODTRL D6T2////6 $ DECOMP D6T2/LD6T2,,////////1.+8/16 $ MATGEN ,/IRR/1/NORSET $ UMERGE USET,DM,IRR/DTR/'T'/'L'/'R' $ MPYAD DT6,DTR,/DT6R/1 $ FBS LD6T2,,DT6R/CRYY $ Endalter @ 33239; 70.5.2; Nonlinear - GPSTRESS, QUAD4 - Incorrect results In a nonlinear analysis with PARAM,LGDISP,1, the grid point stress (GPSTRESS) output for grids attached to CQUAD4 elements with a STRESS(CORNER) request will be incorrect. Values of 1.0E+30 are printed for all components. Avoidance: Do not request STRESS(CORNER) and GPSTRESS in the same run. @ 33242; 70.6; Aeroelastic Analysis - SOL 145 restart - UFM 5423 in AMP module Restarts into SOL 145 or 146 may fail with UFM 5423 in the AMP module. Avoidance: Insert the following DMAP alter: compile phase1dr alter 'call phase1a'(,-1) delete /gpjk,gdjk,gdak,gpak, $ delete /gpgk,gdgk,,, $

compile pfaero alter 'DELETE \/D1JK,D2JK','' DELETE /D1JK,D2JK,wskjf,, $ endalter @ 33272; 70.6; Eigenvalue Analysis - READ, REIGL - Indefinite Mass Matrix, Wron g Answers If the mass matrix is indefinite, then the roots coming out of Lanczos may not be correct. A symptom of this error is the issuance of the following error messages: INERTIA (STURM SEQUENCE) COUNT DISAGREES WITH THE NUMBER OF MODES ACTUALLY COMPUTED IN A (PARTIAL) INTERVAL FOUND xxx MODES BETWEEN aaaa HZ AND bbbb HZ BUT THERE SHOULD BE ONLY yyy MODES. Note that in almost all cases, an indefinite mass matrix is caused by modeling errors. Starting in Version 70.6, the job will terminate with the fatal message below if an indefinite mass matrix is encountered. USER FATAL MESSAGE 4646 (REIG) THE MASS MATRIX INPUT TO THE REAL EIGENSOLVER IS INDEFINITE. USER ACTION: REMOVE THE NEGATIVE MASS TERMS. If GDR is being used (DYNRED) it is possible that poor numerics (very small negative eigenvalues) may lead to the above fatal message in Version 70.6. For these cases the fatal message may be avoided by adjusting the zero tolerance with SYSTEM(60). The default value 6, which gives a zero value of 1.0E-6. Thus a value of 4 will relax the zero tolerance by two orders of magnitude. Avoidance: Modify the model to remove the indefinite mass matrix.

@ 33274; 23

70.6; Coupled Fluid/Structural Analysis - Fluid Superelements - SFM 54 If PARAM,FLUIDSE,>0 and any structural grid points or SPOINTs are assigned interior or exterior to the fluid superelement then SFM 5423 may be issued. Avoidance: Ensure that no structural points appear in the interior and exterior point list for the fluid superelement. This can be verified in the printout of the superelement map. This includes generalized coordinates used for component mode synthesis on the fluid superelement. The seqset1 should be input on fluid points (grid, CD=-1) rather than scalar points.

@ 33285;

; Plotting - Aeroelastic Elements - Missing Plot Output Beginning in Version 70.5 the NASTRAN plotter no longer supports aeroelastic elements. Avoidance: The only avoidance is to use Version 70 or PATRAN Version 8.1 which is due out by the end of this year.

33299;

70.7; P-Elements - QUAD4 - Poor Answers for p=1 The QUAD4 p-element may be too stiff when p=1 on some edges. This is most noticible for coarse meshes. An avoidance is to specify p >= 2.

@ 33307;

70.6; Aeroelastic Analysis - AELINK, TRIM - UFM 5267 The above message was produced by a client whose length units was millimeters and mass units were kilograms. The trim matrix became singular in this case, apparently because of the dynamic range within the equations to be solved. *** USER FATAL MESSAGE 5267 (ASGLD) ZXX MATRIX IS SINGULAR. TRIM AND AELINK ENTRIES SPECIFY AN INADEQUATE CONSTRAINT SET. Avoidance: In this case, the problem was avoided by scaling the ZZX and PZ matrices going into the trim module by 1.0e-4. This avoidance is likely to be problem dependent, so other factors may be required in other cases. The alter to perfrom the scaling is: compile aestatrs alter 'asg'(,-1) add zzx,/zzxs/1.0e-4 $ add pz,/pzs/1.0e-4 delete /zzx,pz,,,/ $ equivx zzxs/zzx/always $ equivx pzs/pz/always $ A more general avoidance is to scale the ZZX matrix so the magnitude of its largest term is unity. Scale the PZ matrix by the same term. This will work in more cases: compile aestatrs list noref alter 'asg ','asg ' type,parm,,rs,n,zzxnorm,xnorm,inorm $ type,parm,,i,n,ncol,nrow $ type,parm,,cs,n,cnorm $ file junk=ovrwrt/zzx2=ovrwrt/pz2=ovrwrt $ norm zzx/junk/s,n,ncol/s,n,nrow/s,n,zzxnorm/1 $ if ( zzxnorm >< 0.0 ) inorm = 1.0 / zzxnorm $ cnorm = cmplx(inorm) $ add zzx,/zzx2/cnorm $ add pz,/pz2/cnorm $ asg casea,zzx2,pz2,xlist,edt,tr,krzx,dit,erhm/ ux,uxdat $ mpyad zzx,ux,pz/balance//1/-1 $ $matprn zzx,zzx2,pz,pz2,ux $ $matprn balance,,,, $ norm balance/junk/s,n,ncol/s,n,nrow/s,n,xnorm/1 $ message //' xnorm of balance is '/xnorm $

@ 33308;

; Sequencing - SEQP - Memory Allocation Error Very large problems that use the metis method of sequencing (system(206)=8) may fail with a memory allocation error as follows:

*** SYSTEM INFORMATION MESSAGE 6916 (DFMSYN) MALLOC FAILED INSIDE METIS. DECREASING THE MEMORY GIVEN TO THE JOB WILL INCREASE THE MEMORY AVAILABLE FOR MALLOC. DECOMP ORDERING METHOD CHOSEN: METIS, ORDERING METHOD USED: MMD Dynamic memory starts at address 000000008BE56940 *** Opencore memory reduced to 49791588 *** Metis Error: ***Memory allocation failed for CreateCoarserGraph: gdata. Requested size: 21237992 bytes time: command terminated abnormally. Avoidance: There is no guaranteed avoidance. Increase the memory or use another sequencing method. Requesting system(206)=72 requests the metis method with a different pre-phase that typically uses 30% less memory. @ 33309; 2001; External Superelements - PARAM, EXTOUT, DMIGPCH - UFM 327 When using the external superelement capabilities with the "param,extout,dmigpch" option in a SOL 103 run, the header entry for the load PAX will also be written out, even though it is not used. This will cause the issue of the following error messages in a subsequent attachment run. *** USER FATAL MESSAGE 327 (IFS2P) BAD DATA OR FORMAT OR NON-UNIQUE NAME. DMIG PAX CARD COUNT = 161 ^^^ USER FATAL MESSAGE 9002 (IFPS) ^^^ ERROR(S) ENCOUNTERED IN THE MAIN BULK DATA SECTION ^^^ SEE MESSAGES ABOVE. ERROR ENCOUNTERED IN MODULE IFP, SORTED

Avoidance: Remove the offending PAX header. This entry is not used in a SOL 103 run.

@ 33334;

2001; Documentation - DEQATN - Increased Limit The number of allowable non-blank characters for the DEQATN entry was increased from 12,500 to 32,000 in Version 70; however, the Quick Reference was not updated to reflect this.

@ 33367;

70.6; Optimization/Design Sensitivity - DESVAR - UFM 316 The message: *** USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA ENTRY DESVAR xx SORTED ENTRY COUNT = yy was produced when a DESVAR punched at the end of an optimization task had an XINIT that was very slightly less than the XLB value. When a new run is attempted, an equality test is used to see if the input satisfies XINIT .ge. XLB and the entry is rejected even though it had been produced in a previous run. Avoidance: Manually reset any punched DESVAR values to satisfy the relations XLB .LE. XINIT .LE. XUB

33401;

; Case Control - SEALL, SEMG, SEKR, etc. - UFM 9015, Poor Message If the SEALL, SEMG, SEKR, SELG, SELR, or SEMR command is specified in a subcase which is not the first subcase then the operation requested by that command is not performed and various messages or wrong answers may result. For example, with SEALL or SEMG then UFM 9015 will be issued. USER FATAL MESSAGE 9015 (SEMODES) THE USET TABLE IS NOT PRESENT. Avoidance: Specify SEALL, SEMG, SEKR, SELG, SELR, or SEMR above all SUBCASE commands.

@ 33424; 70.6; Aeroelastic Analysis - SPLINEi, AESTATRS - UFM, Incompatible Ma trix Multiply An incompatible matrix multiply can occur in calculating the intercept hinge moment coefficients if the aerodynamic forces are splined only to dependent degrees of freedom (e.g., to m-set and r-set points). Avoidance: A DMAP workaround is to include the following alter: compile aestatrs $ alter 'add *uintl','add *intl' $ type parm,,i,n,nointl $ dbstatus uintl//s,n,nointl $ if ( nointl < 1 ) then $ PARAML ALX//'TRAILER'/2/S,N,NOLSET $ MATGEN ,/MUNTL/7/NOLSET/1 $ MATGEN ,/MINTL/7/NOLSET/1 $ else $ ADD UINTL,/MUNTL/-1.0 $ ADD INTL,/MINTL/-1.0 $ endif $ endalter $ @ 33427; 70.6; Aeroelastic Analysis - AELINK, RESTART, MPYAD - SFM 5423 An error in the dependencies results in restarts into solution 144 where only the AELINK entries have been changed failing with SYSTEM FATAL MESSAGE 5423 indicating an attempt to multiply incompatible matrices: SYSTEM FATAL MESSAGE 5423 (DMPYAD) ATTEMPT TO MULTIPLY INCOMPATIBLE MATRICES COLS ROWS FORM TYPE NZWD DENS TRAILER FOR WSKJ IS 101 40 80 2 2 4 250 TRAILER FOR FFAJI IS 102 7 40 2 2 80 10000 TRAILER FOR FFINT IS 103 2 80 2 2 0 0 Avoidance: Do not attempt restarts when only changing AELINK bulk data entries. A simultaneous change in the aerodynamic model geometry will enable the restart, but this would also entail recalulation of the aerodynamics. Alternatively, the following DMAP alter may be inserted as long as there are no other changes to the aero model. compile ifpl alter 'RESTART EDT,EDTR','' endalter @

33454; 2001; Documentation - EIGC, SYSTEM(108) - Block Complex Lanczos is not the Default The more modern and reliable Hessenberg which supports spill was made the default in Version 69 by switching the meaning of system(108)=0 and 1. However, the documentation was never updated to reflect this change. Furthermore, because of poor performance for some problems the new Block Complex Lanczos was not made the default in Version 70.5. Although this decision is reflected in the release notes the decision was made too late to be reflected in the Quick Reference Guide's description of system(108) under the NASTRAN command and EIGC entry. @ 33476; 70.6; Miscellaneous - Elapsed Time Incorrect The elapsed time will be incorrect if more that 24 hours are spent in any one module. @ 33480;

70.6; Postprocessing - SOL 1 through 16, PARAM,POST - Missing output No results will be written to the .op2 or .xdb file if PARAM,POST is specified in SOLs 1 through 16 where SORT2 is the default output format or in all SOLs if SORT2 is explicitly requested. Avoidance: Insert the following DMAP alter in any SOL 1 through 16: compile sol__ $ insert SOL number for __ alter 'IF (POST=0) THEN'(,-1) LABEL LNOSORT1 $ alter 'LABEL LNOSORT1','' endalter

@ 33481;

70.6; Estimate - SPC - Undefined GRID referenced When an SPC references an SPOINT, the estimate utility may indicate that a undefined GRID has been referenced.

@ 33491; 70.6.1; Optimization/Design Sensitivity - CAABSF - UFM 7115 The acoustic absorber element (PAABSF) is not supported in optimization and design sensitivity (SOL 200). If a PAABSF entry is included in the bulk data, even if it is not one of the design variables, the following message is issued, " *** USER FATAL MESSAGE 7115 (DOM2B) PTYPE 1502 15 IS NOT SUPPORTED". Avoidance: Do not include PAABSF entries in an optimization run. @ 33495; xdb

; Postprocessing/Random Analysis - PARAM,POST,0 - Missing Data in The results from a response spectrum spectrum analysis are not written into the ".xdb" file when PARAM,POST is set to 0. Avoidance: Set PARAM,POST,-1 and they are written into the OUTPUT2 file.

@ 33499;

; Elements - PSOLID - Spurious Modes Spurious hourglass modes have been observed when solid elements refer to an anisotropic smeared material property that has a modest shear modulus, but almost zero extensional modulus.

Avoidance: Put ISOP=1 on the PSOLID record for these elements. @ 33500; ; Static Aeroelastic Analysis - Wrong Answers for Multiple Interfe rence Groups If multiple interference groups are used with bodies in any interference groups except the last one and control surfaces exist, the DJX matrix generated by ADG (adgdjx routine) is corrupted. Avoidance: Specify a single interference group. @ 33517; ; Installation - Missing FLEXlm Installation Documentation The Version 70.5 Configuration and Operation Guide for NT does not discuss the installation of the FLEXlm license manager. This software is provided on the Version 70.5 NT CD but must be installed separately. Please see the "install.htm" on the NT-CD itself for installation instructions. @ 33527; 70.6; Iterative Solver - ITER, JACOBI - SWM 6715 An error in the iterative solver for real-unsymmetric problems using the default Jacobi preconditions may lead to poor convergence and the message below: *** SYSTEM WARNING MESSAGE 6715 (SUCGRD) THERE IS A ZERO DIAGONAL TERM IN COLUMN 1 OF DATA BLOCK CEE WHICH WILL BE REPLACED BY 1.0 IN THE PRECONDITIONER. USER ACTION: IT MAY BE BETTER TO USE CHOLESKY PRECONDITIONING (PRECOND=C ON ITER BULK DATA ENTRY) OR THE DIRECT SOLVER (ITER=NO ON NASTRAN STATEMENT) FOR THIS PROBLEM. Avoidance: The error is caused because instead of using the diagonal terms of the stiffness matrix as the Jacobi preconditioner (as intended) the iterative solver is running with an identity matrix as the preconditioner which is the same as running with no preconditioner at all. Use Cholesky preconditioning by setting PRECOND=C on the ITER bulk data entry. @ 33529; 70.6; DBC - CTRIAR, PARAM,POST - UFM 6498 If a run with a ctriar element is execute back to back and the xdb file is not removed, the job will fail with the following message unclear message: *** USER FATAL MESSAGE 6498 (MsgGetStr) Message number is not found PROGRAMMER INFORMATION: Check call to MSGPOP from with the message number 0.0 Severity is INVALID ERROR TYPE (0) Parameter 1 is 3

Parameter 2 is CTRIAR Parameter 3 is CTRR Parameter 4 is 0 USER ACTION: Please contact MSC Client support *** USER FATAL MESSAGE 6498 (MsgGetStr) Avoidance: Delete the xdb and re-submit the job. @ 33533; ; Superelements - Component Modes, SECSET - Possible Wrong Answers If free-free component modes are used (e.g., secset), then the eigenvectors may be incorrect. A symptom of this error is indicated by the discrepancies between the eigenvalues printed in the eigenvalue table versus the eigenvalues printed at the top of the eigenvector output. Avoidance: Do not use free-free components, use the default fixed-fixed component modes. See related error 32365. @ 33543; 70.6; Estimate - Recommends Setting SPARSE=1 The estimate program may incorrectly recommend turning sparse decomposition off (i.e. setting SPARSE=1) for tet models. Sparse decomposition is virtually always better than standard decomposition for decomposing stiffness matrices for real world models. Sparse should never be turned off. @ 33559; 2001; Documentation - PSHELL, STRESS(MAXS) - MAXS is not an Equivalent Stress A remark should be added to the Quick Reference guide description of the STRESS case control command that MAXS for shell elements is not to be confused with an equivalent stress as is the case for von Mises stress or the octahedral stress. @ 33566; 2001; Static Analysis - PARAM,EXTOUT,DMIGPCH - UFM 327 (IFS2P) If param,extout,dmigpch is used in but there are no loads (for example in an eigenvalue analysis) a header record for the "PAX" load matrix is printed to the punch file. When this punch file is included in the bulk data of a subsequent run, the job will fail with the following message: *** USER FATAL MESSAGE 327 (IFS2P) BAD DATA OR FORMAT OR NON-UNIQUE NAME. DMIG PAX CARD COUNT = 52 ^^^ USER FATAL MESSAGE 9002 (IFPS) ^^^ ERROR(S) ENCOUNTERED IN THE MAIN BULK DATA SECTION ^^^ SEE MESSAGES ABOVE. ERROR ENCOUNTERED IN MODULE IFP, Avoidance: @ 33568; M 3007 Edit the punch file and delete the DMIG entry for PAX. SORTED

70.6; Optimization/Design Sensitivity - DSAD, PARAM,NOCOMP, PCOMP - SF When param,nocomps,0 is used in an optimization run with composite stress or strain responses the following message will occur. *** SYSTEM FATAL MESSAGE 3007 (DSAD ) ILLEGAL INPUT TO SUBROUTINE DSAD12 Avoidance: Use param,nocomps,1 which is the default.

@ 33572; 70.6; Installation - MSC_LICENSE_FILE env var supersedes RC file "auth " keyword The precedence of the "authorize" keyword hierarchy in the "MSC/NASTRAN Configuration and Operations Guide", Section 3.2, is incorrect. The actual hierarchy is: 1) 2) 3) 4) 5) "authorize" keyword on command line. MSC_LICENSE_FILE environment variable. "authorize" keyword in an RC or RCF file. <install-dir>/flexlm/licenses/license.dat <install-dir>/conf/authorize.dat <install-dir>/conf/authoriz.dat 6) LM_LICENSE_FILE environment variable.

# Windows NT

Note, the first three items are consistent with all other nastran command keywords. @ 33573; 70.6; Optimization/Design Sensitivity - DOPTPRM - Mislabeled Output The P2 parameter on the DOPTPRM entry is used to control the output of design optimization results. The output that is produced for composite responses is scrambled in that CFAILURE output appears as CSTRESS, the CSTRESS appears as CSTRAIN and the CSTRAIN appears as CFAILURE. Avoidance: The numerical values are correct. It is only the labeling that is in error. @ 33588; 70.6; Optimization/Design Sensitivity - DSAD, CSTRESS, CSTRAIN, CFAILU RE - SFM 3007 The message: *** SYSTEM FATAL MESSAGE 3007 (DSAD) ILLEGAL INPUT TO SUBROUTINE DSAD13 can be produced if there is a combination of CSTRESS, CSTRAIN, and CFAILURE response types on DRESP1 entries and they invoke different sets of elements. Avoidance: Either only use one composite response type or apply the multiple response types to exactly the same set of elements. @ 33600; 70.6; Aeroelastic Analysis - RESTART, MKAERO - Wrong Answers If a restart is selected with additional MKAERO data, only the new MKAERO pairs will be used in the Flutter or Dynamic Aeroelastic analysis and the original data will be ignored. Avoidance: For SOL 145, apply the following dmap alter in the restart run: compile flutter alter 'cntr'(2,-1) endif $ alter 'endif'(2),'endif'(2) For SOL 146, the appropriate alter is: compile seaero

alter 'cntr'(2,-1) endif $ alter 'endif'(7),'endif'(7) @ 33606; CALCV 70.7; SSSALTERs - checka.v705 - SFM 3007, Illegal Input to Subroutine If the sssalter checka.v705 is used on a deck that has no spcs or rigid elements then the message: SYSTEM FATAL MESSAGE 3007, ILLEGAL INPUT TO SUBROUTINE CALCV may be encountered. Avoidance: Put in a dummy grid point and constraint. @ 33626; C) 70.6; Postprocessing - Nonlinear Analysis, PARAM,POST,0 - SWM 6066 (DB In nonlinear analysis with superelements, MSC/NASTRAN Version 70 and 70.5 may fail with messages similar to those below when it attempts to write results to the XDB file. *** SYSTEM WARNING MESSAGE 6065 (DBC) *** DBCMSG ERROR MESSAGE 44 FROM SUBROUTINE DBCAOL OUTPUT LIST DOES NOT MATCH EXISTING IN THE DATABASE, MERGE/RESTART OPERATION IMPOSSIBLE. MSC/XL DATA BASE CONVERSION FINISHED *** SYSTEM WARNING MESSAGE 6066 (DBC) *** DBCMSG ERROR MESSAGE 34 FROM SUBROUTINE DBC THE DBC MODULE HAS BEEN TERMINATED BECAUSE OF THE ABOVE FATAL DBC ERROR. USER INFORMATION: SUBSEQUENT DBC CALLS WILL NOT BE EXECUTED, AND THE DATABASE CONVERSION IS INCOMPLETE. *** DBCMSG ERROR MESSAGE 14 FROM SUBROUTINE DBCROL UNABLE TO OPEN INPUT RELATION NLLF RETURN CODE = 1. *** SYSTEM WARNING MESSAGE 6065 (DBC) *** DBCMSG ERROR MESSAGE 14 FROM SUBROUTINE DBCROL UNABLE TO OPEN INPUT RELATION NLLF RETURN CODE = 1. Avoidance: If PATRAN is being used for post processing the op2 option may be used. Requesting sort2 rather than sort1 output may also solve the problem. @ 33628; ; Superelements - CSUPEXT - UWM 4155 If an image superelement has SPOINTs for component mode synthesis the the following warning message may be incorrectly encountered: USER WARNING MESSAGE 4155---GRID POINTS FOR SECONDARY SUPERELEMENT 82 ARE COLINEAR. CONFORMITY CHECKS NOT MADE. @ 33629;

; Documentation/P-Elements - CBEAM, CQUAD4, CTRIA3 - SFM 6352 If beam and plate p-elements are used in random analyses, the following

uninformative message is printed in the .f06 file: *** SYSTEM FATAL MESSAGE 6352 (SDRP36) LOGIC ERROR: KERROR= 7. USER ACTION: CONTACT MSC CLIENT SUPPORT. PROGRAMMER INFORMATION: SUBCASE ID = 101 IN RECORD NUMBER 29 OF OUGV1 DATA BLOCK DOES NOT MATCH THE SUBCASE ID = 102 IN RECORD NUMBER 2 OF CASEDR DATA BLOCK. 0FATAL ERROR Avoidance: There is no avoidance, p-elements are not supported in random analysis. @ 33644; nswers 70.7; Transient Heat Transfer - Multiple subcases with DLOAD - Wrong A Different DLOADs may not be applied in different subcases of a transient heat transfer analysis. If attempted the load applied in the first subcase will be lost in the second subcase. Avoidance: Loads should be combined in a single DLOAD. If multiple subcases are necessary to reference different tstepnl entries, than the DLOAD case control request should be put above the subcase level. @ 33647; ; Follower Force - FORCE1 - Wrong Answers, Poor Convergence In all linear solutions, the follower stiffness is erroneous when FORCE1 loads are applied. The convergence will be degraded in solution 106. Avoidance: Use PARAM, FOLLOWK, NO to turn off the follower stiffness or apply the loads using another type of force entry. @ 33705; ; Nonlinear Analysis - CELASi - Missing Punch Output A request for elforce(punch) output for celas elements in nonlinear analysis produces no output. Avoidance: Request stress(punch) instead. @ 33725;

; Matrix Methods - MPYAD - System Fatal Message If insufficient scratch space is available for a large mpyad, the job may crash and a meaningful message may not be output. Avoidance: Increase the amount of scratch space available ot the job.

@ 33726; 70.7; Optimization/Design Sensitivity - Aeroelastic Analysis, FLUTTER - SFM 7088 The message: *** SYSTEM FATAL MESSAGE 7088 (DOPR3I) MODE NO. xx ASSOCIATED WITH FLUTTER DRESP1 ID = yy EXCEEDS 0, THE NUMBER OF MODES AVAILABLE FROM THE ANALYSIS.

may be produced in error when NVALUE is not explicitly specified on the FLUTTER entry. Since there is a default value for NVALUE, the input may be valid. The message is also incorrect in that it is (a) listed as a system fatal message when it should have been a user fatal message and (b) the run does not stop despite the supposedly fatal message. Avoidance: If the default value has been used for NVALUE, the message can be ignored. @ 33749; 2001; Composites - PCOMP - Max Strain Ply Failure Index When STRAIN ply results are requested, failure indices are not calculated. Avoidance: Supply a STRESS=n also, where n is the set of elements that will have failure indices calculated, but you get the extra ply stress results in addition to the ply strain results..

@ 33753; 21

2001; Documentation/QRG - Replication example in QRG results in UFM 67 The replication example in the Version 70.5 Quick Reference Guide on page 427 results in USER FATAL MESSAGE 6721. The example: GRID, 101, 17, 1.0, 10.5,,17,3456 =(4),*(1),=,*(0.2),==$ Should be changed to: GRID, 101, 17, 1.0, 10.5,,17,3456 =,*(1),=,*(0.2),==$ =3

@ 33770;

70.7; Documentation - Submittal - submit - Poor Message A Bourne shell cannot be used to process the job file in a user's submit command. If a Bourne shell is used, the following error may occur: -sh: MSC_ASG=/Stemp/TT7293_19.asg: is not an identifier The error message: MSC_ASG=/Stemp/TT7293_19.asg: is not an identifier was written by the Bourne shell, /bin/sh, and was complaining about an export command. Avoidance: Do not specify a Bourne shell. Section 3.11 of the "Configuration and Operations Guide" does specifically state that a Korn shell must be used. However, the example should be changed from "-s /bin/sh" to "-s /bin/ksh". Note, the use of "/bin/sh" is correct on a POSIX-compliant system, but could be misleading to an inexperienced user.

33775;

; Optimization/Design Sensitivity - MODES - Missing Output When an optimization job with ANALYSIS=MODES uses more than one MODES statements in Subcase Control section, the results are printed in *.f06 file only for the first subcase. Avoidance: Do not use the MODES command.

@ 33776;

70.6; Aeroelastic Analysis - AEROF - Null AEROF Output The AEROF output from solution 146 is all zeroes in Version 70.5. For frequency response analysis, the AEROF case control command should produce force and moment data for the the aerodynamic boxes of the DLM. Avoidance: A workaround is to supply a Case Control SET having the desired ids and referencing that SET in the output request: AEROF=<setid>

@ 33779;

70.6; Grid point stress - PSOLID, GPSTRESS, CORDM - Wrong Answers When higher order solid elements are used in Grid Point Stress Volumes, and their PSOLID references a material coordinate system, the local grid point stress will be incorrect. The invariants, such as Pressure or Von Mises are correct. Avoidance: Use the default basic coordinate system.

@ 33797;

2001; Aeroelastic Analysis - GUST - Documentation Clarification The Version 64 Programmer's Manual indicates that the X0 field of the GUST bulk data entry will change the phase relationship (it can be viewed as a time delay) whenever a harmonic gust is used in SOL 146. Also the streamwise location of aerodynamic elements introduces a phase change (a delay) in the aerodynamic forces caused by a harmonic gust. The statement on page 606 paragraph 2 is not completely accurate. The X0 and the aerodynamic element center streamwise location are terms in the equation that calculates the real and imaginary parts downwash for each aerodynamic element. The phase relation represents a time delay for an excitation at X0 to arrive at an aerodynamic box streamwise position.

@ 33809; 70.6; Aeroelastic Analysis - SPLINEi, SUPORT - Incorrect Unrestrained Hinge Moments The unrestrained hinge moments are in error in static aeroelastic analysis if the aerodynamic and structural model coupling splines the displacements from the suport points to the aerodynamic mesh. This error is caused by omitting the suport point displacements from the computation of the aerodynamic mesh displacements for aeroelastic feedback. Avoidance: A DMAP avoidance is: compile aestatrs list noref alter 'mpyad *gdkl.*ukpi','' UMERGE USET,ULPI,URPI/UAPI/'A'/'L'/'R' $ mpyad gdka,uapi,/ukpi $ capture the r-set motion, too @ 33818;

70.7; Buckling Analysis - large problem - READ, Segmentation Fault

On a very large buckling problem, 2.6 million degrees of freedom in this case, depending on the machine, the job may fail with one of the following error messages either in the log or f06 files. The error is caused by trying to solve THE BUCKLING SHIFT SCALE CALCULATION FBS OF a very large problem in core instead of out of core. Trap encountered: SIGSEGV - Segmentation violation Traceback: Offset Offset Offset Offset Offset Offset Offset Offset Offset Offset Offset Offset 0x0000007c 0x00000488 0x00000b50 0x00000040 0x000003f8 0x0000041c 0x00000598 0x0000039c 0x000010c4 0x0000252c 0x000000f4 0x00000958 in in in in in in in in in in in in procedure procedure procedure procedure procedure procedure procedure procedure procedure procedure procedure procedure dbsmr1_ gbkmgr_ ionast_ girdbk_ sfupfc_ ssmf1d_ ssm5d_ bucksd_ lnnrigl_ reig_ xsem06_ mainal

--- End of call chain ------------------------------------------------------------***** RECEIVED SIGNAL 11, ABORTING ... = BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX (FILEX) 0

Avoidance: Possible avoidances are to set shift scale on the eigrl entry or increase memory.

@ 33841; 70.6; Nonlinear Analysis - Inaccurate Stresses using Drucker-Prager Yi eld Criterion The Mohr-Coulomb and Drucker-Prager yield criteria (selected on the MATS1 entry) entry create an inaccurate solution. The error is the result of an incorrect correction factor to the yield stress, which accounts for effects of the internal friction angle. Incorrect factors underestimate the yield stresses, resulting in premature yielding in tension. Avoidance: There is no avoidance, but the error is negligible if the strain is very small. @ 33860; 70.6; Miscellaneous - Execution Summary - "nan" in the f04 Output If MSC/NASTRAN is run on a computer which has more than 68 CPUs. Then the F04 file may have "nan" in the output. 15:44:41 0:00 0.0 0.0 nan nan SEMTRN BGN

Avoidance: There is no avoidance but the "nan" ouput may be safely ignored.

@ 33861; inated

70.7; Optimization/Design Sensitivity - DSAJ, DVGRID - abnormally term Shape optimization jobs may fail in the DSAJ module on machines that enforce boundary alignment. This includes SUN, HP and Digital Alpha. The log file will contain a message of the form; *** TERMINATING /software/msc/nastran705/msc705/solaris/analysis.ultra *** Received signal 10 (SIGBUS) time: command terminated abnormally. and then a machine dependent traceback: from SUN: t@1 (l@1) signal BUS (invalid address alignment) in unpack_ at 0x8fcc58 unpack_+0x1d8: std %f2, [%o1] (dbx) where =>[1] unpack_(0x275, 0x1, 0x1dbb92c, 0x1a494e0, 0x1a49430, 0x0), at 0x8fcc58 [2] dsaj4_(0x18a88d4, 0x244598, 0x18a8680, 0x18a86d8, 0x1f6d010, 0x18a8638), at 0x244810 [3] dsaj_(0x18a8680, 0x18a8690, 0x1358, 0x18a8574, 0x0, 0x18a8578), at 0x243510 [4] xsem02_(0x123, 0x18852b0, 0x133eec4, 0x32, 0x32, 0x1870e70), at 0x11292c [5] MAIN(0x186d400, 0x6b414, 0x0, 0x2, 0x0, 0x7), at 0x6bb68 (dbx) From HP: ( 0) 0x0019b1cc sysrcv + 0x19c [/msc/msc705t2/sppux/analysis.pa2_0] ( 1) 0x011bc1d0 _sigreturn [/msc/msc705t2/sppux/analysis.pa2_0] ( 2) 0x01606ed0 P__FLUSHLINE + 0x437b38 [/msc/msc705t2/sppux/analysis.pa2_0] Stack_Trace: error while unwinding stack From Dec/Alpha I get lots of: Unaligned access pid=8453 <analysis> va=0x145001cfc pc=0x1202e0d74 ra=0x1202e0e14 inst=0x9fe70000 Unaligned access pid=8453 <analysis> va=0x145001d04 pc=0x1202e0d7c ra=0x1202e0e14 inst=0x9fe70008 Avoidance: Once the error is encountered, it may be avoided by adding a dummy design variable to the deck.

@ 33863;

70.7; Buckling Analysis - PARAM,FOLLOWK, TAFF - Poor Performance If follower force stiffness is turned on (which is the default) in a buckling analysis, and there are many load entries, then the calculation of the follower force stiffness will be extremely slow. Avoidance: If the effect of the follower force stiffness is not important, add the following parameter to the bulk data section: param,followk,no

@ 33874;

2001; Documentation - EIGRL, SYSTEM(146) - Incorrect Description The descriptions for system cell 146 in the Quick Reference Guide under the EIGRL entry and the NASTRAN statements are incorrect. The system cell

values are no longer a percentage increase of the memory reserved for the sparse method, rather, it is calculated as (3 * x * buffsize), where x is the value specified for system cell 146. @ 33896; ; Dynamic Analysis - TEMP, Thermal loads - UFM 4017 If multiple LSEQ entries exists in a model and the LSEQs reference different TIDs, then the job will fail with the following unclear messages: *** USER FATAL MESSAGE 4017 (GETEMP) THERE IS NO TEMPERATURE DATA FOR ELEMENT x IN SET yyy 0FATAL ERROR Avoidance: Reference the same TID on all the LSEQ entries. @ 33911; 70.6; Restarts - GKAM - Inefficient Data Recovery Restart for Modal So lutions In a data recovery restart run in SOLs 110, 111, 112, 145, and 146, the GKAM module is unnecessarily reexecuted. Avoidance: Insert the following DMAP alter in the restart run: compile semfreq alter 'call gma','' CALL GMA CASERS ,MATPOOLs,EQDYN , ,MAA USETD ,KAA ,BAA LAMAs ,DIT ,VAFS AGG ,kddd / , ,CMDD MHH ,BHH ,KHH K2MG ,M2MG ,B2MG B2SF / LUSETD /SOLTYP /APP FALSE /FMODE /FALSE NOMSET /fonly $ endalter @ 33913;

,TFPOOL ,GOD ,K4DD ,PHFA ,CK2DD ,PHDH ,K2SF /NOUE /FALSE

, ,GMD ,PHsA ,lamaf ,CB2DD ,VPHFS ,M2SF /NOA /FS

, , , , , , , / /

2001; Documentation - PARAM,CHECKOUT, PARAM,GPL - Missing Output The parameter GPL described under PARAM,CHECKOUT in the Quick Reference Guide should be PARAM,PRTGPL.

@ 33936; sis.

70.5; Aeroelastic Analysis - Mislabeled Force Output from a Trim Analy If there are multiple interference groups (IGID on the CAEROx entry) and the element ID's are not in ascending order with respect to the IGID's, the trim force output can have incorrectly labeled element id's since it uses monotonically ascending values starting from the lowest element ID. Avoidance: Remove the multiple IGID's or make sure that element id's ascend in order with the IGID's.

33938; s.

2001; RESTARTS - Inefficient Restart between binary compatible machine If a database is created on one machine and read into NASTRAN on a different binary compatible machine, the RESTART logic may detect that there are minute differences in the model and everything may be re-calculated. Avoidance: compile phase0 alter 'RMGR,RMG,/','RMGR,RMG,/' The above alter should be used only if your are certain there are no changes to the model.

or superelement manual restart logic could be used. See SERST parameter. @ 33942; ; Documentation/Differential Stiffness - EIGR, EIGRL - Poor Rigid Body Modes If the pressurized modes of a free-free structure are desired, the rigid body modes may not be very clean and may only be separated from the flexible modes by 3 orders of magnitude. Negative values may also be observed. This condition may occur in any model with follower stiffness or rigid elements. Avoidance: No avoidance is necessary. Even though eigenvalue analysis without follower forces typically produces 6 orders of magnitude separation between the flexible and rigid body modes the answers should be acceptable. @ 33943; ; Miscellaneous - PARAM - No Warning Message If a PARAM is entered in the input data but is not used (for example when the param does not exist), no warning message is issued. Depending on the intent of the user this may lead wrong answers. @ 33952; 07 70.7; Component Mode Synthesis - PARAM,METHCMRS Does not work in SOL 1 If residual structure cms is requested and METHCMRS is used to select the eigenvalue solution method in SOL 107, the program will perform a static reduction and will ignore the request for cms. Table 1 in section 6 of the Quick Reference Guide incorrectly indicates that this parameter is supported in solution 107. Avoidance: Use the METHOD command in the CASE CONTROL to request cms for the residual structure. @ 33953; ; Component Mode Synthesis - QSET - Wrong Answers If residual structure cms is used, unused Q-set dof are not automatically removed by AUTOSPC. This may result in unexpected or incorrect answers. Avoidance: Either specify the correct number of Q-set dof (this may be difficult, as it requires prior knowledge), or use PARAM,SMALLQ,0., which will remove unused Q-set dof from the solution. @

33954;

2001; Documentation - PARAM,SMALLQ - Missing Documentation PARAM,SMALLQ is not documented in the Quick Reference Guide. The purpose of this parameter is to remove unused Q-set degrees of freedom from the solution set. If you are performing residual structure qms, then it is recommended to set PARAM,SMALLQ,0. in order to remove the unused Q-set dof.

@ 33955; resses

; Elements - BEAM, BAR, TEMP(INIT), TEMP(LOAD) - Incorrect Beam St In Version 70.5 if TEMP(INIT) references TEMPRB or TEMPP1, then TEMP(LOAD) must reference TEMPRB and TEMPP1 for the same elements, conversely if TEMP(LOAD) references TEMPRB or TEMPP1, then TEMP(INIT) must reference TEMPRB or TEMPP1 for the same elements, else wrong answers will results.

@ 33957;

70.7; Composites - CQUAD8, PCOMP - Poor Results The bulk data description of the CQUAD8 in the Quick Reference Guide needs to be changed as follows: - Remove Figure 1 and change the title of Figure 2 to Figure 1, - The description of the MCID field needs to be changed as follows: "MCID ...................MCID coordinate system (defined by the CORDij entry or zero for the basic coordinate system) onto the surface of the element (see Remark 3). MCID is ignored ........." - Change Remark 3 as follows: "3. ........the line of constant xi." - Add the following paragraph to Remark 3: "The definition of the material coordinate system by projection is used to calculate an angle THETA. Please note that since xi changes direction throughout the element based on element shape, the material coordinate system varies similarly. Because of this an orthotropic or anisotropic material will cause the CQUAD8's stiffness to be biased by both it's shape and grid ordering. Use the QUAD4 element if a constant material coordinate system direction is desired." Editors comment: Note that the "xi"s above need to be replaced with the appropriate greek letter.

33960; Technique

; Documentation/Aeroelastic UG - SOL 110 with Lagrange Multiplier

Section 8.11 in the Aeroelastic User's Guide demonstrates the use of the Lagrange Multiplier Technique with TPL problem ha110a.dat. Currently, solution 110 problems can not be solved reliably with the LMT approach and this example should be removed. @ 33973; 70.6; DMAP - MATMOD Option 25 - Wrong Answers The form of the column vector outputs in MATMOD option 25 is 7 (labeled "ROW VECTOR") and should be 2 (labeled "ROW RECTANG."). Avoidance: Run MODTRL on the output; for example, MATMOD I1,I2,I3,,,/O1,O2/25 $ MODTRL O1// / /2 $ MODTRL O2// / /2 $ @ 33994; 70.6; Optimization/Design Sensitivity - scr=no, DVBSHAP - UFM 1126 In Version 70.5, if shape optimization is attempted with the command line option 'scr=no', then the following fatal message occurs: *** USER FATAL MESSAGE 1126 (GNFIST) DMAP MODULE MDCASE IS ATTEMPTING TO OUTPUT DATABLOCK 201 NAME = CASESTAT, WHICH ALREADY EXISTS. USER ACTION: DELETE THE DATABLOCK, OR USE FILE CASESTAT=OVRWRT Avoidance: Insert the following DMAP alter: compile desopt alter 1 file casestat=ovrwrt $ endalter @ 33997; 2001; Elements - Large Field PSOLID Entry - Stress Results at GAUSS Po ints, not GRID A missing PSOLID large field format continuation line will incorrectly cause stresses to be returned at the GAUSS points instead of the GRID. This will change the output values to those calculated at the GAUSS points instead of the GRID points. This could cause optimistic results to be inadvertently used for design. Avoidances: 1) Include the blank continuation line, (you may want to consider specifying the defaults so someone doesn't delete the blank line thinking it is a useless line), 2) Use small field format. @ 34005; nswers 70.6; Optimization/Design Sensitivity - Aeroelastic Analysis - Wrong A The Version 70.5 release of MSC/NASTRAN enabled the ability to simultaneously perform static and static aeroelastic analysis and design. It has been found that, if the static aero and static subcases both have load requests, the load calculation can become confused and incorrect or no loads can be applied in the static aeroelastic subcases.

Avoidance: The only known avoidance is to not use LOAD requests in static aero subcases in situations where there are also static subcases. @ 34015; 70.7; Rigid Elements - RSPLINE - UFM 4204 The word "WARNING" in "USER WARNING MESSAGE 4204" which warns the user that there are RSPLINE elements with no dependent degrees of freedom was inadvertently changed to "FATAL" in Version 70.5. The run continues and the results are unaffected. @ 34028; 70.5; Elements - CBAR, EMG, PARAM,COUPMASS - SFM 4276, EC 7777 If coupled mass is used with CBAR elements with pinflags, the run may fail with the following traceback in the EMG module: Traceback: Offset Offset Offset Offset Offset Offset Offset 0x00000010 0x00000130 0x000003ac 0x00000a08 0x000003c8 0x00000108 0x0000097c in in in in in in in procedure procedure procedure procedure procedure procedure procedure zerod_ ebarkd_ emgbrd_ emgpro_ emg_ xsem09_ mainal

--- End of call chain --Avoidance: Do not use coupled mass.

@ 34038;

2001; Optimization/Design Sensitivity - DESVAR, DSAJ - SFM 4276, EC 31 When a solution 200 run has both shape and sizing variables and the shape variables have smaller ids than those for the sizing variables, the job may fail in the DSAJ module with "SYSTEM FATAL MESSAGE 4276, ERROR CODE 31". Avoidance: Renumber the DESVAR ids so that the shape DESVAR have larger ids than the sizing variables.

@ 34042;

70.6; Postprocessing - GPSTRESS, PARAM,POST - Missing xdb Output If your run contains multiple subcases in nonlinear static analysis and you have requested strfield, output will only be obtained for the first subcase and subsequent subcase output will be missing. Avoidance: There is no avoidance other than performing RESTARTs for each subcase.

@ 34043; 2001; Documentation - ITER - Iterative Solver Support of SOL 153 not D ocumented The Quick Reference Guide description of the ITER bulk data entry fails to list solution 153 as a supported solution sequence.

@ 34060;

70.6; Superelements - PARAM,METHCMRS - UFM 3032 If superelements are present and PARAM,METHCMRS is specified for the residual structure, then UFM 3032 will be issued stating that the program is unable to find the selected set (EIGRL entry). Avoidance: Insert the following alter: compile cmpmode alter 'methc\=','' if ( seid=0 and methcmrs>0 ) methc=methcmrs $ alter 'read ','' read cmkxx,cmmxx,,,dynamics,uset,cases,vyxw,sils,uset,,,eqexins/ cmlama,cmphixz,cmmi,cmoeigs,eigvmats,/ 'mode'/s,n,nozset//methf/cmset/methc $ endalter

@ 34084;

70.7; Aeroelastic Analysis - AMP, SPC, METHOD - SFM 4276 EC 1044 QOPEN SYSTEM FATAL MESSAGE 4276. QOPEN ERROR CODE 1044 is issued in SOL 145 if BOTH the METHOD and SPC change (or are not included) for each subcase. Avoidance: Include the same METHOD and SPC callouts in all residual subcases or place them above the subcase level. Also include the following DMAP alter: compile sefluttr alter 'modefsrs','' $ change cases to casem call MODEFSRS MR,USET,DM,CASEm,DYNAMICS,MAA,KAA,GPLS,SILS,EED, EQEXINS,,pg,gm,got,kfs,ys,lll,kll/ ,,,,PHA,LAMA,PHA0,LAMA0/ app/NORSET/NOASET/false/ 0/NOASET/METH/0/true/NOQSET/nsub/0 $ alter 'call setq','' CALL SETQ CASE1// SEID/PEID/S,MTEMP/S,K2GG/S,M2GG/S,B2GG/S,MPC/ S,SPC/S,LOAD/S,DEFORM/S,TEMPLD/S,P2G/S,DYRD/s,METH/ S,MFLUID $ endalter

@ 34089;

70.7; Composites - RESTART, PCOMP, DBLOCATE - Infinite CPU Loop If PCOMPs are changed in a RESTART and there is more than one PCOMP entry in the model, then the run may enter an infinite cpu loop in the DBLOCATE module. Avoidance: You can avoid the error with the following DMAP alter. malter 'ept,eptr','ept,eptr' restart ept,,/list/ $ malter 'epts,eptsr','epts,eptsr' restart epts,,/hists/ $ This alter will redo the entire analysis. Only use this alter if there is

a difference in the number of plies between the restart pcomp card and the new pcomp card. @ 34114; ut

70.6; Optimization/Design Sensitivity - PARAM,OPTEXIT,3 - Missing Outp PARAM OPTEXIT 3 is used to exit SOL 200 after evaluating the constraints. However, the results of this evaluation are not printed, making this option of marginal utility. These prints could be of particular benefit if the user has synthesized a response that is not currently available with standard output requests. The user could see the values of the response, but avoid costly sensitivity/optimization calculations. Avoidance: The following dmap alter provides the P2 output (see the DOPTPRM bulk data entry) when PARAM OPTEXIT= 3 compile desopt $ alter 'optexit=3'(,-1) $ if ( descycle=1 ) DOM10 DESTAB,XINIT,,CNTABRG, CVALRG, , DVPTAB,PROPI,,R1TABRG,R1VALRG,, RSP2RG,R2VALRG,,OPTPRMG,OBJTbg,DRSTBLg, tolv,folv,frqrprg,dbmlib,bcon0,bconxi,wmidg// -1/DESMAX/OBJIN//eignfreq $ alter 'dom10','dom10' $

@ 34139;

2001; Dynamic Analysis - LSEQ, STATSUB - Missing UWM 3204 If a dynamic response analysis specifies a static subcase but then in the dynamic subcase the dynamic loads are incorrectly applied such that the LID field of the LSEQ entry references non-existant static load entries, no warning message will be issued and unexpected answers will result. The following user warning message should be issued: *** USER WARNING MESSAGE 3204 (SSG1) UNABLE TO FIND SELECTED SET ( yy) IN TABLE (SLT ) IN SUBROUTINE (SSG1). Avoidance: Other that ensuring that the static load entries exist, there is no avoidance.

@ 34151; 2001; Miscellaneous - Unknown Error Message, Lost some, 2206 of 2209 e liminated When the new reordering methods are used, it is possible that the message "Lost some, 2206 of 2209 eliminated" (the numbers may not be the same) may be written to the f04 and the job will stop. This appears to happen when the EXTREME reordering is used and insufficient memory is available. This has only been observed in two client runs. Avoidance: If the message above is encountered, disable the EXTREME method by setting SYSTEM(206) to 1, 8, or 9 @ 34153; 70.7; Optimization/Design Sensitivity - PCOMP - UFM 6482, Response Typ es Mislabeled

If a composite type response (CSTRESS,CSTRAIN, or CFAILURE) is invoked in a subcase that does not support this type of response (e.g, ANALYSIS=MODES). The designated response type is scrambled; CFAILURE becomes CSTRESS, CSTRAIN becomes CFAILURE, and CFAILURE becomes CSTRAIN. Further, RTYPE CFAILURE is misspelled CFAIURE in USER FATAL MESSAGE 6482. Avoidance: The FATAL message is legitimate, but one needs to infer the correct response type. @ 34200; 70.6; Geometric Nonlinear Analysis - CHEXA, CQUAD4, CTRIA3, TEMP(LOAD) - Wrong Answers In sol 4, ssg1 puts out an incorrect load vector for CHEXA, CQUAD4 and CTRIA3 elements when thermal loads are present along with a global output system. Avoidance: Use solution 106 or 64. @ 34208; 70.7; Optimization/Design Sensitivity - TA1 - SFM 3001 When a design variable overrides an analysis property in a p-element sizing optimization job, it results in the following error message: SYSTEM FATAL MESSAGE 3001 (TA1) THE INPUT DATA BLOCK NAMED EHT AT POSITION 10 DOES NOT EXIST.

Avoidance: Ensure that property values for both the analysis and design model are identical or apply the following dmap alter. compile desopt alter 'luset= ' DBVIEW eht0=eht (WHERE DESITER=0) $ EQUIVX eht0/eht/always $ DBVIEW ehta0=ehta (WHERE DESITER=0) $ EQUIVX ehta0/ehta/always $ DBVIEW gdntab0=gdntab (WHERE DESITER=0) $ EQUIVX gdntab0/gdntab/always $ DBVIEW medge0=medge (WHERE DESITER=0) $ EQUIVX medge0/medge/always $ DBVIEW mface0=mface (WHERE DESITER=0) $ EQUIVX mface0/mface/always $ DBVIEW pval0=pval (WHERE DESITER=0) $ EQUIVX pval0/pval/always $ @ 34211;

70.6; Shape Optimization - DBLOC, DVSHAP, MATGEN - UFM 6962 When performing shape optimization using an external database for basis vectors, if the dblocated basis vectors (UGD) do not have same GSET size as the analysis model, then the following fatal message will occur: USER FATAL MESSAGE 6962 (MATGEN) ILLEGAL VALUE FOR PARAMETER 3 = **** Avoidance: Use an external model displacement vector (UGD) compatible with the current analysis model GSET size.

34236; 70.5; Installation/Authorization - FlexLM Failing on NEC SX4 with SIGA LARM error -5038 There is a problem on the NEC SX4 machine using the FlexLM network licensing manager that may cause large jobs to fail with the message: SIGALARM - Alarm call 5038 error. Avoidance: There is no avoidance, use a node lock authcode instead. @ 34238; 70.7; Elements - PBEAM, PBAR, IFP - SFM 4276, EC 7777 HP systems may have fail with a traceback in IFP when it attempts to read bulk data inputs that contain PBAR and/or PBEAM input with very large property values. Avoidance: There is no avoidance. @ 34239; 2001; Documentation - TLOADi - Missing Documentation A remark under the description of the TLOADi entries in the Quick Reference Guide indicates that if time dependent loads are applied in a frequency response solution, that a Fourier transform is performed to transform the loads to the frequency domain and that the results are converted and reported in the time domain. This remark should refer the user to the Section 5.4 in the Advanced Dynamics User's Guide to obtain more information on the process. @ 34254; e 70.7; Iterative Solver - ITER, BIC - Poor Performance, Poor Convergenc If an initial guess is given to the iterative solver with BIC preconditioning (default) in MSC/NASTRAN V70.5 it will be ignored. avoidance: To be able to use an initial guess with the Version 70.5 set PRECOND=BICWELL or PRECOND=BICILL on the ITER Bulk Data entry and set SYSTEM(241)=2000 on the NASTRAN entry. @ 34259; 2001.0.1; Acoustic Analysis - Superelements - UFM 9101 When a fluid only model with a fluid superelement is run in solution 103, the job may fail with the following messages: ^^^ BULK DATA AND CASE CONTROL PARTITIONING, RESTART CHECKING, AND UNDEFORMED PLOTTING IS ^^^ INITIATED FOR SUPERELEMENT 0 ^^^ USER FATAL MESSAGE 9101 (PHASE0) ^^^ FLUID ELEMENTS ARE DETECTED IN THIS SUPERELEMENT. ^^^ USER ACTION: ENSURE THAT ALL FLUID ELEMENTS ARE ASSIGNED TO SUPERELEMENT yy Avoidances: 1) Do not use a superelements for this analysis, it is not needed. This may require stripping the superelement designation from all of the fluid grids if the model comes from a fluid/structure analysis. 2) Add a dummy structural element similar to the following:

grid, 10000 $ token grid point in se 0 celas2, 10000, 1.0 10000, 1 $ token element param, ascoup, no $ try to avoid error @ 34264;

70.6; Estimate - File Names may not Contain Blanks Blanks in pathnames are not handled properly when "memory=estimate" is specified. Only the part of the directory name after the blank is shown and estimate issues the message "This file does not exist." The job continues, but the default memory or buffsize is used. Avoidance: Specify a pathname that does not include blanks.

@ 34265; eter

2001; Documentation/Superelements - PARAM,MINIGOA - Undocumented Param In MSC/NASTRAN Version 70, an undocumented parameter (MINIGOA) was added to allow reduction in the amount of disk space used when using superelements. When this parameter is used, the transformation matrix (GOA) will contain terms only for the degrees of freedom defined in the U5 (USET, USET1, SEUSET, SEUSET1) set. As the GOA matrix is often one of the largest matrices stored in the database, this can allow for a significant reduction in the amount of disk space used. The limitation of using this approach is that data recovery will be available only for these degrees of freedom and the elements connected to them.

@ 34266;

2001; Documentation - EIGRL - SYSTEM(273) Not Documented in the QRG In Version 70.5, setting SYSTEM cell 273 to 1 allows the selection of the old (pre-v70.5) Lanczos eigensolver. This is mentioned in the Version 70.5 Release Guide, but is not documented in the QRG.

@ 34270;

; P-Elements - Wrong Answers for p-BEAM Element at p=1 BEAM p-elements with p-order of 1 are stiffer in bending than BEAM h-elements due to incorrect residual bending flexibility in BEAM p-elements. Avoidance: The avoidance is to start at a p-order of 3.

@ 34271;

70.6; Optimization/P-Elements - P-BEAM - Wrong Results in OES1 Table Wrong ENDB max and min stresses are written to the OES1 table for the p-Beam element types by Module SDRP. However, the OFP printout of the ENDB max and min stresses is correct since the OES1 table values are not read but computed by the OFP module. This error will affect design optimization results and postprocessors which report these values. Avoidance: There is no known avoidance.

@ 34287; ies

70.6; Aeroelastic Analysis - Incorrect Stability Derivative Sensitivit The UXU and UXR matrices produced by DSARLP may be in error. Consequently, this can produce wrong answers for the sensitivities of

stability derivatives and will produce unpredictable downstream errors. Avoidance: There is no known avoidance. This capability should be used with caution. Zero or unreasonable stability derivative sensitivities may be an indication that this error has been encountered. @ 34290; 316 2001; Documentation/QRG - TABLES1 - Discontinuity is not Allowed, UFM The Quick Reference Guide description for the TABLES! entry should explicity state that discontinuities are not allowed and will lead to a fatal message. @ 34321; ; Installation - FLEXlm on NT needs lmutil.exe program The client installation of FLEXlm on NT needs to also install the lmutil.exe program (and the lmgr325c.dll file) so the user can check the availability and usage of Nastran licenses. The current installation does not allow this. Under the current installation procedure the user would have to get these files from the FLEXlm license server computer. This should be automatically done with the installation. Avoidance: The flexlm directory in the standard installation should have lmutil.exe. Also, the msc705\i386 directory contains lmgr325c.dll. The installer may have to copy the .dll from the i386 directory to the flexlm directory for lmutil to be able to find it. @ 34325; 70.7; P-Elements - GMLOADs on FEEDGEs with p=1 may give bad answers GMLOADs on FEEDGEs with p=1 will give incorrect loads if another edge with p>1 was processed. Avoidance: There is no known avoidance. @ 34362;

70.7; P-Elements - Incorrect p-shell Output with PROJ=F Option If the PROJ=F option is chosen on the OUTRCV entry for p-shells, the moments may be incorrect. In addition, if a coordinate system is specified with the CID=n option, the transformation may be incorrect. Note that the PROJ=F option does not output any data into the f06 file, but it does affect the output in the OP2 and XDB files. The option was created primarily for use by MSC/ARIES.

@ 34393; n 70.5

70.7; Utilities - xdb, DBC - msc2ptc Failed to work for NASTRAN Versio The msc2ptc utility does not run due to a problem with the nast705 script and missing m2p.dsc file. An attempt to convert the xdb file fn.xdb results in the message: fn: executable for fn not found or not executable. Verify that the fn executable has been installed. The full pathname of the bracket executable is: /nastran705/msc705/solaris/fn

Avoidance: The updated files are posted at http://www.macsch.com/support/support.html @ 34415; 70.6; Superelements - EPOINT - Wrong Answers In part superelements with extra points, the sequence of grid points provided in the usetd table, produced from the uset table by adding extra point data, may differ. This will cause plausible but wrong answers. In one example discovered in testing the load vectors were in a different sequence than the corresponding terms in the stiffness matrix. Avoidance: An avoidance which prevents the error is to put in at least one SEQGP entry for a residual structure grid point. It can be input in a manner that will cause no extra resequencing of the model. For example, if the main bulk data file includes grid point 47, the following SEQGP entry will suffice: SEQGP, 47, 47.1

This avoidance may be used with any model, of any size, with negligible effect on cost. The presence of an SEQGP entry forces the code to avoid the branch which contains the error. @ 34424; 2001; File Management Section - DBDICT May Fail to List All Keys NDDL datablocks are uniquely defined by their qualifier values. In addition to the NDDL datablock name, a path, or list of applicable qualifiers, is thus used to "point" to a particular datablock. The DBDICT FMS statement (which can also be called in DMAP) can be used to list qualifier values for datablocks stored on the database, and does so by printing out "keys," or unique lists of qualifiers and their values. A code error may prevent some of the keys from being printed out. Avoidance: If this problem is encountered, the DBDIR FMS statment (also callable in DMAP) can be used to obtain a complete qualifier listing. @ 34429; 70.6; Miscellaneous - Overflow "*****" in the f04. If a job requires more than 999999.9 CPU seconds, then the F04 will print ********. If a job requires more than 999999.9G of I/O then ******** will be printed. Avoidance: There is no avoidance. @ 34459; 70.7; Heat Transfer - Missing RADLST Information in Punch File If large CHBDYG id's exist, then the punch file which contains the RADLST entries will be incomplete and contain a line such as: RADLST 12 1****************

Avoidance: 1) Do not use element ID greater than 10000001 2) Use restart for the subsequent run rather than including the

RADLST punch file because all the radiation matrix information is in the NASTRAN database. @ 34462; 70.6; P-Elements - HEXA - Poor Answers, Element too stiff when modelin g thin plate The HEXA element with p=1 on most edges is too stiff. This was discovered when analyzing the thin flat plate test problem of the MacNeal Harder test set. For most problems the effect is small, but for the thin flat plate model the effect is dramatic. Avoidance: Use a higher p-order. @ 34467; 70.7; P-Elements - CHEXA, CPENTA, CTETRA - Wrong Answers for Transform ed Stresses If stresses and strains are being transformed using the CID= keyword in the OUTRCV bulk data entry, then wrong answers may result. Avoidance: The only avoidance is to request the stresses in the basic (default) coordinate system. @ 34474; 70.6; Security - FLEXlm License may be Refused by the nastran Command The nastran command will refuse to accept an otherwise valid FLEXlm license spec for redundant servers written as: port@host1,port@host2,port@host3 Avoidances: 1) Do not use redundant servers. They are high maintenence with little gain. Multiple servers, specified via "port@host1:port@host:..." on UNIX and "port@host1;port@host;..." are more powerful and less work. 2) Only specify one of the hosts. Each of the hosts in the quorum knows about the other two, so the comma specification is redundant. You could even specify all three hosts as if they weren't redundant, using the ":" or ";" separators as appropriate. @ 34480; 70.6; Eigenvalue Analysis - PARAM,RESVEC,YES - Missing Eigenvalues By simply specifying PARAM,RESVEC,YES, incorporated residual flexibility in the solution sequences making the the resflex sssalter unnecessary. There is an error in the capability which may cause missing modes. The error in the results is obvious because there will be fewer rather than more eigenvalues when you compare the eigenvalues "BEFORE AUGMENTATION OF THE RESIDUAL VECTORS" to those "AFTER AUGMENTATION OF THE RESIDUAL VECTORS". Avoidance : Insert the following DMAP alter: compile resvec alter 'if ( noxset'(3),'' if ( noxset>-1 or append ) then $ endalter @

34482;

70.6; DMAP - Database, DBSTORE - DBSTORE Does not Function on RESTART The automatic restart logic does not schedule modules to run on restart that have no output data blocks. The dbstore subdmap, and the dbstor subdmap which it calls, have this attribute. Avoidance: Use putsys(1,109) before the call statement, or use an alter in the v70.5 sssalter library that includes dummy output data blocks for both modules.

@ 34491;

70.7; Delivery Database - eie layer unable to extend file On Cray systems, if a user has the Delivery Data Base (SSS*) files on a device that does not have a sector size that is a multiple of 4096 words, and if the user specifies ff_io_defaults=eie.sds.diag, then the following message is printed: eie layer unable to extend file "/apps/mscnast/v7052/msc705/unicosts/SSS.MASTERA" to a multiple of 65536 bytes /apps/mscnast/v7052/msc705/unicosts/SSS.MASTERA : open failed : errno =5048 : Unknown error ffopen(/apps/mscnast/v7052/msc705/unicosts/SSS.MASTERA) using alternate layers system opening layer system /apps/mscnast/v7052/msc705/unicosts/SSS.MASTERA : using layers : syscall The job will then continue to run if the the BUFFSIZE is 4097. However, if BUFFSIZE is set to a different value, then the following error is printed: ** ERROR ** UNABLE TO VALIDATE DATABASE FILE FORMAT. DATABASE MAY NOT HAVE BEEN BUILT ON A COMPATIBLE PLATFORM. BIOPNR ERROR ** TIMESTAMP/RECL VALIDATION FAILED. FILX = 6, LOGNAME = MASTERA , NSBUF3 = 32768 FILE = /apps/mscnast/v7052/msc705/unicosts/SSS.MASTERA RECL SHOULD BE 32768, ACTUAL VALUE IS 4096.

Avoidance: 1. Change BUFFSIZE to 4097 2. Turn off FF_IO @ 34517; 70.7; P-Elements - STRAIN, SHELLS & SOLIDS -- Incorrect Principal and von mises strain If a problem has either shell or solid p-elements, then the principal, von Mises, and mean pressure strains may be incorrect. Avoidance: There is no avoidance. @ 34520;

70.7; Acoustic Analysis - METHOD(STRUCTURE) - Wrong Answers In fluid/structure analysis, in modal frequency response (SOL 111), if the fluid modes are calculated using the modal method, i.e. the case control

command METHOD(FLUID)=n, but no METHOD(STRUCTURE) is specified, incorrect results occur. Avoidances: 1. 2. 3. 4. Use use put use SOL 108, or, both a METHOD(FLUID) and a METHOD(STRUCTURE), or, the fluid in a superelement and process it separately, or the following alter for Version 70.5:

compile gma alter 295 partn k1dd,vafse,/k1shh,,, $ partn kddd,vafse,/kdhh,,, $ igh=ig $ call xadd5 k1shh,kdhh,,,/gk1shh/igh/cu/cz/cz/cz $ endalter $

@ 34523;

2001; Optimization/Design Sensitivity - DEQATN - SFM 4276 Optimization runs with large DEQATN entries that are near or exceed the limit of 32000 non-blank charachters may fail on Cray computers with the following message: lib-1190 /u/app/nastran/msc/msc705/unicosc90/NAST705: UNRECOVERABLE library error A numeric input field contains an invalid character. Encountered during a sequential formatted READ from an internal file (character variable) Error initiated at line 656 in routine '_rdfmt'. SIGNAL - Signal SIGABRT(6) caught, job aborting. See signal(2), CRI manual SR-2012 for further information. Beginning of Traceback: Called from line 42 (address 15624627c) in routine 'TRBK'. Called from line 39 (address 132322c) in routine 'RECOVR'. Called from line 112 (address 15073635b) in routine 'sysrcv'. Called from address 15335217c in routine '__handlr'. Called from address 15327333d in routine '_lwp_killm'. Called from line 16 (address 15331121a) in routine 'raise'. Called from line 92 (address 15174456c) in routine 'abort'. Called from line 59 (address 15674675d) in routine '_ferr'. Called from line 656 (address 15732746d) in routine '_rdfmt'. Called from line 346 (address 15665655d) in routine '_xfer_iolist'. Called from line 350 (address 15737620b) in routine '_FRF'. Called from line 56 (address 6106040c) in routine 'PRERR2'. Called from line 132 (address 4610355c) in routine 'IPRSCM'. Called from line 429 (address 4505303c) in routine 'IFSEQU'. Called from line 1816 (address 4454631d) in routine 'IFS6P'. Called from line 1184 (address 4336537b) in routine 'IFPDRV'. Called from line 198 (address 4237106a) in routine 'IFP'. Called from line 293 (address 10526562c) in routine 'XSEM02'. Called from line 311 (address 2064b) in routine 'MAINAL'. Called from line 306 (address 4617d) in routine '$START$'. End of Traceback.

The IBM AIX RS6000 completes but outputs numerous occurances of the following log file message: 1525-097 A READ statement using decimal base input found the invalid digit '' in the input file. The program will recover by assuming a zero in its place. Avoidance: Other than shortening the DEQATN entry there is no avoidance. @ 34535; ; Optimization/Design Sensitivity - PCOMP, MAT8, SDRCOMP - UWM 202 0, UFM 5503 In solution 200, if ply failure indice responses are requested for a composite element whose PCOMP entry references a material entry that does not define the composite failure allowables for one or more of the plies, then the following fatal and/or warning messages are issued and the failure index responses are not used for that element in the optimization analysis. ** USER FATAL MESSAGE 5503 (DSAD13) ELEM NO. 42649 PLY NO. NOT BE FOUND. 1 REFERENCED ON A DRESP1 ENTRY COULD

*** USER WARNING MESSAGE 2020 (SDRCOMP) ONE OR MORE OF THE ALLOWABLE STRESS CONSTANTS WERE INCORRECT FOR ELEMENT 1 PLY 4 ELEM NO. 1 PLY NO. 1 REFERENCED ON A DRESP1 ENTRY COULD NOT BE FOUND. USER ACTION: CHECK THE ELEMENT IDS REQUESTED ON THE DRESP1 ENTRY. IF A CFAILURE RESPONSE TYPE IS REQUESTED, CHECK THAT THE CORRESPONDING PCOMP AND MAT1 BULK DATA ENTRIES HAVE BEEN CORRECTLY SPECIFIED. USER INFORMATION : USE OF UNREFERENCED ELEMENTS IS NOT RECOMMENDED AND MAY INDICATE A USER INPUT ERROR. RESULTS FOR OTHER RESPONSES ARE NOT AFFECTED. Avoidance: Specify the failure indices. If isotropic ply properties are needed desired, use an equivalent MAT8 entry. @ 34548; 70.6; Acoustic Analysis - EPOINT, CAABSF - SFM 6719 (MRGIDX) In frequency response, if EPOINTs exist in a fluid-only analysis with CAABSF elements with no SPCs, the following fatal message will be encountered: *** SYSTEM FATAL MESSAGE 6719 (MRGIDX) MATRICES HAVE DIFFERING NUMBERS OF ROWS: [A] = 48 x 48 [B] = 44 x 48 [D] = 48 x 48 0FATAL ERROR

44

[C] =

48 x

Avoidance: Insert the following DMAP alter or add a dummy grid point and constrain all of its degrees-of-freedom. compile fdrmgen alter 'if ( noa',''

type parm,,i,n,noue $ if ( noa=-1 and noue=-1 ) then $ compile freqrs alter 'if ( noa'(3),'' IF ( noa>-1 or noue>-1 ) then $ endalter @ 34552; 70.6; Acoustics - EPOINTs, GMA, GKAM - Fails with no message If EPOINTs are used in a fluid-only model in solution 111, the run will fail without message after the eigenvalue extraction and before the frequency response output. Avoidance: Use the following alter: compile gma alter 'gkam' if ( fonly and noue>-1 ) then $ matgen ,/iee/1/noue $ paraml phfa//'trai'/1/s,n,nohset $ paraml phfa//'trai'/2/s,n,noaset $ type parm,,i,n,nhe,nae $ nhe=nohset+noue $ nae=noaset+noue $ matgen ,/vhe/6/nhe/nohset/noue $ matgen ,/vae/6/nae/noaset/noue $ merge phfa,,,iee,vhe,vae/phdfh/1 $ merge MFHH,,,,vhe,/mfhh1 $ equivx MFHH1/mfhh/-1 $ merge K1FHH,,,,vhe,/k1fhh1 $ equivx K1FHH1/k1fhh/-1 $ endif $ fonly and noue>-1 @ 34558; 70.7; P-Elements - BEAM, OUTRCV - Wrong Answers at View Points The results output at view points for beam p-elements is wrong if a curved coordinate system is specified as the output coordinate system on the OUTRCV entry. @ 34581; ; Heat transfer - VIEW3D - FATAL, DIMENSION LENGTH x IS SMALLER T HAN "VFZTOL" It is possible running a view factor calculation using the VIEW3D module (this is the default view factor method selected by PATRAN) to encounter the message: *** FATAL *** DIMENSIONAL LENGTH: 1.732432600E-11 IS SMALLER THAN "VFZTOL": 1.000000013E-10 REASSIGN "PARAM,VFZTOL,X.XE-XX" TO BE SMALLER THAN THE DIMENSIONAL LENGTH. Avoidance: Comment out the VIEW3D input so that the older view factor module is used. In the test problem that encountered the poor message, the old view factor module issued the messages below which identified the problem: *** USER FATAL MESSAGE 4528 (VNASIN) ERROR IN CHBDY CARD xx AN AREA4 ELEMENT MUST BE A QUADRILATERAL

CONTAINING NO ANGLE GREATER THAN OR EQUAL TO 180 DEGREES. THE GRID POINTS MUST BE SPECIFIED SEQUENTIALLY AROUND THE QUADRILATERAL. *** USER FATAL MESSAGE 4530 (VNASIN) ERROR IN CHBDY CARD yy THE FOUR GRID POINTS MUST ALL BE IN ONE PLANE @ 34585; 2001; Database - SFM 1041 (GALLEX) A limit in the database directory size may cause large jobs to fail with the message SYSTEM FATAL MESSAGE 1041 (GALLEX). Avoidance: This limitation can be overcome by increasing the buffsize and/or setting system(158)=1000 and system(159)=500 to allow for a larger expansion of the database. @ 34590; 002 GL; Eigenvalue analysis - EIGB, SINV, CYCLIC4 - SFE 4276, FREAD EC 3 In SOL 16, cyclic buckling, if the SINV method is used, and no eigenvalues are found for the current harmonic, then the run fails with the following message in CYCLIC4. *** SYSTEM FATAL ERROR 4276.FREAD Avoidance: Use the Lanczos method. @ 34591; 70.6; Cyclic Buckling - SOLs 116 and 16, CYSUP - SFM 3001 and UFM 1126 The use of CYSUP in SOL 116 causes SFM 3001 in UPARTN. In SOL 16, if eigenvalues are requested at more than one harmonic, then UFM 1126 in CYCLIC3 is issued. Avoidance: Insert the DMAP alter: compile sol16 alter 1 file pvec=save,ovrwrt $ COMPILE MODECYC $ ALTER 'CALL *MODERS'(,-1) $ DELETE /MR1,,,, $ REACT = -1 $ endalter @ 34592; utput ; Aeroelastic Analysis/Random Analysis - EPOINT - Missing Random O Random analysis with an aeroelastic model fails to yield the acceleration random output if any random output requests for EPOINT dofs are made. Avoidance: Make two separate runs, with the epoint data recovery being requested in a separate runs. The runs will be more efficient if restart is used. @ 34602; rs 70.7; Frequency Response - STRESS(BILIN), STRESS(CORNER) - Wrong Answe Complex corner stress output for QUAD and TRIA elements is incorrect in ERROR CODE 3002 PID= 0

MSC/NASTRAN Version 70.5.2. The value for the real part of the stress at the grid point will be overwritten with the imaginary value. The results at the center of the element are not affected. Avoidance: Other than not requesting the corner stress option, there is no avoidance. @ 34604;

70.6; P-elements - Estimate - Over-prediction of Resources MSC/ESTIMATE may predict significantly more resources than required for P-elements. This over-prediction is partially a result of the uncertainty of the P value. Avoidance: Do not use MSC/ESTIMATE for P elements.

@ 34613;

70.6; Cyclic Symmetry - LOADCYN - UFM 4342, Job Aborts If a cyclic symmetry job encounters the fatal message: USER FATAL MESSAGE 4342 (CY1BLD) THE LOAD BULK DATA CARD GIVEN BELOW IS INCONSISTENT WITH SPECIFIED ANALYSIS HARMONICS OR SEGMENT OR SYMMETRY TYPE. then the job will not produce an "END OF JOB" and will issue the traceback below: Traceback: Offset Offset Offset Offset Offset Offset Offset Offset 0xffffffff 0x0000014c 0x00001094 0x00000884 0x0000063c 0x000008c8 0x00000114 0x000009a0 in in in in in in in in procedure procedure procedure procedure procedure procedure procedure procedure WriteUnit _xlfWriteFmt cy1bld_ cy1ccl_ cycl1_ xsem12_ mainal

--- End of call chain --Avoidance: Although the job should not issue a traceback, the fatal message output is correct and the user should correct the LOADCYN input. @ 34652; 2001; Documentation - RFORCE, CTRIAX6 - METHOD 1 and 2 Produce the Sa me Loads Remark 9 in the MSC/NASTRAN V70.5 Quick Reference Guides for the RFORCE bulk data entry states: 9. METHOD=2 should be used for the CTRIAX6. METHOD=1 produces null loads. This is incorrect. Method 1 produces the same loads as method 2. @ 34665; 70.6; Estimate - Nonlinear, HEXA, PENTA, TETRA - Under-predicts Memory In a NonLinear analysis (SOL 66, 99, 106, 129, 153,159) the ESTIMATE

utility erroneously assumes the rotational degrees of freedom are fixed for grids attached only to 3-dimensional elements. The result is that the memory and disk requirements may be under-predicted. @ 34666; 70.7; Optimization/Design Sensitivity or Superelement- GENEL - Possib le Wrong Answers If the number of GENEL's is an odd number greater than 1, the stiffness terms produced by the GENEL's will be in error on the second and subsequent design cycles in SOL 200. Similarly, in other solution sequences, having GENELs in more than one superelement, or in a superelement and residual structure can result in some of the GENELS not generating their respective stiffness matrices which will lead to wrong answers. Avoidance: Insert the following dmap alter. compile semg1 $ alter 1 $ type parm,,i,n,nsize $ alter 'sma3.*nogenl','' $ if ( nokggx<0 ) then paraml kjjz//'trailer'/2/s,n,nsize $ else $ paraml kjjx//'trailer'/2/s,n,nsize $ endif $ matgen ,/knull/7/nsize/nsize $ sma3 gei,knull/kjjy/lusets/nogenl/-1 $ $ @ 34671; 70.7; Beam Library - PBARL, PBEAML - Incorrect Sensitivity for Custom Cross Sections If custom beam cross section are added to the beam library and they are applied in the model without providing analytical sensitivities of the beam properties with respect to the beam dimensions, the sensitivities calculated using finite difference techniques will be too large in magnitude by 50%. Avoidance: There is no easy avoidance except to manually scale the gradients. The sensitivities are of the correct sign and are therefore of some value when optimizing the cross section. @ 34699; ; Heat Transer Analysis - Coupled Convection - SFM 6934 The fatal message below will be encountered if a connection grid is also specified as an ambient point. *** SYSTEM FATAL MESSAGE 6934 (EMGPOM) FOR ELEMENT ID = 100020 AND FILE TYPE = 1 EMGPRO DETERMINES A DIFFERENT COUNT OF ACTIVE CONNECTION SILS THAN DOES CALLER OF EMGPOM EMGPRO COUNT = 5 CALLING ROUTINE COUNT = 6 Avoidance: If the PATRAN interface is being used, make sure there are no overlapping points between application region one and application region two while applying coupled convection LBCs. @ 34706; 70.6; Restarts - RSSCON - UFM 9086

Running a nonlinear analysis restart with RSSCON entries results in USER FATAL MESSAGE 9086. In other solution sequences the restart may be inefficient. The restart fails because of a logic error in subdmap PHASE0. Avoidance: compile phase0 alter 'if ( noxsort )','' equivx geom1/ngeom1/-1 $ endalter @ 34714; 70.6; Optimization/Design Sensitivity - Superelements - UFM 7052, Wron g Answers RTYPE=FREQ was introduced in Version 70 as a "friendly" alternative to RTYPE=EIGN. When RTYPE=FREQ is used with superelements, it can produce an incorrect user fatal message: *** USER FATAL MESSAGE 7052 (DOMPTC) ILLEGAL PROPERTY NAME This error was observed when there was no DESOBJ request. When a DESOBJ request was included, the job ran, but the sensitivity results were incorrect. Avoidance: Use RTYPE=EIGN and not RTYPE=FREQ when there are superelements. @ 34740;

70.7; Postprocessing - Corrupted XDB Access Objects with internal keys exceeding 214747 in solution sequences other than linear statics will corrupt the object maps in the xdb. The internal key id is a 1 to n list of any particular type of output such as grids, bar elements, quad elements etc. In the Version 70.5 system, logic has been added to prevent integer overflow before the access key formation. A FATAL error indicating that translation has stopped is generated if an overflow condition is encountered. Avoidance: Each superelement restarts the 1 to n list for each type of output, thus splitting the model into superelements is one avoidance.

@ 34756; 2001; Documentation/Aeroelastic User's Guide - Inconsistent Descriptio n of REFC Field The Aeroelastic User's Guide should be consistent in referring to the REFC field on the AERO and AEROS entries as being the reference chord. The following places in the Guide incorrectly refer to REFC as the semichord: Section 8.2, page 350, should refer to chord rather than semichord. Section 8.13 (example HA145KR ) incorrectly inputs the reference semichord on the AERO bulk data entry. The REFC should be 40.0 and the same answers could be produced if the reduced frequency values were doubled. Section 9.1, page 578 should refer to chord rather than semichord.

@ 34766; n

70.7; Beam Library - PBARL, NSM, TYPE=HEXA = Incorrect Mass Calculatio For the beam library, PBAR section properties are computed based on the user's PBARL input. Due to a code error, this computation is incorrect for the NSM (non-structural mass) property calculation for section TYPE=HEXA. Therefore, if the mass properties of the BAR are used in the analysis, the answers will be incorrect. Avoidance: The PBEAML/PBEAM computation is correct for TYPE=HEXA, so PBEAML's could be used in place of PBARL's for this section type if the bar is not being designed. The error is benign in a static analysis without inertial relief or gravity loads.

@ 34767;

70.6; SSSALTERS - checka - Incorrect Degree of Freedom Identifiers The output for the checka alter was modified for V69.1. Due to an error in the alter, the degree of freedom identifiers for the matrices REACCOLN and REACCOLA will be incorrect. Avoidance: Change the following line in the alter: MATGPR GPLS,USETB,SILS,REACcoln//'H'/'G'//printol $ to: MATGPR GPLS,USETB,SILS,REACcoln//'H'/'N'//printol $ and change the following line: MATGPR GPLS,USETB,SILS,REACcola//'H'/'G'//printol $ to: MATGPR GPLS,USETB,SILS,REACcola//'H'/'A'//printol $

@ 34774;

70.7; Fluid Elements - GRIDB - UFM 315 In Version 69 and later MSC/NASTRAN releases, when fields 3, 4, and 6 of the GRIDB entry are used the job will fatal with the following error messages: *** USER FATAL MESSAGE 315 (IFPDRV) FORMAT ERROR ON BULK DATA ENTRY GRIDB x

Although the fields are not used, Remark 6 of the Quick Reference Guide's description of the GRIDB entry clearly states: "Fields 3, 4, and 6 are ignored, which facilitates the conversion of GRID entries to GRIDB entries. Note that the fields are the same except for fields 1 and 9 when a cylindrical coordinate system is used." Thus, allowing information in these fields is an important design feature. When blanks are used on fields 3, 4, and 6 of the GRIDB in an attempt to circumvent the above error, if structure with GRIDs is present, then arbitrary errors occur specifying that GRIDs are not present, when they are. Such as:

*** USER FATAL MESSAGE 2007 (GP2) ELEMENT 402 REFERENCES UNDEFINED GRID POINT @ 34781;

1381

70.7; Eigenvalue Analysis - READ - Poor Message, SFM 7340 (LNNHERR) If a Lanczos eigenvalue analysis requests more modes than exist, it is possible that the job will fail with the message: *** SYSTEM FATAL MESSAGE 7340 (LNNHERR) process error reported by subroutine LNNP2CD see LNNP2CD abstract (ier = -729) *** SYSTEM FATAL MESSAGE 7340 (LNNHERR) warning reported by subroutine LNNDRVD see LNNDRVD abstract (ier = 729) This error has been observed only on small problem where some dofs did not have independent mass terms. Avoidance: An avoidance is to set f2 to limit the search region within plausible limits. For small models where all the eigenvalues are desired a Givens method may also be selected.

@ 34783;

; Case Control - MPCFORCE - Incorrect Subcase Labels If mpcforce output is requested in a multiple subcase run, and not all subcases have an mpc callout, then the subcase labels for the mpcforce output will be incorrect. The results, however, are correct. The labels will be in ascending order of the subcase ids. Using the following case control setup, Avoidance: Ignore the subcase labels, the results are correct. Else place the subcases with MPC requests first.

@ 34784;

2001; Documentation - SYSTEM(198), QRG - Incorrect Default Value The MSC/NASTRAN Quick Reference Guide states that the default for SYSTEM(198) (a.k.a MINFRONT) is 8. The default is actually machine dependent is documented in the Configuration and Operations Guide.

@ 34795; 44

; Nonlinear Static Analysis - PARAM,BUCKLE, Thermal Loads - UFM 20 In a PARAM,BUCKLE,1 restart with thermal loads, if the PARAM,LOOPID specifies a loop ID which occurred in a subcase earlier than the previous subcase in the cold start run, then the restart run will fail with UFM 2044 in the NLCOMB module. Avoidance: In the restart run remove all subcases after the subcase in which PARAM,LOOPID occurred. For example, if the cold start contained 4 subcases and PARAM,LOOPID value occurred in the 2nd subcase, then remove the third and fourth subcases in the restart run.

@ 34824; Output

; Nonlinear Analysis - CQUAD4,CTRIA3 TABLEM1 - Wrong Intermediate

All output for the QUAD4 and TRIA3 elements at intermediate steps (NLPARM INTOUT > 1) is wrong when there exists a temperature dependent material property. Avoidance: Other than ignoring the intermediate output there is no avoidance. @ 34833; e 70.5; Frequency Response - Corner Force - Missing or Wrong Corner Forc If corner force output ( stress(corner...)=x ) is requested in a frequency response analysis, the corner force for the QUAD4 will either be missing or incorrect. This error has been corrected for Version 70.5. Avoidance: Use the center stress. @ 34834; 70.5; Eigenvalue Analysis - PARAM,RESVEC - Wrong Answers If a C-set is defined for a superelement, the component modes are passed thru a filtering process to remove any which might be a linear combination of the static transformation. In V70, if residual vectors are not requested, the component modes will not be passed thru the filter. This may lead to either rank-deficient matrices or wrong answers. If the filtering is not performed, the reduced stiffness matrix will incorrectly contain the original eigenvalues on the diagonal and there will be no off-diagonal terms (as there should be if the filtering is not performed). Avoidance: If residual vectors are requested with at lease one loading or U6-set dof defined, then the filtering will be performed. Otherwise, use Version 70.5 or later version. @ 34850; than f06 ; Postprocessing - DBC, PARAM,POST - OPNPFL Message to log Rather

DBC OPNPFL messages may be written to the log file rather than the f06.

@ 34874;

70.7; Eigenvalue Analysis - PARAM,RESVEC, Superelements - Poor Answers If residual vectors are requested in dynamic analysis (PARAM,RESVEC,YES) and none are appended, the original modes will be lost without warning. The user should also see related error 34480. Avoidance: Either do not include residual vectors, make sure that residual vectors are calculated, or include the following DMAP alter: compile resvec alter 'WERE AVAILABLE AFTER FILTERING' equivx phx/aphx/-1 $ equivx lama/alama/-1 $ endalter

@ 34902;

70.7; Optimization/Design Sensitivity - CBUSH, CVISC, CDAMP - UFM 1126

Optimization (SOL 200) produces User Fatal Messages 1126 when damping element properties ( CDAMP, CVISC, CBUSH) are active design variables and superelements are used in the model. Avoidance: Apply the following DMAP alters: compile desopt alter 'type DB' type db belmdsf, bdictdsf, belmdsb, bdictdsb $ alter 'dopr5' type db zuzr01 equivx tabdeq/zuzr01/always $ compile fea list alter 'type parm' type db zuzr01 alter 'dspt1'(2),'dspt1'(2) PARAML zuzr01//'TRAILER'/1/S,N,NDVAR $ no. of des. var. alter 'dspt1'(3),'dspt1'(3) PARAML zuzr01//'TRAILER'/1/S,N,NDVAR $ no. of des. var. matprn zuzr01 // $ alter 'trd2'(2),'trd2'(2) TRD2 CASERS,TRL,NLFT,DIT,KHH,BHH,MHH,PHT,zuzr01,,,/ UHTDS,,TOLSCR/ SOLTYP/NOUE/NONCUP/0 $ COMPILE PSLGDV alter 'emg,'(3),'emg,'(3) IF ( DOFREQ>0 OR DOMTRAN>0 ) EMG, EST,CSTMS,MPTS,DIT,,,ett,,,,BGPDTS,GPSNTS,,,,ditid/ ,,,,BELM,BDICT/0/0/ $ compile super3 alter 'lcgen '(2),'' dbstatus cases//s,n,nocases $ if ( delta and nocases<1 and seid=0 ) then $ LCGEN CASErS,SLT,ETT/CASESX $ else $ LCGEN CASES,SLT,ETT/CASESX $ endif $ @ 34925; 2001; Documentation - QRG - DISPLACEMENT Remark 2b for the DISPLACEMENT Case Control Command should be modified from SORT2 is the default in transient response analysis (structural and heat transfer). SORT2 is not available for real eigenvalue (including buckling) complex eigenvalue, or flutter analysis. If SORT1 is selected for one or more of the commands ACCE, DISP, ENTH, FORC, HDOT, MPCF, OLOA, SPCF, STRA, STRE, and VELO than the remaining commands will also be output in SORT1 format. to SORT2 is the default in transient response analysis (structural and heat transfer). SORT1 is the default in frequency response analysis. SORT2 is not available for real eigenvalue (including buckling), complex eigenvalue, or flutter analysis. If SORT1 is selected in a transient solution for one or more of the commands ACCE, DISP, ENTH, FORC, HDOT, MPCF, OLOA, SPCF, STRA, STRE, and VELO then the remaining commands will

also be output in SORT1 format. If SORT2 is selected in a frequency response solution for one or more of the commands ACCE, DISP, FORC, MPCF, OLOA, SPCF, STRA, STRE, and VELO then the remaining commands will also be output in SORT2 format. @ 34926; 2001; Superelements - SEALL - SEP2DR, Logic error 110 If the SEALL command refers to an undefined superelement, directly or via the SET command, then the run terminates with "SEP2DR LOGIC ERROR 110". Avoidance: Correct the undefined superelement references. @ 34940; ; Nonlinear Heat Transfer - RADBC, NLITER - SFM 4276, Traceback In the radiation dominate problem, the stiffness term is approximately equal to 4*T**3. Therefore, if the guess at the initial temperature value is too low for the problem, the temperature may rapidly exceed an infinite value and the traceback that follows will occur: Trap encountered: SIGTRAP - Trace trap FP overflow Traceback: Offset Offset Offset Offset Offset Offset 0x00000014 0x00000564 0x00001124 0x00000780 0x00000088 0x00000990 in in in in in in procedure procedure procedure procedure procedure procedure zerod_ nrdbcd_ nlemg_ nliter_ xsem15_ mainal

--- End of call chain --Avoidance: Make a better estimate of the initial temperature on the TEMPD bulk data entry. In general it is safer to have a very high initial temperature than too low. @ 34949; ; Plotting - XYPLOT - Bad Scaling of XYPLOT data The XYPLOT capability fails to scale the xy-plots properly when small values of displacement exist in the entire frequency range of interest. Resolution/Avoidance: e model so the response values are not so small. @ 34956; rix ; Aeroelastic Analysis - Panels in XY Plane Produce a Singular Mat The Quick Reference Guide AERO and AEROS descriptions caution against placing a lifting surface in the plane of symmetry. The problems associated with this modelling practice in the XZ plane of symmetry have been removed in Version 70.6. However, the XY symmetry was not addressed so that vehicles flying in ground effect (SYMXY=-1) with any panel or body Use Microsoft Excel to plot the data or scale th

in the XY plane still produce a singular matrix. The user is also cautioned not to attempt to perform a SYMXZ=1 (symmetric) analysis with ALL panels and bodies on the XZ plane. @ 34967; 70.7; Eigenvalue Analysis - PARAM,RESVEC - PARAM,RESVINER Ignored If PARAM,RESVINER,YES is specified to calculate residual vectors for the inertial loads due to the unit acceleration of mass and there are no LOADSET-LSEQ loadings or U6 set degree of freedom, then PARAM,RESVINER,YES will be ignored. This may lead to poor answers. Avoidance: Include the following DMAP alter for V70.5 compile semodes alter 'call modef','' CALL MODEFSRS MR,USET,DM,CASEM,DYNAMICS,MAA,MKAA,GPLS,SILS,EED, EQEXINS,,PG,GM,GOT,KFS,YS,LLL,KLL, bgpdts,scstm,cstms,medge/ ,,,,PHA,LAMA,PHA0,LAMA0/ APP/NORSET/NOASET/FALSE/0/NOASET/METH/METHF/TRUE/ NOQSET/NSUB/0 $ compile modefsrs alter 'subdmap','' SUBDMAP MODEFSRS MR,USET,DM,CASES,DYNAMICS,MAA,KAA,GPLS,SILS,EED, EQEXINS,VAFS,PG,GM,GOT,KFS,YS,LLL,KLL, bgpdts,scstm,cstms,medge/ PHA,LAMA,PHFA,LAMAF,PHSA,LAMAS,PHSA0,LAMAS0/ APP/NORSET/NOASET/FS/NOFASET/NOSASET/METH/METHF/ NASOUT/NOQSET/NSUB/MODETRAK $ alter 'call.*resvec','' CALL RESVEC PG,USET,GM,KFS,YS,bgpdts,scstm,cstms,medge,, PHSA0,LLL,KLL,LAMAS0,MSAA,GOT,KSAA,,VAFS,MIS/ PHSA,LAMAS/ 0/NORSET/0/'A '/FALSE $ $ compile resvec list alter 99 matprn px2,ox2,ox3// $ message //'RESVPGF='/RESVPGF $ alter 'if ( noxset'(3),'' if ( noxset>-1 or append ) then $ alter 'matmod.*ratios.*bigrats' matprn ratios//$ $ $ alter for CSR 34874 alter 'WERE AVAILABLE AFTER FILTERING' equivx phx/aphx/-1 $ equivx lama/alama/-1 $ $ alter 'PARTN.*PAB,,VAFS/PX1','' PARTN PAB,,VAFS/PXaa,,,/1 $ alter 'EQUIVX.*PAB/PX1','' EQUIVX PAB/PXaa/-1 $ alter 40,55 endif $ replace 'else' with endif IF ( RESVINER='YES' OR INRLM<0 ) THEN $ VECPLOT ,,BGPDTS,SCSTM,CSTMS,,,MEDGE,/ VGT///4////ALTSHAPE//SEID $ TRNSP VGT/VG $ $

IF ( NOMSET>-1 OR NOSSET>-1 ) THEN $ UPARTN USET,VG/VF,,,/'G'/'F'/'N'/1 $ ELSE $ EQUIVX VG/VF/-1 $ ENDIF $ if(eigset = 'a')then $ PARAML USET//'USET'//////'F'/S,N,NOFSET if(nofset<>noaset)then $ upartn uset,vf/va,,,/'f'/'a'/'o'/1 $ else $ equivx vf/va/-1 $ endif $ mpyad mxx,va,/paa $ append pxaa,paa/px1 $ else $ UPARTN USET,MFF/MOF,,,/'F'/'O'/'A'/1 $ MPYAD MOF,VF,/POI $ APPEND PO,POI/PX1 $ endif $ ELSE $ EQUIVX PO/PX1/-1 $ ENDIF $ RESVINER='YES' OR INRLM<0 $ endalter @ 34968;

; Eigenvalue Analysis - PARAM,RESVINER - Missing Residual Vectors If residual vectors are requested for the residual structure (or in a model which does not contain superelements), and the model is free-free, then if PARAM,RESVINER,YES is used, the associated residual vectors will not be created. Avoidance: Move everything except one grid point into a superelement and request the residual vectors for the superelement. Then the vectors will be included.

@ 34969; EHEXD

70.5; Optimization/Design Sensitivity - EMG - UFM 4659, UIM 4656 from When superelements are used in a shape optimization task elements, the job may fail in EMG during the sensitivity with the user fatal messages: UFM 4659 and UIM 4656 from Other similar error message may also be issued depending elements being used in the model. Avoidance: Do not use superelements. with solids analysis phase routine EHEXD. on the type of

@ 35007; 9058

70.6; Optimization/Design Sensitivity - PARAM,DDRMM,0 - UFM 5629, UFM If PARAM,DDRMM,0 is specified in SOL 200 with ANALYSIS=MFREQ or MTRAN, then the run will fail with UFM 5269 or UFM 9058, respectively. Avoidance: Insert the following DMAP alter: compile fea alter 'delete \/msf1,bsf1,ksf1,,'(,-1) if ( drmm ) drmm=false $

alter 'IF ( DROPT\=2'(2) delete /msf1,bsf1,ksf1,, $ IF ( DOSPCF ) call CFORCE1 MSF,M2SF,BSF,B2SF,KSF,K4SF,K2SF,,/ msf1,bsf1,ksf1/ APP/APP/false $ delete /MMG1,BMG1,KMG1,, $ IF ( DOMPCF ) CALL CFORCE1 MMG,M2MG,BMG,B2MG,KMG,K4MG,K2MG,,/ MMG1,BMG1,KMG1/ APP/APP/true $ alter 'destrn\=true'(,-1),'destrn\=true'(,2) alter 'ELSE if ( drmm )' destrn=true $ drmm=false $ endalter @ 35029; 70.7; Postprocessing - DBC - LAMA Objects have Incorrect Attributes in the XDB When subcases in optimization refer to eigenvalue or buckling analysis, the LAMA objects in the XDB will have incorrect attributes and this may effect the ability of other products such as MSC/InCheck to process the results. In addition, if a frequency range is specified rather than an explicit number of modes, NASTRAN DBC translation may be incorrectly terminated. @ 35040; ; Static Analysis - SORT2, PUNCH - Punch Output has Zero Results If SORT2 is requested for static analysis results, zero results are written to the PUNCH file. Avoidance: The following DMAP alter uses OFP to write the data in f06 format to an alternate fortran unit thus avoiding the punch unit. $ $ Output of QUAD4 Element Stress data $ for post-processing of printed stress in $ a fortran unit separate from the usual $ f06 file. $ COMPILE SEDRCVR,SOUIN=MSCSOU,NOLIST,NOREF ALTER 'OFP.*OES2S,OSTR2,,,,,','' $ type parm,,i,n,linespp $ linespp=getsys(linespp,9) $ paraml kelm//'trailer'/1/s,n,noelem $ No. of nl elem. in model noelem=20*noelem $ calc. lines per page to 20 x no. of elem. putsys(noelem,9) $ set lines per page putsys(13,2) $ write printed output to unit 13 ofp oes2s,ostr2 // s,n,cardno $ putsys(linespp,9) $ reset lines per page to original value putsys(6,2) $ restore unit 6 for printed output Version 68.2 will give correct result if SYSTEM(210)=1 is set on the NASTRAN entry. @ 35050; 70.6; Miscellaneous - Data Transfer between Window and UNIX

When an input file for MSC/NASTRAN is generated in a Microsoft Windows environment and is moved to a window on a unix computer with the drag and drop feature, it causes input errors. Each line of input has a " " symbol at its end, symbolizing an illegal, unprintable character (carriage return) in the input, leading to many illegal input messages. Avoidance: An avoidance is to use a file transfer utility such as ftp instead. It may require some experimentation to find out if the ftp utility you use avoids inserting these illegal characters. They are carriage return symbols. MSC has had success using the ftp option in the EXCEED product. MSC/NASTRAN itself will delete these spurious carriage returns from its input files, starting with Version 70.6. If other files are transferred this way, such as scripts used to run MSC/NASTRAN, the symptoms of this problem are difficult to detect. When you submit one of these damaged scripts it will return immediately with no action and no diagnostics, unless you have included commands in the script that request diagnostic output. If you list the script on the screen with the "more" command, the spurious carriage returns do not appear on the screen. They can be seen with a vi editor, and removed with that editor or by other means. @ 35054; 70.6; Optimization/Design Sensitivity - INREL - UFM 6143 If inertia relief (PARAM,INREL,-1) is requested in SOL 200 and there are more than 12 static load cases then the run will fail with UFM 6143 in the MERGE module in subDMAP INERSEN. Avoidance: Insert the following DMAP alter in V70 or V70.5: compile inersen alter 'nld6\=' $ type parm,,i,n,nz $ nz=nld6-5 $ alter 'vn6r'(2),'' matgen ,/vn6r/4/1/nld6r//6/nz/i661/0/1 $ alter 'vn6r'(4),'' matgen ,/vn6r/4/1/nld6r//6/nz/i661/0/1 $ endalter @ 35066; 2001; Optimization/Design Sensitivity - DRESP2 - UFM 315 When a DRESP2 bulk data entry is missing a continuation line an error message is issued but the message indicates that a subsequent DRESP2 is in error. The message is similar to the following: *** USER FATAL MESSAGE 315 (IFPDRV) FORMAT ERROR ON BULK DATA ENTRY DRESP2 1001 SORTED ENTRY COUNT = 584 USER INFORMATION : THE FORMAT OF DRESP2 ENTRY HAS BEEN MODIFIED IN MSC/NASTRAN VERSION 68. PLEASE REFER TO THE QUICK REFERENCE GUIDE FOR MORE INFORMATION. USER INFORMATION : PLEASE NOTE THAT THE FORMATS OF THE FOLLOWING DESIGN OPTIMIZATION ENTRIES HAVE BEEN MODIFIED IN MSC/NASTRAN VERSION 68 : DCONSTR DOPTPRM DRESP1 DRESP2 DSCREEN DTABLE DVPREL1 DVPREL2 *** USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA ENTRY DRESP2 1001 SORTED ENTRY COUNT = 584

^^^ USER FATAL MESSAGE 9002 (IFPS) ^^^ ERROR(S) ENCOUNTERED IN THE MAIN BULK DATA SECTION ^^^ SEE MESSAGES ABOVE. ERROR ENCOUNTERED IN MODULE IFP, Avoidance: Look for the DRESP2 with the ID preceding the one listed on the error message. @ 35067; 70.7; Frequency Response - PBUSH, PELAST - UFM 5423, 6719 In SOL 111, if there are frequency dependent elements, and if the a-set equals the g-set, the modal reduction of the frequency dependent matrices is not performed. In a structural analysis, this results in UFM 5423, and in an acoustic (uncoupled) analysis, the result is SFM 6719. Avoidance: Add a dummy SPC. @ 35096; 70.7; Iterative Solver - ITER, DEFORM - SFM 6980 (SB85RD) A job that uses the iterative solver and specifies DEFORM type loads failed with the message: *** SYSTEM FATAL MESSAGE 6980 (SB85RD) LARGEST BLOCK ROW OF THE INITIAL MATRIX ( 85606 WORDS) EXCEEDS THE SPECIFIED MAXIMUM RECORD LENGTH ( 65536 WORDS). USER ACTION: PLEASE CONTACT MSC CLIENT SUPPORT. *** SYSTEM FATAL MESSAGE 3008 (UTMCOR) MODULE = SOLVIT INSUFFICIENT CORE AVAILABLE FOR SUBROUTINE USER ACTION: INCREASE OPEN CORE BY AT LEAST 20070 WORDS PROGRAMMER INFORMATION: LOCATION KEY = M FATAL ERROR Avoidance: Increasing memory may solve the problem but it is likely that the sparse solver will need to be selected. @ 35117; 70.7; Acoustic Analysis - FORCE(PHASE) - Wrong Answers When magnitude/phase force output is requested for fluid elements, real/imaginary outputs are printed in the magnitude/phase format instead. Avoidance: There is no avoidance. @ 35121; R3C) 70.7; Optimization/Design Sensitivity - DRESP1, DRESP2 - UFM 7139 (DOP When an unreferenced DRESP1 entry of RTYPE=FREQ appears in a superelement optimization job the fatal message below may occur: *** USER FATAL MESSAGE 7139 (DOPR3C) ON DRESP2 entry ID = 10, RESP1 ID = AN EXISTING DRESP1 ENTRY. Avoidances: 1) Remove the unreferenced DRESP1 entry using RTYPE=FREQ or 2) Use RTYPE=EIGN instead of RTYPE=FREQ @ 35138; 70.7; Documentation/SSSALTER - prldbckla.v706 - no longer needed 1001 DOES NOT REFERENCE

Buckling with preload has been included in the standard 105 solution sequence in v70.6. Therefore, the prldbcka alter is no longer needed. Avoidance: The new capability is invoked through the STATSUB Case Control command. The documentation for STATSUB in Version 70.6 follows: Format: STATSUB[(BUCKLING)]=n (PRELOAD) Examples: STATSUB=23 STAT=4 STATSUB(PRELOAD)=7 Describer Meaning BUCKLING Subcase id number corresponding to static subcase of buckling or varying load (default in buckling analysis) PRELOAD Subcase id number corresponding to static subcase of pre load or constant load (default in dynamic analysis) n Subcase identification number of a prior SUBCASE specified for s tatic analysis (Integer>0) Remarks: 1. Unchanged 2. Unchanged 3. Remove this remark, the limitation on LOADSET has been removed. 4. In dynamic analysis only one STATSUB command may be specified in each dynamic subcase. In buckling analysis with a preload, both STATUS(BUCKLING) and STATSUB(PRELOAD) must be specified in each buckling subcase. @ 35139; 70.7; Inertia Relief - param,inrel,-1 - UFM 9105 If param,inrel,-1 is included in the input file for a solution sequence that does not use inertia relief (e.g., sol 103), then the job will terminate with the following error messages: ^^^ USER FATAL MESSAGE 9105 (PHASE0) ^^^ A SUPORTI BULK DATA ENTRY IS NOT PRESENT ANALYSIS (PARAM,INREL,-1). FOR INERTIA RELIEF

Avoidance: Remove param,inrel,-1 from the input file since it is not used in a non inertia relief run. @ 35141; 70.7; Frequency Response/Complex Eigenvalues - PARAM,MODACC, CVISCi Wrong Answers If PARAM,MODACC,>-1 is specified in SOLs 110, 111, 146, or 200 with ANAL=MFREQ and the model contains viscous damping then wrong answers will result.

Avoidance: Insert the following DMAP alter: compile gma alter 'call xADD5 K2XDD' equivx b2ydd/b2dd/-1 $ endalter @ 35143; 70.6; Optimization/Design Sensitivity - DSADJ - SFM 1015 (GBKMGR) The fatal message: SYSTEM FATAL MESSAGE 1015 (GBKMGR) may occur in the DSADJ module and has been attributed to an incorrect assessment of the amount of open core available. Avoidance: The message has only occurred when running with a small memory request. A possible avoidance is to increase the memory request. @ 35154; 70.7; Superelements - BEGIN SUPER - SFM 3007 In a superelement analysis using PARTS, if more than 12 points occupy the SAME location at the boundary in the same superelement, then the job will fail with the following messages: *** SYSTEM FATAL MESSAGE 3007 (SEP1X) ILLEGAL INPUT TO SUBROUTINE SEPBSD 0FATAL ERROR Avoidance: Do not define more than 12 points occupying the same location at the boundary for a single part superelement. @ 35158; 70.7; DMAP - LCGEN - SFM 3007, Poor Message If the first input, usually CASES, to LCGEN is purged then SFM 3007 is issued. SFM 3001 should be issued instead. Avoidance: There is no avoidance. @ 35187; 2001; Miscellaneous - UFM 6440 (MODGM2) If double-field entries are used to enter property ID's longer than 8 digits, then it is possible to obtain USER FATAL MESSAGE 6440 indicating that a property id is not found even though it exists in the input file. The cause of this error is the sorting routine which uses only the first 8 digits of the id. The problem will only occur if property ID's which are both 8-digit and longer than 8-digits exist in the input. Avoidance: Either keep the property ids less than 9 digits, or use all property ids greater than 8 digits in length. @ 35188; 70.7; Restarts - SFM 471 The database of large restart runs may be corrupted when data spills from one file to another. In the restart run the error may lead to the message: ** ERROR ** UNABLE TO VALIDATE DATABASE FILE FORMAT. DATABASE MAY NOT

HAVE BEEN BUILT ON A COMPATIBLE PLATFORM. BIOPNR ERROR ** TIMESTAMP/RECL VALIDATION FAILED. FILX = 30, LOGNAME = DBALL5 , NSBUF3 = 4096 FILE = /disk/tmp/test.dball TIMESTAMP OFFSET INVALID. *** SYSTEM FATAL MESSAGE 471 (DGETTS) ERROR ATTEMPTING TO OPEN THE FOLLOWING PHYSICAL FILE NAME (PFILEX= 30) : /disk/tmp/test.dball USER ACTION: 1) CHECK THE READ/WRITE PRIVILEGES GRANTED THIS FILE. 2) CHECK MAXIMUM NUMBER OF FILES THAT MAY BE OPENED SIMULTANEOUSLY. 3) CHECK THAT THIS DATABASE FILE IS COMPATIBLE WITH THIS PLATFORM. 4) ENSURE THAT NO OTHER JOB HAS OPENED THE DATABASE FOR UPDATING. 0FATAL ERROR Avoidance: Increase the buffsize. Using fewer but larger database files may also avoid this error, but if the system is limited to 2gb files, this avoidance may not be possible.. @ 35197; sults 2001; Documentation/Optimization - QRG, PBUSH, DVPRELi - Unexpected Re The Quick Reference Guide fails to document that when a property on a PBUSH entry is designed a negative FID number must be specified on the DVPREL1 and DVPREL2 entries. If a positive FID number is specified on a DVPREL1/2 entry, unexpected results may occur. Avoidance: Use a negative word position number from the element property table for the FID field on the DVPREL1/2 entry. The following table provides the word positions for the PBUSH entry in the element property table. EPT Section for PTYPE = "PBUSH" Word 1 2 through 7 8 through 13 14 15 through 18 @ 35200; Type I R R R R Item Property ID K1 through K6 B1 through B6 GE1 SA,ST,EA,ET

2001; Documentation - QRG - FSLIST Missing from QRG Version 69 and later Quick Reference Guides do not have a description for the FSLIST bulk data entry. Avoidance: See the V68 Quick Reference Guide or the Fluid-Structure Seminar Notes for the description. An abbreviated description follows: FSLIST Defines the fluid points (RINGFL entry) that lie on a free surface boundary. Format:

FSLIST, RHO, IDF1,IDF2,IDF3........ Field Contents

RHO Mass density at the surface. Real>0.0; the default is ta ken from DRHO on the AXIF entry). IDFi Identification number of RINGFL entry. (Integer>0 or Character="AXIS" in first and/or last field only) Remarks: 1. This entry is allowed only if an AXIF entry is also present. 2. The order of the points must be sequential with the fluid on the right with respect to the direction of travel. 3. The word "AXIS" defines an intersection with the polar axis of the fluid coordinate system. 4. If the fluid density varies along the boundary, there must be one FSLIST entry for each interval between fluid points. @ 35206; 70.7; External Superelements - PARAM,EXTOUT - Wrong Answers, UFM 5423 If external superelements are used and restarts are performed, the restart logic will insist on re-processing the external superelements and any downstream superelements whether or not any changes are made to the model. If the superelements are created using PARAM,EXTOUT,DMIGDB, this may result in incorrect answers or UFM 5423 indicating an incompatible matrix multiply. Avoidance: When restarting from a run which read in external superelements which were created using PARAM,EXTOUT,DMIGDB insert the following DMAP alter: compile extin alter 'IF.*NOKAA.*0.*THEN','' IF ( NOKAA>0 and nomatk<0) THEN $

@ 35217; rts

70.7; External Superelements - PARAM,EXTOUT,DMIGDB - Inefficient Resta If external superelements are used and restarts are performed, the restart logic will insist on re-processing the external superelements and any downstream superelements whether or not any changes are made to the model. Avoidance: The following alter will disable the deletions for external superelements. If any changes are made to the external superelements, this alter should not be used. compile phase0 alter 'INVOKE ALL RESTART DELETIONS HERE','' if(setype<>'EXTRNA')RESTART ,,,HISTS/TLIST/TRUE $

@ 35234;

70.7; Miscellaneous - DPD - UFM 2079, Poor Message The text of UFM 2079 is incorrect. The text reads: *** USER FATAL MESSAGE 2079 (DPD) GRID OR SCALAR POINT <nn> HAS AN UNDEFINED COORDINATE REFERENCED ON A <xxxx> ENTRY. Both of the the values "nn" and "xxxx" may be wrong and words "UNDEFINED COORDINATE" should be "ILLEGAL COMPONENT".

@ 35246; s

70.7; Matrix Methods - MPYAD - SFM 4276 EC 3002 or 32 on Large Matrice It is possible that MPYAD may fail with SFM 4276 EC 3002 or 32 when very large, dense matrices are being multiplied. The client model that encountered this error had a very long duration transient analysis (more than 100000 time steps). This model failed during mode-acceleration data recovery in module DDR2 (using MPYAD). Avoidance: It is possible that providing more memory may allow the run to finish. For the problem mentioned above, using residual vectors (PARAM,RESVEC,YES) instead of mode-acceleration allows the run to complete. If you are encountering this problem in a run using mode-acceleration, try using residual vectors instead of mode-acceleration.

@ 35259;

70.7; Nonlinear Transient Response - LSEQ - Wrong answers If LSEQ entries are used in SOL 129 and the DAREA and LID fields have the same ID, then the dynamic loads which are based on the LSEQ entries will be doubled and wrong answers will result. Avoidance: Insert the following DMAP alter: compile nltran alter 'trlg ','' TRLG CASESX2H,USETD,DLT,,BGPDTS,SILS,CSTMS,TRL,DIT, GMD,GOD,,EST,MPTS,/,,,PDT,,TELH,DLT1/ /S,N,DCOL/IMETHOD/RSTIME/BETA/S,N,FAC1/S,N,FAC2/ S,N,FAC3//TABS $ endalter

@ 35261;

70.7; Heat Transfer - CHBDYG, NLCOMB - UFM 4017 If the area of a CHBDY element is less than 1.E-10, than the job may fail with the message: 0*** USER FATAL MESSAGE 4017, ESTNL AND FILE (103) ARE NOT PROPERLY ALIGNED *** SYSTEM FATAL MESSAGE 3007 (NLCOMB) ILLEGAL INPUT TO SUBROUTINE NLCKDT 0FATAL ERROR Avoidance: This error may be caused by a poor selection of units. Change the units to increase the size of the element.

35277;

70.7; Static Analysis - LOAD, TEMP, GRAV - Wrong Answers If a LOAD bulk data entry references a TEMP bulk data entry which has already been referenced by a previous subcase (TEMP(LOAD)), then the temperature loads are ignored and wrong answers are produced for all subcases which contain GRAV or TEMP loads. In addition, then the warning message below, will not be issued: USER WARNING MESSAGE 3204 (SSG1) UNABLE TO FIND SELECTED SET ( (SSG1) 2) IN TABLE (SLT ) IN SUBROUTINE

Avoidance: Do not reference TEMP bulk data entries from the LOAD bulk data entry. In order to combine temperature and mechanical loads, use a LOAD case control entry to specify the static loads, in combination with the TEMP(LOAD) case control entry to specify the TEMP loads. @ 35280; 70.7; Coupled Fluid/Structure Analysis - PARAM,PANELMP - Wrong answers In a coupled modal frequency response analysis (SOL 111), if there is no ASETi or OMITi entries, rigid elements, multipoint constraints nor single point constraints, all panel participation factors are zero. Avoidance: Add 1 dummy grid with all degrees of freedom constrained. @ 35289; 70.7; Miscellaneous - DBLOCATE - UFM 453 DBLOCATE fails if the MASTER dbset is assigned twice under different logical names and the first assignment is not used on a DBLOCATE statement. The following is an example of the input and fatal message: ACQUIRE NDDL ASSIGN ADB_1='test.MASTER.2' DBLOCATE LOGICAL=ADB_2 WHERE(DBSET='ADB' OR DBSET='AEDB') $ ASSIGN ADB_2='test.MASTER.2' DBDICT ENDJOB

*** USER FATAL MESSAGE 453 (LCXLOD) THE DATABASE NUMBER (IDBSNO)= 3 BEING USED IN THIS RUN DOES NOT HAVE ENOUGH INFORMATION TO START FROM. @ 35291; 70.7; NAST/API Toolkit - NDDL Qualifiers not reset to their defaults NDDL qualifiers are not reset to their default values when submitting multiple MSC/NASTRAN jobs via the API Toolkit. Though this was the original design intent, this has been shown to cause errors if the second and subsequent jobs depend on qualifier default values as is the case with restarts. The application programmer should reset these values manually via dmap in such cases.

@ 35295;

; Documentation/Linear Static User's Guide - Inertia Relief In SOL 101, inertia relief is calculated differently than in SOL 1. The description of inertia relief in the Linear Statics User's Guide describes the method used in SOL 1 to calculate the accelerations. In SOL 101, the approach is different. In SOL 101, 6 additional loads are created. These loads are created by applying unit accelerations in each of the 6 dof about PARAM,GRDPNT. The matrix URA is the accelerations required on these 6 new loads to provide static equilibrium when combined with the original loadings. In SOL 1, the accelerations about the SUPORT dof are calculated.

@ 35298; 70.7.2; DMAP - SDR1, UMERGE1, UMERGE - Poor message When two matrices are being merged row-wise in modules SDR1, UMERGE1, SSG4, DDR2, and READ and the number of columns in each matrix do not match, then the following message is issued. This error also occurs in UMERGE and UMERGE1 if the the number of zeros and non-zeros in the partitioning vector do not match the corresponding number of rows and columns in the sub-matrices. 0*** SYSTEM FATAL ERROR 4276.GETSTR along with the following traceback Traceback: Offset 0x00000020 Offset 0x00000020 Offset 0x00000184 Offset 0x0000016c Offset 0x00004000 Offset 0x00000480 Offset 0x000001ac Offset 0x00000ec4 Offset 0x00000218 in in in in in in in in in procedure procedure procedure procedure procedure procedure procedure procedure procedure raise tracbak_ nsexit_ pexit_ gperr_ getstr_ mrgcol_ merge_ sdr1b_ ERROR CODE 31 PID= 0

System Fatal Message 6143 should have been issued instead. Avoidance: There is no avoidance. @ 35314; ; Buckling Analysis - Insufficient memory - SFM 3001,1015, Out of range, Error 924 In a buckling analysis and possibly other eigenvalue solutions, if insufficient memory is allocated, then depending on the machine and number of cpus requested, the job may fail with any one of the following unclear messages: 1. *** SYSTEM FATAL MESSAGE 3001 () THE INPUT DATA BLOCK IN POSITION xx DOES NOT EXIST

2. *** SYSTEM FATAL MESSAGE 1015 (GBKMGR) PHYSICAL FILE NOT FOUND IN FILE ASSIGNMENT TABLE BIOMSG: ERROR 4276 HAS OCCURRED IN ROUTINE GBKMGR , FILE INDEX (FILEX) = 0 STATUS = 0 3. BIORD ERROR - FILX = 0(00000000 HEX) OUT OF RANGE BIOMSG: ERROR 924 HAS OCCURRED IN ROUTINE IONAST , FILE INDEX (FILEX) = 0

BIOWRT ERROR - FILX = 0(00000000 HEX) OUT OF RANGE 0*** SYSTEM FATAL MESSAGE 1022 (XIO) ERROR OCCURED IN BIOWRT 0********* DUMP OF XIO PARAMETERS 0XIO NAME = 407 ATTEMPTED OPERATION = CLOSE FILEXX = 7 UNIRFX = 25908493 B = 018FDF6A CLR = 1060 CBP = 1060 FLAGS = 00000400 Avoidance: Increase memory. @ 35315; ; Parallel Processing - Process remains active after job is killed If a model using parallel on HPUX 10.20 is 'killed' by the system (for example, due to a full disk or exceeding file size limits), it is possible that a process may still remain active, possibly with a high CPU usage. Avoidance: If this occurs, use 'kill -9 pid' (where pid is the process id obtained using ps -ef) to 'kill' the remaining process. @ 35324; ; Submittal - No Output If an MSC/NASTRAN output file is opened by some other program, that other program may prevent MSC/NASTRAN from using the file. If it does, an error message such as the following may be displayed: *** SYSTEM FATAL MESSAGE (pgm: nastran, fn: run_job) Unable to rename: ./d10101d.out to ./d10101d.out.1 Avoidance: Close any MSC/NASTRAN output files in use by other programs. @ 35333; ; Nonlinear Analysis - CBUSH, PBUSHT - Nonlinear Affects Ignored If a nonlinear model contains a CBUSH element, no nonlinear materials, and PARAM,LGDISP is not set, NASTRAN will solve the problem as a linear analysis. Avoidance: Add PARAM,LGDISP to the deck. @ 35370; 70.7.2; Eigenvalue Analysis - READ, EIGRL - Distributed Memory Parallel Hangs MSC/NASTRAN Version 70.5.4 for IBM/SP computers introduces Distributed Memory Parallel Lanczos as a beta capability. There are known problems with the capability that may cause the system to hang. The problems include 1. If the master node does not find any roots the job will fail, and if bat=no is set the job will hang. This problem may be avoided by selecting a different number of processors. This problem may be also be caused by

error 2 below. 2. If dmp=n is specified on the submittal line but nastran numseg=x is not specified in the data deck or rc file, numseg will automatically be set to n. This action changes the shift logic which often causes the master node to find any roots. Error number 1 above is then encountered. An avoidance is to explicitly specify numseg. 3. Distributed Parallel Lanczos hangs the system if multiple subcases with different boundary conditions exist. The only avoidance is to submit the job as two separate runs. 4. Modal Frequency response jobs that use Lanczos are likely to hang in the frrd1 module. There is no avoidance. 5. Restarts are not supported, this is a limitation that may not be addressed for several releases. @ 35410;

; Nonlinear Transient Response - SPC, LSEQ - SFM 6920 in MATMOD In SOL 129, if the number of unique, with respect to DAREA ID, LSEQ Bulk Data entries is less than the number of subcases and static enforced displacements are specified, then SFM 6920 will be issued from the MATMOD module. Insert the following DMAP alter: compile nltran alter 'MATMOD YST','' paraml yst//'trai'/1/s,n,yscol $ if ( yscol>1 ) then $ MATMOD YST,,,,,/YS,/1/1 $ else $ equivx yst/ys/-1 $ endif $ yscol>1 alter 'MATMOD YST'(2),'' paraml yst//'trai'/1/s,n,yscol $ if ( yscol>1 ) then $ MATMOD YST,,,,,/YSt1,/1/1 $ else $ equivx yst/yst1/-1 $ endif $ yscol>1

@ 35411;

70.6; Submittal - news=yes - UFM 756 (XCSA) If an include statement is the first line of an input deck or follows immediately after a NASTRAN statement and news=yes is requested, then the job may fail with the message: USER FATAL MESSAGE 756 (XCSA) THE BEGIN CARD IS NO LONGER SUPPORTED IN THE EXECUTIVE CONTROL SECTION. USER INFORMATION: THE BEGIN STATEMENT IN THE EXECUTIVE CONTROL SECTION HAS BEEN REPLACED BY THE SUBDMAP DMAP STATEMENT. SEE USER MANUAL SECTION 5. ENDDATA *** USER FATAL MESSAGE 765 (XREAD)

END-OF-FILE ENCOUNTERED ON UNIT 1. THE INPUT FILE IS INCOMPLETE. USER INFORMATION: THE FOLLOWING GUIDE LINES SHOULD BE FOLLOWED FOR PREPARATION OF AN INPUT DECK. FILE MANAGEMENT SECTION (FMS) CAN BE ENDED BY EITHER AN "ENDJOB" OR A NON-FMS CARD. EXECUTIVE CONTROL SECTION CAN ONLY BE TERMINATED BY A "CEND" CARD. Avoidance: Specify the Executive control statements in the input file and include the rest of the data. @ 35415; ) 70.7; Optimization/Design Sensitivity - REPCASE, SET - SFM 3003 (DOPR3 The fatal message shown below may be issued in an optimization job if there are REPCASE subcases but no SETt case control commands (all the data recovery requests = ALL) *** SYSTEM FATAL MESSAGE 3003 (DOPR3) ATTEMPT TO READ PAST RECORD IN DATA BLOCK Avoidance: Place a set command in the case control. @ 35460; 2001; Nonlinear Transient Analysis - NLTRD, PARAM,SNORM - SFM 4276 A nonlinear transient run may give a traceback similar to the following when param,snorm is used. +80 +81 +82 +83 +84 +85 +86 Avoidance: @ 35483; o Offset Offset Offset Offset Offset Offset Offset 0x000001c4 0x0000003c 0x00001688 0x00000258 0x00001124 0x000004c0 0x000001a4 in in in in in in in procedure procedure procedure procedure procedure procedure procedure getnml_ gshnod_ nqd4d_ nembrd_ nlemg_ nltrd_ xsem15_

Do not specify param,snorm.

70.7; Performance - SGI Only, Implausible Warning Message with ff_io=n MSC/NASTRAN V70.5(r4) mistakenly omitted support for EAG FFIO in the template file. On some jobs this may lead to a slight degradation in I/O performance and hence longer run times. Avoidance: This can be corrected by updating: <install-dir>/bin/nast705.lcl After line 143, which currently reads: %MSC_ARCH=irix64:%%threads:% %thread_verbose%" the following lines can be added: %MSC_ARCH=irix64:%%ff_io=^[ya]:%%ff_io_opts:% export FF_IO_OPTS='%ff_io_opts%' %MSC_ARCH=irix64:%%ff_io=^[ya]:%%ff_io_defaults:% export FF_IO_DEFAULTS='%ff_io_defaults%' print "MP_SUGNUMTHD_VERBOSE:

%MSC_ARCH=irix64:%%ff_io=^[ya]:% print "EAG FFIO enabled." %MSC_ARCH=irix64:%%ff_io=^[ya]:% print "FF_IO_OPTS: $FF_IO_OPTS" %MSC_ARCH=irix64:%%ff_io=^[ya]:% print "FF_IO_DEFAULTS: $FF_IO_DEFAULTS" %MSC_ARCH=irix64:%%ff_io=^n:% print "EAG FFIO disabled." %MSC_ARCH=irix64:%%ff_io=^n:% unset FF_IO_OPTS %MSC_ARCH=irix64:%%ff_io=^n:% unset FF_IO_DEFAULTS @ 35508; ; Eigevalue Analysis - READ - UIM 4158, Massless Mechanisms Eigenvalue analysis of models with massless mechanisms may fail in the READ module with the following messages: *** USER INFORMATION MESSAGE 4158 (DFMSA) ---- STATISTICS FOR SPARSE DECOMPOSITION OF DATA BLOCK KFF FOLLOW NUMBER OF NEGATIVE TERMS ON FACTOR DIAGONAL = 5 MAXIMUM RATIO OF MATRIX DIAGONAL TO FACTOR DIAGONAL = 1.0E+15 AT ROW NUMBER 4*** SYSTEM FATAL MESSAGE 7340 (LNNHERR) process error reported by subroutine LNNP1ED see LNNP1ED abstract (ier = -729) *** SYSTEM FATAL MESSAGE 7340 (LNNHERR) process error reported by subroutine LNNDRVD see LNNDRVD abstract (ier = -729) Avoidance: Add a small amount of mass to the mechanism or remove the mechanism. @ 35547; 70.5.5; ESTIMATE - Will not work if there is a blank space in path If a directory in the path to the file contains a blank space (for example ('Program Files'), MSC/ESTIMATE will not be able to find the input file and will fail. Avoidance: Do not include blank spaces in directory names. @ 35551; 70.7; Superelements - THRU, SPOINT, EXTRN - UFM 6739 (SEP2MP) If the THRU range is used on the EXTRN entry for a set of SPOINTs, the program will incorrectly issue UFM 6739 stating that the SPOINTs are specified with a nonzero degree of freedom Avoidance: Do not use the THRU option on the EXTRN for SPOINTs. @ 35562; 70.7; Aeroelastic Analysis - TRIM - Wrong Answers Version 70.5 introduced the capability in SOL 144 to trim an aircraft that had more free variables than there are free degrees of freedom. The results from this Generic Control System capability will be in error unless the aerodynamic extra points (AESTAT and AESURF) specified on the TRIM bulk data entry are in ascending ID order. Avoidance: When there are redundant control surfaces, list the known trim variables on the TRIM entry in ascending ID order. The ID's are those entered on the AESTAT and AESURF bulk data entries. @ 35603; ssages 70.7; Parallel Processing - Distributed Memory Parallel - Confusing Me

At submittal, confusing error messages may be issued to the terminal when running MSC/NASTRAN V70.5.3 or V70.5.4 on AIX in distributed mode. The various messages may lead you to believe the job has failed even though it hasn't. The error messages will be displayed if the IBM Parallel Environment for AIX V2.4 or later is installed. Avoidance: The messages may be ignored or they may be avoided by changing the following line in the <install-dir>/bin/nast705.dmp file from: %MSC_ARCH=aix:% to %MSC_ARCH=aix:% print -u3 $host %cpu_use% %adapter_use% print -u3 $host

@ 35615; ; Aeroelastic Analysis - scr=mini, multiple subcases, SOL 145 - Mi ssing output If multiple flutter subcases are specified in a SOL 145 job and the job is run with scr=mini, then the job will only process the first subcase and ignore subsequent subcases. Avoidance: Insert the following DMAP alter: compile super3 alter 'saverstl\=','' compile sefluttr alter 'call super3'(,1) if ( saverst='sedr' ) saverstl=true $ compile sedisp alter 1 file ol1s=ovrwrt $ endalter @ 35619; 70.7; Complex Eigenvalue Analysis - Number of Desired Roots not Found Prior to MSC/NASTRAN 70.7 the Complex Lanczos eigenvalue method would require the user to specify many shifts to find many modes in a lightly damped system. Beginning in Version 70.7, an the system will automatically choose new shifts until the desired number of modes is found. @ 35622; 70.5.2; DMAP - ISHELL may have Difficulties Locating the Called Program In versions prior to V70.5.2, the ISHELL module may have difficulty locating the called program file. This may result in the fatal message: ^^^ PRECLOSING FORTIO UNIT 25 ^^^ ISHELL PROCESS RETURN CODE: ^^^ ISHELL FAILED WITH RETURN CODE 1 1

Starting with V70.5.2, the search logic has been improved to search for different file extensions and thru the path correctly. Avoidance: If you are using ISHELL in a version prior to V70.5.2, use a program name containing the exact name specified in the ISHELL statement (upper-case only, maximum of 8 characters including the extension) and make sure that the program is in the current directory. @

35630; ; Postprocessing/Eigenvalue Analysis - PARAM,POST- Wrong Eigenvect or Labels in XDB If multiple eigensolutions are requested in SOL 103 with PARAM,POST,0, then the eigenvectors will be labeled incorrectly in the xdb file. Avoidance: Use PARAM,POST,-1 or insert the following DMAP alter: compile super3 alter 'EQUIVX CASECC','' EQUIVX CASErs/CASESEP4/-1 $ endalter @ 35633; ; Acoustic Analysis - Missing Output In solution 108, if displacements are requested for fluid grids, and accelerations are requested for structural grids, then no fluid grid pressures will be output.. Avoidance: Solution l111 works correctly. @ 35646; 70.7; Acoustic Analysis/Superelements - CAABSF - UFM 9101 If a fluid superelement is used, and frequency dependent acoustic absorbers (CAABSF) are correctly put in the residual with their fluid grids (in order to keep them frequency dependent), then an error message similar to the following is erroneously invoked causing the run to fail prematurely: ^^^ BULK DATA AND CASE CONTROL PARTITIONING, RESTART CHECKING, AND UNDEFORMED PLOTTING IS ^^^ INITIATED FOR SUPERELEMENT 0 ^^^ USER FATAL MESSAGE 9101 (PHASE0) ^^^ FLUID ELEMENTS ARE DETECTED IN THIS SUPERELEMENT. ^^^ USER ACTION: ENSURE THAT ALL FLUID ELEMENTS ARE ASSIGNED TO SUPERELEMENT 2000 Avoidance: Insert the following DMAP alter: compile phase0 alter '9101'(,-1),'' paraml ects//'trailer'/5/s,n,trwd5 $ bits 161-176 and 65-80 type parm,,logi,n,fluidel $ fluidel = MOD(trwd5/2**6,2)>0 or MOD(trwd5/2**5,2)>0 or MOD(trwd5/2**4,2)>0 $ if ( fluidse>-1 and seid<>fluidse and ((acous>0 and seid>0) or (fluidel and seid=0)) ) then $ endalter $ @ 35647; 70.7; Elements - CBEAM, GP2 - UFM 2007, Undefined Grid Point If a model contains both CBAR and CBEAM elements and the value of BIT is set to 0.0 on a CBEAM entry and p-elements are not being used, the run will fail with UFM 2007 stating that the BEAM element references an undefined GRID point.

*** USER FATAL MESSAGE 2007 (GP2) ELEMENT xx REFERENCES UNDEFINED GRID POINT yyy Avoidance: Do not place a value in the BIT field of the BEAM element unless you are using p-elements. @ 35657; 2001; Documentation/QRG - THERMAL - Clarification of Remark 1 The following remark should be added to THERMAL case control command description in the Quick Reference Guide. 3. In a transient heat transfer analysis the SID on the punched TEMP bulk data entries equal the time step number. @ 35695; pied. ; Restarts - Nastran fails to trap invalid dbset that have been co Nastran fails to trap invalid copies of DBSETs used in RESTART (or reuse) runs. The following sequence demonstrates the problem: > nastran s31498a.dat # cold start run creates s31498a.MASTER and s31498a.DBALL > cp s31498a.MASTER s31498b.MASTER # copy s*a.MASTER to s*b.MASTER; DBALL is not copied > nastran s31498b.dat # run using s*b.MASTER; nastran finds s*a.DBALL using autoassign > nastran s31498a.dat # run using s*a.MASTER; nastran fails to recognize s*a.DBALL was modified above and is no # longer in synch with s*a.MASTER @ 35726; 2001; Elements - CBUSH - UWM 979 (XYTRAN), THE ID COULD NOT BE FOUND Trying to use XYPLOT for CBUSH postprocessing results in the following error message: *** USER WARNING MESSAGE 979 (XYTRAN) AN XY-OUTPUT REQUEST FOR POINT OR ELEMENT ID n -FORC- CURVE IS BEING PASSED OVER. THE ID COULD NOT BE FOUND IN DATA BLOCK 106 @ 35743; meters 70.7; Aeroelastic Analysis - IFP - UFM, TRIM Entries with no Fxed Para If you define a TRIM subcase in MSC/NASTRAN in which all trim parameters (aerodynamic extra points) are FREE, the resultant MSC/NASTRAN execution will fail due to an illegal TRIM Bulk Data Entry.MSC/NASTRAN considers it to be illegal to have a TRIM entry with zero PARAM/VALUE pairs. That TRIM entry would imply that all parameters were FREE, something that is not necessarily illegal, even for regular TRIM. Avoidance: Specify at least one parameter that is fixed at some value. @ 35745; TA1) 2001; Elements - PBARL/PBEAML, SOLs 1-16 - UFM 1126 (GNFIST) or 2011 ( PBARL and/or PBEAML entries do not work in SOL 1. They cause the following error message:

in V69.0 - V70.0; *** USER FATAL MESSAGE 1126 (GNFIST) DMAP MODULE MODEPT IS ATTEMPTING TO OUTPUT DATABLOCK 201 NAME = IEPT , WHICH ALREADY EXISTS. USER ACTION: DELETE THE DATABLOCK, OR USE FILE IEPT =OVRWRT FATAL ERROR or UFM 2011 (TA1) in V70.5; *** USER FATAL MESSAGE 2011 (TA1) NO PROPERTY ENTRY FOR ELEMENT TYPE BAR *** USER FATAL MESSAGE 2011 (TA1) NO PROPERTY ENTRY FOR ELEMENT TYPE BEAM 0FATAL ERROR Avoidance: Insert the following DMAP alter in SOLs 1-16: MALTER 'IFP9 EPT','' IFP9 EPT.7/IEPTX/S,N,NOGOIFP9 $ endalter @ 35754; 70.7; Documentation - Missing Beam Server Documentation in NT C.O.G. The documentation for the beam server is missing from the "Configuration and Operations Guide, Windows NT Edition". Avoidance: The UNIX documentation can be used in its place. @ 35760; GL; Transient Analysis - MPCFORCES - Inconsistent Answers In a transient analysis, the MPCFORCES may appear incorrect because they may vary when the independent and dependent degrees of freedom are reversed. This is due to the inconsistency between the way the "equilibrium" is calculated using the modified Newmark integration scheme (which is evaluated at 3 time steps-- Un+1, Un, and Un-1) and the MPCFORCE data recovery (which is evaluated at Un). Avoidance: Use a smaller time step if mpcforces are desired. The error is negligible if (delta t)**2 * acc << disp @ 35762; 2001; DMAP - MATMOD, P1=4 - Possible Wrong Answers Matmod option 4 (p1=4) does not produce the correct output matrix when using form 2 (LUSETS). Avoidance: Use form 1 (SILS) instead of form 2 @ 35765;

; Nonlinear Analysis - ITER - UFM 3032 (NLITER), Poor Message If the ITER entry uses the free field keyword approach but it is entered incorrectly a misleading fatal message may result. For example the entry "ITER,100,IPAD=3" lead to the following fatal message: *** USER FATAL MESSAGE 3032 (NLITER) UNABLE TO FIND SELECTED SET ( 100) IN TABLE ( 0FATAL ERROR

This entry should have been: ITER,100 ,IPAD=3 @ 35778; ; Superelements - SELOC, SEP1X - Poor Message If the GRID points or POINTs listed on a SELOC entry in the PBi fields do not have the same relative geometry as the GRID points listed in the PAi fields, then the following misleading error message may occur: *** USER FATAL MESSAGE 6790 (SEPO21) THE FOLLOWING SUPERELEMENTS ARE DETACHED: Avoidance: Be sure that the relative geometry is correct. @ 35781; 70.7; Nonlinear Analysis - PSHELL, MID2=-1 - Nonlinear Stresses, SWM 6 065 (DBC) If plane stress elements are mixed with plane strain elements (i.e., PSHELL with MID2=-1) then garbage may appear in the nonlinear stress output and if PARAM,POST,0 is selected then the DBC module will issue the following warning message: SYSTEM WARNING MESSAGE 6065 (DBC) DBCMSG ERROR MESSAGE 22 FROM SUBROUTINE ELMBLK Avoidance: There is no avoidance. @ 35790; ; RFORCE - SOL 103, Followk - Wrong Answers RFORCE follower force effects (param,followk) were included in Version 70.6. However, it was not documented that only RFORCE method 2 is supported. If method 1 is chosen follower force effects are ignored. Avoidance: Use method 2 for the RFORCE. @ 35819; 70.7; Superelement - RROD, RBAR- UFM 6591 (SEP2RP) If RROD or RBAR elements are used to connect interior and exterior degrees of freedom on a PART superelement, the run may incorrectly fail with the message: *** USER FATAL MESSAGE 6591 (SEP2RP) RIGID ELEMENT DEFINES DEPENDENT EXTERIOR POINTS Avoidance: An avoidance which sometimes works is to interchange the ga and gb points, and their related data. This behavior has not been found using other rigid elements, such as the RBE2. @ 35821; s

2001; Documentation/QRG - PCOMP - ELFORCE Not Required for Ply Stresse In the MSC/NASTRAN Quick Reference Guide, Remark 7 under the PCOMP entry which states that an ELFORCE and ELSTRESS request is necessary for composite ply stresses is no longer true. Only a STRESS request is required.

@ 35845;

; Iterative Solver - PARAM CDITER, NASTRAN ITER=YES The iterative solver produces incorrect results for the SPCFORCE RESULTANT when the linear gap (PARAM,CDITER) is used. Avoidance: Use the direct solver or ignore the spcforce resultant output.

@ 35851;

70.6; DMAP - ISHELL - Access Violation on NT A call to the ISHELL module in Version 70.5 on Windows NT may fail with an Access Violation. Avoidance: Two possible avoidance exist. If the problem occurs, use Version 70.6 or higher. Else, use character parameters whose length is 8 characters with no blank spaces. The workaround for Version 70.5 may not always work. Whether or not the access violation occurs depends on the contents of the various data fields and is unpredictable.

@ 35857; ssage

; Nonlinear Transient Analysis - PCOMP, STRESS - SFM 1013, Poor Me Composite materials run in sol 129 that request stress output may fail with the message: *** SYSTEM FATAL MESSAGE 1013 (GBKMGR) CANNOT FIND CORRECT CLUSTER POINTER RECORD BIOMSG: ERROR 4276 HAS OCCURRED IN ROUTINE GBKMGR , FILE INDEX (FILEX) = 0 Avoidance: Do not use the STRESS= case control command since nonlinear transient analysis does not support composite data recovery. Else substitute equivalent PSHELL/MAT2 entries for the PCOMP entries.

@ 35876;

70.7; SSSALTERs - checka - Effective Weight Matrices Incorrect Output for the checka.vxx alter has the row and column output for the modal effective mass and FRACTOT output incorrectly labeled. It should state that each column represents a mode and each row represents an associated degree of freedom. Avoidance: Edit the alter to correct the printout form or request an updated copy from MSC.

@ 35877;

70.7; Rigid Elements - RSPLINE - UFM 5289, Incorrect DOF Identified If an RSPLINE specifies dependent degrees of freedom that have already been specified as dependent by another RSPLINE entry, then USER FATAL MESSAGE 5289 is correctly issued. However, the UFM incorrectly calls out the independent degrees of freedom on the entry as dependent and lists the independent degrees of freedom twice. The dependent degrees in error are also listed but they may be obscured by the incorrect data. Avoidance: Inspect the RSPLINE and other MPC entries to determine which are in error.

@ 35882; f 2

70.7; SSSALTERs - modevala.vxx - Matrix STRAINED is off by a Factor o If PARAM,EVAL,-1 is used in the modevala.vxx alter, the matrix for the modal strain energy (STRAINED) is off by a factor of 2. If the recommended value of PARAM,EVAL,-2, is used this problem does not occur. Avoidance: Modify the alter to divide the MATRIX by 2 as follows: Replace: DIAGONAL STRAINE/STRAINED/ $ with DIAGONAL STRAINE/STRAIND2/ $ add straind2,/strained/(.5,0.) $

@ 35886;

70.7; Loads - LOAD(TEMP), LOAD - No UFM, Unexpected Answers As documented in the Quick Reference Guide, the TEMPERATURE load ID should be unique with respect to all other loads in the model. However, this rule is not enforced, and misleading results can occur under the following situation. 1) The TEMPERATURE load has same set ID as a mechanical (e.g. FORCE, PLOADi) or body (e.g. GRAV) load. 2) The mechanical or body load is scaled by a non-unity value on a LOAD bulk data entry. Although the Quick Reference Guide LOAD card documentation implies that the TEMPERATURE loads are ignored by the LOAD card, the temperature load delta-T will be scaled by the load factor on the LOAD card if there is a mechanical/body load with the same set ID as the TEMPERATURE set. Avoidance: Ensure that the TEMPERATURE set is unique with respect to all other loads in the model. MSC/NASTRAN does not issue any warning or information messages when the TEMPERATURE set is not unique.

@ 35892;

70.7; ACCESS - Sample Program Build Errors The MSC/ACCESS sample programs are built with the command msc705 access build Failing to use the msc705 (msc706, msc71, et al) and access scripts will result in the makefile failing. Alternatively, you could set MSC_BASE and MSC_ARCH to the correct values and directly invoke access build However, under no circumstances can you use make or make build

@ 35893;

70.7; Plotting - OUTPUT(POST) - SFM 4276. BIOMSG ERROR 3012 If OUTPUT(POST) and SETS DEFINITION are incorrectly both included in the case control the following error message is output: 19 OUTPUT(POST) 20 SETS DEFINITIONS BIOMSG: ERROR 3012 HAS OCCURRED IN ROUTINE QOPEN = 0

, FILE INDEX (FILEX)

The message should say the same thing as Remark 1 of the SETS DEFINITION description of the MSC/NASTRAN Quick Reference Guide: "Either SETS DEFINITION or OUTPUT(POST) may be specified but not both." @ 35895; 70.7; Buckling - EMG, TEMP(LOAD) - SFM 3003 EMG ETT If a buckling job, SOL 105, contains multiple static/buckling subcase pairs, and if a temperature set (e.g. TEMP(LOAD)) is called out in the second or later static subcase that does not exist, then the following error occurs: SYSTEM FATAL MESSAGE 3003 (EMG) ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK ETT (FILE 107) IN SUBROUTINE EMGTAB Avoidance: Add the referenced temperature set entries to the bulk data file. @ 35896; ; Optimization/Design Sensitivity - PCOMP, Z0 - Unexpected Answers When optimizing ply thicknesses in Sol 200, the user must remember to recalculate Z0 if the midplane is to remain at the nodes. It is common for users to leave field 3 of the PCOMP entry (Z0) blank, and let MSC/NASTRAN calculate Z0 as -t/2. But in optimization, this field is not updated automatically, which can lead to unexpected results. Avoidance: Either:Create another design variable for Z0 (field 3 of PCOMP) and use DLINK to calculate updated Z0 or use an additional DVPREL1 to calculate z0 = 1/2 ( c0 + x1 + x2 + x3 + ... xn) where co is the thickness of invariant plies and the xi's are designed ply thicknesses) @ 35901; 70.7; Optimization/Design Sensitivity - Rigid Element Sensitivity - UF M 1126 in UMERGE When a shape optimization job includes grid points which are associated with rigid elements and are designed (referenced by DVGRID/DVSHAP/DVBSHAP) and the job is run with the option of SCR=no, it may fail in subdmap respsen with the following error message: *** USER FATAL MESSAGE 1126 (GNFIST) DMAP MODULE UMERGE IS ATTEMPTING TO OUTPUT DATABLOCK = DUG , WHICH ALREADY EXISTS. USER ACTION: DELETE THE DATABLOCK, OR USE FILE DUG=OVRWRT Avoidance: 201 NAME

1) Use SCR=YES 2) Use the following dmap alter if SCR=NO, compile respsen alter '\/ dug ','\/ dug '(,1) umerge uset, uny, dum / dugz / 'G'/ 'N'/ 'M' $ equivx dugz / ugy / always $ alter '\/ dug '(2),'\/ dug '(2,1) umerge uset, upnds1, dum / dugz / 'G'/ 'N'/ 'M' $ equivx dugz / upf / always $ compile update alter 'restart ygp' DBVIEW RMGP=RMG WHERE (DESITER=DESP) $ RESTART RMGP,RMG,//true $ endalter @ 35914; 70.7; Static Analysis - MPCForce output - Wrong Answers In linear static analysis (SOLs 101 and 200), if there exist MPC entries, but no rigid element entries, and some subcases, but not all select those MPCs, then MPCFORCE output may be labeled with the incorrect SUBCASE ID listed in the upper right hand corner of the printed output and the forces will be wrong. Avoidance: Add a rigid element or insert the following alter: compile sedisp alter 'IF ( NOQMG','' IF ( NOQMG>-1 and not(delta) ) THEN $ if ( nomset<0 ) then $ PARAML USET//'USET'//////'g'/S,N,NOgSET $ matgen ,/qmx/7/nogset/nloads $ else $ alter 'call cforce' endif $ endalter @ 35918; 70.7; Acoustic Analysis - PARAM,WTMASS - Inefficient Processing When param, wtmass is used in conjunction with fluid elements it causes the mass matrix to be of form 1 (unsymmetric). This causes an unsymmetric solution, which requires more memory and cpu time than a symmetric solution. Avoidance: The following alter avoids this problem: compile semg list noref alter 'MPYAD.*MJJX1','' $ add sym flag matprn wtmassg// $ MPYAD MJJX1,WTMASSG,/MJJX/////6 $ force symmetry @ 35921; 70.7; Aeroelastic Analysis - SPLINE3 - UFM 2331 (MKSPL) Aeroelastic Analysis with SPLINE3 entries may fail with the following message in Version 70.5: *** USER FATAL MESSAGE 2331 (MKSPL) BOXES PICKED ON SPLINE3 ENTRY ID = CAERO ENTRY ID = yyy xxx HAVE NOT BEEN GENERATED BY

to

USER INFORMATION: CAERO ENTRY ID = zzz.

yyy GENERATES BOXES

yyy

Avoidance: Use SPLINE2 entries if possible. If that is not possible, order GRID components by internal order which can be determined by using PARAM,USETPRT,1 . @ 35931; 70.7; Restarts/P-Elements - PARAM,PVALINIT - Wrong Answers If a restart is performed from a p-version run and PARAM,PVALINIT is not specified, then the restart will default the starting p-value to the first p-value rather than the last p-value computed in the cold start. Avoidance: Specify PARAM,PVALINIT or insert the following DMAP alter:

compile ifps1 alter 'type ','' type parm,,i,n,seid,gnmax,pvalid,pvaltmp,np,adptindx,pval0, oldseq1 $ type parm,nddl,i,n,pvallast $ endalter @ 35932; ; Nonlinear Analysis - NLRGAP - Radial Gap without Friction, UFM 3 16 (IFPDRV) Using the NLRGAP does not allow input of the nonlinear forces without reference to a TABLED1 which specifies O.O for friction. Leaving the field 8 blank yields the following message: *** USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA ENTRY NLRGAP SORTED ENTRY COUNT = 36 for the input item: NLRGAP 20 SID 90 GA 101 GB XY plane 21 tabk tabg tabu radius 22 0.0 20

Avoidance: The avoidance is to specify a table for the TABU field with values of 0.0 for friction. @ 35935; ; Nonlinear Analysis - NLRGAP - UFM 2093 (DPD4) An NLRGAP element that identifies a non-existent grid id produces a fatal message that incorrectly identifies the grid. The following fatal message was produced by an NLRGAP element that referenced a nonexistent GRID id of 90: *** USER FATAL MESSAGE 2093 (DPD4) NOLIN OR NLRGAP SET 20 REFERENCES GRID 100 COMPONENT SIL 61 WHICH IS NOT IN UD SET. *** USER FATAL MESSAGE 2093 (DPD4) NOLIN OR NLRGAP SET 20 REFERENCES GRID 100 COMPONENT SIL 62 WHICH IS NOT IN UD SET. *** USER FATAL MESSAGE 2067 (DPD) UNDEFINED GRID POINT 90 ON NOLIN- ENTRY. Avoidance: Correct the NLRGAP entry. @ 1 2

35937;

; Performance - CPU Time not reported consistently The CPU time reported in the f04 file is not consistently reported across all platforms. Thus, comparisons of CPU time between machines that are reported differently are meaningless. The method of reporting for most systems is shown below: Compaq: Cray: HP: IBM: NEC: SGI: Sun: user + system user user user + system user user user

Windows: user + system Avoidance: To compare CPU times across different platforms, use the time summary at the end of the LOG file. @ 35939; es 70.7; Miscellaneous - INCLUDE - UFM 398, Unexpected Answers or Messag In Version 70.6, if a referenced include file does not exist, the job will issue the following fatal message *** USER FATAL MESSAGE 398 (GETLIN) A REQUESTED INCLUDE FILE WAS NOT FOUND . . THE INCLUDE REQUEST IS SKIPPED but the job will continue. This may lead to unexpected results as the information in the INCLUDE file will not be included. Avoidance: Make sure that the referenced INCLUDE file exists. @ 35993; 70.7; Optimization/Design Sensitivity - DRESP1 - UFM 6482 (DOPR3R) If the DESOBJ references a DRESP1 that refers to a dynamic response (e.g., FRDISP or TACCL) and there are constraints on nondynamic responses (e.g., WEIGHT or DISP) then the message: USER FATAL MESSAGE 6482 (DOPR3R) FOR SUBCASE 0, THE ANALYSIS TYPE IS GLOBAL COMMAND REFERENCES AN INCOMPATIBLE DRESP1 ENTRY WITH RID = n AND RTYPE = FRACCL. 0FATAL ERROR BUT THE DESOBJ

is produced in error and it incorrectly states that the response type is incompatible with the analysis type. Avoidance: Modify the input to have the DESOBJ refer to a DRESP2 and then create a DRESP2 as shown in the example Example, CASE control has: DESOBJ=40001

Bulk data has DRESP1, 40000, ZIEL1, FRACCL , , , dresp2 40001 kludge 40000 dresp1 40000 deqatn 40000 kludge(fraccl) = fraccl @ 35996; ; SSSALTERs - trncomp1 - SFM 3002 (SDR2) If the sssalter trncomp1 is used with PARAM,DDRMM,-1 or in SOL 112 T then the message: SYSTEM FATAL ERROR 3002 (SDR2 ) EOF ENCOUNTERED WHILE READING DATA BLOCK SILS SDR2. (FILE 106) IN SUBROUTINE 2, 140., 6294

is issued whenever a STRAIN case control request occurs. Avoidance: To avoid the error the follwing sdr2 dmap statements must be changed. SDR2 SDR2 CASEDR,CSTMS,MPTS,DIT,EQEXINS,SILS,ETT,OL2,bgpdts, ,,UG,EST,,,,,GPSNTS,,,/,,,,OEF1AX,/ APP1/S,N,NOS/FRCNCMP $ CASEDR,CSTMS,MPTS,DIT,EQEXINS,SILS,ETT,OL2,bgpdts, ,,UG,EST,,,,,GPSNTS,,,/,,,OES1AX,,/ APP1/S,N,NOS/STRNCMP $

to: SDR2 CASEDR,CSTMS,MPTS,DIT,EQEXINS,SILS,ETT,OLt,bgpdts, ,,UGdisp,EST,,,,,GPSNTS,,,/,,,,OEF1AX,/ 'reig'/S,N,NOS/FRCNCMP $ CASEDR,CSTMS,MPTS,DIT,EQEXINS,SILS,ETT,OLt,bgpdts, ,,UGdisp,EST,,,,,GPSNTS,,,/,,,OES1AX,,/ 'reig'/S,N,NOS/STRNCMP $ @ 36000; ; Nonlinear Transient Analysis - NLTRD2 - Performance Degradation In many problems nltrd2 runs slower in Version 70.5 than in previous versions. The investigation continues; however, it has been found that increasing smem from the default of 100 to 1000 significantly improved the performance of nltrd2 to where it exceeded prior versions performance. Avoidance for V69: Nastran command should include keyword "sysfield=mapio=no" in all v69 runs to improve the performance. @ 36014; 70.7; Aeroelastic Postprocessing - FLD&S - Unrestrained Unit Solution not available The static aeroelastic solution sequence (SOL 144) will store the "unit solutions" (displacements and aeroelastically corrected forces associated with the unit perturbation of the aerodynamic extra point) on the XDB if PARAM, POST,0 is specified in the Bulk Data. Due to a typographical error, the variant of the unit solution forces associated with the unrestrained formulation are not correctly passed to the DBC module for storage. Consequently they are not available to be visualized in

SDR2

MSC/PATRAN V8.1+ (which has been enhanced to recognize these data as part of MSC/FLD V1.0). No other results are affected since the data are computed. They are simply not stored for subsequent post processing. Avoidance: The following DMAP alter against V70.6 will correct the problem compile aeopu list noref $ $ Correct the DMAP error for unrestrained unit solutions on $ the structural mesh (FUM==>FURM; IUM==>IURM) $ alter 'dbc *aebgpdt.*prxks','' DBC AEBGPDT, prxks, perxks, peuxks // 'BEPT' /'RM'/'FRM'/'FURM'///////////////// 0/AEDBCP/S,N,CP/'AERO'/0/GEOMU/LOADU/POSTU/ DBCDIAG/DBCCONV/DBCOVWRT/0///// $ alter 'dbc *bgpdts.*perxg','' DBC BGPDTS, PRXG, PERXG, PEUXG, PIRXG, PIUXG // 'BGPDT' /'RM'/'FRM'/'FURM'/'IRM'/'IURM'/////////////// 0/DBCPAE/S,N,CP/'AERO'/0/GEOMU/LOADU/POSTU/ DBCDIAG/DBCCONV/DBCOVWRT/0///// $ @ 36042; 70.7; Optimization/Design Sensitivity - BEAM Library, DVPREL1 - Wrong Answers The design of the PBARL creates internal constraints on the bar dimensions to insure, for example, that the inner radius of a tube does not exceed the outer radius. The calculation of these constraints will be in error if the DVPREL1 that designs a beam dimension that is included in the constraint has a non-zero value for C0. Avoidance: There is no avoidance other than to recast the design task in a way that eliminates the need for a C0 term. @ 36066; 70.7; P-Elements - Incorrect Answers for p>=10 P-Elements with p>=10 may yield incorrect applied loads, boundary conditions, and displacement coordinate systems.

@ 36073; )

2001; Acoustic Analysis - No Component Modes Computed for METHOD(FLUID If an uncoupled acoustic analysis is performed and some of the fluid grids are placed into the a-set because frequency dependent elements are attached to them, then in SOL 108 no component modes are computed when the eigenvalue extraction method is selected using the METHOD(FLUID) Case Control command. Instead the METHOD Case Control command has to be used. This is due to a logic error in SEMRM which does not look for METHOD(FLUID). Avoidance: A workaround is to use METHOD or to use the following DMAP-alter (tested with V70.6): compile semrm

alter 'PARAML CASES' if(method=0) PARAML CASES//'DTI'/1/226//S,N,METHOD $ endalter @ 36075; 2001; Eigenvalue Analysis - PARAM,RESVEC - SFM 2198 in FBS If residual vectors are requested for an unconstrained superelement which has no physical exterior points (no T-set), the run will fail with SFM 2198 in FBS which means that the LOO matrix is missing. Avoidance: Either have at least one exterior point for the superelement, or insert the following alter: compile sekr alter 'IF ( NLKOT>-1 )','IF ( NLKOT>-1 )' IF ( NLKOT>-1 OR RESVEC='YES') THEN $ endalter @ 36076; ; Miscellaneous - INPUTT2 - SFM 6551 with READ ONLY file On NT computers, if the permission of a file used for INPUTT2 is "read-only", the program may fail with SYSTEM FATAL MESSAGE 6551 indicating that the end-of-file was detected incorrectly. Avoidance: Change the file permission to be full access. @ 36078; ; Complex Eigenvalue Analysis - SCR=NO - UFM 9050 (SEKRRS) If while running a Complex Eigenvalue Analysis in MSC/NASTRAN Version 70.6 the run encounters ^^^ USER FATAL MESSAGE 9050 (SEKRRS) ^^^ RUN TERMINATED DUE TO EXCESSIVE PIVOT RATIOS IN MATRIX KLL. and SCR = NO, set SCR = YES this may correct the problem. @ 36096; 2001; Eigenvalue Analysis - GIV, MGIV, HOU, MHOU - UWM 4648,3034 When the tridiagonal eigensolution methods GIV, MGIV, HOU, or MHOU are selected, and the roots are closely spaced, the messages below may result: 0***USER WARNING MESSAGE 4648---- THE MODAL MASS MATRIX IS NOT POSITIVE DEFINITE. 0*** USER WARNING MESSAGE 3034, ORTHOGONALITY TEST FAILED. LARGEST TERM = 1.00E+00 ,NUMBER FAILED = 36 Avoidances: An avoidance is to select the Lanczos method instead. It tests for orthogonality as a requirement for stopping the solution process. Another avoidance when using the tridiagonal methods is to stlightly perturb the system so that the repeated roots are separated. It was found by experiment with a 10 by 10 problem with 9 repeated roots in one cluster that only the first four modes had good orthogonality. If 10 stiffness terms on the diagonal active in a cluster of 9 roots were perturbed so that each differed by 1.E-8 independent, orthogonal vectors were produced. The changes in the eigenvalues were not large enough to appear in the 7 digit format used to print them. @

36104; 2001; Optimization/Design Sensitivity - DEQATN - Infinite cpu Loop, re ad error If a large DEQATN entry is used, then depending on the machine, the job may terminate without any messages in the f06 or f04 files or the run may appear to be caught in an infinite loop. Depending on the machine, any of the following messages may appear in the log file. This situation can occur even if the number of non-zero characters are below the 32,000 limits. 1. fmt: read unexpected character apparent state: unit -1 named last format: (I2,4X,I4) Unit -1 is a (null) formatted internal file *** Execution Terminated (115) ***

2. 1525-097 A READ statement using decimal base input found the invalid digit '' in the input file. The program will recover by assuming a zero in its place. Avoidance: Reducing the size the DEQATN entry. @ 36112; tput ; Transient Response Analsyis - DRMS1 DMRH1 - UWM 5186, Missing Ou DMAP modules DRMS1 and DRMH1 will issue the message: *** USER WARNING MESSAGE 5186. (DRMH1C) A CHANGE IN WORD 2 OF THE OFP-ID RECORDS OF DATA BLOCK 101 HAS BEEN DETECTED. PROCESSING OF THIS DATA BLOCK HAS BEEN TERMINATED. and no output will be produced when the input OUGV1 matrix contains more than 1 type of grid output request (e.g. displacements AND velocities). Avoidance: The easiest avoidance is to only include one type of output request in the case control (e.g. DISP or VELO or ACCE). A possible DMAP avoidance would involve breaking the UG matrix into 3 separate matrices (disp, vel, acc) before running through SDR2 to make 3 separate OUGV1s. @ 36133; 70.7; Optimization/Design Sensitivity - Rigid Element Sensitivity - UF M 1126 in DSVGP5 If rigid element sensitivity is requested in SOL 200 for static and frequency response analysis, then UFM 1126 is issued from module DSVGP5 in subDMAP PSLGDV for data block ENX3. Avoidance: Insert the following DMAP alter: compile pslgdv alter 1 file enx3=ovrwrt/kmmb=ovrwrt/kmntb=ovrwrt/kmnb=ovrwrt/knnb=ovrwrt/ gtkmn=ovrwrt/ym=ovrwrt/egxx=ovrwrt/egx3=ovrwrt/mmmb=ovrwrt/ mmntb=ovrwrt/mmnb=ovrwrt/mnnb=ovrwrt/gtmmn=ovrwrt/gtmmny=ovrwrt/ gtmyg=ovrwrt/gtkmny=ovrwrt/gtbmny=ovrwrt/gtbyg=ovrwrt/gtkyg=ovrwrt/ gtkyg1=ovrwrt/gtk4yg=ovrwrt/gtk4mny=ovrwrt/egx4=ovrwrt $ endalter @ 36136; 70.7; Optimization/Design Sensitivity - PBEAML, PBARL - Incorrect Sens

itivities The sensitivities of the stress recovery locations with respect to the section dimensions are in error for the CHAN and Z section types. This error will result in incorrect stress response sensitivities. Avoidance: There is no avoidance. The magnitude of the error has been under 10% in the examples that have been seen. Thus, the designs produced using the incorrect sensitivities should be adequate. @ 36155; 70.7; Design Sensitivity - DSAMODES, SENSITY - Poor Performance The eigenvector sensitivity capability of the dsamodes module requires the identification of the row of the eigenvector which contains the largest value. Currently this is done by a do loop in the subdamp which checks the magnitude of each term in the vector in turn. This can be very time consuming if the size of the eigenvector is large. Avoidance: Insert the following DMAP alter: compile dsamodes alter 'do while ( mode'(,-1) PARAML USET//'USET'//////'f'/S,N,NOfSET $ matgen ,/row1/4/nofset/1 $ row matrix of 1 2 3 4 .... trnsp row1/rowt $ column matrix of 1 2 3 4 .... alter 'ir\=0','ir\=0'(,5) matmod phj,,,,,/phjn/2////1. $ filter out values less than one add phjn,rowt/phjx///1 $ element-by-element multiply norm phjx/junk///s,n,irr $ find row number ir=int(irr) $ endalter @ 36160; ; User's Modifiable - blank common block - Possible Wrong Answers The blank common block name for the User's Modifiable has been changed for Version 70.5. If the old name common / / is used, the results may be unpredictable. Avoidance: Replace the following common block common / with common /BLNKCM/ @ 36215; 2001; Documentation/QRG - PCOMP - Failure Index Output In the MSC/NASTRAN V70.5 QUICK REFERENCE GUIDE, Remark 7 under the PCOMP bulk data description needs to be changed: from: 7. ELFORCE and ELSTRESS requests must be present for all elements for which ply stress or failure index output is desired. to: 7. In order to get failure index output the following must be present: /

a) ELSTRESS or ELSTRAIN case control commands, b) Sb, FT, and SOUTi on the PCOMP bulk data entry, c) Xt, Xc, Yt, Yc, and S on all referenced MAT8 bulk data entries if stress allowables are used, or Xt, Xc, Yt, Yc, S, and STRN=1.0 if strain allowables are used. @ 36230; 2001.0.1; Aeroelastic Analysis - SPLINEi, SET1, SET2 - Unexpected Answer s If an input file contains both a SET1 and a SET2 Bulk Data entry with the same identification number, one or the other willl randomly be used in the spline(s) that reference that SETi identification number. There is no way to control which entry will be used and no error message is issued that indicates that one of the two inputs is being ignored. Avoidance: There is no real avoidance. You must ensure that your SET1 and SET2 entries do not have common identification numbers. @ 36245; 2001; Documentation - DEFUSET - UFM 316 When DEFUSET References PA or PS Sets If a DEFUSET entry references the PA or PS sets the following fatal message will be encounterd: *** USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA ENTRY DEFUSET ******** SORTED ENTRY COUNT = xx @ 36247; 2001; Documentation - LOADSET - Wrong Answers Multiple loadset entries may not be specified in a dynamics run. If there are loadsets in each subcase the results for the subsequent subcases use the initial loadset and are usually zero. However if the user defines the same assumed DAREA id on each of his LSEQ entries wrong results will occur for the subsequent subcase. The load is calculated from using the initial loadset. A warning or fatal message should be issued if the loadset is ignored. This problem has been fixed in the code in V2001. @ 36250; ; Static Analysis - SUBCOM, SUBSEQ, SDR2 - SFM 3002 If a static analysis has SUBCOMs and SUBSEQ, the job may issue SYSTEM FATAL MESSAGE 3002 from the SDR2 module.

@ 36251; tion

70.7; Beam Library - PBEAML, PBARL - Incorrect Properties for "T1" Sec PBEAML and PBARL produce incorrect I2 and J properteis for the "T1" type section. Avoidance: If possible use the "T" type section.

@ 36257;

70.7; Nonlinear Analysis - UWM 4698 (DCMPD) Nonlinear analysis may identify degrees of freedom large factor diagonal

ratios at every iteration with the UWM 4698 shown below. This may result in a large amount of output. Version 70.7 has been enhanced so that this output may be controlled by setting SYSTEM(60)=-999. USER WARNING MESSAGE 4698 (DCMPD) STATISTICS FOR DECOMPOSITION OF MATRIX KLLRH . THE FOLLOWING DEGREES OF FREEDOM HAVE FACTOR DIAGONAL RATIOS GREATER THAN 1.00000E+07 OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL. @ 36286; 70.7; Loads - RFORCE - Wrong Answers In Version 69 and later systems, RFORCE will produce incorrect moment loads. Moments are produced when CONM2 entries have an offset specified. Moments are also produced for the mass associated with a PBEAM entry when an offset is specified and for the PBEAM entry when a channel cross section is specified. . Avoidance: The only avoidance is to specify an additional GRID at the offset location and use a rigid element to attach it to the main GRID. @ 36287; DOM12 70.7; Optimization/Design Sensitivity - MODTRAK, scr=no - SFM 3001 in When running SOL 200 with mode tracking (MODTRAK) and scr=no, the job may terminate with the following diagnostic message in the f06 file: *** SYSTEM FATAL MESSAGE 3001 (DOM12) THE EXECUTIVE FILE -112 DOES NOT EXIST. PROGRAMMER INFORMATION: SUBROUTINE DOM11C Avoidance: Run with scr=yes or insert the following DMAP alter: compile fea alter 'if ( modetrak'(2) IF ( MODETRAK>0 ) THEN $ delete /mtrak,,,, $ modtrl mtraki//nomod/nomtk $ copy mtraki/mtrak $ endIF $ MODETRAK>0 alter 'delete /mtrak'(,-1),'delete endalter @ 36299;

/mtrak'(,3)

70.7; Miscellaneous - OFP - Inefficient Processing Efficiency has been improved in the OFP module for MSC/NASTRAN Version 70.7 for jobs that request only plot output.

@ 36350;

70.7; FMS - INIT MASTER(NORAM) - SFM 1026 In Version 70.6, if the fms statement INIT MASTER(NORAM) is used on UNIX machines, then the job may terminate with the following error messages: *** SYSTEM FATAL MESSAGE 1026 (XIO) ATTEMPT TO ACCESS A FILE WHICH IS NOT OPENED 0********* DUMP OF XIO PARAMETERS 0XIO NAME = 423

ATTEMPTED OPERATION = FILPSX FILEXX = 23 Avoidance: Remove INIT MASTER(NORAM). @ 36366; ; SSSALTERs - mtacca.vxxx - Wrong Answers SSSALTER mtacca.vxxx assumes that the modes are mass-normalized. If your modes are not mass-normalized, wrong answers will result. Avoidance: Use the default (MASS) eigenvector normalization. @ 36389; 70.7; Miscellaneous - Case Control, MODES - Poor Performance If a very large value of x is requested with the modes=x case control command, then the performance will be very inefficient. An example of this situation is illutrated below: Subcase 1 modes = 20 disp = all Subcase 21 modes = 9000 disp = none Avoidance: Request a reasonable modes value. @ 36391; 2001; Documentation - Superelements/CLOAD - UFM 5423 In the MSC/NASTRAN V70.5 Quick Reference Guide, remark 1 of the CLOAD case control command should read: "1. This command may only appear in the residual structure subcases (see SUPER command) and it must be specified in all of them." @ 36410; es ; Static Analysis - TEMP(INIT) - Wrong Answers for Multiple Subcas In Sol 101, if TEMP(INIT) is not defined above the subcase level, it is ignored and incorrect results occur. Contrary to the Quick Reference Guide documentation TEMP(INIT) is the recommended way for specifying the datum (reference) temperature for static analysis with temperature loads. However, TEMP(INIT) must be made above the subcase level. @ 36428; 70.7; Optimization/Design Sensitivity - DOPR1 - UWM 7602 The following message is correctly issued but is confusing because it states the analyis value which in this case is not greater than the maximum value allowed. The message should state the design value. *** USER WARNING MESSAGE 7602 (DOPR1I) ON PROPERTY ENTRY PBEAM WITH PID= 7610 AND FID= -9, THE PROPERTY VALUE 9.93100E+06 IS GREATER THAN THE MAXIMUM PMAX VALUE 1.00000E+07 SPECIFIED ON THE DVPREL2 761001 ENTRY. @ 36450; 70.7; P-Elements - GMBC, Multiple Edge Boundary Conditions - SFM 6363

If there are multiple edge GMBC's for different coordinate directions using different methodon a p-element, the following error results: *** SYSTEM FATAL MESSAGE 6363 (EVLEDG) THE ENFORCED DISPLACEMENT VECTOR COLUMN DOES NOT EXIST FOR EDGE ID = -21 FOR COMPONENT = 3 USER ACTION: CONTACT MSC CLIENT SUPPORT. PROGRAMMER ACTION: CHECK ENFORCED DISPLACEMENT VECTOR ROUTINE (IERROR=2). Avoidance: A workaround is to use the same method for the multiple boundary conditions. @ 36451; ; Superelements - MATMOD option 16 - Wrong Answers When using a determinate interface with an external superelement, using MATMOD option 16 may lose rigid-body modes due to numeric roundoff. Avoidance: When creating external superelements, do not use a determinate interface for the component models. @ 36463; 190 70.7; Optimization/Design Sensitivity - DVPREL1, PBUSH, PSHELL - UFM 7 When an optimization task includes DVPRELi entries refrencing PBUSH as well as other property entries, it fails in module DOPR1 with the following error message: *** USER FATAL MESSAGE 7190 (DOPR1I) PTYPE PBUSH DEFINED ON A DVPREL1 2 ENTRY CAN NOT BE FOUND ON A PROPERTY ENTRY. USER ACTION: DELETE THE DVPREL1 ENTRY OR INCLUDE AN ADDITIONAL PROPERTY ENTRY. Avoidance: Remove DVPRELi entries which do not reference PBUSH entries. @ 36465; 70.7; Aeroelastic Analysis - Wrong Answers for QHHL in tpl ha145i The tpl deck ha145i fails in Version 70.6 due to an interaction between the incomplete Case Control used in this test case and the changes made to NASTRAN to support multiple aerodynamic symmetries. Under those changes, the AERO/AEROS symmetries (which are single valued for a run) are "replaced" by symmetry in Case Control. However, for upward compatibility, the AERO/AEROS symmetries are used in the absence of any Case Control. A subtle requirement of this upward compatiblity logic is that the Cases must be "aerodynamic" cases for the procedure to function. Unfortunately, HA145I doesn't have a complete Case Control, so the fact that the run is aeroelastic cannot be determined from Case. Consequently, the defaulting rules (whereby the AERO/AEROS would have provided the symmetry) were not applied to this case. Instead, the value from Case Control (-2) that represents the fact that the symmetry is NOT supplied in CASE is mistakenly returned to the DMAP. AMG (which should have failed on illegal symmetry option) responds in a random way to this illegal value. Consequently, the deck gives spurious results. To correct the problem, an FMETHOD entry is needed in the Case Control of

the HA145I test deck. The corresponding FLUTTER entry is not required in the Bulk Data because the test case terminates before needing it. @ 36528; 70.7; Optimization/Design Sensitivity - DEQATN - Infinite CPU Loop If a DEQATN contains a double precision constant (e.g. 10.0D0) in a design optimization task, an infinite loop will be encountered and the job will enter the PREDOM subdmap and never complete. Version 70.7 outputs USER FATAL MESSAGE 224 to correct this problem. Avoidance: Remove the double precision constant from the DEQATN. @ 36530; 70.7; TRANS - MAT5, MAT9, MAT10 - forrtl: error (65): floating invlaid The trans program may fail to convert a binary XDB file to a neutral file when MAT5, MAT9, or MAT10 entries are present in the MSC/NASTRAN input file. This error has only been duplicated on DEC ALPHAs running UNIX thus far. The error message from trans is: forrtl: error (65): floating invlaid. Avoidance: Set PARAM,GEOMU,0 which eliminates all of the undeformed geometry or edit the PHASE0 subdmap DBC statement to eliminate the writing of the MPT table.

@ 36546;

; Nonlinear Buckling Analysis - QUAD4 - Spurious Modes For a rare case of boundary conditions, spurious modes in nonlinear buckling (Sol 106) may occur for coarse meshes of CQUAD4 elements. The rare case of boundary conditions is described as follows. All rotational degrees of freedom in a CQUAD4 are constrained. The transverse displacements are not constrained or only partially constrained.

Avoidance: Refine the mesh. @ 36547; 70.7; Superelements - Limit of the number of superelements - SFM 4276, EC 1044 Models that contain more than 100 superelements may fail with SYSTEM FATAL MESSAGE 4276, ERROR CODE 1044 caused by an integer overflow in routine GETOFF. @ 36563; 70.7; Miscellaneous - SGI, OS level 6.5 - Receiving signal number 10: Bus error On the SGI when running OS level 6.5, if the scratch disk is full, then the job may fail with the following unclear messages in the log file: Receiving signal number 10: Bus error Job aborting.. Avoidance: Allocate more space with the scratch disk.. @ 36564; ; DBC - Restart of an Existing XDB If a cold start run is made but an XDB exists from a prior run. then DBC

may not correctly correlate the number of modes and duplicate results may be stored. @ 36569; 70.7; Miscellaneous - SUBCOM,SUBSEQ - Missing Error Messages, Unexpect ed Answers The number of coefficients for SUBSEQ are limited to 200, which is documented in the QRG. However, if more than 200 coefficients are used--in the case of combining more than 200 subcases--only the first 200 coefficients are used without any warning or fatal messages. Avoidance: If more than 200 coefficients are needed, create a separate subcase @ 36589; 70.7; Performance - MCE1, MPC, RBE - Poor Performance When there are a large number of m-set dofs, the mce1 module may exhibit poor performance. In cases seen to date, the decomp portion of mce1 finishes in a reasonable time, but the next portion, which happens to be an fbs with many right-hand sides, ran for hours without completion. This condition can be detected by looking at the run time for the decomp operation. If the fbs takes more than 10 times this value, the avoidance below may be worth considering. Use the mce1.v706 subdmap from MSC's web site for sssalters. It has been verified that it works correctly for Version 70.5 as well. On the model in question, it reduced mce1 times from over 15 hours to minutes. @ 36611; 2001; Optimization/Design Sensitivity - BNDGRID, DVGRID, MODTRK - UFM 1126 in GP4 If mode tracking is requested with BNDGRID and DVGRID entries present, then the run will fail with the message: *** USER DMAP USET0 , USER FATAL MESSAGE 1126 (GNFIST) MODULE GP4 IS ATTEMPTING TO OUTPUT DATABLOCK 203 NAME = WHICH ALREADY EXISTS. ACTION: DELETE THE DATABLOCK, OR USE FILE USET0=OVRWRT

Avoidance: Insert the following alter: compile desopt alter 'lcgen '(,-3),'' if ( ambs or gmbs ) then $ endalter @ 36617; 70.7.2; Matrix - SOLVE - Possible Wrong Answers with Low Memory If insufficient memory is available for the SOLVE module with SPARSE, then the results may be incorrect. Avoidance: Increase memory. @ 36628;

70.7; Heat Transfer Analysis - RADSET - Unnecessary Calculation of Vie

w Factors According to the MSC/NASTRAN Quick Reference Guide, the RADSET entry "Specifies which radiation cavities are to be included for radiation enclosure analysis." However, in a multiple radiation enclosure model, if an enclosure is NOT referenced on the RADSET entry, the view factors are still calculated. Avoidance: There is no avoidance. @ 36630; dbc 70.7; DBC - Changing tstep or freq between subcases causes failure in If tstep or freq is changed between subcases and param,post,0 is specified to create a dbc database, then dbc will not complete and the following messages will be issued to the f06 file: *** SYSTEM WARNING MESSAGE 6065 (DBC) *** DBCMSG ERROR MESSAGE 44 FROM SUBROUTINE DBCAOL OUTPUT LIST DOES NOT MATCH EXISTING IN THE DATABASE, MERGE/RESTART OPERATION IMPOSSIBLE. *** SYSTEM WARNING MESSAGE 6066 (DBC) *** DBCMSG ERROR MESSAGE 34 FROM SUBROUTINE DBC THE DBC MODULE HAS BEEN TERMINATED BECAUSE OF THE ABOVE FATAL DBC ERROR. USER INFORMATION: SUBSEQUENT DBC CALLS WILL NOT BE EXECUTED, AND THE DATABASE CONVERSION IS INCOMPLETE. Avoidance: The only avoidance is to execute the subcases in different runs. @ 36632; 70.7; Grid Point Stress - PARAM,CURV - Floating point overflow If PARAM,CURV,1 is specified then the run may fail in the CURV module with the following traceback: Trap encountered: SIGTRAP - Trace trap Traceback: Offset 0x00000240 in Offset 0x00001bd4 in Offset 0x00000198 in Offset 0x000000e4 in Offset 0x00000984 in FP overflow procedure procedure procedure procedure procedure pstr2s_ curv3_ curv_ xsem14_ mainal

Avoidance: The input that demonstrated the error contained a STRESS=set id. The error was avoided when this request was changed to STRESS=ALL. @ 36636; NRD) 70.7; Heat Transfer - VIEW, RADSET - UWM 6427, UIM 4534, UWM 6137 (UDS If an initial run is used to generate a radiation matrix punch file and a subsequent run inputs the punch file and defines a RADSET entry that attempts re-calculate the view factor for some but not all of the cavities, then the job will fail with the following messages: *** USER VIEW BECAUSE A USER WARNING MESSAGE 6427 (VIEWD) FACTORS WILL NOT BE COMPUTED FOR RADIATION CAVITY ID= 85 RADIATION MATRIX IS ALREADY DEFINED. INFORMATION: THE RADIATION MATRIX IS DEFINED ON RADLST OR

RADMTX BULK DATA ENTRIES. *** USER INFORMATION MESSAGE 4534 (RMG2) 2 ELEMENTS HAVE A TOTAL VIEW FACTOR (FA/A) LESS THAN 0.99, ENERGY MAY BE LOST TO SPACE. *** USER WARNING MESSAGE 6137 (UDSNRD) INPUT MATRIX IS RANK DEFICIENT, RANK = 10 IERROR USER ACTION: CHECK MODEL *** USER WARNING MESSAGE 5225 (DCMP) ATTEMPT TO OPERATE ON THE SINGULAR MATRIX KLLRH IN SUBROUTINE DCMP. 0FATAL ERROR @ 36645; 70.7; Rigid Elements - RBE3, GP4 - UFM 6490 When an extremely large number of RBE3 elements is used, the job may fail indicating an RBE3 input error: *** USER FATAL MESSAGE 6490 (WRGMTD) FATAL ERRORS HAVE OCCURRED WHILE PROCESSING MPC AND RIGID ELEMENT ENTRIES. PROGRAMMER INFORMATION: SIL NUMBER ******** DOES NOT EXIST IN THE DEPENDENT SIL TABLE. Other symptoms are that run times may RBE3 elements increases. This effect 10,000 RBE3 elements, and becomes the where there are hundreds of thousands increase greatly if the number of begins to be noticeable at about dominant source of cost in runs of RBE3 elements.

The RBE3 code was revised in V70.7 to provide increased efficiency. The original code was made with the assumption that no model would have over 100 RBE3 elements. More modern general software tools for reading long records repeatedly, introduced in V70.7, were used to improve efficiency. Several other enhancements were made at the same time, as described in TAN 4155. The net effect is that a model that runs for 15 hours on a Cray computer in V70.5 now runs in several hundred seconds in V70.7. TANs can be found on the MSC home page by going to the search engine and putting in the appropriate words. At present the TAN number can not be used in the search. Two words that will find this TAN are RBE3 and V70.7. @ 36665; 70.7; Aeroelastic Analysis - PFAERO - Infinite Loop PFAERO If the user attempts to run an aeroelastic solution sequence (SOL 144, 145 or 146 or SOL 200 with aeroelastic analyses) without any aerodynamic data, an infinite loop will occur in the PFAERO subdmap. This is caused by a failure to trap the absence of the requisite aerodynamic data. Subsequently, the PFAERO subdmap (responding to data reuse logic) will endlessly attempt to create the aerodynamic model. Avoidance: You must have at least one aeroelasticity-related bulk data entry to avoid the infinite loop and trigger the "missing data" error traps. @ 36678; 70.7.2; Eigenvalue Analysis - EIGR, EIGRL - SFM 7340 The Lanczos method of eigenvalue extraction may give nonstandard error messages. For example a run may terminate with the following messages. *** SYSTEM FATAL MESSAGE 7340 (LNNHERR) process error reported by subroutine LNNP2CD see LNNP2CD abstract (ier = -725)

*** SYSTEM FATAL MESSAGE 7340 (LNNHERR) warning reported by subroutine LNNDRVD see LNNDRVD abstract (ier = 725) Avoidance: The LNNDRVD or LNNDRVS messages meanings are given below: ier error condition. -723 insufficient workspace revealed during lanczos iteration -724 failure of tridiagonal ql iteration to converge -725 number of eigenvalues computed exceeds internal storage limitations -726 i/o error with lanczos vector file -727 failure of bidiagonal ql iteration to converge -728 convergence failure in gram-schmidt orthogonalization -- probable cause - indefinite m (vibration) or indefinite k (buckling) -729 repeated singular matrix -- probably ill-posed problem -730 number of trust regions too large -731 factorization restore i/o error -732 ran out of computer resources. -733 inconsistent treatment of number of eigenvalues in finite interval -734 could not compute zero eigenvalues for a buckling case @ 36682; 70.7; Miscellaneous - Requesting Huge mem may yield Wrong Answers Reguesting more than 1200mb of memory on an RS6000 system may result in wrong answers or a fatal message.

@ 36685;

70.7; Inertia Relief - INREL, SUPORT1 - SFM 3007 in UPARTN If there are multiple subcases with SUPORT1 command in some, but not all, subcases and PARAM,INREL,-1 is specified in the Bulk Data then the run will terminate with: *** SYSTEM FATAL MESSAGE 3007 (UPARTN) ILLEGAL INPUT TO SUBROUTINE PARTN Avoidance: Insert the following DMAP alter: compile statrs alter 'if ( norset','' if ( norset<6 and norset>0 ) then $ endalter

@ 36701;

2001; Dynamic Analysis - DAREA - Unexpected Answers If on the DAREA entry the coordinate system is specified as "0" for a grid point it will be interpreted as "1". For example: $ DAREA pi 410 3 0 1 ci

is interpeted by the program as DAREA 410 3 1 1

This may cause unexpected answers. Avoidance: Specify the desired degree of freedom.

@ 36706;

; Old Design Senitivity - CBEAM - Possible Wrong Answers If BEAM elements are used in a modal analysis with the old design sensitivity analysis, then the design sensitivity coefficients may be incorrect. Avoidance: 1. Replace the beam elements with bar elements. 2. Use SOL 200 (v707.7) to perform the same task if the user insists using beam elements.

@ 36709; ; Optimization/Design Sensitivity - DSVG1D - UFM 5172 When Materia l has no Density If a designed property (DESVAR/DVPREL1) references a material with zero density, then the job will fail with the following messages: *** USER FATAL MESSAGE 5172 (DSVG1D) THE VARIATIONAL ELEMENT ID IN MDICTDSF INCONSISTENT WITH THE VARIATIONAL ELEMENT ID IN DSPT1, MODULE DSVG1 IS TERMINATED. *** SYSTEM FATAL MESSAGE 3001 (DSVG1) THE INPUT DATA BLOCK NAMED MDICTDSF AT POSITION 1 DOES NOT EXIST. Avoidance: Specify a non-zero density on the material entry. @ 36727; 70.7; Dynamic Analysis w/Differential Stiffness - SUPORTi, STATSUB, sc r=yes - UFM 4407 If differential stiffness is requested in dynamic analyis, supported degrees-of-freedom are specified (see the SUPORT or SUPORT1 entries), and the job is run with scr=yes, then the mass matrix will not be reduced to the supported degrees-of-freedom and UFM 4407 will be issued. Avoidance: Run with scr=no or insert the following DMAP alter: compile phase1dr alter 'or kdiff','' if ( true ) then $ endalter @ 36730; 70.7; Superelements - SEBULK, SECONCT - UFM 6794 When using part superelement, only a single boundary point is allowed at a particular boundary location (within a specified tolerance), this is design intent. However, it has been observed in a model that the job terminates with the following messages: *** USER FATAL MESSAGE 6794 (SEPSEB) DUPLICATE BOUNDARY POINTS DETECTED IN UPSTREAM SUPERELEMENT WHILE SEARCHING 0 AT GRID 7001. LIST FOLLOWS: 13403 49682 1

even though the points identified do not appear to be at the same location within the tolerance. The exact cause of this error is not known at this point. Avoidacne: Use the SEEXCLD command to remove all but a single boundary point at each PART.

@ 36737; 4574

70.7; Optimization/Design Sensitivity - PARAM SOFTEXIT YES, SDR1 - UFM If PARAM,SOFTEXIT,YES is specified with static or frequency response analysis in SOL 200, then the run will fail with UFM 4574 and SFM 3001 in module SDR1 and subDMAP SEDISP on the final data recovery. Avoidance: Insert the following DMAP alter: compile desopt alter 'dbdelete','' endalter

@ 36741;

70.7; Response Spectrum Analysis - RSPEC - SFE 4276 EC 777 If a response spectrum analysis is performed in Version 70.5 on HP computers, then the job may fail with the following message and traceback in the RSPEC module: ( ( ( ( ( ( Process traceback 0) 0x00093e90 1) 0x01126d78 2) 0x004935e8 3) 0x00199024 4) 0x00098d44 5) 0x0008fbc8 19.5 ... sysrcv + 0x2a0 [/nast2/msc705t2/hpux/analysis] _sigreturn [/nast2/msc705t2/hpux/analysis] rspcod_ + 0x740 [/nast2/msc705t2/hpux/analysis] rspec_ + 0xbc [/nast2/msc705t2/hpux/analysis] xsem14_ + 0x2fc [/nast2/msc705t2/hpux/analysis] mainal_ + 0x9b0 [/nast2/msc705t2/hpux/analysis]

real

Avoidance: The only avoidances are to run Version 70 or generate the curves on another computer. @ 36752;

70.7; Miscellaneous - MSCACT - UFM (mscact) Unknown keyword The MSCACT utility incorrectly flags the keyword assignment "summary=jid" as an error and outputs the message shown below. However, "summary=job" is accepted and generates the requested functionality. USER FATAL MESSAGE (mscact) unknown keyword "summary=jid". Avoidance: 1) Use "summary=job" instead of "summary=jid". 2) Alternatively, the error can be corrected by changing line 1180 of <ins-dir>/msc705/util/mscact.c from else if( !strncmp( "job", q, lq ) ) to else if( !strncmp( "jid", q, lq ) ) Once the change is made, rebuild MSCACT as described in the "Configuration and Operations Guide".

@ 36764; 70.7; Acoustic Analysis - FORCE(PHASE), Velocities and Pressure levels - Wrong Answers The acoustic output titled "COMPLEX ACCELERATIONS VELOCITIES AND PRESSURE LEVELS" prints an incorrect format for the element ID when the MAGNITUDE/PHASE option is requested. The incorrect format is that the "PRESSURE(DB)" column is truncated and the phase values are misaligned. Here is an example of the incorrect MAGNITUDE/PHASE output: ( P E A K ) C O M P L E X A C C E L E R A T I O N S V E L O C I T I E S A N D P R E S S U R E L E V E L S (MAGNITUDE/PHASE) ELE-ID EL-TYPE X-ACCELERATION Y-ACCELERATION Z-ACCELERATION X-VELOCITY Y-VELOCITY Z-VELOCITY PRESSURE (DB) 1.4994E-43 HEXPR 1.364242E-11 -2.376055E-10 4.083463E+04 .0 .0 .0 5.298811E+0 .0 .0 .0 .0000 .0000 -1624.7582 Here is the output for real/imaginary: E S A N D C O M P L E X A C C E L E R A T I O N S P R E S S U R E L E V E L S V E L O C I T I

(REAL/IMAGINARY) ELE-ID EL-TYPE X-ACCELERATION Y-ACCELERATION Z-ACCELERATION X-VELOCITY Y-VELOCITY Z-VELOCITY PRESSURE (DB) 107 HEXPR 1.364242E-11 -2.376055E-10 4.083463E+04 .0 .0 .0 5.298811E+01 .0 .0 .0 -5.428147E-13 9.454022E-12 -1.624758E+03 The error only affects the printed output and not the formats of OUTPUT2 and DBC modules. Avoidance: Request REAL/IMAGINARY output. @ 36776; DOM9D) 70.7; Optimization/Design Sensitivity - PBAL, DLINK, DOM9 - UFM 6499 ( An optimization job which designs PBARL dimensions with DLINK entries fails in DOM9 with the following message: THE GRADIENT OF CONSTRAINT NUMBER NNNNN IS ZERO CONSTRAINT IS REMOVED FROM THE ACTIVE SET ** THIS IS A VIOLATED CONSTRAINT ** A FATAL ERROR WILL BE GENERATED **** ONE OR MORE VIOLATED CONSTRAINTS HAS A ZERO GRADIENT **** NO PROGRESS IS POSSIBLE OPTIMIZATION TERMINATED *** USER FATAL MESSAGE 6499 (DOM9D) THE DOT OPTIMIZATION ALGORITHM HAS ENCOUNTERED AN ERROR THAT MAKES FURTHER PROGRESS IMPOSSIBLE. SEE MESSAGES ABOVE. Avoidance: There is no avoidance. @ 36777; 70.7.2; Optimization/Design Sensitivity - ANALYISIS=MODES and MFREQ - SF

M 3001 in DSAH If a model has fluid-structure interaction in SOL 200 and both ANALYSIS=MODES and ANALYSIS=MFREQ are requested, the run may fail with SFM 3001, stating that data block PHG does not exist. Avoidance: The following alter will avoid the FATAL error, but has not been confirmed to give correct answers. compile fea alter 'IF.*DOSUPER3 AND NOT.*FSCOUP','' IF ( DOSUPER3 ) THEN $ @ 36780; 70.7; Submittal - vers= - Does not get right version If a different version than the latest installed (or latest known by the invoked nastran command if under the form nastxxx where xxx is a version number) is requested, the rc files of the requested version are picked up, but not the executable and SSS. @ 36790; ; Heat Tranfer Analysis - VIEW, UDSFA - UFM 6136, estimate=mem MEM=estimate did not yield the correct estimate for a large heat trasnfer radiation problem. The job subsequenty Failed with the message: *** USER FATAL MESSAGE 6136 (UDSFA) INSUFFICIENT CORE FOR NUMERIC PHASE OF SPARSE DECOMPOSITION. USER ACTION: INCREASE CORE BY 5217107 WORDS 0FATAL ERROR @ 36792; 70.7; Aerodynamic Analysis - PARAM,OPGTKG,1 - Missing output PARAM,OPGTKG,1 requests a printout of the aerodynamic splining matrices. The capability was inadvertently deleted from Version 70.6. Avoidance: Insert the following DMAP alter: compile phase1a alter 'np\=seid','' equivx uset0/aeusets/-1 $ equivx bgpdts/aebgpdts/-1 $ dbdict datablk=(aebgpdts,aeusets) select=(name,size, modltype,aeconfig,symxy,symxz) qualcurr $ dbview bgpdtf=aebgpdts where (modltype='aerostrc' or modltype='structur' ) $ dbview usetff=aeusets where (modltype='aerostrc' or modltype='structur' ) $ IF ( OPGTKG>=0 ) then $ MATGPR bgpdtf,usetff,,GPJK//'K'/'G' $ MATGPR bgpdtf,usetff,,GDJK//'K'/'G' $ endIF $ OPGTKG>=0 endalter @ 36815; 2001; Composites - TEMPP3 - Wrong Answers The TEMPP3 thermal loading bulk data entry is obsolete and is no longer supported. The TEMPP3 entry may give wrong answers for composites with

inplane or normal thermal gradients. Avoidance: Use the TEMPP1 bulk data entry instead which works correctly in all cases and includes the same features as the TEMPP3. @ 36821; tivities ; Optimization/Design Sensitivity - RHO - Incorrect Material Sensi

The design sensitivities with respect to the material densities of transient response analysis that has a steady state response may be incorrect. Avoidance: Do not specify the material density as a response for a steady state analysis since the response is not affected by the density. @ 36823; 2001; Nonlinear Transient Analysis - OTIME, SDR1 - BIOMSG: ERROR in BLDPK In nonlinear transient analyis, if the value of an output time on an OTIME= request exceeds the time value of the final time step, the job will fail in the SDR1 module with the message: = BIOMSG: ERROR 0 STATUS = 211 HAS OCCURRED IN ROUTINE BLDPK 0 , FILE INDEX (FILEX) 211

the same job runs with fewer time step intervals for output (NO) on the TSTEPNL card or with fewer OTIME requests in the case control. Avoidance: Do not use the OTIME entry in nonlinear transient analysis. The Quick Reference Guide does not indicate that OTIME is supported for nonlinear transient solutions. @ 36836; ; FMS - INIT - UFM 1257 If a large value is assigned to a DBSET member, e.g., init scratch logical(scratch(50gb)) then the job may fail with the following unclear messages: *** USER FATAL MESSAGE 1257 (DBMAP) THE MAX-SIZE SPECIFIED ON THE INIT FMS STATEMENT OR SMEM VALUE IN THE RC FILE OR SMEM KEYWORD IN THE COMMAND PROCEDURE FOR THE MEMFILE OF DBSET SCRATCH IS 100 BLOCKS AND IS LESS THAN THE RECOMMENDED MINIMUM OF 110 BLOCKS. USER ACTION: SPECIFY A MAX-SIZE OR SMEM VALUE AT OR ABOVE THE RECOMMENDED VALUE. Avoidance: Either increase the buffsize or decrease the dbset value @ 36842; ; Superelements - PARAM,SESEF - UFM 5423 If the fraction of total strain energy (PARAM,SESEF) is requested and external superelements exist, the job will fails with USER FATAL MESSAGE 5423.

Avoidance: Include the following alter: compile postreig, list alter 'MPYAD.*KAA,PHILS','' type parm,,i,n,temp dbview kaanew=kaa, where(seid=temp and wildcard) temp=seid $ paraml kaanew//'trailer'/1/s,n,noaset $ MPYAD KAAnew,PHILS,/FA $ @ 36870; AD) 2001; Nonlinear Static Analyis - MPCFORCE, SE, INTOUT - SFM 5423 (DMPY If mpcforce output is requested in a nonlinear analysis, and the model contains superelements and intermediate output is requested, then the job will fail with the following messages: *** USER INFORMATION MESSAGE 7321 (SEDRDR) DATA RECOVERY FOR SUPERELEMENT 20 IS NOW INITIATED. *** SYSTEM FATAL MESSAGE 5423 (DMPYAD) ATTEMPT TO MULTIPLY INCOMPATIBLE MATRICES COLS ROWS FORM TYPE NZWD DENS TRAILER FOR KMG IS 101108522 1310 2 2 12 3 TRAILER FOR UGJ IS 102 1108522 2 2179608 8275 TRAILER FOR PM IS 103 5 1310 2 2 0 0 TRANSPOSE FLAG IS 0 0FATAL ERROR Avoidance: Insert the following DMAP alter: compile sedisp alter 'PARAML UGI'(,-1) if ( app='nlst' ) then $ equivx ugi/ugj/-1 $ else $ alter 'PARAML UGI'(,2) endif $ endalter or do not request intermediate output. If intermediate output is important, break each subcase into multiple equivalent subcases--one for each load increment. @ 36930; 2001; Miscellaneous - ECHOON, ECHOOFF - Doesn't Always Work Correctly ECHOON and ECHOOFF entries may be used in the Executive, Case Control, or Bulk Data but there are limitations. ECHOOFF must not be the first entry in the Executive Control section else it will be ignored. ECHOOFF does not handle continuation entries (for example SET commands) in the Case Control correctly. If after and ECHOOF command has been placed in the Bulk Data, PARAM entries are the only entries that follow the ECHOON command, they will not printed in the unsorted output. @ 36943;

70.7; Aeroelastic Analyisis - GUST - Missing Output for Gust Analysis

In SOL 146, the gust loads are computed inside the GUST module and then transformed to generalized coordinates. Only the generalized loads are output. This makes it impossible for clients to manipulate these loads in any meaningful way. Avoidance: There is no simple avoidance. @ 36949; 70.7; DMAP - GPFDR - Extra Output The GPFDR module with the third parameter xflag=1 (the cross strain energy output) gives extra output for the first subcase. There should be no results output for the first subcase. @ 36978; Kept 70.6; Optimization/Aeroelasticity - DMI, W2GJ - Downwash Matrices not In SOL 200 with static aeroelasticity, user input downwash matrices (W2GJ) are not kept after the first cycle. Avoidance: Apply the following dmap alter in V70 and V70.5: COMPILE PFAERO ALTER 'dmiin' CALL DBSTORE W2GJ,,,,//0/1/'DBALL'/0 $ $ COMPILE FEA ALTER 'fea' $ TYPE PARM,,I,Y,IFOUND $ ALTER 'pfaero' IF ( DESCYCLE >1 ) CALL DBFETCH /W2GJ,,,,/0/1/0/0/S,IFOUND $ $ @ 36997; 64 70.7; Optimization/Design Sensitivity - Premature Termination - UIM 64 An optimization job starting from a feasible region may terminate for a hard convergence after the first design cycle without showing any improvement in either objective or constraint and showing any changes in the design variables. Here is the termination message: *** USER INFORMATION MESSAGE 6464 (DOM12E) RUN TERMINATED DUE TO HARD CONVERGENCE TO AN OPTIMUM AT CYCLE NUMBER = 1. Avoidance: None. @ 36998; 70.7; SSSALTER - nlgyroa.v706 - Gyroscopic Effects Ignored The version 70.6 nlgyroa.v706 sssalter does not include the gyroscopic effect. Avoidance: Replace nlgyroa.v706 with the following alter. A copy of the revised alter is also available on the MSC web site. $

$ $ $ $ $ $

THIS ALTER IS CONFIDENTIAL AND A TRADE SECRET OF THE MACNEAL-SCHWENDLER CORP. THE RECEIPT OR POSSESSION OF THIS ALTER DOES NOT CONVEY ANY RIGHTS TO REPRODUCE OR DISCLOSE ITS CONTENTS, OR TO MANUFACTURE, USE, OR SELL ANYTHING HEREIN, IN WHOLE OR IN PART, WITHOUT THE SPECIFIC WRITTEN CONSENT OF THE MACNEAL-SCHWENDLER CORPORATION.

$ echooff $ $ $ nlgyroa.v706 - updated 8-13-99 $ $ updated July 7, 1998 - tlr $ $ added PARAM,RFORCE,i - set i to the RFORCE to be used by this alter $ the default is the first RFORCE in the sorted $ bulk data $ $ nlgyroa.v69 - updated for v69 -- 9-25-96 $ $ tested on system built on 3-14-95 VERSION 68.2 $ $************************************************************************ ******$ $ ALTER FOR SOL 106 TO INCLUDE CENTRIPETAL TERMS. $ $ THE ALTER WILL ADD CENTRIPETAL TERMS TO THE TANGENT STIFFNESS MATRIX $ DURING MODES CALCULATION SPECIFIED BY THE PARAMETER NMLOOP. TO USE THIS $ ALTER THE USER MUST SUPPLY A RFORCE FOR THE RESIDUAL STRUCTURE. $ SUPERELEMENTS SHOULD NOT BE USED, THE CENTRIPETAL EFFECT OF UPSTREAM $ SUPERELEMENTS WILL NOT BE CALCULATED. $ THE COORDINATE SYSTEM REFERENCED ON THE RFORCE CARD MUST BE CENTERED ON $ THE AXIS OF ROTATION. THE GRID ID ENTRY IS IGNORED FOR THE CALCULATION $ OF THE CENTRIPETAL TERMS. THE ROTATION IS ASSUMED TO BE ABOUT THE $ REFERENCED COORDINATE SYSTEM (CID) ORIGIN, THE MAGNITUDE (A) AND ROTATION $ VECTOR COMPONENTS (N1, N2, & N3) DETERMINE THE ORIENTATION. $ $ THE PARAMETER NMLOOP SHOULD BE SET FOR THE END OF THE FIRST SUBCASE. $ $ ALTER DEVELOPED BY DAVID BELLA FOR THE MACNEAL-SCHWENDLER CORPORATION, $ AUGUST, 1992. $ $************************************************************************ ******$ $ compile nlstatic $ alter 'NORMAL MODES WITH NONLINEAR STIFFNESS'(,1) $message //'testing xxxxxxxxxxxxxxx' $ $ $ CALCULATE CENTRIPETAL TERMS $ CALL CENTRIP SLT,BGPDTS,EQEXINS,CSTMS,SILS,MGG, geom3/ KCJJ/ $ $ dbstatus kcjj//s,n,iskcjj $

if (iskcjj>0)then $ message //' ' $ message //' message from NLGYRO alter - ' $ message //' centripetal stiffness terms calculated' $ message //' adding into stiffness matrix for eigenvalue'/ ' solution' $ message //' ' $ $ matprn kcjj// $ else $ message //'dmap warning message' $ message //'no centripetal stiffness terms calculated' $ message //' ' $ endif $ $ CALL SEMR2 kcjj,GOT,USETnl,GM/kcAA,kcFF,kcMG,kcSF/ ERROR/NOMSET/NOSSET/NOOSET/NORC/NOQSET/NOTSET $ matprn kcjj,kcaa// $ $ $ ADD CENTRIPETAL TERMS TO STIFFNESS $ alter 'call.*gma.*casesx2.*maa'(,-1) add5 kaa,kcaa,,,/kaanew $ equivx kaanew/kaa/always $ matprn kaa// $ add5 ksf,kcsf,,,/ksfnew $ equivx ksfnew/ksf/always $ add5 kmg,kcmg,,,/kmgnew $ equivx kmgnew/kmg/always $ $ alter 'method=-1' add5 kllrh,kcaa,,,/klltemp $ equivx klltemp/kllrh/-1 $ $ $ ENDALTER $ $ $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ $$$$ $ COMPILE CENTRIP $ $ SUBDMAP CENTRIP SLT,BGPDT,EQEXIN,CSTM,SIL,MJJ,geom3/KCEN/ $ $ $******************************************************************** $ * $ This DMAP sequence calculates Coriolis and Centripetal effect * $ matrices in global coordinates. * $ * $ For node i: * $ * $ [Bcoriolis] = 2[x]m * $ * $ [Kcentripetal] = [x][x]m * $ * $ 0 -Omega(z) Omega(y) * $ [x] = Omega(z) 0 -Omega(x) * $ -Omega(y) Omega(x) 0 * $ * $******************************************************************** $

TYPE PARM,,I,N,SLTRECNO,geom3rec,wordno,sid,id $ TYPE PARM,,RS,N,OMEGA $ TYPE PARM,,CS,N,N1C,N2C,N3C $ TYPE PARM,,CS,N,OMEGAC $ type parm,,i,y,rforce=-1 $ $ SEARCH FOR RFORCE ENTRY $ $tabprt geom3//'geom3'///1 $ $tabprt slt//'slt'///1 $ dbstatus slt//s,n,noslt $ if(noslt<>1)then $ message //'FATAL - FATAL - FATAL' $ message //'slt does not exist in alter - error' $ message //'run terminated by nlgyro alter' $ exit $ endif $ $ if(rforce<0)then $ $ $ take first RFORCE in SLT $ message //' ' $ message //' PARAM,RFORCE not specified, first RFORCE entry '/ 'will be used for nlgyro alter' $ message //' ' $ SLTRECNO= 1 DO WHILE ( SLTRECNO > 0 ) $ $ PARAML SLT//'DTI'/S,N,SLTRECNO/1//S,N,LTYPE/ $ IF ( (SLTRECNO <> -1 ) AND ( LTYPE = 10 ) ) THEN $ $ $ EXTRACT SPIN VECTOR FROM RFORCE CARD. $ PARAML SLT//'DTI'/SLTRECNO/4//S,N,CID/ $ PARAML SLT//'DTI'/SLTRECNO/5/S,N,RPT/ $ PARAML SLT//'DTI'/SLTRECNO/6/S,N,N1/ $ PARAML SLT//'DTI'/SLTRECNO/7/S,N,N2/ $ PARAML SLT//'DTI'/SLTRECNO/8/S,N,N3/ $ SLTRECNO = -2 $ $ ENDIF $ $ SLTRECNO = SLTRECNO + 1 $ $ ENDDO $ $ else $ $ $ param,rforce specified $ message //' ' $ message //' param,rforce specified, searching for'/ ' rforce number'/rforce $ message //' ' $ $ geom3rec=1 $ do while (geom3rec>0) $ $ message //'top of loop - geom3rec='/geom3rec $ PARAML geom3//'DTI'/s,n,geom3rec/1//S,N,ID/ $

$ message //'id ='/id/' geom3rec ='/geom3rec $ if(id=5509)then $ $ found rforce entries wordno = 4 $ do while(wordno>0) $ find rforce requested PARAML geom3//'DTI'/geom3rec/s,n,wordno//s,n,sid $ $ message //'wordno='/wordno/' sid ='/sid $ if(sid=rforce)then $ $ found correct rforce message //' ' $ message //'found requested rforce entry number'/sid $ message //' ' $ wordno = wordno + 2 $ paraml geom3//'DTI'/geom3rec/s,n,wordno//s,n,cid $ message //'cid ='/cid $ wordno = wordno + 1 $ paraml geom3//'DTI'/geom3rec/s,n,wordno/s,n,rpt $ message //'rpt ='/rpt $ wordno = wordno + 1 $ paraml geom3//'DTI'/geom3rec/s,n,wordno/s,n,n1 $ wordno = wordno + 1 $ paraml geom3//'DTI'/geom3rec/s,n,wordno/s,n,n2 $ wordno = wordno + 1 $ paraml geom3//'DTI'/geom3rec/s,n,wordno/s,n,n3 $ message //'n1,n2,n3 ='/n1/n2/n3 $ wordno = -2 $ else $ if(wordno>0)wordno = wordno + 10 $ endif $ enddo $ wordno>0 endif $ id=5509 if(geom3rec>0)geom3rec=geom3rec + 1 $ enddo $ endif $ IF ( RPT = 0. ) THEN $ $ OMEGA= 0. RETURN $ $ ENDIF $ $ $ THESE ARE NEEDED LATER ON $ MATGEN ,/I1X1/1/1/ $ MATGEN ,/I6X6/1/6/ $ $ $ CALCULATE RADIANS PER UNIT TIME $ OMEGA= 2.0 * PI(1) * RPT $ OMEGAC = CMPLX(OMEGA,0.) $ $ $ $ CREATE A COLUMN VECTOR, MJJTSUM, FROM THE DIAGONAL OF MJJD $ TO BE USED FOR PARTITIONING PURPOSES LATER $ DIAGONAL MJJ/MJJC/'COLUMN' $ $ $-------------------------------------------------------------------$ $-------------------------------------------------------------------$ $

$ CALCULATE ROTATION MATRIX IN THE CID SYSTEM $ IF (N1 <> 0.) THEN $ $ DETERMINE FACTOR FOR X-ROTATION (in CID system) $ MATGEN ,/PVC6/6/6/1/1/4 $ MATGEN ,/PVR6/6/6/2/1/3 $ MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $ MPYAD ONE6X6,I6X6,ONE6X6/ROTX/1/-1 $ DELETE /PVC6,PVR6,ONE6X6,,/ $ N1C = CMPLX(N1,0.) $ ENDIF $ $ $-------------------------------------------------------------------$ $ IF (N2 <> 0.) THEN $ $ DETERMINE FACTOR FOR Y-ROTATION (in CID system) $ MATGEN ,/PVC6/6/6/2/1/3 $ MATGEN ,/PVR6/6/6/0/1/5 $ MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $ MPYAD ONE6X6,I6X6,ONE6X6/ROTY/1/-1 $ DELETE /PVC6,PVR6,ONE6X6,,/ $ N2C = CMPLX(N2,0.) $ ENDIF $ $ $-------------------------------------------------------------------$ $ IF (N3 <> 0.) THEN $ $ DETERMINE FACTOR FOR Z-ROTATION (in CID system) $ MATGEN ,/PVC6/6/6/0/1/5 $ MATGEN ,/PVR6/6/6/1/1/4 $ MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $ MPYAD ONE6X6,I6X6,ONE6X6/ROTZ/1/-1 $ DELETE /PVC6,PVR6,ONE6X6,,/ $ N3C = CMPLX(N3,0.) $ ENDIF $ $ $-------------------------------------------------------------------$ $ $ ADD FOR TOTAL ROTATION (in CID system) $ ADD5 ROTX,ROTY,ROTZ,,/ROTTOT/N1C/N2C/N3C $ $ $-------------------------------------------------------------------$ $-------------------------------------------------------------------$ $ $ GENERATE G-SIZE CORIOLIS MATRIX IN THE GLOBAL SYSTEM $ $ $ FIRST, GENERATE G-SIZE MATRIX IN THE CID SYSTEM $ ADD ROTTOT,/OMEGA/OMEGAC $

MATMOD OMEGA,SIL,,,,/OMEGAGG,/5 $ $ $-------------------------------------------------------------------$ $ $ USE THE MJJC VECTOR TO PARTITION OUT THE ZERO MASS TERMS $ ALONG WITH THE SPOINTS (WHICH CAUSE PROBLEMS FOR MATMOD OPTION 5) $ PARTN OMEGAGG,MJJC,/,,,OMEGAXX/ $ $ $-------------------------------------------------------------------$ $ $ TRANSFORM CORIOLIS ACCELERATION MATRIX TO GLOBAL SYSTEM $ FIRST, TRANSFORM FROM CID TO BASIC $ THEN , TRANSFORM FROM BASIC TO GLOBAL $ $ CHECK IF THE CSTM EXISTS. IT WILL NOT EXIT IF THE GLOBAL SYSTEM $ IS THE SAME AS THE BASIC SYSTEM. ERROR IN MATMOD. THEREFORE $ GENERATE THE IDENTITY MATRIX $ $ PARAML CSTM//'PRESENCE'////S,N,NOGTOB=0 $ IF (NOGTOB < 0) THEN DIAGONAL MJJ/GTOBY/'SQUARE'/0. $ ELSE MATMOD CSTM,SIL,BGPDT,,,/GTOBY,/5//-1 $ ENDIF $ $ PARTN GTOBY,MJJC,/,,,GTOB/ $ $ $ CHECK IF THE CID IS EQUAL TO ZERO (BASIC) $ IF (CID = 0) THEN EQUIVX GTOB/CTOG/ALWAYS $ ELSE MATMOD CSTM,SIL,BGPDT,,,/CTOBX,/5//CID $ $ $ USE THE MJJC VECTOR TO PARTITION OUT THE ZERO MASS TERMS $ ALONG WITH THE SPOINTS (WHICH CAUSE PROBLEMS FOR MATMOD OPTION 5) $ PARTN CTOBX,MJJC,/,,,CTOB/ $ MPYAD CTOB,GTOB,/CTOG/1 $ ENDIF $ $ SMPYAD CTOG,OMEGAXX,CTOG,,,/OMEGAYY/3////1 $ MERGE ,,,,OMEGAYY,MJJC,/OMEGAGGX/ $ $ $-------------------------------------------------------------------$ $-------------------------------------------------------------------$ $ $ CALCULATE CENTRIPETAL ACCELERATION MATRIX $ $ MULTIPLY MASS BY CENTRIPETAL ACCELERATION $ TO CALCULATE CETRIPETAL FORCE MATRIX $ SMPYAD OMEGAGGX,MJJ,OMEGAGGX,,,/KCEN/3////////6 $ $ RETURN $ END $ $ $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

$ echooff $ @ 37010; ; Optimization/Design Sensitivity - DRESP1 - SFM 4276, EC 31 When an optimization deck includes a DRESP2 entry which references the same DRESP1 ID more than once, the following system fatal message is issued from module DOM9: = BIOMSG: ERROR 0 STATUS = followed by: 0*** SYSTEM FATAL ERROR 4276.GETSTR ERROR CODE 31 PID= 0 31 HAS OCCURRED IN ROUTINE GETSTR , FILE INDEX (FILEX) 0

Avoidance: Do not duplicate the same DRESP1 ID on a DRESP2 entry. @ 37027; 2001; Aeroelastic Analysis - RESTART - SFM 3001 for QHHA or Inefficien t Restart A read-only restart in aeroelastic solution 145 will unnecessarily recompute and decompose the AJJT for a list of existing Mach numbers and reduced frequency even though the list is a subset of the original list. Another symptom is SFM 3001 for QHHA. Avoidance: If the original MKAERO1 entry is re-entered as it was input on the initial run, then the AMG and DECOMP modules in the PFAERO subDMAP will not be re-run. @ 37031;

70.7; Aeroelastic Analysis - AESYMXZ, AESYMXY, AECONFIG, IFP - UFM 603 Case control commands are free field input and should allow users to add comments at the end of the lines so long as a dollar sign precedes the comment. The new Version 70.6 aeroelastic analysis case control commands do not allow comments to be specified on the same line as the command. The simple input file below fails in IFP due to the comments onthe AE* case control commands. sol 144 cend aesymxz=symmetric $ this is a comment on the aesymxy line aesymxy=asymmetric $ this is a comment on the aesymxy line aeconfig=csr $ this is a comment on the aeconfig line spc = 1 $ this is a comment on the SPC line begin bulk param,checkout,yes enddata

@ 37032;

2001; Eigenvalue Analysis - PARAM,RESVEC - Possible Wrong Answers When residual vector is requested, it has been observed on some models that the additional augmented eigenvalue is lower than the highest calculated mode. This has been observed on a very small problem and also on problems using virtual mass.

Avoidance: Insert the following DMAP alter: compile resvec alter 'matmod oxf','' type parm,,rs,y,resvoxf=.001 $ matmod oxf,,,,,/ox5,/2////resvoxf $ filter again alter 'append=','' append=(noox5<0 or ox5r<10000) $ set flag for below endalter or inspect the augmented eigenvalue table, if the additional mode is lower than the highest calculated frequency, use one of the following avoidances: 1. Specify F2 on the EIGRL entry. 2. Specify a smaller maxset. 3. If it is a small problem, use AGIV or AHOU @ 37042; 70.7; Matrix Methods - MPYAD - SFM 4276 (GINO) If working memory is small relative to problem size, data is stored past the end of working memory, destroying executive tables. The result will ultimately be a confusing error message, such as: *** SYSTEM FATAL MESSAGE 4276 (GINO) ERROR CODE 100 PID= 0 *** USER INFORMATION MESSAGE 4276 (GINO) TO OBTAIN A NASTRAN DUMP RESUBMIT JOB WITH DIAG 44 INSERTED IN THE EXECUTIVE CONTROL SECTION. 1 * * * END OF JOB * * *

*** SYSTEM FATAL MESSAGE 1123 (GDBSNO) DATABASE NUMBER: -10855138 IS OUT OF RANGE. DBSPTR = -1085513983 USER ACTION : MAKE SURE ALL ATTACHED DATABASE(S), INCLUDING THE DELIVERY DATABASE, ARE VERSION (e.g. 69) COMPATIBLE. 0FATAL ERROR Avoidance: Increase memory. @ 37043; 70.7; Rigid Elements - RBE3, GP4 - Poor Efficiency When there are more than 50,000 rbe3 elements run times in the GP4 module, where they are processed, may be very long. Avoidance: Avoidances are to reduce the number of rbe3 elements, or put them in several superelements. Each superelement will call GP4 for only the elements within it. @ 37051; 70.7; Aeroelastic Analysis - TRIM, ASG - SFM 4276, EC 100 The message "SYSTEM FATAL ERROR 4276 ERROR CODE 100" is issued from the ASG module in an aeroelastic trim analysis when there are no SUPORTed degrees of freedom but there are aerodynamic extra points (AESTAT and/or AESURF entries) that are not constrained on the TRIM entry. Avoidance: Constrain or remove the aerodynamic extra points or add to the SUPORT degrees of freedom. @ 37055; 70.7; Eigenvalue Analysis - Component Modes - Excessive resource usage

If component modes are requested and the number of q-set points is not exactly equal to the number of component modes then the reduced stiffness and mass matrices and the resulting residual structure matrices will contain many small off-diagonal terms which should be exactly zero. This condition contributes to inefficiencies in subsequent matrix operations. Avoidance: Insert the following DMAP alter: compile semr3 alter 'dotrip\=','' alter 'dotrip\='(2),'' alter 'nullkqq' nullkqq=0 $ alter 'cmmat' if ( noqset1>noqset ) then $ partn cmmat,,vq1qx/xxx,,,/1 $ equivx xxx/cmmat/-1 $ else if ( noqset1<noqset ) then $ merge cmmat,,,,,vq1qx/xxx/1 $ equivx xxx/cmmat/-1 $ endif $ endalter @ 37059; 70.7; Miscellaneous - NEC SX-5, HPIO - Job Hangs It has been reported that when HPIO is turned on the NEC SX-5 and the disk space is full, then nastran will not terminate properly and the job will hang. Avoidance: Kill the job manually. @ 37078; 2001; Random Response - CBAR, PLOAD1 - SWM 3002, SFM 4276, missing PSD output If PSD element force or stress output for a BAR element is requested in a random analysis, and the loading involves PLOAD1 entries, then the requested random output will be missing. Depending on the machine, the job will issue one of the following messages: SWM 3002 "EOF ENCOUNTERED WHILE READING DATA BLOCK OES2 (FILE 107) IN SUBROUTINE RAND2" SFM 4276 in the SDRX module. Avoidance: There are two possible avoidances: !. Replace the BAR elements with BEAM elements. 2. If the BAR element must be used, replace PLOAD1 with equivalent point loads. @ 37080; 2001; Optimization/Desgin Sensitivity - PBAR - Wrong Answers When an optimization job includes CBAR elements which are loaded by PLOAD1 entries, the end-B maximum stress, end-B minimum stress and end-B Safety margin in compression for the static case are incorrectly selected from the stress data file. Avoidance: Use item code + 1 for selecting end-B maximum stress and end-B

minimum stress. However, there is no avoidance for getting the correct end-B safety margin. @ 37088; 70.7.2; External Superelements - param,extdrout,dmigop2 - Wrong answers When using the enhanced external superelement capability, if the boundary points have a different order in the assembly model than in the component model, then the boundary solution written out for PARAM,EXTDROUT,DMIGOP2 will be in the incorrect order, resulting in incorrect results for data recovery on the component. Avoidance: Insert the following dmap alter. compile super3 alter 'call.*extdrout','call.*extdrout'(3,) CALL EXTDROUT ULS,OL1M,EQEXINS,MAPS,uset/ DUM/APP1/' '/TRUE $ CALL EXTDROUT UH2,OL1R,EQEXINS,MAPS,uset/ DUM/APP/SOLTYP/TRUE $ ELSE $ CALL EXTDROUT ULS,OL1M,EQEXINS,MAPS,uset/ DUM/APP1/SOLTYP/TRUE $ alter 'call.*extdrout'(4,),'call.*extdrout'(6,) CALL EXTDROUT UG,OL1M,EQEXINS,MAPS,uset/ DUM/APP1/' '/TRUE $ CALL EXTDROUT UH2,OL1R,EQEXINS,MAPS,uset/ DUM/APP/SOLTYP/TRUE $ ELSE $ CALL EXTDROUT UG,OL1M,EQEXINS,MAPS,uset/ DUM/APP/SOLTYP/TRUE $ compile extdrout, list alter 1,1 SUBDMAP EXTDROUT UG,OL,EQEXINS,MAPS,uset/ DUM/APP/SOLTYP/EXTER $ alter 'matmod','' paraml uset//'uset'//////'g'/s,n,lgset/'a'/s,n,noaset $ matgen eqex/intext/9/0/lgset $ $if(lgset<>noaset)then $ $ upartn uset,intext/intexta,,,/'G'/'A'/'O' $ $else $ equivx intext/intexta/-1 $ $endif $ copy eqex/eqexinxxx $ copy x_ul/x_ulc $ call ted intexta,x_ulc,eqexinxxx/dummy $ $ alter 76,76 OUTPUT2 dummy,X_OL//0/EXTDRUNT///'MATPOOL'/XNAME2 $ tabprt dummy//'matpool'///1 $ tabprt x_ol//xname2///1 $ $ compile ted subdmap ted intexta,x_ul,eqex/dumout mpyad intexta,x_ul,/extdb/1 $ matmod extdb,eqex,,,,/EXTDBx,/16/0 $ X_MATUL copy extdbx/dumout $

return end $ @ 37089; GL; Aeroelastic Analysis/Superelements - SPLINEi - AECOMP Datablock Does not Exist If external superelements are used in an aeroelastic solution and all splines are contained within the external superelements, the run will fail claiming that datablock AECOMP does not exist. @ 37102; 2001; Superelements - Buckling, SPCD - Wrong Answers If buckling (SOL 105) is run with superelements and there are enforced displacement on the upstream superelements, then the static answers will be incorrect and hence the eigenvalues will be wrong. Avoidance: Insert the following DMAP alter: compile phase0 alter 'ISOLAPP\=NSOLAPP','' noop() $ alter 'ISOLAPP\=ISOLAPP'(,-1) IF ( SEID>0 ) ISOLAPP=NSOLAPP $ endalter or place all enforced displacements on the boundary into the residual superelement. @ 37106; 70.7; P-Elements - TEMPP1 - QUADp Gives Wrong Answers for TEMPP1 Loads The QUAD p-element gives wrong results if it is loaded with temperature gradients (TEMPP1 entry). Avoidance: There is no avoidance. @ 37126;

2001; Documentation - XYPLOT - LONG Command Default The MSC/NASTRAN V70.5 Quick Reference Guide case control description in the XYPLOT section for the LONG command says that the default is YES. This is not correct. The default is NO.

@ 37128; 70.7.2; External Superelements - PARAM,EXTDROUT,DMIGOP2 - SFM 3007 (UPAR TN) If the original component model for an external superelement has one or more superelements defined, and the residual structure is equal in size to the ASET chosen for the reduced matrices, then the data recovery run on the component will fail with the following messages if the PARAM,EXTDROUT,DMIGOP2 option is used. *** SYSTEM FATAL MESSAGE 3007 (UPARTN) ILLEGAL INPUT TO SUBROUTINE CALCV Avoidance: Insert the following alters.

compile extdrin alter 'upartn.*uset.*ul','' paraml uset//'uset'//////'g'/s,n,lgset/'a'/s,n,noaset $ if(lgset<>noaset)then $ UPARTN USET,EXTDB/UL,,,/'G'/'A'/'O'/1 $ else $ equivx extdb/ul/-1 $ endif $ @ 37130; 01 2001; Optimization/Design Sensitivity - ANALYSIS=FLUTTER, SE - SFM 30 In SOL 200, if ANALYSIS=FLUTTER and superelements exist, the analysis results will be wrong if the aerodynamic model is splined to an upstream superelement. Further,the job may fail with SFM 3001 in DVSG1 indicating that datablock DUG does not exist. See also CSR 38501 Avoidance: Run the model without using superelements, or use this alter for V70.7. compile respsen alter 'dsflte','dsflte' DSFLTE KHHx,BHHx,MHHx,QHHL,FLUTAB,r1tabr,casem/ CPHFL,CPHFR,VTQU $ alter 'dug0','vtqu0'(2) IF ( NODUD>0 ) CALL xSEDRDR, DUD,EMAP,CASEXX,SLT,ETT,MAPS,bgpdts,USET, GOAT,GOAQ,GM,CASEs,casem/ DUG/ RSONLY/'DYNAMICS'/false $ IF ( NODVD>0 ) CALL xSEDRDR, DVD,EMAP,CASEXX,SLT,ETT,MAPS,bgpdts,USET, GOAT,GOAQ,GM,CASEs,casem/ DVG/ RSONLY/'DYNAMICS'/false $ alter 'dugx'(3),'vtqux'(3) alter 'doflutd'(6) IF ( NOT(IMAGE) AND NOPSLG>-1 ) THEN $ alter 'cdelb1'(5) endif $ not(image) and nopslg>-1 alter 'dsfltf','dsfltf' if ( seid = 0 ) then $ dsfltf vtqu,cdelk,cdelb,cdelm/delfl $ else $ dsfltf vtqu,cdelk1,cdelb1,cdelm1/delfl $ endif $ compile phase1a alter 'ISOLKNT=NSOLAPP+1','ISOLKNT=NSOLAPP+1' if (not desopt) isolknt=nsolapp+1 $

@ 37132;

70.7; SSSALTERs - sedampa doesn't supposrt r-set and c-set If R- or C-set dof exist in a superelement and the sssalter sedampa is used, the associated superelement will not have modal damping applied. The alter is intended to terminate the run with a message stating that C- and

R-set dof are not supported by the alter. Avoidance: Do not use R- or C-set dof with the alter. The alter has been replaced by a standard feature, which does allow R- or C-set dof to exist in Version 70.7. To correct the alter, replace paraml uset//'uset'////s,n,nousetxx//'R'/s,n,norset/ 'C'/s,n,nocset $ with paraml uset0//'uset'////s,n,nousetxx//'R'/s,n,norset/ 'C'/s,n,nocset $ @ 37146; trices

70.7; Aeroelastic Analysis - RESTART, GUST - SFM 5423, Incompatible Ma If an aerelastic gust analysis attempts to change the number of retained modes in a restart, then "SYSTEM FATAL MESSAGE 5423" is issued indicating an "ATTEMPT TO ADD INCOMPATIBLE MATRICES (SADD5)" Avoidance: Perform a cold start.

@ 37151;

2001; Documentation/QRG - Plot Command AXES is miss typed as AXIS In case control section of Quick Reference Guide the Plot command AXES page header and format spell it as AXIS which is incorrect.

@ 37152;

70.7; Optimization/Design Sensitivity - MODTRK - SFM 3001 "SYSTEM FATAL MESSAGE 3001" occurs in the MODTRK module when all three of the following are present: 1. shape design variables, 2. at least one subcase with mode tracking 3. and at least one subcase with ANALYSIS=STATICS Avoidance: Place the MODTRAK request above the subcase level

@ 37169;

2001; Documentation/QRG - ADAPT - Typographical Error The Quick Reference Guide description of the ADAPT entry contains the typographical error: TY=PEUNIP, SIGT=OL22 it should be TYPE=UNIP, SIGTOL=22

@ 37173; 2001; Optimization/Design Sensitivity - PARAM,SOFTEXIT - Incomplete De sign History

In SOL 200, if PARAM SOFTEXIT YES is used and soft convergence is achieved, the DESIGN CYCLE HISTORY does not include the results from the final exact analysis. Avoidance: The information that should be in the DESIGN CYCLE HISTORY can be obtained using the P1 and P2 parameters on the DOPTPRM entry. @ 37178; 70.7; Miscellaneous - Missing UFM 315 - Unexpected Answers If a negative sign is the only character in a field of a bulk data entry it may be ignored. This may lead to unexpected answers. For example if the negative sign was intended for the value in the following field as in the following entry: CBAR 70101 ++000363 ++000363 +000364 51001 464 1322 0. .238 0. 1.0 .238

Avoidance: Specify "ECHO=PUNCH" in the Case Control and perform a search (grep) on negative signs followed by blanks ("- "). @ 37190;

70.7; Optimization/Design Sensitivity - PBARL - Wrong Answers Use of PBARL with "T" section in an optimization task may cause an infeasible result with Version 70.5 or earlier release of MSC.Nastran. Avoidance: 1. Use equivalent PBAR with DVPREL2 equations defining the PBAR properties. 2. Use a new version such as v70.7.

@ 37195; 70.7; Optimization/Design Sensitivities - Rigid Elements - Incorrect S ensitivities When a shape optimization job has locations of rigid elements being designed it may result in incorrect sensitivities or may have difficulty converging. This may lead to excessive time in the DSVGP5 module. Avoidance: There is no avoidance. Version 70.7 corrects this error and shows a dramatic improvement in performance of the DSVGP5 module for some problems. @ 37222; ng ; Optimization/Design Sensitivity - RESTART - Inefficient Processi Restarts into SOL 200 will not be efficient if the design model is not included in the cold start run. Avoidance: 1. Include the design model in the cold start run or 2. Insert the following alter in the cold start run: compile prtsum alter 1

DBEQUIV WHERE (DESITER=0 and WILDCARD) , CONVERT (DESITER=1), OVRWRT $ endalter @ 37224; essage 2001; Optimization/Design Sensitivity - PCOMP, DSAD - SFM 4276, Poor M In solution 200, when using the cfailure criteria force item code 7 (interlaminar stress), NASTRAN will either fail in the DSAD module with system fatal 4276 or give wrong answers if the code is specified for the last lamina (ply). This fatal occurs because there are no interlaminar stresses for the last lamina since interlaminar stresses are between laminae. Also, using the cfailure criteria as force item code 8 (max of FP and FB) will fail unless the lamina (ply) number specified on the DRESP1 is the final one. @ 37227; 70.7.2; Nonlinear Analysis - Creep Law May Cause Overflow - SFM 4276 in nmcrd Some forms of creep law may cause an overflow condition in nonlinear creep analysis. The error causes an overflow in the nmcrd subroutine which will be identified in the log file. Avoidance: A different creep law may avoid the error. @ 37247; ; Transient Analysis - TEMP(LOAD) - Wrong Answer In transient analysis, thermal strains are not subtracted from the total strains before computing the state of stress. This leads to wrong answers. Avoidance: This is a limitation, there is no avoidance. @ 37253; 2001; Cyclic Symmetry - Virtual Mass, MFLUID - Wrong Answers Virtual mass will cause wrong results when used with cyclic symmetry if a fluid segment is modeled. This path should be blocked with a fatal error. Fully enclosed fluids in each segment will give correct answers. Avoidance: Use acoustic fluid elements instead. @ 37263;

; Virtual Mass - MFLUID - PARAM,VMOPT,1 Does Not Function Properly When param, vmopt, 1 is present, the virtual mass should be added to the structural mass at the g-set, that is, before any constraints are applied. In Version 70.6 and 70.7, setting this parameter has no effect and the solution method adds virtual mass after all constraints are applied. This may or may not affect the results Avoidance: The alter below will make vmopt, 1 take effect. compile semg list noref $ Alter 'if ( vmopt=1 and','' $ always do what user requests IF ( VMOPT=1 ) THEN $ Compile semrm list noref $

Alter 'if ( vmopt=0 or','' $ IF ( VMOPT=0 ) ) THEN $ Note that the original dmap would over-rule the user's request if it did not seem to "make sense". The alter always honors the vmopt request. Note that if vmopt=1 is used on medium- and large-size models, run times and computer resource requirement may be very large. sssalter vma.f707 is scheduled to be available on 15 Oct 99. It allows a more accurate answer when vmopt=0 (the default value) is used. @ 37264; 70.7.2; Performance - MPYAD - MPYAD Method 1, Storage 2 not selected Under certain condition, the MPYAD Method 1, Storage 2 method, which is optimized for dense matrices, will not be correctly selected. This happens when all of the B and D matrices just fit into working memory. Avoidance: The work around is to either increase or decrease memory. @ 37272; ; Elements - CHEXA - Stress Output at Gauss Points does not work w ith IN=2 Element Stresses at the Gauss integration points cannot be produced for the 8-point CHEXA element when the integration option, IN=2 is used. Instead, the stresses at the corner points are generated. The description in the QRG for the PSOLID bulk data entry states that if the stress flag is "1" or "GAUSS" the integration points will be the location of the stress output, this is incorrect. Avoidance: There is no avoidance for this integration option. @ 37276; rge ; Nonlinear Transient Heat Transfer - TEMPC, TRAN - Fails to Conve If an nonlinear transient heat transfer analysis has enforced temperature boundary conditions, that is temperature boundary condition that are a function of time, MSC/NASTRAN uses the large conductance approach to enforce the temperature boundary condition. This is achieved by using a CELAS2 entry with stiffness equal to 1.0E10, and an SLOAD entry on the node with a temperature of 1.0E10. Thus, the temperature is equal to u = F/K or SLOAD divide by the K value. In some case this hard coded value for the spring stiffness may be to large and it may adversely affect convergence. Avoidance: 1. Replace the TEMPBC,TRAN with SLOAD and a CELAS2 entry, and set the stiffness smaller. 2. Use convergence on P only on the iterations where convergence difficulties are encountered. @ 37284; 70.7; Utilities - MSGCMP.EXE, Windows NT - Missing Utility The msgcmp.exe executable is missing from the Version 70.5 NT release of MSC/Nastran. Avoidance: Download msgcmp.exe from the MSC ftp site or contact your local support person. The file is located in the /pub/outing directory. Since this is an executable, use the BIN option for downloading. @

37290;

2001; Documentation/QRG - PARAM,POST - Patran may use the xdb The Quick Reference Guide documentation of parameter POST,0 does not mention the possibility to use the xdb database with MSC.Patran. This should be added and it should be clarified that the parameters MODEL and SOLID are not supported by MSC.Patran.

@ 37291;

70.7; Optimization/Design Sensitivity - GE - Wrong Sensitivities In a dynamic optimization job, if the design variables include damping coefficients and the adjoint method path is invoked, then the dynamic response sensitivity coefficients with respect to all design variables will be incorrect. This error has been corrected in Version 70.7. Avoidance: Turn off the adjoint method using the following dmap alter compile descon alter 'dsad'(1,-1) pexist = true alter 'dsad' pexist = false

@ 37295;

2001; Documentation/QRG - MATHP - Poor Convergence The following sentence need to be added to the Quick Reference Guide description of the TAB1 parameter of the MATHP entry: Stresses are negative for compression and positive for tension. If this convention is not followed the job may fail to converge or issue a traceback.

@ 37307;

; Nonlinear Analysis - TEMP(MAT), HEXA - Wrong Equivalent Stress The "EQUIVALENT STRESS" for the nonlinear HEXA element is wrong if there are temperature dependent material properties. Avoidance: There is not avoidance.

@ 37309;

70.7; DMAP - APPEND - UFM 4276 EC 1039 The APPEND module may fail and issue the message "USER FATAL MESSAGE 4276.OPENDV EC=1039" Avoidance: There is no avoidance. The error is fixed in Version 70.7.

@ 37319; 70.7.2; Optimization/Design Sensitivity - DESGLB - Unexpected Answers DESGLB is intended for the selection of design constraints at the global level (e.g., weight, volume). If DESGLB is used above the subcase level with the intent to apply the constraints to all the subcases, the constraints will be ignored for the second and subsequent subcases. Avoidance: Use of the DESSUB above the first subcase applies it to all subsequent subcases until a new DESSUB request is entered. DESGLB is

reserved for subcase independent response only. @ 37335; 70.7.2; Restart - NMLOOP, LOOPID - Inefficient Restart, Modes Recalculat ed In a restart from SOL 106 to SOL 111, the stiffened modes are being recalculated unnecessarily. Avoidance: Insert the following dmap avoidance. compile semfreq $ alter 'call modefsrs'(,-1) $ equivx lama/lamas/-1 $ equivx pha/phsa/-1 $ if ( false ) then $ alter 'call modefsrs' $ endif $ endalter @ 37336; DEPEND=-1 ; SSSALTERs - SEGYRO, CYGYR0 Alters - Wrong Results with Parameter

The SEGYRO alter puts the gyroscopic terms in the damping matrix and not in the mass matrix for param, depend,-1 and this leads to wrong answers. The assumption eigenvalue lambda = i*Omega is not valid in a rotating reference frame. The Corillis forces are equal to the mass times the cross product of the rotation vector times the velocity vector in the ROTATING SYSTEM. ie. f=m(omega X V ). The assumption that lambda is proportional to i*omega is only good for a limited set of response cases where the applied load has the same frequency as the rotations. An example would be an unbalanced turbine. The problem would diverge at the critical rotational frequency. @ 37341; 2001; Static Analysis - PARAM,CDITER and INREL - Wrong Answers If PARAM,INREL,-1 is specified along with PARAM,CDITER,n then no message is issued stating that PARAM,INREL will be ignored The Version 70.5 Release Guide and QRG indicate that PARAM,CDITER and INREL are mutually exclusive. A fatal message will be issued in Version 2001. @ 37344; 2001.0.1; Acoustic Analysis - ACSRCE, DAREA - Wrong Answers When using an ACSRCE power input during an acoustic frequency response analysis, if the scale factor on the DAREA is increased, then QDOT is erroneously factored instead of power. Avoidance: Scale the values on the TABLEDi instead. Also, the Quick Reference Guide needs to document the equation used: P(f)=A*C(f) @ 37356; 70.7.2; Nonlinear Transient Heat Transfer - DLOAD - SFM3002 (TAH1) If transient thermal analysis loads (QVOL, QHBDY) are combined with temperature dependent material properties (MATT4), then the analysis will fail with SYSTEM FATAL MESSAGE 3002 in the TAHT module.

Avoidance: There is no avoidance. @ 37382; ; Nonlinear Analyis - CBUSH1D, PARAM,LGDISP - Wrong Answers Small initial values of stiffness for the CBUSH1D element, i.e. K=1, may lead to wrong results when PARAM, LGDISP,1 is specified. The problem occurs because the compression displacement is not limited to the length of the element. Avoidance: Specify a larger value for the stiffness. @ 37392; 70.7.2; Acoustic Analysis - CELAS1, CAABSF - UFM 5423 If frequency dependent elements are present in both the fluid (CAABSF) and the structure (CELAS1) then the run will fail with UFM 5423. Avoidance: Insert the following DMAP alter: compile freqrs alter 'if ( fs )'(2),'' if ( false ) then $ alter 'if ( fs )'(3),'' if ( false ) then $ alter 'if ( fs )'(4),'' if ( false ) then $ alter 'if ( fs )'(5),'' if ( false ) then $ endalter @ 37394;

; Eigenvalue Analysis - EIGRL, EIGR - Incorrect Termination Method The Lanczos method for Real Eigenvalue analysis may not find all desired roots when ND is specified on the EIGRL bulk data entry. Lanczos may report the reason for termination as "REQUIRED NUMBER OF EIGENVALUES FOUND IN RANGE" even though the number of desired eigenvalues is not output.

Avoidance: Specify a frequency range on the eigrl entry. @ 37399; 70.7.2; DMAP - OUTPTR Module does not Process STRAIN Requests The OUTPRT module introduced in Version 70.7 does not process STRAIN requests. As a results the partitioning vector that is generated will not include these points unless they are requested with other output requests. Avoidance: An avoidance is to also request STRESS for the same output set. At present, this module is used only in sssalter spdra.v707. @ 37400; 70.7.2; P-Elements - RFORCE - Wrong Answers Wrong answers may be obtained when RFORCE is applied to p-elements where p>2. Avoidance: The only avoidance is to use the origin of the basic coordinate system to locate the rotation vector. @

37402; 70.7.2; Optimization/Design Sensitivity - BNDGRID - UFM 6499(DOM9), UFM 9050(SEKRRS) When the C field on a BNDGRID entry in a shape optimization task is left blank or takes the value of zero, this input error is not detected. Therefore, the job may fail in one of the following ways: 1. UFM 6499 (DOM9) due to a zero gradient of a violated constraint 2. UFM 9050 (SEKRRS) due to excessive pivot rations in Matrix KLL. 3. Wrong Answers Avoidance: Specify a non-zero component on the C field. @ 37407; ; Superelements - SEP1X, DTI, SETREE - SFM 4276 EC 3001 If external superelements are used in a multi-level model containing PART superelements (BEGIN SUPER), and the multi-level tree is defined using a DTI,SETREE entry, the run may fail in the SEP1X module with SYSTEM FATAL MESSAGE 4276. The fatal message will indicate that "ERROR 3001 HAS OCCURRED IN GREAD". Avoidance: Replace the DTI,SETREE entry with an SETREE entries. @ 37411; 70.7.2; Parallel Processing - DMP - FATAL, Infinite CPU Loop If MSC.Nastran is run with DMP=n (n > 1) on an NEC system, and there is a FATAL message, then the processors which do not get the FATAL hang. They accumulate CPU time, but never finish. Avoidances: Avoidances are given below; however, this and several other problems related to DMP have been corrected in Version 70.7.2. DMP users should install this later version. 1. Modify the $MSC_BASE/bin/nast707.lcl file: insert after line 294: %MSC_ARCH=superux:% errorcode=$? insert at the end of the file: %MSC_ARCH=superux:% exit $errorcode 2. Or run without DMP being specified. @ 37437; 70.7.2; Frequency Response - dmp>1 and multiple FREQ subcases - Infinite loop A job with multiple frequency subcases submitted with Distributed Memory Parallel, dmp>1, in solution 111 will hang. Avoidance: Apply the following dmap alter: compile freqrs alter 'ckdd,cbdd'(,-1) $ putsys(1,287) $ alter 'ckddf,cbddf'(,-1) $ putsys(1,287) $ compile semfreq alter 'fs and'(,2),'' jump endloop $ alter 'if ( loopflag'(,-2)

label endloop $ endalter @ 37454; 70.7.2; Parallel Processing - DMP, ITER=YES, SPCF - SFM If distributed memory parallel processing is specified with the iterative solver and spcforces are requested, the the job will fail. Avoidance: Either do not request spcforces, distributed processing, or the iterative solver. @ 37457; 70.7; Optimization/Design Sensitivity - PBARL, Smax, Smin - Possible W rong Answers In an optimization run, if PBARL is used and only Smax or Smin is specified instead of the four stress data recovery points, then wrong answers may occur. Avoidance: There are two possible avoidances: 1. Specify all four stress data recovery points instead of using Smax or Smin. or 2. Use PBAR instead of PBARL. @ 37458; 2001.0.1; DBC - DBC, PARAM,POST - Job Crashes during BLKWRT A job with thousands of subcases may fail in DBC during a BLKWRT. The following message was encountered: MSC.XL 0 0 0 0 DATA BASE CONVERSION STARTED PROJECT 0 PATHQUAL 0 PATHLINK 0 PATHINT 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

POST-PROCESSING DATA RECOVERY INITIATED SUBCASES 1 0 0 0 0 0 ILLEGAL DATA BASE NUMBER PASSED TO BLKWRT 0FATAL ERROR

Avoidance: There is no avoidance. @ 37462; 70.7.2; Parallel Processing - DMP>1, CBAR, CBEAM, SEQP - SFM 4276 If a statics job is submitted with Distributed Parallel Processing, dmp>1, and the job contains BAR or BEAM elements that specify disjoint GRID points for the orientation vector then the job may fail in the SEQP module or hang. Avoidance: There are several avoidances:

1. Do not use distributed parallel processing. 2. Modify the CBAR, CBEAM entries to explicityly define a coordinate direction. 3. Modify the CBAR, CBEAM entry to reference an attached GRID point. 4. User Version 70.7.2 where this error has been corrected. @ 37483; 70.7; Nonlinear Analysis - SFM 3007 (NLCOMB)/ILLEGAL INPUT TO SUBROUTI NE NLCXGR MSC.Nastran may fail with SYSTEM FATAL MESSAGE 3007 (NLCOMB), ILLEGAL INPUT TO SUBROUTINE NLCXGR on Sun/Solaris systems when run with MEM => 2Gb. Avoidance: Either: 1. Reduce memory below 2Gb 2. Run with version 70.7 or higher. @ 37485; 70.7.2; Optimization/Design Sensitivity - Incorrect Sensitivities for St atic Aero Incorrect sensitivities and subsequent poor optimization performance occurs on long word machines, such as the CRAY, for static aeroelastic sensitivity analysis when there are more design variables than there are AESTAT and AESURF entries. Avoidance: Run the job on a short word machine, reduce the number of design variables, or increase the number of AESTAT/AESURF entries to satisfy the above inequality. @ 37486; 70.7.2; SSSALTERs - GPSP, PARAM,AUTOSPC - be-set dofs may be autospcd The be-set is a new set introduced in Version 70.7. It is not used in any solution sequences at present. Its only use is in the rflaga.v707 alter package. The be-set should not be autospcd, even when singular. This is analogous to the suport entry. It sometimes is autospcd, although the circumstances when this happens are not well understood yet. Avoidance: An avoidance is to always used an spcoff entry for every be-set point. It may not be necessary, but it should not cause problems either. Example uset, be, 17, 3 spcoff, 17 3 @ 37487; 70.7.2; Miscellaneous - USET - UFM 2104 (GP4), Poor Message The be-set is new in Version 70.7. When a degree-of-freedom is placed in a mutually exclusive set, such as the s-set, and the be-set, the following truncated message appears: *** USER FATAL MESSAGE 2101 (GP4) SCALAR POINT 1 ILLEGALLY DEFINED IN SETS The phrase "UBE" should also appear. If you see a message of this type look for be-set entries on the same dof . These dofs are entered on USET, BE-type inputs. US

@ 37495; 70,7; Documentation - Hoffman's Theory incorrect in Composites Seminar Notes (NAS113) The Equation for Hoffman's Failure Theory in Chapter 5 of the NAS113 (Analysis of Composites Materials) semninar notes is incorrect. There is a sign error for one term. The term: (SIG1*SIG2) / (Xt*Xc) should be preceded by a negative sign in the equation. @ 37496;

2001; Frequency Response - CELAS2 , STRESS - Wrong Answers In a direct frequency response analysis, the stress output for the CELAS2 element is incorrect. Avoidance: Use the CBUSH element instead,

@ 37503; 2001.0.1; Acoustic Analysis - Misleading UWM 5462 (IFS5P) The presence of acoustic elements always leads to the following misleading warning message: *** USER WARNING MESSAGE 5462 (IFS5P) OPTIONS IN PSOLID ID = 800000 FOR FLUID ELEMENT ARE NOT ALLOWED. DEFAULT VALUES WILL BE TAKEN. Avoidance: Ignore the warning message. @ 37508; ; GPFORCE - PARAM,NOELOF - Incorrect Format on PUNCH Output If PARAM,NOELOF,1 is specified the punch output format for GPFORCE is incorrect. Avoidance: There is no avoidance for this option. The f06 output is correct. @ 37585;

707; Case Control - SET - Duplicate Output If a SET command is used to request grid point output, and the set contains a thru range referencing non-existing grid points, then duplicate output will be printed for the last existing grid point. For the following example, no grid point exists between the range of 10001 and 10010, the displacement for grid point 12 is printed twice. SET 1 = 11,12 , 10001 THRU 10010 $ DISP(PHASE,SORT2) = 1 Avoidance: Remove the reference to the non-existing grid points in the set.

@ 37590; 70.7.2; GPFORCE - GPFDR - Output does not include MPC forces in SOL 144 When Grid Point Forces (GPFORCE) are requested in SOL 144 (Static Aero),

the forces of multipoint constraint (MPCFORCE) should be included in the output, but are not. Avoidance: Add the MPCFORCE=ALL or MPCFORCE(PLOT)=ALL output request to Case Control to cause the MPC forces to be computed. Once computed, they will be included in the GPFORCE output. @ 37593; 70.7.2; Superelements - CSUPER, STATSUB, SPC - UFM 1126 (GNFIST) In a normal modes analysis with image superelements (CSUPER), if the structure is preloaded (STATSUB) and spcs are applied to the interior of the primary superelement, then the job will fail with the following unclear error messages: *** USER FATAL MESSAGE 1126 (GNFIST) DMAP MODULE MATREDU IS ATTEMPTING TO OUTPUT DATABLOCK 202 NAME = KDSF , WHICH ALREADY EXISTS. USER ACTION: DELETE THE DATABLOCK, OR USE FILE KDSF=OVRWRT Avoidance: Replace the following nddl statement PATH KDSF with PATH KDSF seid,PEID,MPC,SPC,MTEMP,NLOOP,DESITER,PVALID, PEID,MPC,SPC,MTEMP,NLOOP,DESITER,PVALID, APRCH,HIGHQUAL,K2GG,M2GG,AUXMID,STATSUB,P2G $

APRCH,HIGHQUAL,K2GG,M2GG,AUXMID,STATSUB,P2G $ @ 37597; 70.7.2; SSSALTER - delmodea.vxx, Part Superelements - Modes Not Deleted When using the delmodea.vxx alter with PART superelements, and the DMI,SKPMODE entry is placed in the main bulk data section, then selected modes will not be deleted. The following warning messages will be issued: ^^^ USER WARNING MESSAGE: MODE PARTITIONING MATRIX NOT FOUND IN BULK DATA ^^^ PLEASE CHECK INPUT DECK FOR DMI, SKPMODE ENTRIES ^^^ ^^^ THE MODE DELETION ROUTINE CANNOT CONTINUE - NORMAL PROBLEM SOLUTION CONTINUING WITH ALL MODES... Avoidance: Either request a copy of the updated alter from your local support engineer or put the DMI, SKPMODE in the last partitioned bulk data section. @ 37598; 70.7.2; Parallel Processing - DMP, Component Modes - Infinite loop If distributed parallel processing is selected by setting dmp>1 on a component modes analysis the job may hang. Avoidance: Insert the following DMAP alter: compile cmpmode alter 'xofp' if ( getsys(np,265)>1 ) then $ delete /cmlama,cmphixz,cmmi,cmoeigs,eigvmats $ disutil ,,,,/cmlama///3 $

disutil ,,,,/cmphixz///3 $ disutil ,,,,/cmmi///3 $ disutil ,,,,/cmoeigs///3 $ disutil ,,,,/eigvmats///3 $ paraml cmphixz//'trai'/1/s,n,nozset $ else if ( getsys(np,265)=1 ) then $ disutil cmlama,,,////3 $ disutil cmphixz,,,////3 $ disutil cmmi,,,////3 $ disutil cmoeigs,,,////3 $ disutil eigvmats,,,////3 $ endif $ compile resvec alter 'read ','' read kzz,mzz,,,,,,,,,,,/ ALAMA1,APHx1,Mx,ox,Ex,/ 'MODES'/S,N,NV////-1/'LAN'//1.e8//999/0////prtsum $ if ( getsys(np,265)>1 ) then $ delete /alama1,aphx1,mx,, $ disutil ,,,,/alama1///3 $ disutil ,,,,/aphx1///3 $ disutil ,,,,/mx///3 $ paraml aphx1//'trai'/1/s,n,nv $ else if ( getsys(np,265)=1 ) then $ disutil alama1,,,////3 $ disutil aphx1,,,////3 $ disutil mx,,,////3 $ endif $ endalter @ 37600; ; Nonlinear Analysis - PSHELL - Incorrect NLSTRESS Format The NLSTRESS format is wrong if both plane-strain and plane-stress shell elements exist in the same input file. That is if multiple PSHELL records exist, some with MID2>-1 and some with MID2=-1. Avoidance: This output can be turned off by requesting NLSTRESS=NONE. Another avoidance is not to put both types of shell elements into a single nonlinear run. The other results (including the stresses using the linear format) are not affected. @ 37602; 70.7.2; Static Analysis - DMP, dmp>1 and iterative solver - SFM 4276 (UN PACK) If a SOL 101 run is made with Distributed Memory Parallel and the iterative solver is selected, then the job may crash with the following message: *** SYSTEM FATAL MESSAGE 4276 (UNPACK) ERROR CODE 20 PID= 0 Avoidance: Do not use DMP. The error is corrected in Version 70.7.2 @ 37606; 70.7.2; Parallel Processing - DMP>1 and mergeresults=yes, OLOAD - Wrong Answers In static analysis with domain decomposition (a.ka., Distributed Memory Parallel, dmp>1) and mergeresults=yes on the nastran submittal line, the OLOAD output and OLOAD maximums will be wrong for the boundary points if distributed loads like PLOAD and GRAV are applied.

Avoidance: Insert the following DMAP alter: compile sedisp alter 'pj0s','pj0s'(,1) upartn usetx,pj/pax,pga,,/'g'/'a'//1 $ merge ,,pax,,,,spcpart/pag0/1 $ Loads at boundary due to local $ distributed loads DISUTIL pag0,,,/pg0///8 $ sum of all boundary dist. loads disutil pg0,,,/pg00///3 $ broadcast to slaves if ( getsys(np,265)>1 ) equivx pg00/pg0/-1 $ PARTN pg0,,SPCPART/,pay,,/1 $ get local boundary dist. loads add5 pay,pjxl,,,/pj0s $ add in boundary point loads umerge1 usetx,pj0s,pga,,/pj1/'g'/'a'//1 $ combine endalter @ 37615; 2001.0.1; Eigenvalue Analysis - EIGRL - UWM 4698 (DCMPD), SFM 7340 When Lanczos encounters a massless mechanism it will first provide the following message: *** USER WARNING MESSAGE 4698 (DCMPD) STATISTICS FOR DECOMPOSITION OF MATRIX (NONE) . THE FOLLOWING DEGREES OF FREEDOM HAVE FACTOR DIAGONAL RATIOS GREATER THAN 1.00000E+15 OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL. RATIO GRID POINT ID DEGREE OF FREEDOM MATRIX DIAGONAL 1 3.37236E+00 R2 MATRIX/FACTOR DIAGONAL

1.00000E+15

It may then issue a fatal message of the form: *** SYSTEM FATAL MESSAGE 7340 (LNNHERR) process error reported by subroutine LNNP2CD see LNNP2CD abstract (ier = -729) *** SYSTEM FATAL MESSAGE 7340 (LNNHERR) warning reported by subroutine LNNDRVD see LNNDRVD abstract (ier = 729) *** SYSTEM FATAL MESSAGE 3003 (READ) ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK SCRATCH (FILE 315) IN SUBROUTINE REIGL3 Avoidance: An avoidance is to add an spc to the DOF listed in the first message. @ 37633; ; Restart - EIGRL, EIGR - Inefficient Restart In a restart run, if an unreferenced EIGRL (or EIGR) is added to the restart run input, then the modes will be recalculated even though the entry is not referenced. Avoidance: Remove the unreferenced EIGRL/EIGR entry. @ 37635; ; Nonlinear Transient Analysis - RESTART, PARAM,G, GE -Different

Answers In a nonlinear transient restart analysis, if large displacement is turned on (param,lgdisp,1), and param W3 damping exists in the model (with PARAM,G ), the results may be different than the one in the coldstart run. The difference is due to the possible non-syncronization of the stiffness update between the coldstart and restart runs. Avoidance: Recommended damping in nonlinear transient (SOL 129) is param, W4 damping with GE value in MAT1. If stiffness update syncronization is desired with param W3 damping, set KSTEP=1 to force the stiffness update. Care must be exercise when there is adaptive gap element since the penalty stiffness can be fictituously high inducing an unrealistic high damping force. @ 37642; ; Restarts - CBUSH - SWM 2072 for card type 2601 If a model contains CBUSH elements, then upon RESTART SYSTEM WARNING MESSAGE 2072 will be issued by the RESTART module when it compares the GEOM2 data block. Avoidance: No avoidance is necessary, ignore the message. @ 37646; 2001; Documentation/QRG - CSUPER - Improve Remark 3 of CSUPER Entry Remark 3 of the description of the CSUPER bulk data entry in the QRG should be improved to read: 3. For identical or mirror image superelements, the grid point IDs, GPi, may appear in any order. However, if they are not in the same order as the external GRIDs of the primary superelement, then the SEQSEP entry is also required. In case of external superelements, the GRID IDs must be in the order that the terms in the associated matrices occur in. @ 37653; 2001; Documentation - CREEP - Thermal Correction The Quick Reference documentation for the CREEP entry should be changed to further explain what turns on the thermal correction for the creep strain rate. If the T0=0.0 or blank, then the thermal correction is not utilized. However if PARA,TABS is used, TABS is added to T0 and then the thermal correction is utilized even if T0 is blank or 0.0. If thermal correction is utilized then, EXP must be meaningful and have the consistent units. @ 37661; 70.7.2; Superelement - Repeated SUPER Command - UFM 7323, Possible Wron g Answers If the SUPER case control command is repeated incorrectly as shown below . subcase 100 super = 0,1 . subcase 200 super = 0,2 . subcase 300 super = 0,1 .

then following unclear fatal message is issued *** USER FATAL MESSAGE 7323 (SEDRA) SUBCASES FOR SUPERELEMENT 0 HAS A LOAD SEQUENCE NUMBER = WHICH IS OUT OF ORDER. The job will continue to execute and the results are unpredictable. Avoidance: Correct the offending SUPER case control command. @ 37700; 70.7.2; Parallel Processing/Statics - Distributed Memory Processing - SF M 7492 If DMP is specified on the command line for a SOL 101 run, the following error message may appear: *** SYSTEM FATAL INCONSISTENT ACCORDING TO ACCORDING TO USER ACTION: @ 37719; MESSAGE 7492 (PRESOL) VALUE FOR NUMBER OF LOCAL GRIDPOINTS ON PROCESSOR 1: SIL TABLE, THE NUMBER OF LOCAL GRIDPOINTS IS 100763 GEQMAP, THE NUMBER OF LOCAL GRIDPOINTS IS 101073 PLEASE CONTACT MSC.SOFTWARE CLIENT SUPPORT. 1

Avoidance: Do not use DMP. This error is corrected in Version 70.7.2. ; Documentation - V70.7 Release Guide Clarification Section 2.7 of the Version 70.7 Release Guide documents a new keyword "DCMPSEQ". This section of the guide does not give examples and it contains a typo. The Guide states that this keyword is the same as SYSTBM(206), this is a typo, it should be SYSTEM(206). It states that the new option is invoked by adding 64 to the value of the keyword; however the default of SYSTEM(206) is not documented. The default value of SYSTEM(206) is 4 for all systems except NEC, the default for NEC is 0. This keyword must be set on the NASTRAN command line in the input deck. For all systems except NEC the input to activate this option is: NASTRAN DCMPSEQ=68 or NASTRAN SYSTEM(206)=68 @ 37748; t 2001; Miscellaneous - V70.7, New Log File Messages - Wrong DOF Outpu The following new messages are sent to the log file in Version 70.7: 10:32:43 Finite element model generation started. 10:32:43 Finite element model generated 1 degrees of freedom. 10:32:43 Finite element model generation successfully completed. The second message above always reports 1 DOF which is in almost all cases wrong. @ 37764; 70.7.2; Parallel Processing/Frequency Response - DMP, PARAM,RESVEC - UFM Distributed Memory Parallel frequency response analysis will fail when

PARAM,RESVEC,YES is used to specify residual flexibility. Avoidance: Either: 1) Run using a newer version (V70.7.2 or newer) 2) Do not use residual vectors with dmp 3) Insert the following DMAP alter: compile resvec alter 'read '(,-1) $ For dmp>1, deactivate parallel np=getsys(np,197) $ capture NUMSEG if( np>1 ) putsys(0,197) $ set NUMSEG=0 alter 'read ' if( np>1 ) putsys(np,197) $ restore NUMSEG compile cmpmode alter 'read ' if ( getsys(np,197)>1 and getsys(np,265)>1 ) then $ $ Since the master node has all modes, then $ then we will generate a dummy mode on the $ slaves to avoid hang-ups. delete /cmlama,cmphixz,cmmi,cmoeigs,eigvmats $ nozset=1 $ paraml cmkxx//'trai'/1/s,n,noxset $ matgen ,/cmphixz/4/nozset/noxset $ matgen ,/lmat/7/3/nozset $ matgen ,/cmmi/nozset/nozset $ lamx lmat,/cmlama $ modtrl cmlama// /0/0/0/0/0 $ endif $ getsys(np,197)>1 ) then $ endalter @ 37773; 70.7.2; Matrix Methods - PARTN/MERGE - Poor Performance The PARTN and MERGE modules perform poorly when the partitioning vectors are purged. Avoidance: There is no avoidance but this has been corrected in Version 70.7.2. @ 37775; 70.7.2; Miscellaneous - Q4SKEW,Q4TAPER, TETRAAR, and T3SKEW do not work in Unix rcfile On unix workstations, if you attempt to set SYSTEM cells 189, 190, 191, or 218 in the "rcfile" using the cell name (Q4TAPER, Q4SKEW, TETRAAR, T3SKEW), the default values will be used for these cells with no warning messages issued, Avoidance: Use the SYSTEM cell number in the "rcfile" for these cells. For example: nastran nastran nastran nastran @ 37793; system(189)=1. system(190)=1. system(191)=100. system(218)=1.

2001; Nonlinear Normal Modes - RESTART, PARAM,FOLLOWK - Wrong Answers Nonlinear normal mode analysis by a restart in SOL 103 from SOL 106

database gives erroneous solution in presence of follower forces in Version 70.7. The follower stiffness is not included in this restart runs. This error may be avoided if the normal mode analysis is performed in SOL 106 during the coldstart run, or SOL 103 is used with a statsub option for the preload. However, a restart procedure is required if free-free modes are subject to a nonlinear deformation under the preload. Avoidance: For free-free modes, insert the following DMAP alter into the SOL 103 restart compile semg alter 'call semg1' type parm,,char8,y,followk='yes' $ TYPE PARM,,RS,N,stol=0. $ type db slt $ dbstatus slt//s,n,noslt $ PARAML CASES//'DTI'/COLNO/4//S,N,LOADID $ if ( followk='yes' and noslt>0 and loadid>0 and seid=0 ) then $ TAFF SLT,BGPDTS/ESTF,GPECTF/ LUSETS/LOADID/0/1. $ EMG ESTF,CSTMS,MPTS,DIT,,UGni,ETT,EDT, DEQATN,DEQIND,BGPDTS,GPSNTS,,,,/ KFELM,KFDICT,,,,/ 1/0/0//1/////////-1/1. $ EMA GPECTF,KFDICT,KFELM,BGPDTS,SILS,CSTMS,,/ KFGG0,/-1 $ dbview cstm0=cstms where (peid=0) $ emakfr bgpdts,cstms,slt,mjj,cstm0,scstm/ kfrfgg/loadid $ rforce stiffness ADD5 KFGG0,kfrfgg,,,/Kfgg $ PARAML KFGG//'TRAILER'/3/S,N,UNSYM//S,N,NOKFGG $ IF ( UNSYM<>6 AND NOKFGG>-1 ) THEN $ $ Symmetrize stiffness matrix if unsymmetric TRNSP KFGG/KFGGT $ ADD5 KFGG,KFGGT,,,/KF/1./-1. $ norm kfgg/dummy///s,n,xnorm/1 $ stol = xnorm*1.e-11 $ tolerance recommended matmod kf,,,,,/kfa,/2////stol $ paraml kfa//'null'////s,n,notnull $ if ( notnull=-1 ) modtrl kfgg////6 $ symmetric ENDIF $ UNSYM<>6 AND NOKFGG>-1 add kfgg,kjj/kfjj $ equivx kfjj/kjj/-1 $ ENDIF $ FOLLOWK='YES ' endalter @ 37797; 70.7.2; Documentation - diag= keyword does not work for all diags The "diag" command line keyword does not work for all possible diag flags. In particular, diag flags 5 and 6 cannot be set from the command line or an RC file. Avoidance: Set these diags in the bulk data file using the standard DIAG statement, DIAG 5,6 @ 37799; 2001; Optimization - DRESP2, DRESP1 - SFM 3007 (DOPR3)

In an optimization run, if a DRESP2 references both DRESP2 and DRESP1, and the DRESP1 references a property PTYPE (e.g., PBAR) which points to more than one element, then the job will fail with the following error messages: *** SYSTEM FATAL MESSAGE 3007 (DOPR3) ILLEGAL INPUT TO SUBROUTINE DOPR3J 0FATAL ERROR *** SYSTEM INFORMATION MESSAGE 1156 (XCLNUP) GINO FILE 110 WAS LEFT OPEN AT PROGRAM TERMINATION. Avoidance: There are two possible avoidances: 1. Avoid referencing a DRESP2 with a DRESP2, reference a series of DRESP1 instead. 2. Reference ELEM instead of property on the PTYPE field on the DRESP1 entry. @ 37805; 70.7.2; Optimization/Design Sensitivity - ANALYSIS=MFREQ - UFM 9068 The message "USER FATAL MESSAGE 9068 NO DRESP1 OR DRESP2 ENTRIES EXIST FOR DESOBJ=n" may be issued in error for a solution 200 analysis that contains only frequency or transient analysis subcases and the objective is a WEIGHT or VOLUME response. The fatal will only occur if the DRESP1 ID that is just below the ID called out by the DESOBJ request has a blank ATT1 field on the DRESP1 bulk data entry; i.e., the response is RTYPE = EIGN, FREQ or LAMA. Avoidance: Provide an ID on the DRESP1 entry for the WEIGHT or VOLUME objective that is lower than any other DRESP1 ID or ensure that the ID that is just below the objective ID has a non-blank ATT1 value. @ 37806; ; Documentation - V70.7 Release Guide Release Notes, p50, "File Size Reduction for Modal Frequency Response" The efficiencies described in this section require the use of sssalter spdra.v707, available on the msc internal web site. Contact your local MSC Technical Representative to get a copy of this dmap alter. @ 37812; 70.7.2; P-elements - InterFace Elements - Incorrect EPSILONs The EPSILON table produced when interface elements are used is incorrect in the NEC Version 70.7.0 release of MSC.Nastran The EXTERNAL WORK is correct however. Avoidance: There is on known avoidance. The error has been corrected in Version 70.7.2. @ 37832; 70.7.2; Heat Transfer - TSTRU, TEMP(INIT) - SFM 3001(TA1) If TEMP(INIT) is set to the same ID as TSTRU then the job will fail with the message: *** SYSTEM FATAL MESSAGE 3003 (TA1) ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK ETT (FILE 106) IN SUBROUTINE TA1ETD

Avoidance: Specify a different set ID for TEMP(INIT). @ 37833; 70.7.2; Installation - INSTALL.txt and INSTALL.htm on CD state incorrect MPT version The files INSTALL.txt and INSTALL.htm on the V70.7 CD state that for running distributed parallel MSC.Nastran on SGI R8K/ R10K/R12K, MPT 1.5 is required. This is a typo, MPT 1.3 is required. @ 37834; 70.7.2; Estimate - buffsize=estimate keyword broken in V70.7 In Version 70.7, if the keyword buffsize=estimate is used, then the job will fail with the following error messages in the f06 file. NASTRAN BUFFSIZE=ESTIMATE $(/USERS/SWAN/.NAST707RC[26]) *** USER FATAL MESSAGE 17 (LOOKUP) SYMBOL "ESTIMATE" IS UNDEFINED followed later by: *** USER WARNING MESSAGE 782 (SYSCHK) THE NASTRAN BUFFSIZE PARAMETER CANNOT BE SMALLER THAN MACHINE DEFAULT = 2049 USER INFORMATION: NASTRAN BUFFSIZE PARAMETER IGNORED. USER ACTION: USE THE INIT FMS STATEMENT TO SET THE DESIRED BUFFSIZE FOR EACH INDIVIDUAL DBSET. *** USER INFORMATION MESSAGE 480 (SYSCHK) JOB TERMINATED DUE TO ABOVE ERROR(S) IN THE NASTRAN STATEMENT PROCESSOR. FATAL ERROR Avoidance: Specify a specific buffsize. @ 37835; able 70.7; Installation - HP-UX, Install Script Reports No Disk Space Avail Under some circumstances on HP-UX systems, mscsetup may report "0 MB" disk space available on NFS or automounted filesystems. The full circumstances are not yet known. Note: The error is a direct result of incorrect information being returned from the df(1) command on HP-UX 10.20 and 11.00. The specific command used by mscsetup is "df -k", and sample output showing the error is: $ df -k /net/trinity/scratch /tmp_mnt/net/trinity/scratch (trinity:/scratch allocated Kb Kb 0 used allocated Kb 45 % allocation used Another invocation of df, that does not return sufficient information, appears to have better results: ) : 0 total 0 free allocated

$ df /net/trinity/scratch /tmp_mnt/net/trinity/scratch (trinity:/scratch): 2032506 blocks -1 i-nodes Finally, an alternative command yields the correct results: $ bdf /net/trinity/scratch Filesystem kbytes used avail %used Mounted on trinity:/scratch 1981000 862335 1016253 46% /tmp_mnt/net/trinity/scratch Avoidance: Ignore the disk space warning messages and force mscsetup to install the s/w if you know sufficient space is available to install the selected products. @ 37837; 70.7.2; Documentation/Estimate - Minimum Values The ESTIMATE utility will apply a minimum value to estimates for the following: Parameter --------memory disk sdball scratch scr300 Keyword Default Value (MB) ---------- -----------------memmin 16 dskmin 1 dballmin 1 scratchmin 1 scr300min 1

The minimum value is applied after the conservative value (i.e., "memco", "dskco", "dballco", "scratchco", and "scr300co" respectively). Thus, setting "memco=0.5" to reduce all memory estimates by 50% will have *no* effect if the estimate falls below the minimum value. @ 37843; 70.7.2; Loads - PLOAD4, Large Field Format - UFM 4577 (GP3C) If a PLOAD4 references an element with large field format, the job may fail with the message: USER FATAL MESSAGE 4577 (GP3C) PLOAD4 LOAD SET= 1 REFERENCES ELEMENT ID =******************* WHICH WAS NOT FOUND AMONG TRIA3, TRIA6, QUAD4, QUAD8, HEXA, OR PENTA ELEMENTS Avoidance: There is no avoidance other than renumbering the elements. @ 37844; 2001; Accounting - XSEMDR - Can not Determine Solution Sequence In releases prior to Version 70.7 the f04 Execution Summary ended with the last module specific to the solution. Beginning in Version 70.7 the f04 Execution Summary ends with the last line XSEMDR END. For example: 13:19:02 0:13 527 EXIT BEGN 13:19:02 0:13 END 212.0 212.0 0.0 0.0 12.8 12.8 0.0 0.0 NLTRAN XSEMDR

This prevents the accounting program provided my MSC which reads the last line from determining the solution sequence.

@ 37859; 70.7.2; Miscellaneous - Missing TPL file mcso42.dat in V70.7 The file mcs042.dat was missing from the Version 70.7 tpl. It has been added to the Version 70.7.2 tpl. @ 37862; GL; Miscellaneous - FLEXlm can not be used with system checkpoint ca pabilities MSC.Nastran jobs (as well as any other program) that use FLEXlm licensing or any other socket-based communications cannot be checkpointed by system-level checkpoint/restart capabilities, such as SGI's cpr utility. If attempted messages similar to the following will be issued in the log file: Checkpointing id 259527 (type PID) to directory /tmp/test.cpr CPR Error: Uncheckpointable fd, pid 259527 fd 10 type socket Checkpoint failed. Restarted all processes back to their running states CPR Error: ckpt_create child failed: errno 47 (Checkpoint/Restart error) Failed to checkpoint process 259527 --------------------------------Avoidance: Use a node-lock authorization code for licensing and avoid the use of any socket connections by MSC.Nastran. @ 37872; 2001; SSSALTERs - checka, PARAM,LANSUP - SFM 3007 If PARAM,LANSUP,1 is used in the checka sssalter in version 70.5, 70.6, or 70.7, the run will fail with SYSTEM FATAL MESSAGE 3007 indicating an illegal operation while merging matrices. Avoidance: Replace: UMERGE USET,DMI,Dar/DM1/'A'/'R'/'L' $ DN changed to DAR with: equivx dar/dm1/-1 $ @ 37873;

70.7; Old Design Sensitivity - SFM 3001, MPTS Does Not Exist If the old sensitivity (SOL 103) is used and no material entries are present, the run will fail with SFM 3001 indicating that data block MPTS does not exist. Avoidance: 1) use SOL 200 to perform similar task or 2) or add a dummy material to the input file if the old dsa is used.

@ 37880; 70.7.2; Parallel Processing - DMP - SPCForce Output with Multiple Subcas es - UFM 5423 Distributed Memory Parallel jobs, dmp>1, with multiple subcases will fail with the following message if spcforce output is requested.

*** USER FATAL MESSAGE 5423 (SADD5) ATTEMPT TO ADD INCOMPATIBLE MATRICES, (SADD5) TRAILER FOR QG0S TRAILER FOR QGI 0FATAL ERROR @ 37886; AKFR COLS ROWS FORM TYPE NZWD DENS IS 1 450 2 2 0 0 IS 2 450 2 2 24 267

Avoidance: Either do not select dmp or do not select spcforces. 2001; Nonlinear Analysis - RFORCE, PARAM,FOLLOWK - Poor Performance EM Nonlinear analysis of large problems with RFORCE as a follower force may exhibit poor performance in EMAKFR. Avoidance: There is no avoidance. @ 37892; 70.7.2; Documentation/COG - V70.7 smaster=estimate fails The "MSC.Nastran V70.7 Configuration and Operations Guide" incorrectly states that "smaster=estimate" is valid, it is not. Avoidance: Set an actual size if the MASTER DBSet default size needs to be increased. @ 37898; 2001; Superelements - SDAMPING; PARAM,SESDAMP,YES - NORM=MAX or POINT, Wrong Answers Superelement modal damping (PARAM,SESDAMP,YES) produces incorrect answers when MAX or POINT normalization is selected on the EIGR or EIGRL entries. Avoidance: Use MASS normalization which is the default. @ 37907; 70.7.2; Optimization/Design Sensitivity - DRESP1, PCOMP - UFM 5503 The message: UFM 5503 ELEM NO. n PLY NO.y REFERENCED ON A DRESP1 ENTRY COULD NOT BE FOUND may be issued in error for composite responses (DRESP1 RTYPE=CSTRESS,CSTRAIN or CFAILURE) if the SOUTi requests on the underlying PCOMP entries have different values for the different layers (i.e., some are YES and some are NO or blank). Avoidance: Make all the SOUTi requests on a given PCOMP entry the same if design responses are to be imposed on elements associated with the PCOMP. @ 37913; 70.7.2; Submittal - nastran Command does not recognize LSF Submittal The nastran command does not automatically recognize a submittal via LSF. For NQS, the nastran command will recognize it is running in a batch job and set the defaults batch=no and notify=no. In addition, if submitted as a DMP job by LSF, the nastran will not recognize the list of selected

hosts. Avoidance: The problem may be avoided by: 1. set notify=no and batch=no in an RC file or on the command line when running under LSF. 2. Set hosts=$(print $LSB_HOSTS sed 's/^ *//;s/ *$//;s/ */:/g') on the command line when running a DMP job under LSF @ 37918; 2001; Documentation/DMAP - DISUTIL not documented The MSC.Nastran DMAP Modules and Datablocks CD for Version 70.7 mentions, but does not document the module DISUTIL The following is the documentation for this module: Broadcast data blocks between slave and master processors for parallel processing. Additional processing may occur on the master processor. Format: Format for DISMETH=1 or 2: DISUTIL B,X,R,EQMAP/ EPSSE/ NSKIP/S,N,EPSI/1 or 2 $ Format for DISMETH=3 on master processor: DISUTIL DB,,,////3 $ Format for DISMETH=3 on slave processor: DISUTIL ,,,,/DB///3 $ Format for DISMETH=4 or 5: DISUTIL RESMAX,RESMAX0,CASECC,HEADCNTL////4 or 5 $ Format for DISMETH=6: DISUTIL UG,SPCPART,GEQMAP,/UGG///6 $ Format for DISMETH=7: DISUTIL PG,SPCPART,GEQMAP,/PGG///7 $ Format for DISMETH=8 or 9 on master processor: DISUTIL MATS,,,/MATM///8 or 9 $ Format for DISMETH=8 or 9 on slave processor: DISUTIL MATS,,,////8 or 9 $ Input Data Blocks: B Rectangular matrix which is the local load vector (with local values in local a-set). X Solution of the equation [A][X]=[B] which is the local solution matrix (with global values in local a-set). R Residual matrix which is local (with local or global values in local a-set) EQMAP Table of degree-of-freedom global-to-local maps for domain decomposition. DB Any data block to be broadcast from the master to the slave processors. RESMAX Resultant or maxima matrix. RESMAX0 Resultant or maxima matrix for residual structure. CASECC Table of Case Control command images. HEADCNTL List of integer codes for header print control UG Displacement matrix in g-set for the current processor (local). SPCPART Partitioning vector for domain decomposition. GEQMAP Table of grid based local equation map indicating which grid resides on which processors/partitions for domain decomposition. MATS Any matrix on slave processors.

Output Data Blocks: EPSSE Table of epsilon and external work. UGG Displacement matrix in g-set for all processors (global). PGG Force matrix in g-set for all processors (global). DB Any data block to be broadcast from the master to the slave processors. MATM Any matrix on master processor. Parameters: NSKIP Input-integer-default=1. Record number in CASECC corresponding to the first subcase of the current boundary condition. EPSI Output-integer-default=1. Static solution error ratio flag. Set to 1 if the error ratio is greater than 1.E-3. DISMETH Input-integer-default=1. Method. 1 Compute epsilons and external works assuming a-set components of RUF are local values (which occurs in a direct solution).. 2 Compute epsilons and external works assuming a-set components of RUF are global values (which occurs in a in iterative solution). 3 Broadcast table or matrix from master to slaves. 4 Broadcast VECPLOT resultant output from slaves to master and combine on master. 5 Broadcast VECPLOT maxima output from slaves to master and combine on master. 6 Broadcast displacement matrices from slaves to master and merge into global displacement matrix on master. 7 Broadcast force matrices from slaves to master and add/merge into global force matrix on master. 8 Broadcast any matrix from slaves to master and add all matrices on master. 9 Broadcast any matrix from slaves to master and append columnwise all matrices on master. Remarks: 1. All executions of DISUTIL must be synchronized across all processors. 2. EPSSE may be purged. EPSSE contains: a. Sequential number of subcases b. Superelement (domain) identification number c. Epsilon error ratio d. Strain energy (external work)3. RESMAX0 may be purged, on the slave processors. 4. UGG may not be purged on the master processors. 5. For DISMETH=8 or 9 then MATS and MATM must be in machine precision. @ 37925; 2001; Optimization/Design Sensitivity - PCOMP - UFM 5629 (DSAD1C) If RTYPE=STRESS or STRAIN in a composite optimiztion deck with NOCOMPS=1, the following fatal message is issued. *** USER FATAL MESSAGE 5629 (DSAD1C) THERE ARE DRESP1 ENTRIES FOR RESPONSE TYPE = NO CORRESPONDING OUTPUT RESPONSES. STRESS BUT THERE ARE

Avoidance: There is no avoidance. @ 37932; 2001; Optimization/Design Sensitivity - DVMRELi, DVCRELi, scr=no - Wro ng Answers

In an optimization run, if the design variables involve material properties (DVMRELi) or connectivity (DVCRELi) only, and scr=no is used, then the run will converge prematurely. A symptom of this problem is the unexplainable job termination after the first design cycle with no change in the objective function. Avoidance: Run the job with scr=yes or insert the following alter: compile update alter 'RESTART EPTS,EPTSP' DBVIEW mPTSP=mPTS WHERE (DESITER=DESP) $ RESTART mPTS,mPTSP,/UHIST $ DBVIEW geom2SP=geom2S WHERE (DESITER=DESP) $ RESTART geom2S,geom2SP,/UHIST $ endalter @ 37944; ; Documentation - VECPLOT - DMAP Module and Datablock Dictionary The Version 70.7 DMAP Module and Data Blocks Dictionary has incorrectly indicates that the VECPLOT module has only one output datablock. Starting in Version 70.7 the VECPLOT module has two output datablocks. This is correctly documented in the V70.7 Release Notes. @ 37952; 70.7.2; Documentation/COG - Estimate Program, Precedence of Keywords The precedence of ESTIMATE processed keywords in the MSC V70.7 Configuration and Operations Guide listed in Section 6.2 should be as follows: 1) ESTIMATE keywords, which are also NASTRAN keywords that are recognized in the bulk data file (Example: NASTRAN BUFFSIZE=4097) 2) ESTIMATE keywords on the command line, Example: msc707 estimate myjob.dat memco=2.5 Note: NASTRAN only keywords will cause an error (example: auth=1700@server) 3) ESTIMATE keywords, which are also NASTRAN keywords, contained in the nastran INI and/or RC files (example: buffsize) 4) ESTIMATE keywords contained in data-file-directory/.estimaterc on UNIX, or data-file-directory\estimate.rcf on NT, where data-file-directory is the directory containing the input data file 5) ESTIMATE keywords contained in $HOME/.estimaterc on UNIX, or %HOMEDRIVE%%HOMEPATH%\estimate.rcf file on NT 6) estimate.ini file in the install-dir/msc707/arch on UNIX (example: /msc/msc707/hpux/estimate.ini) There is no equivalent on NT. @ 37969; ; Superelements - Poor Message, UFM 6790 (SEPO21)

UFM 6790 shown below is output when a list superelement has no points that can be identified as boundary points. The fatal message is followed by table print which may be ignored. *** USER FATAL MESSAGE 6790 (SEPO21) THE FOLLOWING SUPERELEMENTS ARE DETACHED: 1 3 62000000007 52000000008 212000000002 222000000003 232000000004 242000000005 252000000006 61101529088 71102053376 81102577664 91103101952 101103626240 11 22000000001 42000000007 52000000008 212000000002 22 2000000003 232000000004 242000000005 252000000006 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 @ 37978; 2001; Nonlinear Analysis - HEXA - Poor Convergence for Small Loads Geometric nonlinear analysis (SOL 106 with PARAM LGDISP > 0) with solid elements and extremely small loads may fail to converge. @ 37981;

2001; Nonlinear Analysis - UWM 5426 for DEFORM Entry Should be a FATAL Using a DEFORM entry in the case control for SOL 106 (nonlinear statics) will produce: *** USER WARNING MESSAGE 5426, DEFORM BULK DATA CARDS REFERENCED BY DEFORM CASE CONTROL COMMAND ARE NOT DEFINED. THE REQUEST FOR STRESSES AND/OR FORCES IS IGNORED. This will cause data recovery for stresses and/or forces to be ignored. It will also lead to misleading results, because the DEFORM entry is not applicable to Sol 106. Recommendations: 1) change Warning 5426 to Fatal for sol 106 2) Add the following remark to Deform Bulk data section: 4. The deform entry, when called by the deform case control command, is applicable to linear static, inertia relief, differential stiffness, and buckling (Solutions 101,105,114, and 200) and will produce fatal messages in other solution sequences. Use SPCD to apply enforced displacements in solution sequences for which DEFORM does not apply. 3) Modify remark 2 in the Case control FROM: DEFORM is only applicable in statics, inertia relief, differential stiffness, and buckling problems. TO: DEFORM is only applicable in linear statics, inertial relief, differential stiffness, and buckling problems (Sols 101, 105, 114, and 200), and will produce a fatal message in other solution sequences.

37990; 4549

; Nonlinear Analysis - Beam with Warping Failed to Converge - SFM A nonlinear analysis of a BEAM with the warping degree of freedom specified may results in failed convergence or SFM 4549 followed by a traceback: SYSTEM FATAL MESSAGE 4549, SINGULAR (C+KBRR) MATRIX ENCOUNTERED FOR BEAM ELEMENT Traceback: Offset Offset Offset Offset Offset Offset Offset Offset 0x00003e28 0x000009d0 0x000022dc 0x0000017c 0x00001124 0x00000790 0x00000088 0x00000990 in in in in in in in in procedure procedure procedure procedure procedure procedure procedure procedure nmepd_ nmatd_ nbem2d_ nembrd_ nlemg_ nliter_ xsem15_ mainal 1001

@ 38012;

2001; Documentation/QRG - PARAM,FOLLOWK The Quick Reference Guide incorrectly states that PARAM,FOLLOWK is available only in solutions 103, 105, and 107. It is also available in solutions 108, 109, 111, and 112.

@ 38013;

; DMAP - PARTN - TYPE is Ignored Specification of TYPE is ignored in the PARTN module. This may cause a problem in user dmaps. The avoidance is dependent on the intent of the DMAP.

@ 38016;

; Superelements - SESET failures in SEP1X using THRU range entry It has been observed that the SESET entry may cause failures when a THRU range is specified where the low GRID ID is less than the starting GRID ID of the model. Avoidance: An avoidance is to make the starting grid equal to the low the low grid identification.

@ 38021; 70.7.2; Documentation/COG - Fujitsu, XDB, TRANS The V70.7 Configuration and Operations Guide states that the Fujitsu VPP,Vx UXP/V systems require that the XDB file be translated. This information is incorrect. The correct line would be similar to that of HP HP-UX right underneath the Fujitsu line. @ 38023; 2001; Documentation/QRG - PBEAML - NSM(B) not included on PBEAML entry The Quick Reference Guide incorrectly omits the field for the NSM (non-structural mass) on end B of the PBEAML entry. The last field on this entry is the NSM for end B. @

38031;

; Nonlinear Transient Thermal Analysis - Missing Results from XDB A multiple subcase transient thermal analysis will be missing temperature results in the OP2 and/or XDB files for the 2nd and subsequent subcases. Avoidance: A DMAP avoidance is under investigation.

@ 38041; 70.7.2; Optimization/Design Sensitivity - DESGLB, GRAV, Multiple Subcase s - UFM 5423 In an optimization run, if the DESGLB global constraint is used, and there are two or more static subcases with gravity loads, then the job may fail with the following error messages: *** USER FATAL MESSAGE 5423 (SADD5) ATTEMPT TO ADD INCOMPATIBLE MATRICES, (SADD5) COLS ROWS FORM TYPE NZWD DENS TRAILER FOR EGM 1830 114 TRAILER FOR EGKL 10976 681 0FATAL ERROR Avoidance: Use the DESSUB constraint callout instead. @ 38046; 2001; Miscellaneous - OUTPUT2, TABPRT - Incorrect Format Field Output datablocks (OFP tables) contain information in the header record which describes the information contained in the datablock. Word 9 is used to indicate if the information is (1) REAL, (2) REAL/IMAGINARY, or (3) MAGNITUDE/PHASE. If the data is REAL/IMAGINARY, this flag is incorrectly set to 1. Avoidance: There is no avoidance. @ 38055; 70.7.2; Parallel Processing - DMP - Incorrect SPCforces for constrained & loaded points In a distributed memory parallel execution, dmp>1, SPCforces are incorrect for constrained and loaded interior DOFs. Avoidance: Insert the following DMAP alter: compile disph1b alter 'pss' call dbstore pss,,,,//0/0/' '/0 $ compile sedisp alter 'kfs,uox,pss'(,-1) call dbfetch /pss,,,,/0/0/0/0/0 $ endalter @ 38056; 2001; Beam Library - PBARL ILLEGAL INPUT TO SUBROUTINE DOMPFTN IS 1 80532 2 2 IS 2 80532 2 2

When the generalized PBARL is designed using the beam library server, the job fails in DOPR1 with the following message:

"ILLEGAL INPUT TO SUBROUTINE DOMPFTN". This message is caused by designing the sixth property (FE) in the EPT table which is currently not used. Avoidance: In the bsbrid subroutine, change the Do 110 i = 1,nprop do 100 j = 1,ndim sentyp(i,j)=3 100 continue 110 continue to Do 110 i = 1,nprop if ( iprop .eq. 6 ) go to 110 do 100 j = 1,ndim sentyp(i,j)=3 100 continue 110 continue @ 38061; 2001; SSSALTERS - cygyroa.vxx, TEMP(LOAD) - SFM 3003 (EMG) The following message will occur in with the SSSALTER CYGYROA if thermal loads are used. SYSTEM FATAL MESSAGE 3003 (EMG) ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK ETT (FILE 107) IN SUBROUTINE EMGTAB The error occurs when the differential stiffness matrix is being generated, and hence the actual output (e.g. displacements, stresses) is correct. However, since differential stiffness is not generated, restarts off this database will not include the desired gyroscopic effects. The error occurs because of the unique way thermal loads are called out in Cyclic solution sequences. The TEMP(LOAD) case control command points to a bulk data LOADCYN entry which inturn points to a temperature set. The EMG module issues the fatal error because it is searching for the temperature set called out by the temp(load) command rather than the temperature set called out on the appropriate LOADCYN entry. Avoidance: Use this modified version of the cygyroa.V705 alter. Note that this version requires the addition of PARAM,TEMPSET. See instructions at top of alter. $ V70.5 $ modified 12/16/99 by Jack Castro to fix error if TEMP(load) is used. $ Directions: Insert PARAM,TEMPSET in bulk data with the value of the $ temperature load set which is called out on the appropriate LOADCYN entry. $ $ THIS ALTER IS CONFIDENTIAL AND A TRADE SECRET OF THE $ MACNEAL-SCHWENDLER CORP. THE RECEIPT OR POSSESSION OF $ THIS ALTER DOES NOT CONVEY ANY RIGHTS TO REPRODUCE OR $ DISCLOSE ITS CONTENTS, OR TO MANUFACTURE, USE, OR SELL $ ANYTHING HEREIN, IN WHOLE OR IN PART, WITHOUT THE $ SPECIFIC WRITTEN CONSENT OF THE MACNEAL-SCHWENDLER CORPORATION. $ $

$ $ cygyroa.v69 - updated for v69 -- 9-25-96 $ $ tested on system built on 3-14-95 VERSION 68.2 $ $************************************************************************ ******$ $ ALTER FOR SOL 114 TO INCLUDE GYROSCOPIC TERMS. $ $ THE ALTER WILL ADD CENTRIPETAL, CORIOLIS, AND DIFFERENTIAL STIFFNESS $ TERMS TO THE SUPERELEMENT STIFFNESS AND DAMPING MATRICES. TO USE THIS $ ALTER THE USER MUST SUPPLY A RFORCE FOR EACH SUPERELEMENT, INCLUDING $ THE RESIDUAL STRUCTURE ( SEID=0 ). THE COORDINATE $ SYSTEM REFERENCED ON THE RFORCE CARD MUST BE CENTERED ON THE AXIS OF $ ROTATION FOR THE SUPERELEMENT. THE GRID ID ENTRY IS IGNORED FOR THE $ CALCULATION OF THE GYROSCOPIC TERMS. THE ROTATION IS ASSUMED TO BE $ ABOUT THE REFERENCED COORDINATE SYSTEM (CID) ORIGIN, THE MAGNITUDE (A) $ AND ROTATION VECTOR COMPONENTS (N1, N2, & N3) DETERMINE THE ORIENTATION. $ $ ONLY THE ZERO HARMONIC SHOULD BE USED IN THIS STEP OF THE PROCEDURE. $ OTHER HARMONICS MAY BE USED IN THE RESTARTS. $ $ AFTER THIS INITIAL STATIC RUN THE MODEL CAN BE RESTARTED IN ANY $ CYCLIC SOLUTION SEQUENCE AND THE GYROSCOPIC EFFECTS WILL BE INCLUDED. $ $ THE ABOVE PROCEDURE RESULTS IN A RESTART INWHICH REFORMULATION OF $ THE SUPERELEMENT MATRICES IS NOT PERFORMED. REFORMULATION IS NORMALLY DONE $ WHEN THE USER CHANGES THE GEOMETRY, ELEMENT OR MATERIAL PROPERTIES, OR $ ADDS GRIDS OR SCALAR POINTS. IF REFORMULATION IS DONE ( SEMA ) THEN $ THE GYROSCOPIC TERMS WILL NOT BE INCLUDED IN THE SUPERELEMENT MATRICES. $ $ IF SPEED DEPENDENT DIFFERENTIAL AND GYROSCOPIC TERMS ARE REQUIRED IN THE $ RESTART, THEN PARAMETER DEPEND SHOULD BE SET TO -1. IF THIS OPTION IS $ ACTIVATED, THEN THE RESTART MUST NOT HAVE MASS DEPENDENT LOADS (I.E. NO $ GRAVITY LOADING) BECAUSE THE MASS MATRIX IS MODIFIED AND THE RESULTING $ LOADS WILL BE INCORRECT. $ $ ALTER DEVELOPED BY DAVID BELLA FOR THE MACNEAL-SCHWENDLER CORPORATION, $ AUGUST, 1992. $ $************************************************************************ ******$ $ echooff $ compile cycmode $ alter 1 $ putsys (0,192) $ $ COMPILE CYCSTATX $ $ALTER 44 $ V67 $ALTER 51 $ V68 $alter 'if.*(.*dbdict\>=1.*).*call.*dbdict1' alter 'call.*super3'

$ TYPE DB SLIST $ TYPE PARM,,I,N,LPFLG=0,NOUP $ TYPE PARM,,I,N,SCNDRY=+1,EXTRN=+1 $ TYPE PARM,,I,Y,DEPEND= 0 $ TYPE PARM,,RS,N, OMEGA= 0. $ TYPE PARM,,CS,N, SCALE $ $ $ PREPARE DELETE LIST $ FILE HISTS=APPEND $ MATGEN ,/TMPMGG/1/2 $ MATGEN ,/TMPBGG/1/2 $ MATGEN ,/TMPKGG/1/2 $ $ RESTART TMPMGG,,/HISTS $ RESTART TMPBGG,,/HISTS $ RESTART TMPKGG,,/HISTS $ $ $ LOOP ON SUPERELEMENTS $ SEID=-1 $ RE-INITIALIZE FOR SEP2DR $ DO WHILE ( LPFLG >= 0 ) $ $ IF ( RSONLY ) THEN $ CHECK FOR EXISTENCE OF SUPERELEMENTS $ LPFLG=-1 NOUP=-1 SEID=0 $ ELSE $ DO SEP2DR WHEN SUPERELEMENTS EXIST. $ SEP2DR SLIST,EMAP//S,N,SEID/S,N,PEID//S,N,LPFLG///// S,N,NOUP/S,N,SCNDRY/S,N,EXTRN/S,N,NOMR/'ALL' $ $ ENDIF $ $ MESSAGE //' '/ $ MESSAGE //' USER DMAP INFORMATION MESSAGE - DIFFERENTIAL'/ $ MESSAGE //' STIFFNESS AND GYROSCOPIC EFFECT GENERATION'/ $ MESSAGE //' AND ASSEMBLY FOR SUPERELEMENT'/ SEID $ MESSAGE //' '/ $ $ IF ( SCNDRY = -1 ) THEN $ $ MESSAGE //' '/ $ MESSAGE //' USER DMAP WARNING MESSAGE - SECONDARY'/ ' SUPERELEMENTS'/ $ MESSAGE //' MUST BE IDENTICAL ELEMENTS (NOT MIRROR) '/ ' WITH LOADING'/ $ MESSAGE //' AND DISPLACEMENTS IDENTICAL TO THE PRIMARY'/ ' SUPERELEMENT'/ $ MESSAGE //' '/ $ $ ELSE $ $ $ SET QUALIFIERS $ CALL SETQ CASES//SEID/PEID/S,MTEMP/S,K2GG/S,M2GG/

S,B2GG/S,MPC/S,SPC/S,LOAD/S,DEFORM/S,TEMPLD/ S,P2G/S,DYRD/S,METH/S,MFLUID $ $ PARAML GPECT//'PRES'////S,N,NOGPECT $ IF ( NOGPECT >= 0 ) THEN $ $ $ $ CALCULATE CENTRIPETAL AND CORIOLIS TERMS $ CALL CORICEN SLT,BGPDTS,EQEXINS,CSTMS,SILS,MJJ/ KCJJ,BRJJ/S,OMEGA $ $ IF ( OMEGA = 0. ) THEN $ $ MESSAGE //' '/ $ MESSAGE //' USER WARNING: NO RFORCE SPECIFIED,'/ $ MESSAGE //' GYROSCOPIC TERMS WERE NOT'/ $ MESSAGE //' ADDED TO STRUCTURAL MATRICES.'/ $ MESSAGE //' '/ $ $ ENDIF $ $ $ $ CALCULATE DIFFERENTIAL STIFFNESS $ $message //'before emg...TEMPLD='/templd $ type parm,,i,y,tempset $ $message //'before emg...TEMPSET='/TEMPSET $ EMG EST,CSTMS,MPTS,DIT,,CYUG,ETT,EDT,,,BGPDTS,GPSNTS,,,,/ KDELM,KDDICT,,,,/ 1/0/0//1//TEMPSET/DEFORM//////1/ $ PARAML $ IF ( NOKTJJ < 0 ) THEN $ $ MESSAGE MESSAGE MESSAGE MESSAGE MESSAGE MESSAGE $ ENDIF $ $ $ ADD DIFFERENTIAL AND CENTRIPETAL TERMS $ ADD5 KTJJ,KCJJ,,,/KJJGYRO/ $ PARAML KJJGYRO//'NULL'////S,N,NOGYRO $ $ IF ( NOGYRO < 0 ) JUMP ENDLOOP $ $ IF ( (DEPEND < 0 ) AND (OMEGA <> 0.) ) THEN $ $ SCALE= CMPLX( -1.0/(OMEGA*OMEGA), 0. ) $ ADD MJJ,KJJGYRO/MJJTEMP//SCALE $ EQUIVX MJJTEMP/MJJ/ALWAYS $ //' '/ $ //' USER WARNING: NO LOADING SPECIFIED,'/ $ //' DIFFERENTIAL STIFFNESS TERMS'/ $ //' WERE NOT ADDED TO STRUCTURAL'/ $ //' MATRICES.'/ $ //' '/ $ EMA GPECT,KDDICT,KDELM,BGPDTS,SILS,CSTMS,,/KTJJ,/ $ KTJJ//'NULL'////S,N,NOKTJJ $

EQUIVX $

MJJ/MJJTEMP/NEVER $

SCALE= CMPLX( 1.0/OMEGA ) $ ADD BJJ,BRJJ/BJJTEMP//SCALE $ EQUIVX BJJTEMP/BJJ/ALWAYS $ EQUIVX BJJ/BJJTEMP/NEVER $ $ MESSAGE //' '/ $ MESSAGE //' USER INFORMATION: SPEED DEPENDENT TERMS'/ ' ADDED'/ $ MESSAGE //' TO STRUCTURAL MATRICES'/ $ MESSAGE //' '/ $ $ ELSE IF ( (DEPEND < 0) AND (OMEGA = 0.) ) THEN $ $ MESSAGE MESSAGE MESSAGE MESSAGE MESSAGE MESSAGE MESSAGE $ ELSE $ $ ADD EQUIVX EQUIVX KJJ,KJJGYRO/KJJTEMP/ $ KJJTEMP/KJJ/ALWAYS $ KJJ/KJJTEMP/NEVER $ //' '/ $ //' USER WARNING: SPEED DEPENDENCE REQUESTED,'/ $ //' BUT NO RFORCE SPECIFIED.'/ $ //' DIFFERENTIAL STIFFNESS OR'/ $ //' GYROSCOPIC TERMS WERE NOT '/ $ //' ADDED TO STRUCTURAL MATRICES.'/ $ //' '/ $

$ $ ADD CORIOLIS TERMS TO SUPERELEMENT DAMPING $ ADD BJJ,BRJJ/BJJTEMP/ $ EQUIVX BJJTEMP/BJJ/ALWAYS $ EQUIVX BJJ/BJJTEMP/NEVER $ $ MESSAGE //' '/ $ MESSAGE //' USER INFORMATION: CONSTANT SPEED TERMS'/ ' ADDED'/ $ MESSAGE //' TO STRUCTURAL MATRICES'/ $ MESSAGE //' '/ $ $ ENDIF $ $ ENDIF $ NOGPECT>=0 $ ENDIF $ SCNDRY <> -1 $ $ TRIGGER REASSEMBLEY AND REREDUCTION OF SUPERELEMENT MATRICIES $ IN RESTART BY DELETING ASSEMBELLED AND REDUCED MATRICES $ RESTART ,,,HISTS/TLIST/TRUE $ $ LABEL ENDLOOP $ $ ENDDO $ LPFLG>=0 $ $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ $$$$ $ COMPILE CORICEN $

$ SUBDMAP CORICEN SLT,BGPDT,EQEXIN,CSTM,SIL,MJJ/KCEN,BCOR/OMEGA $ $ $******************************************************************** $ * $ This DMAP sequence calculates Coriolis and Centripetal effect * $ matrices in global coordinates. * $ * $ For node i: * $ * $ [Bcoriolis] = 2[x]m * $ * $ [Kcentripetal] = [x][x]m * $ * $ 0 -Omega(z) Omega(y) * $ [x] = Omega(z) 0 -Omega(x) * $ -Omega(y) Omega(x) 0 * $ * $******************************************************************** $ TYPE PARM,,I,N,SLTRECNO $ TYPE PARM,,RS,N,OMEGA $ TYPE PARM,,CS,N,N1C,N2C,N3C $ TYPE PARM,,CS,N,OMEGAC $ $ $ SEARCH FOR RFORCE ENTRY $ SLTRECNO= 1 DO WHILE ( SLTRECNO > 0 ) $ $ PARAML SLT//'DTI'/S,N,SLTRECNO/1//S,N,LTYPE/ $ IF ( (SLTRECNO <> -1 ) AND ( LTYPE = 10 ) ) THEN $ $ $ EXTRACT SPIN VECTOR FROM RFORCE CARD. $ PARAML SLT//'DTI'/SLTRECNO/4//S,N,CID/ $ PARAML SLT//'DTI'/SLTRECNO/5/S,N,RPT/ $ PARAML SLT//'DTI'/SLTRECNO/6/S,N,N1/ $ PARAML SLT//'DTI'/SLTRECNO/7/S,N,N2/ $ PARAML SLT//'DTI'/SLTRECNO/8/S,N,N3/ $ SLTRECNO = -2 $ $ ENDIF $ $ SLTRECNO = SLTRECNO + 1 $ $ ENDDO $ $ IF ( RPT = 0. ) THEN $ $ OMEGA= 0. RETURN $ $ ENDIF $ $ $ THESE ARE NEEDED LATER ON $ MATGEN ,/I1X1/1/1/ $ MATGEN ,/I6X6/1/6/ $ $

$ CALCULATE RADIANS PER UNIT TIME $ OMEGA= 2.0 * PI(1) * RPT $ OMEGAC = CMPLX(OMEGA,0.) $ $ $ $ CREATE A COLUMN VECTOR, MJJTSUM, FROM THE DIAGONAL OF MJJD $ TO BE USED FOR PARTITIONING PURPOSES LATER $ DIAGONAL MJJ/MJJC/'COLUMN' $ $ $-------------------------------------------------------------------$ $-------------------------------------------------------------------$ $ $ CALCULATE ROTATION MATRIX IN THE CID SYSTEM $ IF (N1 <> 0.) THEN $ $ DETERMINE FACTOR FOR X-ROTATION (in CID system) $ MATGEN ,/PVC6/6/6/1/1/4 $ MATGEN ,/PVR6/6/6/2/1/3 $ MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $ MPYAD ONE6X6,I6X6,ONE6X6/ROTX/1/-1 $ DELETE /PVC6,PVR6,ONE6X6,,/ $ N1C = CMPLX(N1,0.) $ ENDIF $ $ $-------------------------------------------------------------------$ $ IF (N2 <> 0.) THEN $ $ DETERMINE FACTOR FOR Y-ROTATION (in CID system) $ MATGEN ,/PVC6/6/6/2/1/3 $ MATGEN ,/PVR6/6/6/0/1/5 $ MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $ MPYAD ONE6X6,I6X6,ONE6X6/ROTY/1/-1 $ DELETE /PVC6,PVR6,ONE6X6,,/ $ N2C = CMPLX(N2,0.) $ ENDIF $ $ $-------------------------------------------------------------------$ $ IF (N3 <> 0.) THEN $ $ DETERMINE FACTOR FOR Z-ROTATION (in CID system) $ MATGEN ,/PVC6/6/6/0/1/5 $ MATGEN ,/PVR6/6/6/1/1/4 $ MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $ MPYAD ONE6X6,I6X6,ONE6X6/ROTZ/1/-1 $ DELETE /PVC6,PVR6,ONE6X6,,/ $ N3C = CMPLX(N3,0.) $ ENDIF $ $ $-------------------------------------------------------------------$

$ $ ADD FOR TOTAL ROTATION (in CID system) $ ADD5 ROTX,ROTY,ROTZ,,/ROTTOT/N1C/N2C/N3C $ $ $-------------------------------------------------------------------$ $-------------------------------------------------------------------$ $ $ GENERATE G-SIZE CORIOLIS MATRIX IN THE GLOBAL SYSTEM $ $ $ FIRST, GENERATE G-SIZE MATRIX IN THE CID SYSTEM $ ADD ROTTOT,/OMEGA/OMEGAC $ MATMOD OMEGA,SIL,,,,/OMEGAGG,/5 $ $ $-------------------------------------------------------------------$ $ $ USE THE MJJC VECTOR TO PARTITION OUT THE ZERO MASS TERMS $ ALONG WITH THE SPOINTS (WHICH CAUSE PROBLEMS FOR MATMOD OPTION 5) $ PARTN OMEGAGG,MJJC,/,,,OMEGAXX/ $ $ $-------------------------------------------------------------------$ $ $ TRANSFORM CORIOLIS ACCELERATION MATRIX TO GLOBAL SYSTEM $ FIRST, TRANSFORM FROM CID TO BASIC $ THEN , TRANSFORM FROM BASIC TO GLOBAL $ $ CHECK IF THE CSTM EXISTS. IT WILL NOT EXIT IF THE GLOBAL SYSTEM $ IS THE SAME AS THE BASIC SYSTEM. ERROR IN MATMOD. THEREFORE $ GENERATE THE IDENTITY MATRIX $ $ PARAML CSTM//'PRESENCE'////S,N,NOGTOB=0 $ IF (NOGTOB < 0) THEN DIAGONAL MJJ/GTOBY/'SQUARE'/0. $ ELSE MATMOD CSTM,SIL,BGPDT,,,/GTOBY,/5//-1 $ ENDIF $ $ PARTN GTOBY,MJJC,/,,,GTOB/ $ $ $ CHECK IF THE CID IS EQUAL TO ZERO (BASIC) $ IF (CID = 0) THEN EQUIVX GTOB/CTOG/ALWAYS $ ELSE MATMOD CSTM,SIL,BGPDT,,,/CTOBX,/5//CID $ $ $ USE THE MJJC VECTOR TO PARTITION OUT THE ZERO MASS TERMS $ ALONG WITH THE SPOINTS (WHICH CAUSE PROBLEMS FOR MATMOD OPTION 5) $ PARTN CTOBX,MJJC,/,,,CTOB/ $ MPYAD CTOB,GTOB,/CTOG/1 $ ENDIF $ $ SMPYAD CTOG,OMEGAXX,CTOG,,,/OMEGAYY/3////1 $ MERGE ,,,,OMEGAYY,MJJC,/OMEGAGGX/ $ $

$-------------------------------------------------------------------$ $ $ MULTIPLY MASS BY CORIOLIS ACCELERATION $ TO CALCULATE CORIOLIS FORCE MATRIX $ MPYAD MJJ,OMEGAGGX,/CJJ1 $ MPYAD OMEGAGGX,MJJ,/CJJ2 $ ADD CJJ1,CJJ2/BCOR $ $ $-------------------------------------------------------------------$ $-------------------------------------------------------------------$ $ $ CALCULATE CENTRIPETAL ACCELERATION MATRIX $ $ MULTIPLY MASS BY CENTRIPETAL ACCELERATION $ TO CALCULATE CETRIPETAL FORCE MATRIX $ SMPYAD OMEGAGGX,MJJ,OMEGAGGX,,,/KCEN/3////////6 $ $ RETURN $ END $ $ $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ $ echoon $ @ 38062; 2001; Cyclic Symmetry - differential stiffness, modes, tempi - UFM 434 2, SFM 3001 In a cyclic symmetry run that uses the differential stiffness (e.g., buckling and normal modes with preload), the applied loads must be uniform. To enforce this requirement, the loads must be applied to segment one only. If loads are applied to segments other than one, then the job will fail with following error messages: *** USER FATAL MESSAGE 4342 (CY1BLD) THE LOAD BULK DATA CARD GIVEN BELOW IS INCONSISTENT WITH SPECIFIED ANALYSIS HARMONICS OR SEGMENTS OR SYMMETRY TYPE. LOADCYN xxxx yyy zzz To avoid this message, remove loads from segments greater than one. Furthermore, if the preload consists of thermal loads, then the job will fail with the following error messages: *** SYSTEM FATAL MESSAGE 3001 (CYCLIC4) THE INPUT DATA BLOCK NAMED PAC AT POSITION EXIST. Avoidance: Insert the following dmap alter. compile phase1dr alter 'lcgen casers'(,1) if ( app='bkl0' and cyclic ) equivx casers/casew/-1 $ endalter @ 38070; 70.7.2; Parallel Processing - DMP - System error message at completion o f job DMP jobs may report the following error text at job completion: 4 DOES NOT

ERROR 0031-636 User requested child or EOF termination of pm_command This message does not affect the MSC.Nastran job. Avoidance: Delete the following line from <install-dir>/bin/nast707.dmp: %MSC_ARCH=aix:%export MP_NEWJOB=yes @ 38073; ; Cyclic Symmetry - RESTART, STATSUB - Inefficient Restart If a restart from SOL 114 (Cyclic static) into SOL 115 (Cyclic modes) is performed to calculate pre-stressed modes of the model, the static solution will be re-calculated in SOL 115. There is no avoidance. @ 38096; ; Miscellaneous - Accounting file, Solution type and CPU time are not written The Version 70.7 accounting program does not properly identify the solution name, CPU, or elapsed times in the account files. Avoidance: Use the following procedure to modify and install a new accounting logging program. 1) Change the current working directory to <install-dir>/msc707/util 2) Change lines 1440 through 1443 in mscact.c from if( (f04 = fopen( f04f, "r" )) == NULL ) { (void) strcat( f04f, ".f04" ); f04 = fopen( f04f, "r" ); } to (void) strcat( f04f, ".f04" ); f04 = fopen( f04f, "r" ); That is, delete the "if" statement and its matching closing brace. 3) Build a new version of mscact with the command msc707 util mscact 4) Install the new account logger program with the command cp -p mscact ../<arch>/acct chmod a+rX ../<arch>/acct where <arch> is the architecture of the current platform, i.e., aix, alpha, hpux, irix, irix64, solaris, sppux, superux, unicos, unicosc90, unicosts, or uxpv. @ 38097; 70.7; Installation - FLEXlm - /etc/inittab file is wrong. mscsetup for MSC.Nastran Version 70.7 does not set the /etc/inittab line

with the full path of the flexlm utility. If the system is subsequently rebooted FLEXlm will not run automatically. Avoidance: If FLEXlm is already installed, and you want the new version of FLEXlm, install it but do not have mscsetup set the /etc/inittab file; if FLEXlm is being installed for the first time, modify the /etc/inittab file after the installation is complete and add the full path for the <install-dir>/bin/flexlm utility. @ 38100; 70.7; Installation - No Output When installing MSC.Nastran on Intel computers, the last step in the installation asks if you would like to enter the authorization code. If you choose not to enter the code, the program does not create the authorize.dat file, but the nastxxx.rcf file will incorrectly point to the non-existant file. If you submit a Nastran job, no output files will be created. Avoidance: Either: 1) enter the authorization code when installing. 2) edit the nastxxx.rcf file to remove the "AUTH=" line 3) when you try to run MSC.Nastran, enter the following on either the command line or in the "optional keywords" box pause=yes If you do this, the command window trying to run the program will stay open and you will obtain the error message indicating that the authorize.dat file does not exist. @ 38115; 70.7; Installation - Missing library libxlsmp.a[smprt.o] IBM AIX clients that don't have the Fortran compiler installed will be unable to run MSC.Nastran V70.7.0 and V70.7.2 until the libxlsmp.a library is installed on their system or in the <install-dir>/msc707/aix directory. The following message is issued in the log file: MSC/NASTRAN started Mon Jan 03 09:46:03 EST 2000 Could not load program /tmp/vxlq/n707/msc707/aix/analysis Could not load library libxlsmp.a[smprt.o] Error was: No such file or directory Avoidance: Install the missing library. @ 38116; ; Superelements - CSUPER, PARAM,INREL - SFM 5423 (DMPYAD) If SOL 101 with inertia relief is used with external superelements, the external superelement data recovery run will fail with: *** SYSTEM FATAL MESSAGE 5423 (DMPYAD) ATTEMPT TO MULTIPLY INCOMPATIBLE MATRICES indicating that UOO and CRX are not correct. This is due to an error which results in CRX not being written to the EXTDRUNT. Avoidance: Include the following DMAP alter in the residual structure run:

compile super3 alter 339 dbview crx0=crx (where seid=0 and wildcard) $ equivx crx0/ol1m/-1 $ @ 38119; 2001; Postprocessing - SOL 105, PARAM,POST,0 - Missing Output Under PARAM,POST,0 in SOL 105, the buckling eigenvalues and eigenvectors will not be available in MSC.Patran. Avoidance: Use PARAM,POST,-1 or insert the following DMAP alter: compile super3 alter 'CASEdr,ol2s'(,-1) app=app1 $ alter 'CASEdr,ol2s' app='reig' endalter @ 38123; 2001; Optimization/Design Sensitivity - DEQATN, DRESP2 - Wrong Answers If multiple DRESP2's refer to the same DEQATN, only the first (lowest numbered DRESP2 is checked to see if the number of DRESP2 arguments agrees with the number DEQATN arguments. If a subsequent DRESP2 has a mismatch with the number of arguments, the error goes undetected and unpredictable answers are produced. Avoidance: There is no avoidance other than not sharing a multiple DRESP2's. If this is too tedious, the user must checking that the number of arguments on the DRESP2 agree of arguments on the left hand side of the DEQATN in every the lowest numbered one. DEQATN id on be vigilant in with the number DRESP2 except

@ 38127; 2001; Optimization/Design Sensitivity - DRESP2, DSAPRT - Incorrect Sen sitivity Data When the formatted prints of sensitivity data (case control command DSAPRT) for DRESP2 responses list a frequency value of the response, the value has been incorrectly scaled upward by a factor of ten. A similar error exists for time values. Also, values that are greater than 100. appear as *******. Avoidance: Frequency/time values for DRESP1 responses are printed correctly. The user must be aware that printed DRESP2 frequency/time values are 10 times greater than they should be. It should also be noted that when a DRESP2 response spans several frequency, only the lowest frequency is printed. @ 38137; ; SSSALTERs - errors in alter1ga.vxx and alter9ga.vxx SSSALTERs alter1ga.vxx an dalter9ga.vxx have errors in them when using OTM (Output Transformation Matrices) with OUTPUT2 or in frequency response. alter1ga.vxx does not correctly write the tables identifying the terms in the OTM if OUTPUT2 is used. alter9ga.vxx does not correctly use the OTM if it is used in frequency response. Avoidance: Contact MSC and request an updated version of the alters or get them from the MSC web page at:

http://www.mechsolutions.com/support/prod_support/nastran/sssalters/index html @ 38141; 70.7; Miscellaneous - OUTPUT2, Year 2000 - Incorrect Year Printout When running nastran jobs in year 2000+, on some unix machines, the output2 trailer information will print out the date incorrectly. For example, when using param,post,,-1, the trailer information for OEF1 coming out right after UIM 4114 with a job ran on January 6, 2000 will appear as follows: 1001 63 1 6 100 0 1

instead of the correct value 1001 63 1 6 0 0 1

In this case field 5 is the year. This problem is fixed in Version 70.7 Avoidance: There is no avoidance. Note that the results are not affected by the date printout. @ 38143; 2001; Documentation - SYSTEM(173) - Old QUAD4 Formulation SYSTEM(173) allows the selection of either the pre-V68 QUAD4 formulation or the V68-V70.5 formulation. This system cell is incorrectly documented in the Version 70.7 Release Guide. The documentation should state: 1: Requests pre V68 QUAD4 Formulation 2: Requests V68-V70.5 QUAD4 Formulation Likewise the Version 70.7 Release Guide documentation for the keyword OLDQ4K incorrectly states that this keyword is equivalent to setting SYSTEM(173)=2, the keyword is equivalent to setting SYSTEM(173)=1. @ 38167; ons ; Installation - nastran command: fails with multiple 70.7.X versi If V70.7.2 of MSC.Nastran is installed in the same directory as V70.7.0, then both cannot run. The intent was to allow users to use "nast7070" for V70.7.0 and "nast707" or "nast7072" for V70.7.2. This feature was broken in the V70.7.x releases. Avoidances: 1. Install V70.7.2 in a directory with a different MSC_BASE than V70.7.0 2. Use V70.7.2 and not V70.7.0 @ 38168; ; Superelements - Incorrect OUTPUT2 units for different SEs w/ PAR AM OUNIT2 If PART superelments are used, and different values are provided for PARAM,OUNIT2 in different PARTs, the program will use these values incorrectly. 1) Setting distinct PARAM, OUNIT2 values in each of the Partitioned Bulk Data Sections is ignored - only the setting for the residual structure is used and the OUNIT2 values set under the BEGIN SUPER=xxx are ignored.

2) Setting distinct PARAM, OUNIT2 values PER SUBCASE using expanded Case Control causes: A) B) C) D) All geometry for all SE to be written to the 1st OUNIT2 specified End of Data record on that 1st OUNIT2 Results for the residual are then written to that 1st OUNIT2 file The 1st OUNIT2 file is rewound and re-labeled (wiping out the geometry and SE0 results) and the correct SE results are written. E) Labeling, results and End-of-Datas are written out for the remaining SEs. There is no avoidance. @ 38180; 2001; Frequency Response - DMP>1, SOL 108, fluid-structure decks - Wr ong answers If distributed memory parallel, dmp>1, is used with a fluid-structure in SOL 108, then wrong answers result. Avoidance: Insert the following DMAP alter: compile freqrs alter '197','' if ( acsyml and adjflg<2 and not(GETSYS(np,265)>1 and GETSYS(np,197)>0 and soltyp='modal') ) then $ endalter @ 38195; 2001; Plotting - .plt File Contains Bad Date String The plot file will contain an invalid date string on any file generated after Dec 31, 1999. The error is present in all versions of MSC.Nastran, up to and including V70.7.2. On Version 70.7 the year 2000 was written simple as "0", for example 1-10-0; however other strings are possible. This error does not cause any other problems with the plot data. Avoidance: There is no known avoidance. @ 38207; 2001; Design Sensitivity/Optimization - Static Aero, METHOD - UFM 1126 for M1RR If the Case Control for a multi-subcase static aeroelastic solution in SOL 200 contains a METHOD command that varies from subcase to subcase, a FATAL error can occur in which the M1RR matrix is recomputed. Since the M1RR matrix is not declared OVRWRT, this causes UFM 1126 shown below. The error occurs when the METHOD command is the ONLY boundary difference between the two sets of subcases. *** USER FATAL MESSAGE 1126 (GNFIST) DMAP MODULE MPYAD IS ATTEMPTING TO OUTPUT DATABLOCK 201 NAME = M1RR , WHICH ALREADY EXISTS. USER ACTION: DELETE THE DATABLOCK, OR USE FILE M1RR=OVRWRT 0FATAL ERROR Avoidance: Since static aero analysis in SOL 200 is not affected by METHOD selection, remove the extraneous METHOD selection(s) from the Case Control or insert the following alter:

compile desinit alter 'case '(,-1) if ( isolapp=3 ) then $ case casers,/casem/'comm'/s,n,nskip/////// 'mpc'/'spc'/'supo' $ else $ alter 'case ' endif $ endalter @ 38221; 2001; Documentation - PSHELL, MAT2 - Missing Stress Output for Plate E lements When a plate element references a MAT2 entry with an MID larger than 99999999, it is assumed that this MAT2 entry has automatically been created from a PCOMP entry (see QRG, MAT2 entry, Remark 13). Thus, stress output will be missing unless PARAM, NOCOMPS, -1 is specified. Avoidance: Do not use MIDs of MAT2 entries larger than 99999999, or use PARAM, NOCOMPS, -1. @ 38237; 2001; Documentation - slaveout=yes, mergeresults=no - Clarify Document ation for DMP Distributed Memory Parallel executions may specify the keywords slaveout and mergeresults on the submittal line. These keywords are documented in the V70.7 Release Guide. The Release Guide should have clarified and emphasized that the non-default values of these keywords, slaveout=yes and mergeresults=no are typically only used for debug purposes. The following should also be noted: 1. If mergeresults=no is specified in a static run the results of the individual domains will not be sent back to the master and the system solution will not be obtained. 2. The keyword mergeresults has no affect on a solution 103 or 111 run. 3. In solution 108, if mergeresults=no is specified and slaveout=yes is not specified, then the results of the slave processors will be lost. 4. In solution 108, it is possible to get a through-put advantage by saving communication between the master and slaves when mergeresult=no and slaveout=yes is specified. @ 38238; 2001; Cyclic Symmetry - SFM 3001 (READ), Poor Message In SOL 115, if the stiffness and mass matrix have common null-columns the following message will be issued: *** SYSTEM FATAL MESSAGE 3001 (READ) THE INPUT DATA BLOCK IN POSITION 13 DOES NOT EXIST.

Avoidance: Make sure degrees of freedom with no stiffness and no mass are constrained. In Version 70.7 parameter AUTOSPC will do this, but in Version 70.5 AUTOSPC would not correctly constrain grids on CYJOIN entries. A better error message is implemented in Version 2001.. @ 38254; 2001; Optimization/Design Sensitivity - PBAR - UFM 5626 (DSAD17) Optimization analysis does not support item codes 18 and 19 for PBAR for complex stress. If specified the following fatal message is issued:

*** USER FATAL MESSAGE 5626 (DSAD17) ITEM CODE 18 REFERENCED ON DRESP1 ENTRY 16 WITH RTYPE = FRSTRE IS INVALID FOR ELEMENT TYPE BAR 34. USER ACTION: CHECK THE SPECIFIED ITEM CODE WITH THOSE LISTED IN APPENDIX A OF THE MSC/NASTRAN QUICK REFERENCE GUIDE. Avoidance: There is no avoidance. @ 38257; 2001; DMAP - FBS - Wrong Answers, BIOMSG: ERROR 31 The FBS module may return poor fatal diagnostics (BIOMSG: ERROR 31) or wrong answers when a left hand fbs is requested incorrectly. An example of a left hand fbs follows. The conventional (right hand) equation solved by the FBS module is: Given the square matrix A, and a collection of vectors X, find B, where A*X< = B $ FBS module A is first decomposed into its factors L and U. $ DECOMP OR DCMP modules. Typically, X and B have many fewer columns than A. There are applications where X has many more rows than A. That type of problem is most naturally posed as X*A<= B The unknown vectors X have been moved to the left hand side, hence the name LH FBS. One method to solve this problem is to transpose it. Assume, for the sake of simplification, that A is symmetric, that is, A=A'. Then A*X' = B' This form of solution requires transposing the input before solution, and the output after solution. The present lh fbs option still requires transposing the input X, but the output B need not be transposed. It also requires use of system cell 72. Tests to date prove that this option works correctly. An investigation is underway to determine what other solvers can provide LH FBS operations. Another complication results when a forward-only or backward-only solution is requested. See the DMAP Dictionary for a definition of these terms. The definition of forward and backward is somewhat ambiguous in the context of a transposed equation. The definition used in MSC.Nastran is given by example below. It is a good practice when using these advanced forms to verify that you are obtaining correct answers, at least while debugging your alter. The revised equation, and a dmap verified by testing to provide correct answers, is: $ A*X = B', or X'*A = B Decomp A/L,,//1 $ 1= CHOLESKY, A=L*L', A must be symmetric psd $ demonstrate solution with RH FBS for comparison message //'rh solution, requires two extra transposes' $ trnsp b/bt

Fbs L,,Bt/X1t///1 $ 1=fwd only Trnsp X1t/X1 $ Message //'lh solution follows.' $ Type parm,,I,n,c72 $ used to store current value of lh fbs flag C72=GETSYS(C72,72) $ capture current value PUTSYS(1,72) $ required for LH Fbs L,,Bt/X2/2//1 $ 2=LH, 1=fwd only PUTSYS(C72,72) $ reset default for lh fbs selection. Always reset, $ or wrong answers may occur in downstream FBS module or subroutine $ calls Add X1,X2/error//-1.0 $ error = X1 - X2 Norm error/screrr///s,n,maxerror $ Message //'maximum error in lh fbs. Should be binary zero'/ Maxerror $ There have been reports of errors when using system cell 74, a function described in the Numerical User's Guide. It is recommended that dmaps that use this feature be checked for correct answers. There are alternate procedures where one can avoid the use of this system cell by tranposing the input and/or ouput set of vectors seen by the FBS module. @ 38263; 2001.0.1; Superelements - Multi-Master, Data Recovery - Missing Output In a dynamic response analysis using the mulit-master technique, no output will be generated for the superelements. This error only occurs in Version 70.5, 70.6, and 70.7. Avoidance: Insert the following dmap alter. compile super3 alter 'equivx cases \/casedr1','' equivx casedr/casedr1/-1 $ endalter @ 38268; 70.7.3; Transient Response - TRLG - Loss of Accuracy Transient loads are computed in single precision in module TRLG. One effect that has been observed is the accumulated time after many delta-T steps is in error due to numeric underflow. This round-off error typically only occurs if the duration, T is greater than 100000 times delta-T. Loads such as COS(W*T) may appear to have a slight frequency shift. This problem is more likely to occur on 32-bit computers than on 64-bit computers. @ 38276; 2001; Documentation/QRG - XYPLOT - XMIN Incorrectly Described In the Quick Reference Guide, remark 2 on the XMIN (XYPLOT) command incorrectly states that if XMIN is not specified, it is set to the lowest value of "Y". This should state that it is set to the lowest value of "X". @ 38282; ; Elements - TETRA, SYSTEM(213), DISTORT - UFM 4659, 5271 Using NASTRAN system 213=1 to override the element distortion fatal

message, USER FATAL MESSAGE 4659, for TETRA elements does not work. Avoidance: Set a high value for the maximum acceptable aspect ratio on system(191). @ 38283; le

; Parallel Processing - DMP - SGI Only, arrayservices not availab When running distributed parallel (DMP) on the SGI, the job may fail with the following error messages in the log file: 'MPI: asgetnetinfo_array('(null)') failed : array services not available S L A V E 1 Avoidance: Change the authentication method in the /usr/lib/array/arrayd.auth from AUTHENTICATION NOREMOTE to AUTHENTICATION NONE and make sure you start the array services with '/usr/etc/arrayd start' as root. Furthermore, make sure that the user account can do a 'rsh shu01irx date' on the shu01irx machine without running into a 'permission denied' problem.

@ 38293; 2001; Aeroelastic Analysis - Superelements, SPLINE, AMP- SFM 4276 EC 1044 QOPEN In Version 70.7, a flutter analysis (SOL 145) with superelements may fail with SFM 4276, EC 1044 (QOPEN) in the AMP module if the superelement reduction is applied using separate subcases AND the separate subcases do not include the FMETHOD bulk data entry .The problem occurs because the splining to upstream superelements, combined with work in Version 70.6 to support multiple aerodynamic symmetries has coupled the spline matrix data to both the aerodynamic configuration (AECONFIG) and to the superelement/boundary condition qualifier paths. To ensure that the correct linkage among configurations and boundary conditions is performed, the Case Control is examined for aerodynamic cases to determine the AECONFIG name of each subcase. In the case of superelement reduction subcases, these fictitious cases are examined and, since the FMETHOD entry is not present, they are ignored as aerodynamic cases. As a result, the spline matrix reduction to the superelement was not done. Thus, in AMP, these data are missing which results in this error. Avoidance: Insert the following DMAP alter: compile phase1a alter 'AEMODEL.*CASES,EDT','' dbview case0=cases where (seid=0 and wildcard) $ AEMODEL CASE0,EDT// S,N,NSKIP/S,N,LPFLG/ S,N,AECONFIG/S,N,SYMXY/S,N,SYMXZ $ compile phase1dr alter 'call phase1b'(,-1) dbview case0=cases where (seid=0 and wildcard) $ equivx case0/case00/-1 $

call dbstore case00,,,,//0/0/' '/0 $ compile phase1b alter 'AEMODEL.*CASES,EDT','' call dbfetch /case00,,,,/0/0/0/0/0 $ AEMODEL CASE00,EDT// S,N,NSKPAE/S,N,LPFLG/S,N,AECONFIG/S,N,SYMXY/S,N,SYMXZ $ endalter or make sure that all subcases in the SOL 145 superelement run contain the FMETHOD selector or that the FMETHOD appears above the first subcase. The splining to upstream superelements, combined with work in V70.6 to support multiple aerodynamic symmetries has coupled the spline matrix data to both the aerodynamic configuration (AECONFIG) and to the superelement/boundary condition qualifier paths. To ensure that the correct linkage among configurations and boundary conditions is performed, the Case Control is examined for aerodynamic cases to determine the AECONFIG name of each subcase. In the case of superelement reduction subcases, these fictitious cases are examined and, since the FMETHOD entry is not present, they are ignored as aerodynamic cases. As a result, the spline matrix reduction to the superelement was not done. Thus, in AMP, these data are missing which results in this error @ 38295; 2001; Superelements - RESTART, SEBULK, CSUPER - Wrong answers or fatal If external superelements are used in a solution, then a change is made to the external superelement (new matrices generated) for use in a RESTART run, the updated matrices will not be used. Avoidance: If you are using external superelements created using PARAM,EXTOUT,DMIGOP2, the following alter (for V70.5) will provide support for restart changes to the matrices: compile extin, list alter 'FORTIO.*EXISTS.*EXTUNIT.*IOSTAT ' if(iostat=0)then $ dbdelete datablk=(extdb), where(xname=*) $ endif $ @ 38304;

70.7; Frequency Response - STRAIN(SORT2) - Missing Output If STRAIN(SORT2) output is requested in a frequency response run and there are no other output requests, there will be no strain output produced. Avoidance: Add any other (SORT2) output request - for example DISP(SORT2,PLOT)=ALL - to the case control.

@ 38306;

2001; Dynamic Analyis - DAREA, LSEQ - Unpredictable Answers Non-unique Set-ID/DAREA-ID pairs in LSEQ data lead to unpredictable results. This is because the user cannot be sure as to which of the static loads specified as part of the non-unique data are actually applied and which ones are ignored. Accordingly, this scenario must be regarded as a fatal error. The user can avoid this situation by using and referencing the combination LOAD bulk data instead of non-unique Set-ID/DAREA-ID pairs.

38308;

; Heat Transfer Analysis - CHBDYE, QUAD4 - UFM 6285 If a heat transfer analysis attempts to calculate view factor from the edge of plate elements using CHBDYE with sides from 2 to 5, then you will get UFM 6285 Avoidance: Define the line segment with CHBDYP, LINE boundary elements.

@ 38317; range

2001; ACCESS - Utilites error forttl:severe(25) record number outside After a successful build, the following error may occur during execution of any of the sample ACCESS utilities.. The error may occur for both v70.5 and v70.7. "forttl:severe (25) record number outside range, unit 40, file c:\filename.xdb" Avoidance: Download new dbio object libraries, libdbio_vc.lib and libdbio_irc.lib, from MSC's ftp site.

@ 38323;

2001; Utilities - BUILDSSS - UFM 701 PRCMPL In Version 70.7, when using the buildsss utility to create a new delivery database with more than one added user dmap, the following error occurs: *** USER FATAL MESSAGE 701 (PRCMPL) THE NUMBER OF SUBDMAPS BEING COMPILED EXCEEDS THE MAXIMUM ALLOWED NUMBER OF 200 FOR THIS VERSION Avoidance: Compile the additional dmaps in a separate run and add them to the delivery. Contact MSC Support for details on how to perform this.

@ 38337; ; Submittal - Large physical memory (>8GB) may cause failure of na stran command Machines with physical memory in excess of 8GB may cause the nastran command to fail with a message similar to: MSC/NASTRAN V70.7 (SUN SPARC SunOS 5.7) Wed Jan 26 14:23:19 2000 *** SYSTEM INFORMATION MESSAGE (pgm: nast707, fn: validate_buffsize_memory) Assuming "memory=-2147483648" as specified by "memorymaximum=0.8*physical". In some instances, an error will not occur, instead the nastran command may generate an incorrect memory request. Avoidance: Set the maximum memory size in a system RC file to the value listed in the "MSC.Nastran V70.7 Configuration and Operations Guide", Table 4-2, pg 90. For example, on Solaris: memorymaximum=4GB @ 38349; 2001; P-Elements - UFM 9050 DCMP (70.5) or UFM 3001 VECRMA (70.7) It is [possible that in a model containing both p- and h-elements that the

run may fail with either UFM 9050 or UFM 3001. Avoidance: This problem is under investigation. The model demonstrating this problem executed successfully when more memory was allocated or the BUFFSIZE was increased. @ 38365; 2001; Elements - PARAM,SNORM - Wrong Answers The maximum allowed value of the SNORM parameter is 45 degrees. Higer values are automatically set lower. For shells meeting at a large angle, such as a right angle, with a different number of elements on each side, the calculated normal may be greater than 45 degrees from one side. This is legitimate, but will fail the error check. @ 38373; 2001; Rigid Elements and Multipoint Constraint - RBEi, MPC - UFM 5289 If a dof is listed as depednt more than once on an MPC or RBEi entry, Nastran will issue UFM 5289, indicating an error. Unfortunately, the message may point to the wrong dof and may not point to the entry which has the error. Avoidance: There is no avoidance other than inspecting the MPC and rigid element entries. @ 38376; 2001; Static Analysis - DMP>1, GPSTRESS and PARAM,POST,0 - Missing out put in xdb If grid point stresses are requested with dmp>1 and param,post,0 then the xdb file will be missing the grid point stresses for the slave domains. Avoidance: Insert the following DMAP alter: compile sedrcvr alter 'GPLS,EGPSTR'(2),'' DBC OPG1,OUGV1,OEF1X,OES1X,OQG1,GPL,EGPSTR,EGPSF,GPDCT, ELDCT,error1,,,,,,,,,// 'OPG'/'OUG'/'OEF'/'OES'/'OQG'/'GPL'/'GPS'/'SVF'/ 'GPDCT'/'ELDCT'/'perror'////////// -1/DBCPATH/S,N,CP/APP1/ICYCLIC/GEOMU/LOADU/POSTU/ DBCDIAG/DBCCONV/DBCOVWRT/DESCYCLE/////adptindx/lusets $ endalter @ 38378; 2001; DMAP - MATMOD option 12 - Wrong answers, Missing Fatal Message If the first input, I1, to the MATMOD module, under option 12, is purged then the module will incorrectly report that NONULL=0 and the O1 output is purged without warning or fatal even if the second and third inputs (I2 and I3) are not purged and contain null columns. The documentation for MATMOD option 12 contained in the "DMAP Module and Data Blocks" guide states that only I2 and I3 may be purged, which implies that a fatal message should be issued if I1 is purged. Avoidance: Rearrange the inputs so that the first input is not purged. @ 38381; 2001; Optimization/Design Sensitivity - DRESP1 - UFM 6482 from DOPR3R USER FATAL MESSAGE 6482 indicates that an invoked DRESP1 entry in a SOL

200 subcase is not compatible with the ANALYSIS type for that subcase. Although the message points to a real user input error, it is possible that the identified DRESP1 type is not the type that caused the error. Avoidance: The UFM has the correct RID for the DRESP1. Use the RID and not the response type to identify the incorrect input. @ 38396; ; Superelements - NASTRAN LOCBULK=2 - SFM 4192, SFM 3062 in SEMA If the multi-master method is used on a Cray and LOCBULK is set to 2, the run may fail with SFM 4192, indicating undefined grid points. *** SYSTEM FATAL MESSAGE 4192 (SEMA) GRID POINT .... IN SUPERELEMENT BGPDT FOR DOWNSTREAM SUPERELEMENT and 0*** SYSTEM FATAL MESSAGE 3062 MODULE = SEMA FAILURE OCCURRED IN UTILITY MAPBTI CALLED IN SUBROUTINE SEMA REQUESTED EXTERNAL ID(S) NOT FOUND. Avioidance: If possible, do not use the multi-master technique, run using a different computer type, or run using LOCBULK=1 @ 38398; 2001; Optimization/Design Sensitivity - DSAPRT, DRESP1 - SFM 4276 If the fomatted sensitivity option is used and there are no retained responses associated with DRESP1 entries, but there are retained DRESP2 responses, SYSTEM FATAL MESSAGE 4276 and a traceback in dsaprt is produced. Avoidance: Use the unformatted sensitivity option or force the retention of a DRESP1 response. @ 38401; ers 2001; Restarts - Superelements - Changes in SPCs may yield Wrong answ If in a superelement restart run, changes are made to the constraints (spcs) of only one superelment, then wrong answers may result. Avoidance: Insert the following DMAP alter: compile phase0 alter 'IF ( LRESTART )'(2)' IF (KCHNG) CALL CHNGFLG //KCHNG/SEDWN/SEID/PEID $ IF (MCHNG) CALL CHNGFLG //MCHNG/SEDWN/SEID/PEID $ IF (BCHNG) CALL CHNGFLG //BCHNG/SEDWN/SEID/PEID $ IF (PCHNG) CALL CHNGFLG //PCHNG/SEDWN/SEID/PEID $ IF (K4CHNG) CALL CHNGFLG //K4CHNG/SEDWN/SEID/PEID $ endalter @ 38403; ; Postprocessing/Aeroelastic Analysis - TRANS utility does not wor k with Aero The trans utility prorgram delivered with Version 70.6 and 70.7 will not convert a solution 145 xdb file to a ndb file. If executed the utility yields a message similar to the following: . 0 IS NOT DEFINED IN

TRANS - V70.7 (23-NOV-1998) XDB FILE: mlgo_flutter2_test.xdb NEUTRAL FILE: mlgo_flutter2_test.ndb MSC DATA BASE TRANSMIT PROGRAM - VERSION 70.7COMPILE DATE: 30-AUG-1999 due: [22] Invalid argument logical unit 42, named 'fort.42' lately: reading direct unformatted external IO Abort (core dumped) Avoidance: The only avoidance is to execute the run on an ieee compatible system where the post-processing will be performed so that trans is not necessary. @ 38411; 2001; Optimization/Design Sensitivity - DVPREL2, APRCOD, DOM9 - SFM 42 76, EC1404 If an input file for optimization has DVPREL2 entries and APRCOD is set to 3 on the DOPTPRM entry, the run may fail and issue SYSTEM FATAL MESSAGE 4276 ERROR CODE 1404 in the DOM9 module. . Avoidance: 1) If the file contains DVPREL2 entries, do not set APRCOD to 3 or 2) If the deck contains DVPREL1 entries, force them to be retained using a DSCREEN entry. @ 38412; GL; Submittal - "pwd: cannot access parent directories" error Under not-fully-understood circumstances a small subset of C shell users may encounter the following error message: $ /nast/deliv/msc707/unix/bin/msc707 /nast/deliv/msc707/unix/bin/msc707[42]: pwd: cannot access parent directories /nast/deliv/msc707/unix/bin/msc707[97]: cd: bad argument count when submitting a nastran job. Avoidance: One avoidance is to use the Korn shell (ksh) as the login shell. @ 38428; ; Aeroelastic Analysis - PAERO4, AEFACT - UFM 6171 (APD4) If a PAERO4 entry has fewer strips that requested, the program will issue USER FATAL MESSAGE 6171 (APD4) and terminate. Unfortunately, the fatal message will incorrectly state that the error is on an AEFACT entry with the same id as the PAERO4. Avoidance: @ 38443; Provide the correct number of strips on the PAERO4 entry.

2001; Documentation - DEQATN - Unexpected Results The documentation for the DEQATN entry does not mention that the arithmetic is carried out using the type of the input data. For example, if a term is raised to a fractional power:

X=A**(1/2) since both values in the exponent are integers, the value returned for the exponent is calculated using integer arithmetic or 1/2=0. Avoidance: Excersize caution when writing equations. If possible, avoid integer terms. @ 38449; 2001; GPFORCE - MPCFORCE, TEMP(LOAD) - Missing APP-LOAD Output If there are thermal loads and mpcforces, GPFORCE output may be missing the applied load (APP-LOAD) output. Avoidance: There is no avoidance. @ 38455; y on HP-UX ; Documentation - Configuration & Operations Guide - Maximum Memor

Tables 4-2 and C-7 of the "Configuration and Operations Guide" incorrectly state the maximum memory size on HP-UX systems is 2 GB. The maximum memory size on HP-UX systems is: 1 GB (256 MW) on HP-UX 10.20 2 GB (512 MW) on HP-UX 11.00, PA-RISC 2.0 The actual value usable by the "memory" keyword is reduced from this by approximately 35 MB. @ 38458; wers ; Nonlinear Analysis - QUAD4 - Plastic In-Plane Bending, Wrong Ans If a nonlinear analysis is performed with QUAD4 elements with elastic, perfectly plastic material and the elements are loaded in pure shear and twise (no bending, no axial load), then the results may incorrectly indicate that the model can sustain increasing loads even after the elements in the critical load path have reached the inplane shear plastic limit. This problem only occurs for coarse meshes subjected to pure shear loads. It is not likely to affect real world models.

@ 38468; 70.7.3; Miscellaneous - GP4, Possibly Other Modules - Infinite Runtime Sort routines in nastran have an upper limit that may be exceeded if models have many of one type of element. The deck that encountered this error had several hundred thousand rigid elements. Avoidance: There is no avoidance. The limit has been removed in MSC.Nastran V2001. @ 38478; 2001; Documentation/QRG - CONNECT, PBARL, PBEAML - Non-Existant Develo per's Guide The MSC/NASTRAN Quick Reference Guide documentation for the FMS CONNECT statement, remarks :

5. Custom beam cross-section evaluator libraries developed by users should comply with the MSC/NASTRAN Beam Cross-section Evaluator Developer's Guide. There is no such document. Avoidance: The remark should instead, refer to: MSC/NASTRAN V69 Release Guide, Section 3.1 Beam Cross-Section Library Appendix C: Adding Your Own Beam Cross Section Library @ 38501; 2001; Optimization/Design Sensitivity - Flutter, Superelements - UFM 1 126 for FPHP USER FATAL MESSAGE 1126 occurs if a flutter analysis is attempted with superelements in SOL 200 and complex eigenvector output is requested. See also CSR 37130. Avoidance: Remove the eigenvector request or make the following dmap changes: compile fea alter 1 type db fphp endalter @ 38502; ; Optimiztion/Design Sensitivity - DEQATN - SFM 233 (PRSPOP) In SOL 200, if the DEQATN bulk data entry uses certain keywords such as "OR" as variable names in the equation argument list, an error message similar to the following will occur: *** SYSTEM FATAL MESSAGE 233 (PRSPOP) UNDERFLOW OCCURRED DURING PROCESSING OF OPERAND STACK DURING DEQATN EQUATION PROCESSING. PROGRAMMER INFORMATION: STACK POINTER = 0 WITH MINIMUM = 1 *** COMPILATION DIAGNOSTIC INFORMATION *** 1. STATEMENT TYPE: DEQATN EQUATION 2. EQUATION ID NUMBER: 14 3. EQUATION: TWALL(OR,IR)=OR-IR *** SYSTEM FATAL MESSAGE 233 (PRSPOP) The error message needs to be improved to add the following: Possible cause: variable name is a keyword such as "OR". Avoidance: Change the variable name to something different such as "ORR". @ 38519; 2001; Documentation/QRG - PACABS - Units The Quick Reference Guide description for the PACABS entry should state that the units of the B, K, and M fields are per unit area. @

38539;

2001; Distributed Memory Parallel - DMP, DISDCMP - Infinite CPU Loop MSC.Nastran may get caught in an infinite cpu loop when distributed memory parallel is used with inertia relief. Avoidance: Do not use ineria relief with distributed memory parallel in V70.7.

@ 38542;

2001; Miscellaneous - VECPLOT - SWM 3062, Insufficient Memory Module VECPLOT does not check for insufficient memory. This may lead to the following unclear message: 0*** SYSTEM WARNING MESSAGE 3062 MODULE = VECPLOT FAILURE OCCURRED IN UTILITY GTXYZD CALLED IN SUBROUTINE DMTRXD . PREXYZ FLAG ARGUMENT WAS SET WITH AN INCORRECT VALUE TO GET THE REQUESTED DATA RETRIEVAL. *** USER WARNING MESSAGE 3199 (VECPLOT) NON-FATAL MESSAGES MAY HAVE BEEN LOST BY ATTEMPTING TO QUEUE MORE THAN 100 MESSAGES.

@ 38547; 2001; Aeroelastic Anaysis - Multiple Double Lattice Groups yield Wrong Answers If a doublet lattice model (CAERO1 and/or CAERO2) is created in SOL 144 for use in static aeroelasticity, and the model has more than one interference group, the downwashes representing unit perturbations of the control devices (the DJX matrix) are incorrect under most circumstances. The problem is caused due to the erroneous introduction in Version 70.5 of an assumption that the coordinate system ID and the CAEROi ID are in the same order. This is true for a single interference group, but not for multiple groups. As a result, the wrong coordinate system may be used in the computation of the downwash. Avoidance: There is no avoidance other than to convert the model to have a single interference group. Note, also, that unsteady aerodynamics (SOL 145 and SOL 146) are NOT affected. @ 38555; 70.7.2; Installation - NFS Mounted - HP-UX 10.2 and 11.0 The original MSC.Nastran V70.7 (r1) CD may fail to install correctly on some HP-UX 10.20 and 11.00 systems if the installation base directory is NFS mounted from another system. The problem is due to incorrect information being returned from the df(1) utility, this should be visible from the output of the command df -k <install-dir> Avoidance: Choose one of the following avoidances: 1) Use the MSC.Nastran V70.7.2 (r2) release, that works around the HP-UX utility error. 2) Install on the computer exporting the filesystem. If this is not an HP-UX system, you can use the "Custom" installation to deselect the local

computer's binaries and select the HP-UX binaries. @ 38559; 2001; Optimization/Design Sensitivity - DSAD, DRESP2 - SFM 3007 An optimization job may fail in module DSAD with SFM 3007 when the following conditions are met: 1. The DRESP2 entry only references the DNODE term, 2. The job has more than one subcases, 3. The DESGLB entry references the DRESP2 entry with DNODE terms. The avoidance is to replace DESGLB with the DESSUB entry and place it in the last subcase. @ 38564; 70.7.2; Documentation - Linux: how to enable large files (> 2 GB) The "Configuration and Operations Guide" incorrectly states that large files are available on Linux. They were not in the version of either the Linux kernel (2.0.36) or libc (2.0) that MSC.Nastran was built with. Note: The support required for this was not built into MSC.Nastran, so using a system with support for large files will not help. Large file support will be available in a new release of MSC.Nastran Version 70.7.2 on Linux that is to have a build date and release date of June 2000. Avoidance: Use INIT and ASSIGN statements to specify multiple physical files for any DBSet that exceeds 2GB. @ 38574; 2001; Optimization/Design Sensitivity - DISPRS - Poor Performance It is possible, when performing modal frequency analysis in SOL 200, to recover results to the g-set in the DISPRS subDMAP twice with each call performing identical operations. For large problems, this can entail significant CPU resources with no benefit. Avoidance: The following dmap alter removes the redundant call: compile fea alter 'if ( psldf'(2) if (delta) then $ alter 'if (psldf'(2,4) endif $ endalter @ 38595; 2001; Composites - PCOMP, PARAM,S1 PARAM,POST,0 - Wrong Answers in xdb If stresses are to be sorted (PARAM,S1) and composite stresses are requested, then the xdb file will contain incorrect composite stresses. Avoidance: Insert the following DMAP alter: compile sedrcvr alter 'DBC OES1C','' DBC ES1C,,,,,,,,,,,,,,,,,,,// 'OESC'//////////////////// -1/DBCPATH/S,N,CP/APP1/ICYCLIC/GEOMU/LOADU/POSTU/

DBCDIAG/DBCCONV/DBCOVWRT/DESCYCLE/////adptindx $ endalter @ 38608; n 2001; Documentation/QRG - FLFACT - Remark 4 contains incorrect equatio The formula of Remark 4 on the FLFACT bulk data entry of the Quick Reference Guide V70.5 has two omissions and a typographical error. It should read as follows: F1 (FNF-FMID) (NF-i) + FNF (FNF-F1) (i-1) Fi = ----------------------------------------------------------------------(FNF-FMD) (NF-i) + (FMID-F1) (i-1) The above formual was taken from the Aeroelastic Supplement and was compared to output generated by Version 70.5. @ 38610; ect 2001; Documentation/QRG - EIGRL - System Cell 229 Documentation Incorr The documentation for SYSTEM cell 229 incorrectly indicates that if it is given a value of 1, then the scratch database usage for Lanczos will be decreased. This system cell applied only to the old Lanczos solver and will not have an impact on the current Lanczos solver in MSC.Nastran Version 70.7 (The old Lanczos solver is available by setting SYSTEM(273)=1). @ 38614; 2001; Nonlinear Analysis - PARAM NMLOOP may give poor answers Nonlinear SOL 106 with PARAM NMLOOP may give poor answers. The eigenvalues, displacements, and stresses for modes including prestresses and nonlinear geometry may be based upon matrices that do not reflect a converged solution. This error started in version 70.7 when a different version of the history file was used to construct the required matrices. The use of NLPARM method ITER-1 gives fairly good results. Avoidance: An avoidance ALTER is available from MSC to improve results. When using HEXA elements in nonlinear, include NASTRAN SYSTEM(309)=1 , see CSR 38899. @ 38620; 2001; Superelements - DMIG - UFM 327 DMIG matrices in part superelements result in the following fatal message *** USER FATAL MESSAGE 327 (IFS2P) BAD DATA OR FORMAT OR NON-UNIQUE NAME. DMIG also for DMI

The message above occurs when two part superelements adjacent in the processing order have dmigs or dmis with the same name. Avoidance: Add a second dmig or dmi matrix. It does not need to be used anywhere, it is only necessary to have more than one dmig or dmi name. @ 38627; ; Subcases - SPC, DEFORM - UFM 9058 In static analysis, if DEFORM and SPC entries are changed between subcases the job will fail with the message:

UFM 9058 (SUPER3) - THE SOLUTION FOR THE RESIDUAL STRUCTURE DOES NOT EXIST Avoidance: Execute the subcases as two separate runs. @ 38629; Output ; Nonlinear Analysis - TEMP(LOAD), PARAM,LGDISP - Incorrect OLOAD MSC.Nastran considers the fictitious forces for thermal expansion to be loads. These are output by a case control command OLOAD. The output is for the initial geometry, even when geometric nonlinear effects are requested. Since only the OLOAD output is wrong, other items like displacements, element forces and SPCFORCEs are not affected. Other load types (like PLOAD and FORCE1) are correctly updated to the new geometry. @ 38632; cture 2001; Acoustics Analysis - KJJZ of fluid is wrong for prestressed stru An acoustic analysis of a prestressed structure yields an incorrect fluid part of KJJZ. The values are off by a factor of 2. Avoidance: Apply the following dmap alter to Version 70.5 or 70.7: compile semg alter 'add5','add5' add5 kjjl,kjjnl,kdjjnl,,/kjjz $ endalter @ 38634; 2001; SSSALTERs - trncompa.v707, SORT1 - Missing Output If sort1 output is requested while using the trncompa sssalter, no ply output will be obtained. Avoidance: Only request sort2 output. @ 38645;

70.7; Plotting - XYPLOT - Missing Data Points An XYPLOT of grid points versus temperature results in a heat transfer analysis results in missing data points. Avoidance: There is no avoidance.

@ 38653;

2001; Mass Matrix - SYSTEM(303), CUTOFF - UFM 17(LOOKUP) The MSC.Nastran Version 70.7 Release Guide states that the name for SYSTEM(303) is CUTOFF. If SYSTEM CUTOFF=1 is specified, the program will issue: USER FATAL MESSAGE 17 (LOOKUP) SYMBOL "CUTOFF" IS UNDEFINED Avoidance: Use the system cell number rather than the name.

@ 38654;

2001; Superelements/Nonlinear Analysis - PARAM,POST,0 - SWM 6065 (DBC)

If superelements are used in SOL 106 and PARAM,POST,0 is used, it is possible that SWM 6065 may be issued, followed by SWM 6066, indicating that the DBC module has been terminated. If this occurs, no additional information will be written to the .xdb file for post-processing. Avoidance: Insert the following DMAP alter into the file: compile super3 $ remove ol2s from DBC list alter 'dbc.*casedr.*ol2s.*casecc.*oltype','' DBC CASEDR,,,,,,,,,,,,,,,,,,,// 'CASECC'//////////////////// -1/DBCPATH/S,N,CP/APP/ICYCLIC/GEOMU/LOADU/POSTU/ DBCDIAG/DBCCONV/DBCOVWRT/DESCYCLE/////ADPTINDX $ @ 38676; 70.7.2; Acoustic Analysis - Missing Panel Participation Factor Output Prior to Version 70.7.2 fluid/structure problems that have no s-set (no constraints) will not allow panel participation factors to be printed with PARAM,FLUIDMP or STRUCTMP. Avoidance: Use AUTOSPC and an unattached GRID to force a degree of freedom in the s-set. @ 38682;

2001; Nonlinear Analysis - NMLOOP - Possible Wrong Answers The use of NMLOOP yields incorrect prestressed modes in Version 70.7 if the user does not enforce a stiffness update. The prestressed modes are correct in Version 70.5 because the stiffness is updated automatically at the end of the subcase. Avoidance: There are three possible avoidances in V70.7, we recommend the first option. 1. Insert the following dmap alter:

ECHOOFF $ SWITCH ESTNL, REPLACE MODULE EMAKFR WITH SUBDMAP CENTRIP COMPILE NLSTATIC $ VER 70.7 BY R.L.H. MARCH 29, 2000 ALTER 'NMLOOP.*CNMLOOP=2','' $ IF ( NEWP>-1 AND NMLOOP<>0 ) THEN CNMLOOP=2 $ MESSAGE //' DELETE/COPY ESTNL $ CSR 38614' DELETE /MESTNL,,,, $ COPY ESTNLH/MESTNL $ ENDIF COMPILE NLKR $ ALTER 'EMAKFR','' $ TYPE DB,EQEXINS $ CALL CENTRIP SLT,BGPDTS,EQEXINS,CSTMS,SILS,MJJ/KFRFGG $ COMPILE SEKDR $ ALTER 'EMAKFR','' $ CALL CENTRIP SLT,BGPDTS,EQEXINS,CSTMS,SILS,MJJ/KFRFGG $ COMPILE CENTRIP NOREF NOLIST $ SUBDMAP CENTRIP SLT,BGPDT,EQEXIN,CSTM,SIL,MJJ/KCEN/ $ $ $********************************************************************

$ * $ This DMAP sequence calculates Coriolis and Centripetal effect * $ matrices in global coordinates. * $ * $ For node i: * $ * $ [Bcoriolis] = 2[x]m * $ * $ [Kcentripetal] = [x][x]m * $ * $ 0 -Omega(z) Omega(y) * $ [x] = Omega(z) 0 -Omega(x) * $ -Omega(y) Omega(x) 0 * $ * $******************************************************************** $ TYPE PARM,,I,N,SLTRECNO $ TYPE PARM,,RS,N,OMEGA $ TYPE PARM,,CS,N,N1C,N2C,N3C $ TYPE PARM,,CS,N,OMEGAC $ $ $ SEARCH FOR RFORCE ENTRY $ SLTRECNO= 1 DO WHILE ( SLTRECNO > 0 ) $ $ PARAML SLT//'DTI'/S,N,SLTRECNO/1//S,N,LTYPE/ $ IF ( (SLTRECNO <> -1 ) AND ( LTYPE = 10 ) ) THEN $ $ $ EXTRACT SPIN VECTOR FROM RFORCE CARD. $ PARAML SLT//'DTI'/SLTRECNO/4//S,N,CID/ $ PARAML SLT//'DTI'/SLTRECNO/5/S,N,RPT/ $ PARAML SLT//'DTI'/SLTRECNO/6/S,N,N1/ $ PARAML SLT//'DTI'/SLTRECNO/7/S,N,N2/ $ PARAML SLT//'DTI'/SLTRECNO/8/S,N,N3/ $ SLTRECNO = -2 $ $ ENDIF $ $ SLTRECNO = SLTRECNO + 1 $ $ ENDDO $ $ IF ( RPT = 0. ) THEN $ $ OMEGA= 0. RETURN $ $ ENDIF $ $ $ THESE ARE NEEDED LATER ON $ MATGEN ,/I1X1/1/1/ $ MATGEN ,/I6X6/1/6/ $ $ $ CALCULATE RADIANS PER UNIT TIME $ OMEGA= 2.0 * PI(1) * RPT $ OMEGAC = CMPLX(OMEGA,0.) $

$ $ $ CREATE A COLUMN VECTOR, MJJTSUM, FROM THE DIAGONAL OF MJJD $ TO BE USED FOR PARTITIONING PURPOSES LATER $ DIAGONAL MJJ/MJJC/'COLUMN' $ $ $-------------------------------------------------------------------$ $-------------------------------------------------------------------$ $ $ CALCULATE ROTATION MATRIX IN THE CID SYSTEM $ IF (N1 <> 0.) THEN $ $ DETERMINE FACTOR FOR X-ROTATION (in CID system) $ MATGEN ,/PVC6/6/6/1/1/4 $ MATGEN ,/PVR6/6/6/2/1/3 $ MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $ MPYAD ONE6X6,I6X6,ONE6X6/ROTX/1/-1 $ DELETE /PVC6,PVR6,ONE6X6,,/ $ N1C = CMPLX(N1,0.) $ ENDIF $ $ $-------------------------------------------------------------------$ $ IF (N2 <> 0.) THEN $ $ DETERMINE FACTOR FOR Y-ROTATION (in CID system) $ MATGEN ,/PVC6/6/6/2/1/3 $ MATGEN ,/PVR6/6/6/0/1/5 $ MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $ MPYAD ONE6X6,I6X6,ONE6X6/ROTY/1/-1 $ DELETE /PVC6,PVR6,ONE6X6,,/ $ N2C = CMPLX(N2,0.) $ ENDIF $ $ $-------------------------------------------------------------------$ $ IF (N3 <> 0.) THEN $ $ DETERMINE FACTOR FOR Z-ROTATION (in CID system) $ MATGEN ,/PVC6/6/6/0/1/5 $ MATGEN ,/PVR6/6/6/1/1/4 $ MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $ MPYAD ONE6X6,I6X6,ONE6X6/ROTZ/1/-1 $ DELETE /PVC6,PVR6,ONE6X6,,/ $ N3C = CMPLX(N3,0.) $ ENDIF $ $ $-------------------------------------------------------------------$ $ $ ADD FOR TOTAL ROTATION (in CID system) $ ADD5 ROTX,ROTY,ROTZ,,/ROTTOT/N1C/N2C/N3C $

$ $-------------------------------------------------------------------$ $-------------------------------------------------------------------$ $ $ GENERATE G-SIZE CORIOLIS MATRIX IN THE GLOBAL SYSTEM $ $ $ FIRST, GENERATE G-SIZE MATRIX IN THE CID SYSTEM $ ADD ROTTOT,/OMEGA/OMEGAC $ MATMOD OMEGA,SIL,,,,/OMEGAGG,/5 $ $ $-------------------------------------------------------------------$ $ $ USE THE MJJC VECTOR TO PARTITION OUT THE ZERO MASS TERMS $ ALONG WITH THE SPOINTS (WHICH CAUSE PROBLEMS FOR MATMOD OPTION 5) $ PARTN OMEGAGG,MJJC,/,,,OMEGAXX/ $ $ $-------------------------------------------------------------------$ $ $ TRANSFORM CORIOLIS ACCELERATION MATRIX TO GLOBAL SYSTEM $ FIRST, TRANSFORM FROM CID TO BASIC $ THEN , TRANSFORM FROM BASIC TO GLOBAL $ $ CHECK IF THE CSTM EXISTS. IT WILL NOT EXIT IF THE GLOBAL SYSTEM $ IS THE SAME AS THE BASIC SYSTEM. ERROR IN MATMOD. THEREFORE $ GENERATE THE IDENTITY MATRIX $ $ PARAML CSTM//'PRESENCE'////S,N,NOGTOB=0 $ IF (NOGTOB < 0) THEN DIAGONAL MJJ/GTOBY/'SQUARE'/0. $ ELSE MATMOD CSTM,SIL,BGPDT,,,/GTOBY,/5//-1 $ ENDIF $ $ PARTN GTOBY,MJJC,/,,,GTOB/ $ $ $ CHECK IF THE CID IS EQUAL TO ZERO (BASIC) $ IF (CID = 0) THEN EQUIVX GTOB/CTOG/ALWAYS $ ELSE MATMOD CSTM,SIL,BGPDT,,,/CTOBX,/5//CID $ $ $ USE THE MJJC VECTOR TO PARTITION OUT THE ZERO MASS TERMS $ ALONG WITH THE SPOINTS (WHICH CAUSE PROBLEMS FOR MATMOD OPTION 5) $ PARTN CTOBX,MJJC,/,,,CTOB/ $ MPYAD CTOB,GTOB,/CTOG/1 $ ENDIF $ $ SMPYAD CTOG,OMEGAXX,CTOG,,,/OMEGAYY/3////1 $ MERGE ,,,,OMEGAYY,MJJC,/OMEGAGGX/ $ $ $-------------------------------------------------------------------$ $-------------------------------------------------------------------$ $ $ CALCULATE CENTRIPETAL ACCELERATION MATRIX

$ $ MULTIPLY MASS BY CENTRIPETAL ACCELERATION $ TO CALCULATE CETRIPETAL FORCE MATRIX $ SMPYAD OMEGAGGX,MJJ,OMEGAGGX,,,/KCEN/3////////6 $ $ RETURN $ END $ $ $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ $ 2. Change the iteration strategy on NLPARM to AUTO,1 to force a stiffness upon convergence. If that is too costly, apply the preload in two subcases, the first subcase with default iteration strategy, the second subcase with METHOD case control command and NLPARM with with AUTO, 1. 3. Apply the preload in a SOL 106 coldstart run and restart in SOL 103 to calculate the prestressed modes using PARAM,NMLOOP. CSR 38614 is related to this CSR. @ 38684; 2001; Documentation - PSHELL - Remark 14, (ts)10e8(ts) incorrect. Remark 14 under the PSHELL description in the MSC/NASTRAN Verson 70 and 70.5 Quick Reference Guide description is incorrect. It says: 14. If MIDi is greater than or equal to (Ts)10^8(Ts), then parameter NOCOMPS is set to +1 indicating that composite stress data recovery is desired. (MIDi greater than (Ts)10^8(Ts) are generated by PCOMP entries.) It should say: 14. If MIDi is greater than or equal to 10^8, then parameter NOCOMPS

is set to +1 indicating that composite stress data recovery is desired. (MIDi greater than 10^8 are generated by PCOMP entries.) The v70.5 online encyclopedia has the correct remark as does pre v70.0 Quick Reference Guides. @ 38722; 2001; Optimization/Design Sensitivity - PCOMP - UFM 315, UFM 316 In Version 70.7 it optimization run. the PCOMP entries using the updated will fail with the is possible to output updated propertied in an If a model contains PCOMP entries and one or more of has an odd number of layers, then any subsequent run PCOMP entries generated from the previous puch file following error messages: X X

*** USER FATAL MESSAGE 315 (IFPDRV) FORMAT ERROR ON BULK DATA ENTRY PCOMP *** USER FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA ENTRY PCOMP

Avoidance: Punch out the PCOMP entries using the small field format (e.g., param,despch1,x) where x can be -1, -2, -3, -5 -,6, or -7. The default is param,despch1,6. @ 38740; 2001; DMAP - DRHM3 corrupts OSE1 table It has been observed that on Version 70.5 SGI systems that the OES1 table produced bythe DRHM3 module is corrupted and can not be printed. @ 38743;

2001; Optimization/Design Sensitivity - SPCFORCE, DSCREEN - UFM 315 Specifying the response type (RTYPE) of SPCFORCE on the DSCREEN entry in SOL 200 will result in UFM 315 ILLEGAL DATA ON BULK DATA ENTRY DSCREEN. Note that the SPCFORCE response type was first added in V70.7. Avoidance: Create an equivalent DRESP2 for each DRESP1 of type SPCFORCE that will be constrained. Then specify the response type (RTYPE) of EQUA on the DSCREEN entry.

@ 38750;

2001; Aeroelastic Analysis - RESTARTS, MKAERO - Inefficient Processing The dynamic aeroelastic restarts are dependent on the sorted order of the MKAEROi (i=1 or 2) bulk data entries whenever multiple entries are present. This creates the situation of regenerating the aerodynamic matrices when the /,1,99999 method of deleting and reinputting the bulk data is used. The MKAEROi entries on the restart MUST be in the same order as the coldstart run otherwise SOL 145 or 146 will rerun AMG/DECOMP even though the MKAEROi entries are identical to the coldstart but in a different sorted order. Aviodance: To avoid this problem the restart should only delete the specific bulkdata entries required NOT the entire bulkdata input.

@ 38769;

; Materials - CQUAD4, MAT8, G12 - UFM 4297 In Version 70.7, if the G12 field on the MAT8 entry is blank, then the job will fail with the following misleading error messages: *** USER FATAL MESSAGE 4297 (EQD4D) QUAD4 ELEMENT WITH ID = 1 HAS AN INTERIOR ANGLE POSSIBLY GREATER THAN OR EQUAL 180 DEGREES. Avoidance: There are two possible avoidances: 1. Add the G12 field to the MAT8 entry if appropriate. or 2. Set SYSTEM(173)=2

@ 38780;

2001; SSSALTERs - mmecha.v707 - possible errors It has been observed that the mmecha alter may fail due to a singularity problem. This problem is due to an incorrect maxratio test in the presence of 2x2 pivots. This type of pivot can be caused by the Lagrange Multiplier Method.

Avoidance: The revised alter for Version 70.7 is available on the MSC.Software website. If you have trouble getting the updated alter, contact your local support representative. @ 38788; 70.5; Static Analysis - STRESS, CORNER, PUNCH - Missing Grid ID When requesting punch output for element forces or stress at the corners , the punch file does not print the grid ID. It does not know wich fields are integer and which fields are floating point Avoidance: There is no avoidance. @ 38815; ; Distributed Memory Parallel - DMP - Sigbus Error on SGI SGI distributed memory parallel jobs that request 8 or more processors may fail with a "SIGBUS: Bus error" as follows: Core from signal SIGBUS: Bus error (dbx) where > 0 _realloc(0x100193e8, 0x190, 0x19450152, 0x0, 0x110, 0x9436c5a, 0x94369bd, 0x94369d2) ["/xlv43/6.5.7m/work/irix/lib/libc/libc_n32_M4/gen/malloc.c":367,0xfb2a1f ] 1 parse_file(0x111, 0x1, 0x19450152, 0xfb51ef8, 0x110, 0x9436c5a, 0x45,0x94369d2) ["/home/poppy08/kak1/mpt/1.3/mpi/lib/libxmpi/libxmpi_n32_M3/all/all_parse c":264,0x9406214] 2 parse_global_opts(0x9425388, 0x1, 0x19450152, 0xfb51ef8, 0x110, 0x9436c5a, Avoidance: The following changes need to be made to the nast707.dmp file: From: %MSC_ARCH=irix64:% print -nu3 -- "${hosts[$i]} -np 1 ${rcmds[$i]} -c$i -d%r.debug% jid=$jid $args" To: %MSC_ARCH=irix64:% print -u3 -- "#" %MSC_ARCH=irix64:% print -nu3 -- "mpirun ${hosts[$i]} -np 1 ${rcmds[$i]} -c$i -d%r.debug% jid=$jid $args" From: %MSC_ARCH=irix64:%mpirun -f $cmdfile 2>%out%.%j.unique%.mpirun To: %MSC_ARCH=irix64:%chmod a+x $cmdfile %MSC_ARCH=irix64:%$cmdfile 2>%out%.%j.unique%.mpirun @ 38839; 2001; Superelements - SEBULK, Collectors - SWM 3001 (READ) If component modes are requested on a collector superelement then the run will fail with SWM 3001 in READ. Avoidance: Insert the following DMAP alter: compile phase0 alter 'call ifps1' if ( setype='colltr' ) equivx endalter @ 38841;

dynamic/dynamics/-1 $

70.7; Installation - Poor Error Message

If MSC.Nastran is NOT installed on a remote node properly, then the following error may occur: *** USER WARNING MESSAGE (pgm: nastran, fn: validate_dmp_hosts) beowulf21(LAM n1): No response from MSC.Nastran remote query: /scratch/msc/bin/nast707 -d0 -rN70.7 frs=/scratch/DMP/dmp_108.dat drwx=/tmp fw=/scratch/DMP/dmp_108.T6084_15 drwx=/scratch/DMP drwx=dmp_108 drwx=. Segmentation fault (core dumped) Avoidance: Install MSC.Nastran correctly on the remote node. If there is not an authorization file, then the following error may occur: *** USER FATAL MESSAGE (pgm: nastran, fn: validate_authorize) authorize="???????" (LM_LICENSE_FILE) This file does not exist. Avoidance: Install the an authorizatoin file. @ 38869; t be 0 ; Documentation -- DMAP Module Dictionary - APPEND 1st param canno Both the Version 68 DMAP Module Dictionary and the current Version 70.7 online pdf documentation indicate that the 1st parameter for the APPEND module may be either a "0" or a "1" in order to create an output datablock [IN1:IN2]. However, if "0" is used then "UFM 5005 INVALID VALUE OF OPTION PARAMETER DETECTED 0" is issued, followed by "SFM 3007 MODULE = APPEND ILLEGAL INPUT TO SUBROUTINE APPEND". Avoidance: Set the 1st parameter to "1" instead of "0". @ 38889; 2001; GPFORCE - TEMP(LOAD) - Wrong Answers The applied load contribution of the GPFORCE output is incorrect in TEMP(LOAD) is applied to only a portion of the model. Avoidance: The only avoidance is to put a small thermal load on ever element. . @ 38892; 2001; Optimization/Design Sensitivity - SYSTEM(2), DOPTPRM - Undocumen ted Feature If the IPRINT parameter on the DOPTPRM entry is used to obtain intermediate prints from the DOT optimizer, the output is always written to unit 6 regardless of which unit other "f06" output is directed to by NASTRAN SYSTEM(2). Avoidance: Undocumented paramter IWRITE can be used on the DOPTPRM entry to direct the intermediate prints to the unit of the users choice. That is IWRITE=4 places this output in the .f04 file. @ 38895; 70.7.2; DMAP - PARAML - U1 Set not Recognized Degree of freedom Set U1 is not recognized by DMAP modules dealing with

degree of freedom sets, e.g. PARAML, VEC, etc. Avoidance: The only avoidance is not to use the U1 set. @ 38898; 70.7.3; Performance - SYSTEM(107) - Poor Performance If Parallel processing is selected with SYSTEM(107) (a.k.a. keyword PARALLEL) on MSC.Nastran Version 70.7 on the Cray J-90 executable cache memory will be disabled and poor performance results. Avoidance: Other than not selecting parallel there is no avoidance. @ 38899; ; Nonlinear Analysis - HEXA - Poor Convergence Poor convergence may be experienced for Nonlinear jobs with HEXA elements in MSC.Nastran Version 70.7. The problem occurs because the new HEXA formulation introduced for linear analysis in Version 70.7 is used on the first iteration and it shouldn't be. Avoidance: Set SYSTEM(309)=1 to revert to the pre-Version 70.7 formulation for all iterations. @ 38907; ; Complex Eigenvalue Analysis - CEAD - UWM 5411 (REIGL8) The message below may appear many times when using the complex eigensolver module CEAD. *** USER WARNING MESSAGE 5411 (REIGL8) NEGATIVE TERM ON DIAGONAL OF MASS(VIBRATION) OR STIFFNESS(BUCKLING) MATRIX DETECTED. ROW NUMBER = 80 VALUE = -1.20000D+02 Avoidance: This message may be safely ignored when encountered in the CEAD module. @ 38925; 2001; Nonlinear Static Analysis - slideline contact, PELAST - SFM 3001 in EMG Job will fail in SOL 106 when line contact problem includes nonlinear springs in the model. For example, any SLIDE-LINE model with PELAST and/or PBUSHT, which refer to TABLEDi bulk data cards, will be terminated during contact nonlinear stiffness generation in SOL 106. Avoidance: Please insert the following DMAP alter into SOL 106 decks to avoid this difficulty COMPILE NLSTATIC $ ALTER 'CALL NLKR'(3),'' $ CALL NLKR MESTNL ,CSTMS BGPECT ,BGPDTH , DEQATN ,DEQIND / BKELM ,BKDICT , S,NOBKGG/-1 NOSSET /NOOSET

,BTOPSTF ,SILS ,GPSNTS ,ETT

, ,USETNL ,USET1 ,

, ,GMNL , , ,BKAA / /MPCF1 /0

, , , , , / /

,BKGG ,BKFS , ,BKMG /K6ROT /LANGLES /'CONTACT'/0

NOOSETX /LOADFAC UNSYMF $ ALTER 'CALL NLKR'(4),'' $ CALL NLKR MESTNL ,CSTMS BGPECT ,BGPDTH , DEQATN ,DEQIND / BKELM ,BKDICT , S,NOBKGG/-1 NOSSET /NOOSET NOOSETX /LOADFAC UNSYMF $ @ 38927; ATRIX

/LSETID /PLSETID /FOLLOWK / ,BTOPSTF ,SILS ,GPSNTS ,ETT , ,USETNL ,USET1 , , ,GMNL , , ,BKAA / /MPCF1 /0 /FOLLOWK , , , , , / / /

,BKGG ,BKFS , ,BKMG /K6ROT /LANGLES /'CONTACT'/0 /LSETID /PLSETID

2001; Elements - CTETRA, PSOLID - Poor Answers, UWM 4683 (READ) MASS M The mass matrices for mid-side-noded CTETRAs are not calculated correctly if the defaults values for IN and ISOP fields on the PSOLID are not used. This may lead to UWM 4683 from the READ module and poor answers. Avoidance: Leave the IN and ISOP fields blank on the fluid PSOLIDs in order to use the defaults.

@ 38939;

; Licensing - FlexLm - Problem with FlexLm on Windows 2000 In some instances, the authorization process on NT and Windows 2000 may take almost a minute to return with the authorization. If this is happening, add the following line to your .rcf file SYMBOL=LM_NO_SNMP=1

@ 38949;

2001; Submittal - hosts keyword on nast707 command is ignored If the hosts keyword is defined in any of the RC files, then you will be unable to override it's value on the command line. Avoidance: A work around is to put the keyword in a local RC file in the directory where the input file resides. This value will override any values set in the other RC files.

@ 38951; 2001; Submittal - Execution on remote node may cause input and output to be deleted. If a Nastran analysis is run on a remote system using the the node keyword, and the output file directory is visable to the remote node, then all the output files and the input files will be deleted after the job has run. Avoidance: There are two avodiances. 1) Run the job in a directory that is not visable to the remote node 2) Make the following change near the end of the install_dir/bin/nast707.rmt file. 127,128c127

< %s.rsh% %node% -l %username% /bin/rm -f "%r.out%*" < %r.outvis=y:%/bin/rm -f "%r.out%*" --> %r.outvis=n:%%s.rsh% %node% -l %username% /bin/rm -f "%r.out%*" @ 38955; 2001; DMAP - PARAML - Overflow Exit, Floating Point Exception When the PARAML module is used to extract an element from a matrix with the DMI option, and the element is larger than the range allowed by single precision parameters, a machine-dependent overflow error exit occurs. On NT MSC/NASTRAN, the output from the f06 file is: 0THE NUMBER OF NON-ZERO TERMS IN THE DENSEST COLUMN = 1 0THE DENSITY OF THIS MATRIX IS 100.00 PERCENT. MAIN: The FPU has been reset after floating point exception. MAIN: For reference, A(MAINAL) = 405290, A(/SYSTEM/) = 1777f60 [similar data is presented in the .log file] There is no mention in any of the diagnostic output that the PARAML module was being executed when the overflow occurs. One diagnostic method is to temporarily put a message statement in front of every suspected module call where the overflow may have occured, and isolate the offending dmap statement in that manner. Avoidance: An avoidance in Version 70.7, once the statement is isolated, is to divide the offending matrix by 1.e30, then multiply the output from paraml by the same number, but with all parameters at this point converted to double precision. In Version 2001 the scalar module will provide a double precision output parameter that avoids this problem @ 38960;

; Documentation - QRG, PELAST, PBUSHT - Unexpected Answers If the initial slope of the nonlinear force-displacement relationship defined on a PELAST or PBUSHT entry in a nonlinear solution does not match the corresponding Ki value on the PELAS or PBUSH entry, then the results may be unpredictable. Avoidance: Make sure that the initial slope as defined on the TABLEDi entry matches the corresponding Ki value on the PELAS or PBUSH entry.

@ 39031;

; Documentation - DMAP Module Dictionary - MATMOD Option 5, Form 3 Example 3 under the DMAP Module Dictionary's description of MATMOD option 5 should be modified from Transform KGGBASIC coordinate system 10. MATMOD TRNSP SMPYAD CSTM,SIL,,,,/TRANS10,/5//10 $ TRANS10/TRANST10T $ TRANS10,KGGBASIC,TRANS10T,,,/KGG10/3 $ to Transform KGGBASIC to coordinate system 10.

MATMOD TRNSP SMPYAD

CSTM,SIL,,,,/TRANS10,/5//10 $ TRANS10/TRANST10T $ TRANST10T,KGGBASIC,TRANS10,,,/KGG10/3 $

@ 39041; 2001; Optimization/Design Sensitivity - PARAM,NOCOMPS -1 not supported on DRESP1 PARAM NOCOMPS -1 is used to recover shell element stresses for an equivalent homogeneous element. The use of this capability has been blocked in SOL 200 so that if the user invokes the equivalent PSHELL on the DRESP1 entry to obtain stress/strain responses of the element, the corresponding PSHELL will not be found. Avoidance: There is no avoidance. Version 2001 will remove this restriction. @ 39046;

; File Management Section - DBDICT - Missing Dictionary If the DBDICT FMS command is used in a run as follows, the dictionary will be blank. assign master='oldrun.MASTER' DBDICT endjob Avoidance: Use a read-only restart similar to the following to list the database dictionary. assign run1='oldrun.MASTER' restart, logical=run1 dbdict sol 101 cend begin bulk enddata

@ 39056; 2001.0.1; Grid Point Stress - GPSTRESS- 4276 (FREAD) If GPSTRESS output is requested, then following error may occur. SYSTEM FATAL MESSAGE 4276 (FREAD). Avoidance: Ensure that case control requests for stress and force specify the same type of output. If center and grid point output is specified for one, it should also be specified for the other. @ 39067; 2001; Distributed Memory Parallel - DMP, READ - Poor Performance Distributed Memory Parallel eigenvalue jobs may experience poor performance in READ on Compaq Alpha clusters. Avoidance: There is no avoidance. @ 39088; ; Cyclic Modes - SFM 3001 (DCMP)

The following message may occur in a cyclic symmetry modal run (SOL 115) when requesting a tridiagonal solution method such as AHOU or AGIV. The message is output from the DCMP module if the USET and SIL tables are mismatched with the input matrices. *** SYSTEM FATAL MESSAGE 3001 (DCMP) THE INPUT DATA BLOCK IN POSITION 3 DOES NOT EXIST.

Avoidance: An avoidance is to use the Lanczos method which does not perform this operation. Another avoidance is to use the DECOMP module instead: Compile moders list noref $ Alter 'dcmp.*kww1'(,-1) $ If (cyclic) then $ Decomp kww1/lww1,, $ Else $ Alter 'dcmp.*kww1' $ Endif $

@ 39098;

2001; Documentation/QRG - PARAM,USETPRT and PARAM,SETSTR1 In Version 70.7 param,usetsel has been changed such that values greater than 0. are not acceptable. If USETSEL is set greater than 0 the following message occurs: ^^^ USER WARNING MESSAGE 9109 (SEKR0) ^^^ USETSEL IS GREATER THAN ZERO WHICH IS OBSOLETE. USETSEL WILL BE RESET TO -1 TO PRINT ALL SETS. ^^^ USER ACTION: SPECIFY PARAM,USETSEL,0 AND SET NAMES ON PARAM,USETSTR1 THRU USETSTR4. Avoidance: The syntax for the r USETSTRi paramerter is: PARAM,USETSTR1,M:R for more info see the description under TABPRT in the Release Guide for Version 70.7.

@ 39109;

; Eigenvalue Analysis - GIV, MGIV, HOU, MHOU - Poor Answers When using methods GIV, HOU, or their variants, if there are repeated roots their eigenvectors may not be orthogonal. A small test problem with 9 repeated roots produced only three orthognal vectors. Avoidance: An avoidance is to use the LAN method.

@ 39125;

2001; Elements - CRAC3D, MAT9 - UFM 3005 If a model has CRAC3D elements that referene MAT9 entries, , then the job may fail with the following error messages: **** USER FATAL MESSAGE 3005 (SNGERR) Attemp to operate on singular matrix AI in subroutine SD92ND. Avoidance: There is no known avoidance.

39129;

70.7; DMAP - DIAGONAL - UFM 5250 DIAGNL Negative non-integer exponent If the diagonal terms of a matrix are raised to a negative power using the DIAGONAL command with the 'SQUARE' option, if there are negative off-diagonal terms, the program will fail with: UFM 5250 DIAGNL Negative non-integer exponent Avoidance: First extract the diagonal terms using a POWER of 1.0, then raise the resulting matrix to the desired negative power.

@ 39150; 2001; Documentation - DSAPRT, TRIM, STABDER, - Formatted Output not Su pported The formatted sensitivity prints produced using the DSAPRT case control command are not supported for aeroelastic responses (DRESP1 RTYPE = TRIM, STABDER or FLUTTER Avoidance: Use the unformatted sensitivity prints. @ 39168; ; DMAP - Access Violation Error A dmap had a purged input data block of the form call xyz a,,c . . . It in turn called another dmap, with the format call wyz b// $, where b is the purged input. If an attempt was made to test for the presence of b with paraml or to print it with matprn, an access violation occurs. If all inputs to a subdmap are purged, an executive system error occurs. Avoidance: An avoidance is to add another non-purged datablock to the calling argument. This purged argument does not need to be used in the subdmap. Its presence avoids the error that results when all inputs to the subdmap are purged

MAIN: "Access violation" (C0000005) exception encountered. @ 39172; 2001; DMAP - paraml /dbname/'name' - Datablock is discarded If the paraml option 'name' is used on a data block in the calling argument of a subdmap, and the datablock used as input to paraml is an NDDL-scratch data block that has previously been equivalenced, the data block is discarded at the end of the subdmap. Avoidance: There is no known avoidance. @ 39191; xist

2001; Aeroelastic Analysis - SDR1 - UFM 4574, SWM 3001 UXAX Does not e Aeroelastic analyses that result in no a-set elastic deformations for unit perturbations of the controllers will result in the following error

message: *** USER FATAL MESSAGE 4574 (SDR1) DATA RECOVERY STARTED BUT NO DISPLACEMENT VECTORS AVAILABLE *** SYSTEM WARNING MESSAGE 3001 (SDR1) THE INPUT DATA BLOCK NAMED UXAX AT POSITION 3 DOES NOT EXIST. The problem is that UXAX is purged because there are no elastic deformations. This is a somewhat rare event, that can happen if you attempt to model a "rigid" airplane. There are a number of ways in which a rigid airplane can be better modelled. In particular, in the TRIM specification, you can set the AEQR field to 0.0. That will result in a flexible vehicle with no elastic feedback to modify the aero loads. This is the typical "rigid" behavior that is desired. However, there is a DMAP avoidance that can be used: compile aeopu alter 'mpyad *gdka.*uxax'(,-1) type parm,,i,n,nou $ DBSTATUS UXAX//S,N,NOU $ CHECK IF THERE ARE RESTRAINED SOL'NS if ( nou > 0 ) then $ alter 'sdr1 *uset.*uxax.*uergx' endif $ nou This alter will avoid the problem and still supply the rigid and elastic increment forces/moments onto the aerodynamic and structural meshes for unit solutions and trim. These data can be placed on an XDB file (PARAM, POST, 0) and visualized in the "Results" tool of MSC.Patran 8.1.. @ 39198; 2001; DMAP - MATGPR - UWM 3059 If the 2nd parameter is unspecified on the MATGPR DMAP module, then UWM 3059 is issued. This warning can be safely ignored. Avoidance: Specify the 2nd parameter. @ 39203; 2001; Nonlinear Analysis - PARAM,POST - Missing Postprocessing Output PARAM,POST does not work correctly in the Case Control of a Version 70.7 Nonlinear run. Avoidance: Insert the following DMAP alter; COMPILE NLSTATIC alter 'PVT PVT,CASESX2','' dbstatus casesx2//s,n,nocase2 $ if ( nocase2>0 ) PVT PVT,CASESX2/ $ endalter @ 39208; 2001; Plotting - XYPLOT, XYTRAN - Missing Plot Output If a non-existant grid point is requested by the XYPLOT command, the XYTRAN module makes a premature exit. i.e. when a displacement plot command cannot find the grid point, all of the acceleration and element stress/force plots could be skipped. @ 39213; 2001; Dynamic Anaysis - FRLG - Wrong Answers.

In MSC.Nastran it has been found that FRLG can produce incorrect loads when insufficient memory is given. Avoidance: If the answers are suspect, run the analysis at only the suspected problem frequencies and see if the answers change. If the answers change increase memory and rerun the job. @ 39215; 2001; Mass Matrix - UFM 4646, Grids and Components not Identified The indefinite mass matrix check may not correctly identify the grid point and degree of freedom. AN INDEFINITE MASS MATRIX DOF DETECTED AT GRID: 0 COMPONENT: 0

Avoidance: The run may be allowed to continue by setting system(303). MSC's elastic elements do not produce an indefinite mass matrix. Massless mechanisms may cause this error, dmig input, or improperly coded rigid elements. @ 39219; ; Complex Eigenvalue Analysis - EIGC - Missing Eigenvalues In MSC.Nastran Version 70.7 the new Block Complex Lanczos Eigenvalue analysis method automatically reverts to real operations when no damping matrix exists. This may cause roots to be missed in acoustic/fluid problems. Avoidance: Two possible avoidances are: 1. Introduce a false damping matrix by applying the following alter: compile gma alter 'acsyml\=','' acsyml=(acsym='yes' and fs) $ endalter 2. Force the method to always use complex arithmetic by setting system(108)=16384. @ 39224; 2001; Optimiztion/Design Sensitivity - PBEAM - UFM 7055 When an optimization task tries to design K1, K2 and other quantities specified on the last two continuations of a PBEAM entry, then USER FATAL MESSAGE 7055 (DOPR1C/DOPR1Z) will be printed. Avoidance: Do not design K1, K2 and others on the last two continuations of the PBEAM entry. @ 39235; 2001; Optimization/Design Sensitivity - UFM 3001 (DSAL) In a SOL 200 senstivity analysis with multiple frequency subcases, an error can result if an initial subcase has no responses that require a direct (as opposed to adjoint) sensitivity analysis and a subsequent subcase does. Avoidance: A possible avoidance is to arrange the subcases so that any that the initial subcase requires a direct analysis. Another avoidance is to force the use of the direct senstivity calculation for all subcsases with PARAM AUTOADJ NO.

@ 39248; 193

2001; Optimization/Design Sensitivity - DVPREL1, PBEAML, PBARL - UFM 7 The above UFM message can be issued in error when there are both PBARL and PBEAML entries being designed. The error does not occur when a single PBARL is designed. Avoidance: The only avoidance is to replace either the designed PBARL's or the PBEAML's with PBAR's of PBEAM's and use the original dvprel1/dvprel2 techniques to design these properties.

@ 39269; 3021

2001; Optimization/Design Sensitivity - DRESP2 - SFM 4027 (WRITE), EC An optimization task may fail with the SYSTEM FATAL MESSAGE 4027 when the following two conditions are met: 1) a DRESP1 response is referenced in the first subcase and 2) a DRESP2 response is retained in the second subcase where the DRESP2 response references to another DRESP2 entry. Avoidance: Do not use a DRESP2 referencing a DRESP2.

@ 39272;

2001; Optimization/Design Sensitivity - DRESP2 - Wrong Answers An optimization task may generate incorrect type 2 response values when the nested referenced DRESP2 entries are used in multiple subcases. Avoidance: There is no avoidance.

@ 39273; 2001.0.1; Elements - HEXA, Corner Stresses, SDR2 - Possible Wrong Answer s In Version 70.7, if the HEXA element is highly distorted, then the corner stresses may be incorrect. The element center stresses are correct. Avoidances: There are two possible avoidances. 1. Use the version 70.5 HEXA formulation by adding nastran system(309)=1 to the top of the input file 2. Do not use the corner stress values. @ 39275; GREAD) 2001; Optimization/Design Sensitivity - MODTRK, FREQ3, 4, 5- SFM4276 ( In SOL 200, SYSTEM FATAL MESSAGE 4276 may be issued if mode tracking is requested and there is a modal frequency subcase that uses the FREQ3, FREQ4, or FREQ5 entries. in the specification of the excitation frequencies. Avoidance: The use of FREQ3, 4 or 5 is not recommended with sensitivity and optimization analysis since the excitation frequencies are not updated when the structural model is changed. Further, the sensitivity

calculations ignore the effect of the change on the excitation frequency on the design response and are therefore inaccurate, at best, and may cause the design to go in the wrong direction. @ 39280; 2001; Documentation - QRG - SET description under OUTPUT(POST) Missing Keywords The SET description on page 244 of the V70.5 QRG for OUTPUT(POST) is the missing keywords [ELEMENTS, GRID POINTS]. @ 39285; NIT

; Optimization/Design Sensitivity - PARAM EIGD, DSAPRT(EXPORT), IU SUBDMAP EIGVECDS does not contain the parameter iunit. Thus, when the DSAPRT(EXPORT) Case Control command is used, OUTPUT4 fails because IUNIT is undefined. Avoidance: Apply the following alter: INIT MASTER(S) assign output4='a200.op4' unit=21 formatted delete $ ID MSC, SWRO TIME 100 DIAG 5,6,8,56 SOL 200 $need to rewind iunit before anything else malter 'file.*append' output4 //-1/iunit $ $ need to modify output2 from rewind to do nothing malter 'output2.*dscmcol','output2.*dscmcol' OUTPUT2 DSCMCOL,,,,//0/IUNIT/'DSCMCOL'/OMAXR $ compile eigvecds,nolist alter 'nddl' TYPE PARM,NDDL,I,Y,iunit $need iunit id subdmap eigvecds alter 'output4' message // 'iunit='/iunit $ alter 'output4'(2,0),'output4'(2,0) $ remove rewind of iunit $exit

@ 39288;

2001; Elements - CTETRA, PSOLID - UFM 9050 (SEKRRS) If a CTETRA with midside nodes references a PSOLID that has the IN field equal to "TWO", the following error message occurs: ^^^ USER FATAL MESSAGE 9050 (SEKRRS) ^^^ RUN TERMINATED DUE TO EXCESSIVE PIVOT RATIOS IN MATRIX KLL. ^^^ USER ACTION: CONSTRAIN MECHANISMS WITH SPCI OR SUPORTI ENTRIES OR SPECIFY PARAM,BAILOUT,-1 TO CONTINUE THE RUN WITH MECHANISMS. NASTRAN should give a warning message saying that IN = "TWO" is not allowed and that the suggested IN = "THREE" is used instead, and the run should continue.

Avoidance: Use IN equal to "THREE". @ 39294; 2001; Optimization/Design Sensitivity - DRESP2, DNODE - Incorrect Sens itivities When an optimzation task includes a DRESP2 entry which also references a DNODE and another DRESP2 entry, incorrect sensitivity coeffcients for the DRESP2 entry may be produced. Avoidance: Do not reference both DNODE and DRESP2 on a DRESP2 entry. @ 39304; ; Cyclic Normal Mode/Plotting - Incorrect Results in XDB Cyclic Symmetry Modal Analysis produces incorrect XDB outputs. Several calculated modes are truncated due to an apparent error in counting logic.The printed f06 outputs are correct. @ 39309; 2001; P-elements - GP0 - GMLOAD applied to GMCURV for CBEAMs For p-elements, if a GMLOAD is applied to a GMCURV in order to load CBEAMs, a fatal error with no additional information occurs in GP0. Avoidance: The avoidance is to apply the GMLOAD to the individual FEEDGEs for the CBEAMs. @ 39314; ; Nonlinear Analysis - UIM 9005, Saved LOOPID is not correct Even when an increment in nonlinear statics (SOL 106) cannot converge, it still prints the following message, where "n" is the loopid of the unconverged increment: ^^^ USER INFORMATION MESSAGE 9005 (NLSTATIC) ^^^ THE SOLUTION FOR LOOPID= n IS SAVED FOR RESTART This message is not true since increment "n" has not been saved for restart. Avoidance: Ignore the message since only increments up to n-1 have been saved. @ 39326; ; Database - SFM 1041 (GALLEX) A limit in the database directory size may cause large jobs to fail with the message SYSTEM FATAL MESSAGE 1041 (GALLEX). Avoidance: This limitation can be overcome by increasing the buffsize and/or setting system(158)=1000 and system(159)=500 to allow for a larger expansion of the database. @ 39342; 2001; DMAP - DEFUSET - Entry Ignored without Warning The DEFUSET function was removed in Version 70.7. Any such bulk data entries are now ignored, without warning. Avoidance: There is no known avoidance. @

39375;

; FMS - ASSIGN, scr=yes - Database Not Deleted In Version 70.7, if an ASSIGN statement is specified for the database, the database will not be deleted at the end of the job even if scr=yes is specified on the command line. Avoidances: There are two possible avoidances. 1. Add the following FMS statement to the input file init master(s)

2. Add a ",temp" to the end of the assign statement. @ 39378; ; Beam Library - CONNECT - Incorrect Properties for Multiple Beam Servers When more than one beam server is used, it may produce incorrect the properties. This error occurs whe the Group names or evaluator class names show certain pattern in which the first portion of the n-th group name is identical to the name of the (n-1)-th group where n is => 2. In addition the CONNECT entry including (n-1)-th group must precede the CONNECT entry containing n-th group name. For example, the CONNECT entries given here will produce the incorrect properties for the beamsrv1 group: CONNECT BEAMEVAL beamsrv locbml and CONNECT BEAMEVAL beamsrv1 locbmls. Avoidance: Do not use the problematic pattern as the group name for multiple beam servers. @ 39386; 2001; Documentation/QRG - PARAM,S1 - Incorrect Documentation for SOL 1 09 and 112 Stress sorting was implemented in SOL's 109 and 112 in Version 69. The QRG incorrectly shows that the associated parameters are not supported for these solutions. . @ 39387; 2001; Optimization/Design Sensitivity - DRESP2 - SFM 3007 DOM9 An optimization task may fail with the 'SFM 3007 DOM9, SUBROUTINE DOM91' message when the DRESP2 response references another DRESP2 entry. Avoidance: 1) Convert the offending DRESP2 entries into a single DRESP2 entry or 2) Use the following dmap alter: For v70.7 COMPILE DESCON alter ' DSAD ' DBSTATUS RR2IDR//S,N,NORR2ID $ IF ( NORR2ID>0 ) THEN PARAML RR2IDR//'TRAILER'/1/S,N,NRR2 $ TYPE PARM,,I,N,NRR2I NRR2I = NRR2 PARAML RR2IDR//'DTI'/1/ S,N,NRR2I//S,N,R2ID $ IF (R2ID = 0) THEN MODTRL RR2IDR //0 $ ENDIF ENDIF

@ 39391; 2001; Optimization/Design Sensitivity - EIGN, FREQ, FLUTTER - Wrong Se nsitivities Incorrect sensitivities can be produced in SOL 200 when all of the following pertain: a. There are superelements b. A global response is selected as the objective c. The response is not WEIGHT or VOLUME. An example of responses are EIGN, LAMA, FREQ and FLUTTER. Correct answers are provided when the response is constrained. It is only when it is invoked by the DESOBJ case control command that incomplete sensitivity information is provided. @ 39421; ; Nonlinear Analysis - GPFORCE, ESE - Loadstep Not Identified The GPFORCE and ESE output in SOL 106 does not identify the currrent LOADSTEP. Avoidance: The loadstep can only be determined by the order of output.. @ 39422; 2001; Distributed Memory Parallel - Execution fails with large value o f "dmp" Nastran runs with large values of the "dmp" parameter will fail in the setup of the run. The exact value that is allowed is a function of various factors like the size of the current path name. @ 39426; 2001.0.1; Optimization/Design Sensitivity - Multiple Eigenvalue Subcases - Missing Output If two eigenvalue subcases with different boundary conditions are run in SOL 200, the second subcases eigenvalues are printed twice but the first is not printed. The eigenvector results are printed correctly and the optimzation is performed correctly. Avoidance: Setting NASPRT=1 will produce correct eigenvalue results at each design iteration, including the last. Version 70.5 can also be used to get the correct prints. @ 39433; 2001; Inertia Relief - PARAM,INREL,-2 - UWM 4698 When a matrix to be decomposed requires 2 by 2 pivots, its rows may be identified as having negative factor terms by maxratio and USER WARNING MESSAGE 4698 may be issued. 2 by 2 pivots most commonly occur when using Lagrange Multiplier techniques (LMT). The max ratio test is not valid for this circumstance, so this warning message may be ignored. LMT methods are used in sol 101 for param, inrel, -2. @ 39436; 2001; Aeroelastic Analysis - PARAM,POST,0 - Complex Eigenvectors Missi ng from xdb

If complex eigenvalues/vectors are extracted using the pk method, they cannot be written to the xdb file (PARAM POST 0) due to an error in the datablock labelling Avoidance: The op2 file (PARAM POST -1) can be used to display complex eigenvectors in MSC.Patran @ 39439; 2001; Inertia Relief - PARAM,INREL,-2 - Poor Performance When the automated suport capability for inertia relief is selected (param, inrel, -2) the problem may run many times slower than when other methods of inertia relief are used. Avoidance: A dmap avoidance is given in the sssalter library with the name ir2a.v707. @ 39440; 2001; Optimization/Design Sensitivity - PCOMP - Wrong Answers If any thickness of the various plies on a PCOMP entry is defined as a design variable, Z0 is blank, and a DVPREL1 or DVPREL2 has not been used to define ZO, then wrong answers may occur. This error occurs because the the initial Z0 is used for the entire design optimization task even though Z0 changes as the ply thicknesses change.. Thus, sensitivities and optimization results are wrong. Avoidance: A DVPREL1 or DVPREL2 entry must be used to define Z0 ( Z0 =-0.5 times the element thickness). Otherwise, the initial Z0 will be used for the entire design optimization task even though Z0 changes as the ply thicknesses change.. Thus, sensitivities and optimization results are wrong for Z0=blank without DVPREL1 or DVPREL2 definition. @ 39449; g 2001; Documentation/Plotting - CTETRA iem codes for nonlinear are wron The Stress-Strain Item Codes, in Appendix A, of nonlinear HEXA, PENTA and TETRA are wrong in the current Quick Reference Guide, The errors and corrections are listed as follows: 1) The last row of Nonlinear CHEXA (Code=93), "19-146 Items 3 through 17 Repeated for 8 corners", should be changed to "19-146 Items 3 through 18 Repeated for 8 corners ". 2) The last row of Nonlinear CPENTA(Code=91), "19-146 Items 3 through 17 Repeated for 8 corners", should be changed to "19-114 Items 3 through 18 Repeated for6 corners". 3) The last row of Nonlinear CTETRA(Code=85), "19-146 Items 3 through 17 Repeated for 8 corners", should be changed to "19- 82 Items 3 through 18 Repeated for4 corners". @ 39452; 2001; External Superelements - ASET, USET, Residual Vectors - Possible Wrong Answers If an explicit ASET and a user USET (i.e. U1, U2, U3, U4, U5, or U6) are specified, then the USET table is corrupted. The ASET will be reduced by the dof which are common with the user set (u1-u6). and wrong answers may occur.

@ 39473; 2001; Nonlinear Analysis/Hyperelastic Elements - PLSOLID, MATHP, PCOMP - SFM 3003 When hyperelastic material properties (PLSOLID and MATHP) are used together with composites (PCOMP) or the alternate form of beam property (PBCOMP) in a nonlinear analysis, Nastran fails in IFP6 if MATHP refers to any TABLES1 entry. Avoidances: This error can be avoided if user specifies the material constants instead of using tables. @ 39482; ; Distributed Memory Parallel - DMP - Misleading f04 mesages When running jobs using DMP (distributed memory parallel) with more than 2 cpus, the f04 file may only show the execution time of 2 cpus. The log file does have the correct information. The results are correct. Avoidance: Refer to the log file. @ 39488; ; Performance - DECOMP - Solver sequencer causes poor performance The automatic sequencing options in the DCMP and READ solver modules go from a grid point based sequencing to a DOF sequencing under some circumstances. Avoidance: TAN 4593 describes the reasons for this transition, and methods to avoid the transition. It has been observed that DOF sequencing can be more than 10 times slower than grid point sequencing. @ 39490; ; Nonlinear Transient Analysis - NLTRD2 - SFM 4276 UNPACK It is possible that a nonlinear transient analysis may fail with BIOMSG 33, followed by SYTEM FATAL MESSAGE 4276, indicating an UNPACK error. Avoidance: The exact cause of this problem is under investigation. One potential cause is epoints. Replacing epoints with spoints may solve the problem. @ 39502; ; Nonlinear Heat Transfer - RADBC - UFM 4551 (NCONVG) If a nonlinear heat transfer analysis models radiation across a gap using the RADBC entry the job is likely to have trouble converging and will issue the message: *** USER FATAL MESSAGE 4551 (NCONVG) *** STOPPED PROBLEM DUE TO FAILED CONVERGENCE Avoidance: Use a CONV entry to model the radiation across the gap. Avoidance: Model radiation across gap using convection with exponent equal to 4.0 Or run the model in transient analysis until it reached steadty state solution @ 39508; 2001; Nonlinear Analysis/Hyperelastic Elements - PLSOLID, MATHP Nastran job may fail in nonlinear analysis when the specified number(s) of NA (order of polynomial function of the distortional strain energy) and/or

ND (order of polynomial function of thethe volumetic strain energy) in MATHP card are higher than the real rank(s). The real ranks are computered from user specified material constants (Aij and Di) or tables (TABLES1) in MATHP bulk data card by NASTRAN automatically. Avoidances: This error can be avoided if user controls the number of order of NA and ND - don't make them higher than the possible ranks based on the user's input of Material constants (Aij) or Tables (TABLES1) in the MATHP bulk data card. @ 39509; ; Transient Analysis - PARAM,ADSTAT, LFREQ, STATSUB - UFM 6143 If a modal transient analysis is run using PARAMs ADSTAT and LFREQ and the case control command STATSUB, then the run will fail with UFM 6143 indicating incorrect input to the MERGE module. Avoidance: Insert the following DMAP your file: compile super3 alter 'matmod.*lm1.*mass' type parm,,rs,y,lfreq if(lfreq>0.)then $ $ $ correct error if ADSTAT, LFREQ, STATSUB are all used $ this is a simple correction and will not work for the general case $ paraml freq//'trailer'/2/s,n,nmodes $ type parm,,i,n,rmmodes rmmodes=nmodes-NRUH1 $ number of modes removed by LFREQ - NOTE this alter does $ not check for HFREQ or NMODES matgen ,/lfmodes/6/nmodes/rmmodes/NRUH1 $ partitioning vector based on removed modes partn freq,,lfmodes/,nfreq,,/1 $ partn mass,,lfmodes/,nmass,,/1 $ equivx nfreq/freq/-1 $ equivx nmass/mass/-1 $ endif $ @ 39510; ; Nonlinear Analysis - Incorrect Loadstep ID after UIM 6189 If a Nonlinear analysis issues the message: *** USER INFORMATION MESSAGE 6189 (NCONVG) *** BEST ATTAINABLE SOLUTION HAS BEEN FOUND *** Then for the next load increment the 3 permitted bisections are going to take place without convergence. The displacement output for the "BEST ATTAINABLE SOLUTION" will then be labeled with a load factor that includes the increment from the final step with 3 bisections. Avoidance: Set the intermediate output (INTOUT) field 9 on the NLPARM entry to YES. @ 39556; 2001; Eigenvalue Analysis - EIGRL, READ - UFM 5401 USER FATAL MESSAGE 5401 is issued when a Lanczos job does not find any roots. Possible causes are given. One cause that is not given in the

message is a large maximum ratio indicating a modeling error. @ 39560; 2001; Optimization/Design Sensitivity - ANALYSIS=FLUTTER SFM 4276 (tbk ) at DSAPRT,DSAD A design optimization analysis may fail in the DSAD or DSAPRT modules if ANALYSIS=FLUTTER *** SYSTEM FATAL MESSAGE 4276 (tbk) ERROR CODE 7777 PID= 0 *** USER INFORMATION MESSAGE 4276 (tbk) @ 39572; 2001; Normal Modes - SOL 103 - Bad Eigenvalues Bad eigenvalues may occur on the Version 70.7.2 NEC system with a build date of December 15, 1999. This error has been corrected in Version 70.7.2 on the R3 CD. The error was fixed without changing the Version. The build date of the corrected executable is June 28, 2000. Avoidance: There is no avoidance. @ 39609; ; Miscellaneous - SGI ffio may result in a traceback The FFIO feature may cause problems on some SGI operating systems. The jobs may stop wih a traceback in ffopen. Receiving signal number 5: Trace/BPT trap Avoidance; Setting ff_io_opts = 'casecon.DB* casecon.db* casecon*SCR* casecon*scr* may solve the problem. Else turn off FFIO. @ 39612; rix ; DMAP - ORTHOG Module Incorrectly Declares a Matrix to a Null Mat The ORTHOG module may incorrectly declare a matrix with small terms to be null and the following message is then issued. *** SYSTEM WARNING MESSAGE 6467 (ORTHRD) A NULL COLUMN WAS ENCOUNTERED AT COLUMN NUMBER = ORTHOGONALIZATION @ 39624; 2001; P-Elements - MATTi - Wrong Answers If temperature-dependent material properties are being used with p-element shells,and the temperature field is not constant, the answers may be incorrect. Avoidance: There is no avoidance. @ 39630; ; Modal Frequency Response - multiple subcases, element output - M issing Output If multiple subcases exist in a modal frequency response analysis, and element-type output is requested (e.g., force) in the second or subsequent subcases without requesting it in the first subcase, then the element output will be missing. 1 DURING

Avoidance: There are two possible avoidances. The first one is recommended. 1. Add the same request in the first subcase. If the printed output is not desired, use the (plot) option. 2. Use param,ddrmm,-1. This method is not recommended as it may cause substantial additional runtime. @ 39633; 2001; Nonlinear Buckling Analysis - PARAM,BUCKLE,2 - UWM 8592 (REIG) Nonlinear buckling analysis in a cold start run was made possible in Version 70.7 by specifying param, buckle, 2. A buckling analysis should be able to be selectively perfromed for each subcase that has a method = n specified. However, if a method request is not made in each subcase the job will fail with the following message: *** USER WARNING MESSAGE 8592 (REIG) THE METHOD RECORD IN CASE CONTROL IS MISSING OR INCORRECT. THIS WILL CAUSE THE FOLLOWING FATAL ERRORS. USER ACTION: PUT A METHOD ENTRY IN CASE CONTROL WHICH POINTS TO AN EIGR, EIGRL, OR EIGB RECORD IN BULK DATA. *** USER FATAL MESSAGE 3032 (READ) UNABLE TO FIND SELECTED SET ( 0) IN TABLE (EED ) IN SUBROUTINE (REIG). 0FATAL ERROR Avoidance: Insert the following DMAP alter or make the method request above the subcase level. compile nlstatic alter 'cbuckle=-1'(2,1), ''(2,2) else if ( newp>-1 and buckle=2 ) then $ PARAML casesx2//'DTI'/-1/5//S,N,METHod $ if ( method=-1 ) cbuckle=2 $ endalter @ 39636; 2001; Aeroelastic Analysis - SPLINE - UFM 7358 In an aeroelastic analysis, USER FATAL MESSAGE 7358 may be issued in error. This has been observed to occur when the the structural grid points used by the spline are separated by a length that is much less than one. Avoidance: Recasting the dimensions of the problem so that the distances are greater than one should provide a workaround. Another alternative is to select additional gridpoints that are outboard or inboard of the selection. This error does not occur in Version 70.5 so this version could be used as an avoidance. . @ 39643; 2001; Optimization/Design Sensitivity - PCOMP, DSAD13 - Read Past End of Record In a Nastran optimization job (sol 200) the message ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD (DSAD13)

is produced with composite responses (DRESP1 RTYPE = CSTRESS,CSTRAIN or CFAILURE) if the lamina number requested in the ATTA field is greater than the number of lamina on the associated pcomp entry. Avoidance: Correct the incorrect lamina number request @ 39663; 2001; Beam Library - DOM9 - Segmentation Error A large optimization job that utilized the external beam library capability gave a segmentation error in the DOM9 module. Avoidance: Use Node Locked instead of Flexm @ 39682; 70.7; Aeroelastic Analysis - GUST, RMS - Wrong Answers In a symmetric aeroelastic analysis, if an aerodynamic panel lies on the plane of symmetry wrong GUST RMS values or a rank deficient matrix result. Avoidable: For symmetric analysis do not define panels that lie on the plane of symmetry. Version 70.7 allows aerodynamic panels to lie on the plane of symmetry by implementation of the AESYMXZ and AESYMXY commands. @ 39695; GL; Eigenvalue Analysis - PARAM,RESVEC,YES - Traceback in READ PARAM,RESVEC,YES may cause a floating point error in READ. This occurs when the number of modes plus residual vectors exceeds the rank of the mass matrix. This problem typically only occurs in very small test models where all modes may be unexpectedely calculated. Avoidance: Modify the EIGR(L) record so not all modes are computed. If there are disjoint submodels, do not get all of the modes for any submodel for which there will be residual vectors. @ 39697; M 3001 2001; SSSALTER library - pstmac2a.v707, missing eigenvalue table - SF When using the pstmac2a.v707 alter, if an eigenvalue table is not provided, then the job will fail with SFM 3001 Avoidance: There are two possible avoidances: 1. Provide an eigenvalue table or 2. Modify the following statement from lamx to lamx lamdum1,/lamat/nummodes $ @ 39705; 2001; Virtual Mass - MFLUIED, ZFS - UWM 5411 When using virtual mass, if any grid point of an element listed on an ELIST record lies very near the free surface, a nonpositive definite mass matrix may occur. This may result in the following message: ,,lamdum1/lamat $

*** USER WARNING MESSAGE 5411 (REIGL8) NEGATIVE TERM ON DIAGONAL OF MASS(VIBRATION) OR STIFFNESS(BUCKLING) MATRIX DETECTED. ROW NUMBER = 56 VALUE = -3.60212D+01 The number of modes calculated may not match the Sturm sequence output. Avoidance: Change the free surface level slightly so that no grid point lies near the free surface. Near means a distance less than one percent of the square root of the area. @ 39713; 2001; Miscellaneous - DMIG - UFM 2070 (MTRXIN) DMIG matrices may be selected by the (K2gg, M2gg, and B2gg) commands, and the (K2pp, M2pp, and B2pp) commands. If the same matrix is requested twice, for example K2gg = AA $ M2gg = AA $ or K2pp = AAX $ B2pp = AAX $ Then the following misleading fatal message is output: *** USER FATAL MESSAGE 2070 (MTRXIN) REQUESTED DMIG MATRIX AA IS UNDEFINED. Avoidance: @ 39714; Duplicate the DMIG matrix and rename.

2001; Case Control - TFL - Poor Message, BIOMSG, ERROR 3001 MTRXIN If the Case Control contains the command TFL = [sid] but there is no TFL entry in bulk data, a non-standard exit is taken. For Version 70.7 on the NT, the following exit was taken in the MTRXIN module: BIOMSG: ERROR 0. STATUS = 3001 HAS OCCURRED IN ROUTINE FREAD 0 , FILE INDEX =

Avoidance: Correct the input data. @ 39718; 2001; Case Control - MPRES - Missing Output Labels If SORT2 format output is requested with the MPRESS Case Control command, the associated "PRESSURES FOR ELEMENTS ON FLUID STRUCTURE BOUNDARY" output will have no element labeling on it. Avoidance: Request SORT1 output @ 39724; 2001; SSSALTERs - modevala.v707 -MAX Eigenvector Normalization causes Wrong Ans. SSSALTER modevala.vxx assumes that the eigenvectors are normalized to unit generalized mass. If they are not, incorrect results will occur.

Avoidance: Always normalize your eigenvectors to MASS (the default). @ 39742; 2001; Frequency Response - STATSUB - Missing Output If there is a static subcase in SOL 111 to apply a prestress followed by a dynamic subcase, and if output requests are only in the dynamic subcase no output is obtained. Avoidance: Put the output request above all subcases or insert the following DMAP alter: compile super3 alter 'if ( desopt ) then','' if ( true ) then $ endalter $ @ 39771; 2001; Optimization/Design Sensitivity - PBEAML requires SOUT=YES If a tapered PBEAML is being designed and the user has a value of NO for one of the SO stress requests, incorrect sensitivity values will be created. The optimizer will then be unable to make progress. A fatal message may be issued or the job may indicate a hard convergence at design cycle 1. Avoidance: Do not use SO(j) or SO(B) = NO. @ 39773; 2001; Optimization/Design Sensitivity - PBEAM, PBEAML - UFM 7052, 7197 Solution 200 jobs may fail with USER FATAL MESSAGE 7052 or USER FATAL MESSAGE 7197 when the field name of an intermediate section of a PBEAML is referenced on a DVPREL1 entry. Avoidance: Split a PBEAM with multiple intermediate sections into multiple CBEAM/PBEAML. For the new PBEAMLs, provide properties for end A and B only. @ 39786; 6721 2001; Documentation/QRG - Continuation Entry Replication produces UFM The "=(C)," immediate continuation format for card replication as defined on page 428 of the v70.5 Quick Reference Guide no longer works. Avoidance: Use "," instead. For instance, instead of: =(C), , *7, *7, *7, *7, *7, *7, *7 use: , , *7, *7, *7, *7, *7, *7, *7 @ 39791; 2001; DMAP - SDR2 - SFM 3001, Wrong Datablock Position Number The DMAP message: *** SYSTEM FATAL MESSAGE 3001 (SDR2) THE INPUT DATA BLOCK IN POSITION 10 DOES NOT EXIST.

may incorrectly identify the missing data block position. Avoidance: Use DBSTATUS before the offending line to verify the status of the input datablocks. @ 39806; 2001; Miscellaneous - Missing "BEGIN BULK" causes part of input to be ignored Starting with Version 70.7, if an input deck contains a "BEGINBULK" entry (no space between "BEGIN" and "BULK") and the bulk data section contains INCLUDE statements, then bulk data cards that appear BEFORE the "INCLUDE" statement will be ignored. @ 39808; ; SSSALTERs - alter1ga - QGDYN is Output Twice If MPCFORCE output is requested in a run using the alter1ga.v707 alter the run will fail with a message warning that datablock QGDYN is being output twice. Avoidance: Either remove the MPCFORCE request or request an updated version of the alter from MSC. @ 39847; 2001; Documentation - QRG - DCONADD and DCONSTR When the same DCID specified by a DESSUB command is referenced by DCONADD and DCONSTR entries, the constraints defined with the DCONSTR entry are ignored. Avoidance: Two avoidances exist: 1. Do not reference the same DCID with the DCONADD and DCONSTR entries. 2. If one insists on referencing the same DCID with both entries, include that DCID in the DCONADD entry as the DCONSTR identification number. @ 39871; ; Distributed Memory Parallel - DMP, MPCs - signal caught, mpi_ab ort mpi bug When using Distributed Memory Parallel (DMP) and there are many MPCs, then a large amount of memory is needed. If insufficient memory is allocated, then the job will fail with the following messages in the log file ---------SIGNAL CAUGHT: Signal number = 15 Signal code = -1 Signal pid = 3426 Signal uid = 9450 Signal tid = 1 Fault addr = 00000d62 FP inexact = 0x0 FP divide0 = 0x40000000 FP underflow = 0x0 FP overflow = 0x20000000 FP invalid = 0x0 ---------TRACEBACK: Avoidance: If there are large number of MPCs, use a single processor

instead of DMP. @ 39885; 2001; Materials - MAT2 - UFM 2042 (EMG) Starting with Version 70.5, if a bulk data file includes PSHELL/MAT2s with 9-digit MIDs where any of the MIDs are greater than 400,000,000 then the deck fails with the message: USER FATAL MESSAGE 2042 (EMG) MISSING MATERIAL TABLE ****** FOR ELEMENT Such PSHELLs/MAT2s are routinly constructed from the PUNCHed echo of a PCOMP entry but they must not be mixed with other PCOMP entries. The message has been improved in Version 2001. @ 39889;

2001; Distributed Memory Parallel - DMP, EMG - Job Hangs A distributed memory parallel Statics, SOL 101, job will hang if one of the tasks gets a fatal error in EMG. Avoidance: Other than correcting the input there is no avoidance.

@ 39892;

2001; DMP - Missing Library - libmpi.so not found The distributed version of MSC.Nastran Version 70.7 on linux requires a file libmpi.so.1. This file was not placed on the CDROM. There are two possible avoidances: 1. The file may be created by using the lam-6.3.1.tar.gz file which is in $MSC_BASE/msc707/linux with the following commands: configure --with-fflags=-fno-second-underscore --with-shared make install These commands create /usr/local/lam-6.3.1/lib/libmpi.so.1. Please copy this file into $MSC_BASE/msc707/linux. 2. The file may be downloaded from the MSC ftp site from: ftp://ftp.macsch.com/pub/msc-products/NASTRAN/V70.7/linux/ or: ftp.mscsoftware.com/pub/msc-products/NASTRAN/70.7/linux/

@ 39896;

2001; Eigenvalue Analysis - READ, EIGRL, RBE3 - SFM 7340 (LNNHERR) Lanczos eigenvalue problems that ran in Version 70.5 may fail in 70.7 with the message: *** SYSTEM FATAL MESSAGE 7340 (LNNHERR) process error reported by subroutine LNNP2CD Avoidance: This error may be related to the re-formulated RBE3 elements, Setting rotational dof set as independent, appears to be the cause of the problem. If this is the case, either setting:

NASTRAN SYSTEM(310)=1 or removing all rotational dof from the independent set may fix the problem. @ 39901; ; TRANS/RECEIVE - Missing libF77.so.4 It is possible that the fortran libraries cannot be found for the NASTRAN utilities 'trans' and 'receive'. This is evidenced by an error message like: ld.so.1: receive: fatal: libF77.so.4: open failed no such file or directory Avoidance: A possible solution is as follows (example for SUN SOLARIS): Ensure the following three libraries( libF77.so.4; libM77.so.2; libsunmath.so.1) exist in directory <installation-dir>/msc/msc707/solaris. Then set the environmental variable LD_LIBRARY_PATH to that directory. @ 39904; 2001; Superelement - heat transfer - UFM 6599 Superelement heat transfer works in SOL 101 by specifying However, it only works when there is no convection in the NASTRAN heat transfer applies convection by first defining element (CHBDYG or CHBDYE) and then applying CONV with the on that boundary element. For example: chbdyg,101,,area4 ,5,6,17,16 conv,101,1,,,99 However, there is a limitation in superelements that all element IDs must be unique else the following fatal message is encountered: *** USER FATAL MESSAGE 6599 (SEPUEC) DUPLICATE ELEMENT IDS CONV SUPERELEMENT 0 0FATAL ERROR Avoidance: Do not use superelements. @ 39931; 2001; DMAP - FIles Required to Create Delivery Database Missing on NT If you are using Windows NT, the files necessary to create a delivery database are not in the "del" directory. Avoidance: Contact MSC and request a copy of the necessary files @ 39942; 101 AND CHBDYG 101 IN APP HEAT. problem. a boundary same element ID

; Nonlinear Normal Modes - PARAM,FOLLOWK - Wrong Answers Using SOL 106 to perform a nonlinear static analysis of a rotating beam with followk effects leads to wrong results when the modes are calculated by restarting with SOL 103 using the NMLOOP parameter. Restarting SOL 106 from a converged solution with just the NMLOOP parameter and a

METHOD case control request yields onty the results of the nonlinear static analysis, no normal modes are computed. Avoidance: To get the modes from a restart with SOL 106, add a static subcase with the appropriate values of SUBID and LOOPID set with the NMLOOP parameter and METHOD case control command, then the nonlinear normal modes could be obtained. @ 39962; 2001; Superelements - PARAM,RESVEC - Wrong Answers When using superelements or residual vectors, there is a hard-coded limit of 999 eigenvectors in the program for any single component. If a component has more than 999 modes calculated, the program will incorrectly truncate it to 999 modes with no warning. Avoidance: Insert the following DMAP alter to remove the limit: compile resvec nolist noref alter 'read.*kzz,mzz','' paraml kzz//'trailer'/1/s,n,zzsize $ READ KZZ,MZZ,,,,,,,,,,,/ ALAMA1,APHX1,MX,OX,EX,/ 'MODES'/S,N,NV////-1/'LAN'////zzsize/0////PRTSUM $ @ 39963; 2001; Nonlinear Analysis - CELAS1 - Incorrect Sign The element force output for CELAS1 in nonlinear solutions has an incorrect sign in MSC.Nastran Version 70.7. @ 39994;

; Documentation - QRG, YTAXIS - UFM 676 If the XYPLOT command YTAXIS is used, the program will incorrectly issue UFM 676 indicating that it is not recognized. *** USER FATAL MESSAGE 676 (IFP1XY) YTAXIS IS NOT RECOGNIZED ON ABOVE CARD. Avoidance: Remove the command

@ 39995; NKs

; Documentation - DLINK - UFM 7011 (DOPR1B) with Concatenated DLI The following remark show be added to the DLINK bulk data description in the QRG. i Remark 6. If a design variable is specified as dependent on a DLINK entry, then it cannot be specified as independent on another DLINK entry.

@ 39997;

; Case Control - PARAM,INREL, REPCASE - UFM 6143 If conventional static solutions are mixed with inertia relief in the same run and REPCASEs are used, the run may fail with UFM 6143 indicating

incompatible matrices. Avoidance: Do not use REPCASEs when conventional statics and inertia relief are both being executed in the same run.. @ 40001; ; Optimization/Design Sensitivity - SDAMPING - Wrong Answers The BHH value for modal damping is computed as Bi = 2 * Mi * (omega) * (zeta) . Mi and omega can vary with the design variable and therefore affect the modal damping sensitivity. However, these perturbations are ignored in the current sensivity analysis with ANALYSIS = MFREQ. The discrepancy associated with neglecting this term is most pronounced at resonant peaks. Avoidance: Alternative forms of damping, such as PARAM G, cvisc and cdamp elements and material damping are supported and could be substituted for the modal damping input. @ 40007; 2001; Distributed Memory Parallel - SUPER - Fails in DISFBS A distributed memory parallel job with SUPER Case Control commands will fail in DISFBS. Version 2001 traps this invalid input earlier in SEQP and issues the following fatal message *** USER FATAL MESSAGE 7567 (SEQP) THE SUPER CASE CONTROL COMMAND HAS BEEN SPECIFIED. USER INFORMATION: THE SUPER COMMAND IS NOT PERMITTED WITH THE AUTOMATIC MODEL PARTITIONER. USER ACTION: CONSOLIDATE AND REWRITE THE SUBCASES WITHOUT THE SUPER COMMAND. @ 40008; 2001; Case Control - K2GG, K2PP, M2GG, M2PP, P2G - Unexpected Answers DMIG entries are selected by the K2GG, M2GG, P2G, and K2PP, M2PP case control commands. If a case control command selects a bulk data DMIG entry, and there are no DMIG or TF entries in the bulk data file, the module that reads this matrix returns without any error messages, and the run may go to completion although it is missing the selected DMIG information. Avoidance: An avoidance is to put in at least one legal DMIG matrix. The program will provide good diagnostics and a fatal error exit if any selected matrices are missing from bulk data. @ 40037; ; Heat Transfer Analysis - VIEW3D - Wrong Answers The View factor calculation may have trouble detecting elements when the elements are very close. This may lead to wrong answers. Avoidance: Decrease the ETOL error estimate from default 0.1 to a small number such as 0.0001 on the VIEW3D bulk data entry , so that the VIEW3D program will go and calculate the view factor using the contour integration method @ 40040; 2001; P-Elements - TEMP(LOAD), ADAPT - SFM 3003 When the ADAPT module is retrieving the thermal loads to calculate the error estimators, the following fatal error may occur:

*** SYSTEM FATAL MESSAGE 3003 (ADAPT) ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK ETT (FILE 109) IN SUBROUTINE GETEMP Avoidance: The error appears to occur with multiple element types. Currently the only workaround is to specify temperatures using TEMPD. @ 40077; ; Documentation - DMAP Module Dictionary - SUBDMAP NOT EXECUTED If a SUBDMAP is called with no output datablocks or parameters, then it may not execute. Symptoms will be missing output and a statement in the .f04 file "SUBNAME NOT EXECUTED." where SUBNAME is the name of the subdmap. Avoidance: Add an output block to the calling sequence, or add a parameter. @ 40088; ; SSSALTERs - mtacca - Wrong Answres SSSALTER mtacca.vxxx correctly calculates the rigid-body acceleration terms based on the initial conditions, however it does not include the effect of the initial rigid-body acceleration on the rigid-body velocity and displacement terms. Avoidance: The flexible results are correct. If the relative solution is obtained (positive value of PARAM,RBMODES), the answers are correct.. @ 40094;

2001; Transient Response Analysis - K2PP - SFM 3007 (TRD1) If a complex matrix is input using the K2PP command, a transient response run will fail with the message: *** SYSTEM FATAL MESSAGE 3007 (TRD1) ILLEGAL INPUT TO SUBROUTINE TRD1H Avoidance: Use only real matrices in transient analysis.

@ 40127;

2001; Heat Transfer - RADBND - Wrong Answers In a steady state heat transfer analysis if RADBND is used to model the radiation across a gap then wrong answers may result. Avoidance: Use RADBC to model radiation across gap instead of view factor.

@ 40133; 2001.0.1; Frequency Response - TLOAD1, TLOAD2 - Wrong Answers TLOAD1 and TLOAD2 entries define a time dependent load. In frequency response a Fourier analysis is used to transform the time-dependent load to the frequency domain. The TYPE field on the TLOAD1 and TLOAD2 entry which specifies whether the dynamic excitation is and enforced displacement, velocity, or acceleration, is ignored in frequency response. Avoidance: There is no avoidance. @

40134;

2001; Miscellaneous - ESTIMATE crashes when PROJ statement is present The estimate program fails when a PROJ file management statement is present. Avoidance: Remove the PROJ statement.

@ 40139; 2001.0.1; DMAP - MATGPR, USET - Bad DOF Labels When USET entries are present in the bulk data section that place degrees of freedom in user defined displacement sets, Form 1 (MATGPR GPL,USET,SIL, ...) usage of the MATGPR module may cause the degree of freedom output identifier code to change from the expected T1/T2/T3/R1/R2/R3 to simply the component direction number (1 thru 6). This will occur when the degree of freedom internal number associated with the grid/components defined on the USET entry is less than the number of grid points in the bulk data. External grid identification numbers and printed matrix values are correct. Only the component code identifiers may be displayed differently. The identifier (whichever one is displayed) is correct. Avoidance: A possible avoidance is to use the second form of the MATGPR module (MATGPR BGPDT,USET,,.....). @ 40141; 2001.0.1; Optimization/Design Sensitivity - Optimizer has difficulty mak ing progress It has been observed that the Method of Modified Feasible Directions (Method=1 on the DOPTPRM entry) can sometimes fail to provide a better design even though it seems that it should. This occurs in situations when there are many more active constraints than there are design variables Avoidance: The SQP algorithm (Method=3 on the DOPTPRM entry) may make progress in these situations. @ 40156; 2001; Acoustic Analysis - fluid-only models - Missing Acoustic Pressur e Output If a model contains only fluid elements then the "COMPLEX ACOUSTIC RESULTS" output will not be printed. Avoidance: Insert a dummy structural element or the following DMAP alter: compile sedrcvr alter 'fscoup' fscoup=true $ endalter @ 40157; 2001; Nonlinear Analysis - PARAM,BUCKLE,1, RESTART - SFM 6036 Nonlinear buckling analysis by a restart in SOL 106 with param, buckle, 1 will encounter SFM 6036 (divide by zero) if the number of increments for the restart subcase is different from 2 or the bisection is triggered. Avoidance: Specify INTOUT=YES on the NLPARM Bulk Data entry in the last subcase or insert the following DMAP alter: compile nlstatic

alter 'IF ( BUCKLE=2 ) POUTF=1','' if ( buckle>0 ) poutf=1 endalter @ 40158; 2001; Documentation/Rigid Elements - RBE3 - Different Answers The reformulated RBE3 element in V70.7 can result in large differences in answers between V70.5 and V70.7. This has only been observed on models where rotations have been selected on the connected grid points. The changes are most noticeable when the distances between connected grid points are much greater than 1.0. The changes in answer between versions do not occur when only translations are used for the independent dof. Avoidance: To obtain same answers as in Versions prior to 70.7 set: NASTRAN SYSTEM(310)=1 to use the old RBE3 formulation or do not use rotations on the "connected" grid points on RBE3 elements. The answers will then be identical between v70.5 and later versions. @ 40167; 2001; Optimizaton/Design Sensitivity - CBUSH - Wrong Answers on NEC In MSC.Nastran Version 70.7 the NEC system has special code to enhance the performance of design sensitivity tasks that utilize the adjoint method. This code yields wrong answers when there are no coordinate systems in the bulk data file and a CBUSH element is being designed. Avoidance: Add a dummy coordinate system. For small problems where performance is not a large concern, setting PARAM AUTOADJ NO will avoid the problem by avoiding the adjoint method. @ 40171; 2001.0.1; Eigenvalue Analysis - Maxratio Control in READ module is Indir ect. Parameter MAXRATIO is used by the DCMP module in statics analysis to control the degree of near-singularity that is allowed when computing solutions. It is input to the DCMP module with a user parameter. This parameter is not an input to the eigensolver, the READ module, although READ makes tests based on its value. If the DCMP module has not run first the value READ uses is set to 1.E7. If DCMP has run prior to READ the value input by the user for DCMP is also used for READ. Avoidance: DCMP is called in modal analysis when SUPORTs exist, when superelement reduction is used, when residual vectors are computed, and for several other reasons. However, it is not always called before the READ module, that behavior is problem dependent. Look in the .f04 file for the word DCMP before READ to determine what value has been used. If you want to change the value that the READ module uses you could schedule a small matrix to run in DCMP with dmap instructions, and set the value there. @ 40184; 2001; Optimization/Design Sensitivity Analysis - PLOADi, CBAR - Incorr ect Sensitivitie In solution 200, sensitivities are incorrect if a distributed load is applied on CBAR and/or CBEAM elements and there are more than one design variable.

Avoidance: There is no avoidance

@ 40228;

2001; Superelements - CSUPER - Floating Point Exception in SEP1X A model bringing in an external superelement may fail with a floating point exception in module SEP1X if the GRID points defined on the EXTRN entry do not exist. Avoidance: Always make sure that the GRID points on the EXTRN entry exist in the model.

@ 40257;

2001; Dynamic Analysis - PARAM,RSPECTRA, STATSUB - UWM 4568 (INTERR) Response spectra analysis does not work with prestressing (STATSUB Case Control command). If attempted the following message is issued: .*** USER WARNING MESSAGE 4568 (INTERR) THE NUMBER OF SPECTRUMS SELECTED ON THE, DLOAD CARD DOES NOT MATCH THE NUMBER OF SUPPORT POINTS., CALCULATIONS FOR THIS SUBCASE TERMINATED. Avoidance: Insert the following DMAP alter: compile semodes alter 'call postreig'(,-1) call dbstore casem,,,,//0/0/' '/0 $ compile postreig list alter 'equivx ' call dbfetch /casem,,,,/0/0/0/0/0 $ PARAML CASEm//'TRAILER'/1/S,N,ICASE1//S,N,NOCASem $ PARAML CASECC//'DTI'/1/176//S,N,SUPER $ super=-1 means super=all $ no need to append. IF ( NOCASem>-1 AND NOT(RSONLY) AND SUPER>-1 ) THEN $ APPEND CASECC,CASEM/CASESEP4/1 $ PARAML CASECC//'TRAILER'/1/S,N,ICASE $ PARAML CASECC//'TRAILER'/3/S,N,MAXL $ PARAML CASECC//'TRAILER'/4/S,N,IPLT $ PARAML CASEM//'TRAILER'/3/S,N,MAXL1 $ PARAML CASEM//'TRAILER'/4/S,N,IPLT1 $ ICASE=ICASE+ICASE1 $ MAXL=MAX(MAXL,MAXL1) $ IPLT=MAX(IPLT,IPLT1) $ MODTRL CASESEP4//ICASE/ICASE/MAXL/IPLT $ ELSE $ EQUIVX CASECC/CASESEP4/-1 $ ENDIF $ NOCASRS1>-1 alter 'sep4','' SEP4 CASEsep4,PCDB,EMAP,XYCDB,,,/DRLIST/ 'REIG'/'SEID'/S,N,NODR/S,N,SEID $ alter ' sedr ','' IF ( (NOCASem>-1 AND SEID=0) ) THEN $ EQUIVX CASEm/CASEX/-1 $ ELSE $ EQUIVX CASECC/CASEX/-1 $ ENDIF $ SEDR EMAP,CASEx,PCDB,DRLIST,XYCDB,,,

MAPSF,PHGD,BGPDTSD,GDNTAB/ PHILS,CASEDR,PCDBDR,XYCDBDR/ 'REIG'/SEID/S,N,NOUP/S,N,NOSRT1//S,N,NOOUT/ S,N,NOPLOT/S,N,NOXYPLOT/'SEID'/0 $ endalte @ 40261; ; Part Superelements - PARAM,OUNIT2 is ignored The Nastran Quick Reference Guide indicates that PARAM,OUNIT2 may be specified in either the Case Control or the Bulk Data. If it is specified in the Bulk Data after a BEGIN SUPER entry it will be ignored. Avoidance: Specify the parameter in the case control.. @ 40263; ; Transient Response Analysis - EPOINTS, PARAM,ADSTAT - UFM 6143 If a model has EPOINTs and PARAM,ADSTAT is specified on a pre-stressed structure in a modal transient response analysis, the run will fail with UFM 6143 indicating that incompatible matrices were attempted to be merged. Avoidance: Insert the following alter: compile super3 alter 'append.*freq0,mass' if(n1<>nruh1)then $ type parm,,i,n,ted1 $ ted1 = n1 + 1 $ matgen ,/vone2/6/ted1/1/n1 $ merge ,,f0m,,,,vone2/f0m1/1 $ $ else $ n1=nruh1 $ alter 'merge.*f0m,,,,vone' endif $ @ 40275; 2001; DMAP - MATMOD Option 20 - Wrong Answers If the first input to MATMOD Option 20 is purged, then wrong answers result. Avoidance: Rearrange the inputs so that the first input is never purged. @ 40276;

2001; Restarts - SCR=MINI - UFM 2071 in DPD In any dynamic solution with static loads used to define dynamic loads, if the initial run is made using SCR=MINI, the restart run may fail with the following message: *** USER FATAL MESSAGE 2071 (DPD2) will be issued indicating that the DAREA set is not defined. Avoidance: Insert a dummy DAREA or the following DMAP alter in the restart run only: compile (insert main subdmap name; e.g., sedtran for SOL 109, sedfreq for

SOL 108, semfreq for SOL 111, semtran for SOL 112) alter 'dpd ','' DPD DYNAMICS,GPLS,SILS,USET,,/GPLD,SILD,USETD, TFPOOL,,PSDL,,NLFT,TRL,EED,EQDYN/ -1/S,N,LUSETD/0/0/0//0/0/1/123/S,N,NOUE $ endalter @ 40280; ; Nonlinear Analysis - CBUSH, PBUSHT ,CBUSH1D, PBUSH1D - SFM 3003 In SOL 106, the use of CBUSH or CBUSH1D elements with a nonlinear force-deflection with a lot of zero slope around the origin may lead to the following unclear fatal message SYSTEM FATAL MESSAGE 3003 (NLITER) ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK MESTNL (FILE 208) IN SUBROUTINE NLINIT This is due to the fact that all the nonlinear stiffness is zero for this iteration. This should rarely occur-usually with single element decks. If it does add a dummy nonlinear element that will never have a zero slope as for example: $ put in a mnl $ 2 CBUSH 6666 PBUSH 333333 PBUSHT 333333 TABLED1 1212 -2.3 GRID 777777 GRID 888888 dummy element 3 4 5 6 333333 777777 888888 K .1 KN 1212 -9.-9 0. 0. 1. 0. 0. 0. 2.3 0. 0. 7 8 9 0

9.-9

ENDT 123456 123456

@ 40284;

; Postprocessing - Shock Response results not written to XDB File Running shock response analysis using Sol 103 with PARAM,POST,0 will generate an XDB file which contains only the model but no results. Avoidance: Set PARAM,POST,-1 and the results will be written to the .op2 file.

@ 40296;

2001; Optimization/Design Sensitivity - DESPCH - Incorrect Punch File When parameter DESPCH is set to a non-zero value in an optimization task, the updated property entries written in the punch file do not correspond to the current design cycle. In fact, the updated property entries for the previous design cycle will be written. Avoidance: use the following dmap alter where the original inputs of ept, geom2 and mpt are replaced by eptn, geom2n and mptn. compile desopt alter 'dom12 '(2),'dom12 '(2) DOM12 XINIT,XO,CVALO,PROPIF,PROPOF,OPTPRMG,HIS,DESTAB, geom1n,coordo,,,eptn,geom2n,mptn,epttabf,dvptabf,,,,,,,/ HISADD,NEWPRM,,NEWDES/ DESCYCLE/OBJIN/OBJOUT/S,N,CNVFLG/cvtyp/OPTEXIT// MDTRKFLG/DESPCH/DESPCH1/MODETRAK/eignfreq//

PROTYP///fsdcyc $ @ 40307; 2001; DMP - UFM 4276 EC 923, Infinite loop on master+slave(s) In a DMP (distributed memory parallel) analysis if a slave runs out of disk space then the job may hang and become caught in an infinite cpu loop.

@ 40310;

2001; Optimization/Design Sensitivity - DSAPRT If a design variable has more than 6 digits in its id, the formatted sensitivity output will replace the design variable id with ******. Avoidance: Do not use design variables with more than 6 digits in the id.

@ 40315;

; Documentation - DMAP Module Dictionary - DIAGONAL The DMAP Module Dictionary description of the DIAGONAL Module's WHOLE option is incorrect for complex matrices. In case of a complex matrix, the option WHOLE gives the following result: POWER = 1.: Magnitude of the complex matrix elements POWER != 1.: The real part of the complex matrix element is raised to the power

@ 40358;

2001; Acoustic Analysis - PACABS - Possible Unexpected Results SYNTH=-YES requests the calculation of B, K, and M from the specified tables. However if a value is given for B, K, or M on the PACABS entry, then any table input is ignored and the data in the B, K, and M fields is used. If SYNTH is specified as NO and no data is specified for B, K, or M then TIDi is used.

@ 40369;

; Aeroelastic Analysis - HESS, KE-Flutter - SFM, BIOMSG ERROR 3002 The Hessenberg eigenvalue extraction method used in the ke flutter analysis may fail to converge. When this happens, a misleading message such as BIOMSG: ERROR is produced. Avoidance: This problem has only been observed on decks where the flutter input data is questionable. If this message is encountered it is suggested that the k-method of flutter analysis be chosen combined with the complex lanczos method. 3002 HAS OCCURRED IN ROUTINE FREAD , FILE INDEX = 0

@ 40394;

2001; GPFORCE - GPFDR, DEFORM, REPCASE, TEMP - Wrong Answers Grid Point Force gives incorrect answers when thermal (TEMP) loads are combined with DEFORM loads in the same subcase.

@ 40395;

; GPFORCE - GPFDR, CQUAD8 - Wrong Answers The applied load may be incorrect for Grid Point Force output when thermal loads are applied to CQUAD8 elements. Avoidance: There is currently no avoidance.

@ 40399; NFIST)

2001; Optimization/Design Sensitivity - DRESP1,RTYPE=ESE - UFM 1126 (G If RTYPE=ESE is specified on the DRESP1 entry then the job may fail with USER FATAL MESSAGE 1126. Avoidance: compile respsen alter 'IF ( DOWTVOLD\<=0 ) ELTPRT', ''

@ 40408;

; Damping - GE, MAT8 - Possible Wrong Answer The value specified for the GE, material damping, field on the MAT8 entry is being ignored if it is referenced by a PCOMP entry. Avoidance: If the GE field is important, replace the PCOMP/MAT8 with equivalent PSHELL/MAT2/MAT1 entries.

@ 40438;

; User Modifiable Nastran - Not Enough Space The executable built by User Modifiable Nastran can not use memory larger than 115mb else the following error may be encountered: Error (errno=12) in opencore memory allocation of xxxx words MSC.NASTRAN (ocalloc.c): Not enough space This error may be caused by: Insufficient swap space. The maximum possible size was exceeded. Avoidance: This problem is caused by a typo in the link script (umnld). In this file, there is an enviroment variable MAXDATA. For non-dmp it should be set as MAXDATA=0x80000000 and for dmp should be set as MAXDATA=0x70000000 (it should have 7 zeros).

@ 40444;

2001; Optimization/Design Sensitivity - RBEi, RBAR - Wrong Answers When an optimization job includes both sizing variables, and shape variables that affects the grid locations associated with the r-type elements, the response sensitivities with resepcet to a sizing variable may be null. Avoidance: use the following dmap alter: compile desopt alter 'rgsens=' rgsens = false alter 'dopr5 ' rgsens = false compile respsen alter 'rgsens=('

rgsens = false compile pslgdv alter 'dsvgp4' rgsens = false @ 40465; ; Complex Eigenvalue - EIGC, Massless DOF - SFM 3002 In MSC.Nastran Version 70.7, if the model contains massless degrees of freedom and the old lanczos single vector method (system(108)=2) is used, then the job may fail with the following error messages: *** SYSTEM FATAL MESSAGE 3002 (CEAD) EOF ENCOUNTERED WHILE READING DATA BLOCK SCRATCH (FILE 309) IN SUBROUTINE CLASP 0FATAL ERROR Avoidance: Use the new default block lanczos method. @ 40475; 2001; Nonlinear Elements - CBUSH1D, SDR2RD - SFM 6985 A FORCE=ALL Case Control request will cause an analysis with CBUSH1D elements to fail with the message: *** SYSTEM FATAL MESSAGE 6985 (SDR2RD) PREVIOUS ELEMENT-ID = 9010 DID NOT REQUEST CORRECT NUMBER OF PHASE-1 OUTPUT WORDS. Avoidance: Request STRESS=ALL instead or both FORCE and STRESS @ 40479; 2001; DMAP - CEND, ENDALTER - UFM 601 (IFP1D) The automatic generation of a CEND entry introduced in MSC.Nastran V70.7 causes jobs with user DMAP that does not end with an ENDALTER to fail with the following message: *** USER FATAL MESSAGE 601 (IFP1D) THE KEYWORD ON THE ABOVE CARD TYPE IS ILLEGAL OR MISSPELLED. Avoidance: Input an ENDALTER statement @ 40497; sd ; Nonlinear Buckling Analysis - EIGRL - Traceback in lnnrigl, buck In a nonlinear buckling analysis, if the differential stiffnes matrix is null then a traceback may occur in the READ module. Traceback: Offset 0x000003c0 in procedure Offset 0x000005fc in procedure Offset 0x0000149c in procedure Offset 0x000028b8 in procedure Offset 0x000000f4 in procedure Offset 0x00000134 in procedure Offset 0x0000005c in procedure Offset 0x00000984 in procedure --- End of call chain --@ chkjacd_ bucksd_ lnnrigl_ reig_ xsem06_ xsemdr_ nastmon_ mainal

40503;

2001; Transient Response - TRLG - Poor Performance MPYAD picks inefficient methods. It tends to pick method 1 nt, when that is almost always the slowest method, usually requiring multiple passes when there are many time steps. An alter that avoids this error is: Compile semtran list Alter 'trlg.*ppt'(,-1) $ before type parm,,I,n,oldsys66 $ oldsys66=getsys(oldsys66,66) $ message //'oldsys66'/oldsys66 $ putsys (1,66) $ deselect 1 nt Alter 'trlg.*ppt' $ after putsys (oldsys66,66) $ restore original value

@ 40514;

2001; Nonlinear Analysis - PARAM,BUCKLE,2 and scr=yes - Wrong Answers Nonlinear buckling analysis in SOL 106 with PARAM,BUCKLE,2 may give poor answers if a scratch data base is used (scr=yes) and there are linear elements in the model. Avoidance: Insert the following DMAP alter: COMPILE NLSTATIC ALTER 1 FILE KDJJ=SAVE,OVRWRT/KDGG=SAVE,OVRWRT/KDSF=SAVE,OVRWRT/ KDSS=SAVE,OVRWRT/KDAA=SAVE,OVRWRT/KDMG=SAVE,OVRWRT/ KDLL=SAVE,OVRWRT $ alter 'call sekdr '(,-1) dbdelete datablk=(kdaa,kdmg,kdsf,kdjj,kdgg,kdss,kdelm,kddict) , where(wildcard) $ endalter

@ 40545;

2001; Optimization/Design Sensitivity - DEQATN - Traceback in IFS6P It has been observed on Cray computers that DEQATN input that exceeds 10000 characters can cause NASTRAN to fail during the processing of the input bulk data deck. This is well below the 32000 character limit that is documented. Symptons are erroneous echos of the deqatn input and a traceback in the log file which includes subroutines PRERR2 and IFS6P. Avoidance: Run on a computer other than the CRAY. Reduce the number of characters in a DEQATN to less than 10000. In Version 70.7 a DRESP2 can reference a DRESP2 so that a single DEQATN could be broken up into a string of DEQATN's.

@ 40547;

2001; P-Elements - Missing Output If there are both h- and p-elements in the model, and the output requests differ, for example stress=print for h-elements and stress=print,plot for p-elements, the p-element output may be missing.

@ 40565; 2001; Optimization/Sensitivity - PARAM,EIGD,YES - Wrong Sensitivity Co efficients If eigenvector sensitivity is requested (param,eigd,yes) and there is no

constraint in the model (e.g., spc, rbe, etc.), then the eigenvector sensitivitiy coefficients may be incorrect. Avoidance: Add a dummy grid point and constrain the dummy grid point or apply the following dmap alter: compile eigvecds alter 'ssg2'(,-1) if (omit = -1 and nomset = -1 and single = -1 ) then equivx fg/ff/always $ else $ alter 'ssg2'(,6) endif $ no spc, mpc or omit @ 40579; 12I/T**3 ; Documentation/Optimization - DVPREL1 - PSHELL Property Name for

The DVPREL1 entry has the ability to reference property field id names; however, the PSHELL id field 12I/T**3 and other fields that begin with integers can not be referenced in this manner. Avoidance: Refer to the field by the field position. @ 40582; in DPD 2001; Design Sensitivity/Optimization - PARAM,SOFTEXIT,YES - UFM 2109 The message: USER FATAL MESSAGE 2109 will be issued from the DPD module with an accompanying "NO GRID, SCALAR, OR EXTRA POINTS DEFINED." when PARAM SOFTEXIT set to yes. Avoidance: Do not specify PARAM SOFTEXIT YES @ 40620; 2001; Optimization/Design Sensitivity - DSAPRT, PARAM,SOFTEXIT - UFM 4 276 EC 3001 When an optimization task includes the DSAPRT formatted request and PARAM,SOFTEXIT,YES it may fail in subroutine dr2dx with the message SYSTEM FATAL MESSAGE 4276 ERROR CODE 3001. Avodiance: 1) Use either dsaprt(unformatted,...) option or use optexit=4. 2) Use the following dmap alter (for Version 70.7) compile desopt alter 'l exitopt'(8),'l exitopt'(8) CALL EXITOPT XINIT,DESTAB,HIS,,,R1TABRG,RSP2RG,DRSTBLG, FOL1,TOL2,FRQRPRG,DPLDXI,CONSBL,R1VALRG,R2VALRG, DSCMG,DEQATN,PLIST2,DEQIND,DXDXIT,DCLDXT,DVPTAB,TABDEQ, R1MAPR,R2MAPR,DPLDXT,EPTTAB,DBMLIB,,,PROPI,PROPO,,,,, CASEXX,EDOM,MTRAK,HISADD,WMIDG,dnodel,rr2idrg,EPT,MPT,GEOM2/ DOM11SCR/ CNVFLG/DESCYCLE/DESMAX/OBJSID/ DESOBJ/OUTOPT/0.0/DSPRINT/DSUNFORM/DSEXPORT/TRUE/

MODETRAK/MDTRKFLG/EIGNFREQ/DSAPRT/PROTYP $ $ @ 40628; ; Nonlinear Analysis - CROD, TEMP(LOAD) - Wrong Stress Results for Linear Elements If a thermal load is applied in one subcase in a nonlinear analysis and then is removed in a subsequent subcase, the total thermal load is removed in one step from linear elements. This may lead to incorrect stresses for intermediate output. The output at the end of each subcase is correct. Avoidance: Insert MATT1 for the linear element using a constant property, then the stress will be correct for the linear elements at the immediate output. @ 40644; 2001; SSSALTERs - alter1ha and alter2ha - Incorrect Results It is possible that the results (printed output) from alter2ha may be incorrect. This results from the internal sort of the exterior (connecting) dof being different between the alter1ha run and the alter2ha run. Avoidance: Verify that the exterior dof are in the same order between the runs or request an ujpdated version of the alters for V70.7 (also available on the MSC web page). @ 40647; 2001; Loads - DMIG UACCEL, SPCFORCE - UWM 3059 SDR1 In Version 70.7, when using DMIG UACCEL to apply inertia loads in SOL 101 and SPCFORCE = ALL is specified, the following error may occur: *** USER WARNING MESSAGE 3059 (SDR1) SET IDENTIFIER DOES NOT EXIST. ERROR DETECTED IN SUBROUTINE GETBOL 0FATAL ERROR Avoidance: Either use GRAV/RFORCE to apply the loads, or remove the SPCFORCE=ALL case control. @ 40667; 2001.0.1; Random Analysis - SFM 3008/SWM 3002 (RANDOM) Insufficient Memo ry for RAND5A If SORT2 output is requested, and PSDF, ATOC, or RALL output are requested, and the SET being used has a THRU range, the RANDOM module will fail with SFM 3002, indicating insufficient memory. Avoidance: This only occurs if all 3 of the above are done simultaneously. Removing the SORT2 request or using a SET which does not contain a THRU will allow the problem to run. @ 40670; 2001; P-elements - Missing Output If the p-element output requests are varied by subcase, for example stress only for the first subcase and stress and displacement for the second, the output for the earlier subcases may be missing and/or mislabeled.

Avoidance: An avoidance is to request the same output for all subcases. @ 40685; 2001; Plotting - IFP1 - Access Violation on NT Systems When requesting XYPLOT output, several commands require that an equal (=) sign be used. On NT systems, if any of these commands (XTGRID for example) do not have the equal sign, the program will fail with an access violation in IFP1. On other systems, it will terminate with "USER FATAL MESSAGE (IFP1X) 676 IS NOT RECOGNIZED ON ABOVE CARD" Avoidance: Be sure to include the equal sign on the appropriate commands. @ 40686; 2001; Acoustic Analysis - RLOADi, ACSRCE - Wrong Answers If solution 111 is run with multiple subcases loading the fluid structure, all loading on the fluid structure is lost after the first subcase. Avoidance: Apply the following dmap alter: $ begin lost loads fix compile freqrs,list $ alter 'MATMOD FOL' $ if (ncol > nfreq) then $ $ $ ensure that omega contains nfreq for all subcases $ file omegax=append $ type parm,,i,n,nsols $ nsols = ncol/nfreq $ trnsp omega/omegat $ do while (nsols > 0) $ append omegat,/omegax/2 $ nsols = nsols - 1 $ enddo $ (nsols > 0) delete /omega,,,, $ trnsp omegax/omega $ endif $ (ncol > nfreq) $ end lost loads fix @ 40700;

; Superelements - SWM 6065 (DBC), UWM 4698 (DCMPD) When using PART superelements (BEGIN SUPER), error messages 6065 and 4698 may use internal numbers, rather than the original ids, to identify GRID points. *** SYSTEM WARNING MESSAGE 6065 (DBC) *** DBCMSG ERROR MESSAGE 19 FROM SUBROUTINE DBCEQG EXTERNAL GRID ID 2000000334 DOES NOT MEET UPDATE ASSUMPTIONS, CHECK MODEL BEFORE PROCEEDING. Avoidance: To guarnatee a specific order boundary grids are biased with the value of system cell 233. The default value for this system cell is 2,000,000,000. Subtracting this value from the value identified in the message will obtain the external grid id. *** USER WARNING MESSAGE 4698 (DCMPD)

STATISTICS FOR DECOMPOSITION OF MATRIX (NONE) . THE FOLLOWING DEGREES OF FREEDOM HAVE FACTOR DIAGONAL RATIOS GREATER THAN 1.00000E+15 OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL. RATIO GRID POINT ID DEGREE OF FREEDOM MATRIX DIAGONAL 2000000892 1.00000E-10 6 MATRIX/FACTOR DIAGONAL 1.00000E+15

@ 40725; 2001; P-elements/Postprocessing - PARAM,POST - Strains overwrite Stres s in XDB When stress and strain output is selected in an analysis with P-Elements, the P-Element strains will over-write the P-Element stresses on the xdb file. Avoidance: When saving P-element stresses on the xdb, make stress requests only. Strain can be saved when no stresses are requested but will be identified incorrectly on the xdb. @ 40732; 2001.0.1; Acoustic Analysis - CHACAB, CHACBR - Missing Fatal Message, Wr ong Answers The CHACAB and CHACBR elements are structural elements and all faces must connect to structural points. The face on the fluid side should connect to structural grids that have all degrees of freedom constrained (spc'd) except the dof that is normal to the fluid. These grids should coincide with the fluid grids but be separate from them. If the fluid side of the element attaches directly to fluid points a bad solution will result with no warning. This can happen when both the CHACAB or CHACBR element and the fluid points are in the residual. @ 40733; ; Postprocessing/Buckling - PARAM,POST,0 - Missing ESE Results In a solution 105 buckling analysis the Element Strain Energy results are written to the xdb file for the static subcase but not the buckling subcase.

@ 40734; 2001; Superelements - SEBNDRY and SECONCT with THRU - UFM 315, Wrong a nswers, UFM 6782 The THRU format on the SEBNDRY and SECONCT Bulk Data entries causes various failures or wrong answers: 1. THRU ranges on SEBNDRY continuation entries cause UFM 315. 2. The second THRU range on the SEBNDRY parent entry is ignored with no warning or fatal message issued and produces wrong answers. 3. Multiple SEBNDRY entries causes UFM 311.

4. THRU ranges with gaps in numbering on the SECONCT entry causes UFM 6782. Avoidance: For SEBNDRY use only one THRU range. For SECONCT, specify

separate THRU ranges without any gaps. @ 40773; 2001.0.1; Miscellaneous - ASSIGN, OUTPUT2, OUTPUT4 - SIGNAL 11 ERROR, ac cess violation If there are many output2/output4 assign statements, then depending on the machine, the job may fail with any one of the following messages: SIGNAL 11 ERROR MAIN: "Access violation" (C0000005) exception encountered. MAIN: For reference, A(MAINAL) = 428ce0, A(/SYSTEM/) = 1781f80 MAIN: Exception occurred at address 005A04DD. MAIN: Program attempted to store data at location CB31AF28. MAIN: Context Flags 0001003F Avoidance: Removing one or two of output2/4 statements avoided the error in this case. @ 40805; 2001; Superelements - PARAM,FIXEDB,-1 - Incorrect Eigenvectors for Com ponent Modes If output is requested for component modes using PARAM,FIXEDB,-1 and Cor R-set degrees of freedom exist, the output eigenvectors and eigenvalues printed will be incorrect. Note that all other output will be correct. Avoidance: problem. Insert the following alter into SOL 103 to correct the

compile semodes alter 1 putsys(0,192) $ compile super3 alter 'call.*sedisp.*cmpho','' type db,cmpho0,cmlama0 CALL SEDISP, GOT ,GOQ ,LOO ,LAO UOX ,USET ,PJ ,UL CASESX ,CASEDR ,ULS ,EDTS GM ,PSS ,KFS ,KSS CMPHO0 ,OL1M ,CMLAMA0 ,CMPHA PS1 ,CRX ,CASE0 ,BACK QRG ,PG ,DYNAMICS,GPLS SLT ,DIT ,KSF ,KMG MSF ,BMG ,BSF ,K4MG K2MG ,M2MG ,B2MG ,K2SF B2SF ,UH1 ,OL1R ,BGPDTS QSGF ,QMGF ,FRLR ,CASES LCDVEC ,SPCPART ,QAG0 / UG ,QG ,QMG ,OL1S FIXEDB /APP /APP1 /SEID ALTRED /NOQG /NLHEAT /DROPT RSONLY /GW3 /RW4 /C1G NOQMG /NOPJ1 /NCUL /PEXIST PARDCMP $ @ 40821;

,UOO ,CRJ ,YS ,QR ,PP1 ,FB ,SILS ,MMG ,K4SF ,M2SF ,CSTMS ,PVTBC ,PJ1 /DLOOP /NONLNR /NOW3W4 /FREQDEP

, , , , , , , , , , , , / / / / /

2001; Superelements - BEGIN SUPER, RESTART - Wrong answers

When using part (BEGIN SUPER) superelements with any type of RESTART then the data recovery may be incorrect. Avoidance: Insert a dummy (unreferenced) FORCE Bulk Data entry in the main Bulk Data section of the restart run. @ 40826; ; DMAP -- OUTPUT2 with MATRIX neutral format - Wrong answers On short-word computers (all except Cray and Nec), if OUTPUT2 is used to output a single precision matrix in neutral format then the INPUTT2 of the neutrally formatted matrix will produce wrong answers. Avoidance: Use the ADD or ADD5 module to convert the matrix to double precision before passing it to OUTPUT2. @ 40831; NLTRD2) ; Nonlinear Transient Heat Transfer - UWM 6137(UDSNRD), UFM 3007 (

If a nonlinear transient heat transfer analysis, solution 159, model has extra grid points that are not attached to any elements and are not constrained, then the job may fail with the message: *** USER WARNING MESSAGE 6137 (UDSNRD), INPUT MATRIX IS RANK DEFICIENT, RANK = USER ACTION: CHECK MODEL 0*** UNSUCCESSFUL MATRIX DECOMPOSITION. *** SYSTEM FATAL MESSAGE 3007 (NLTRD2) ILLEGAL INPUT TO SUBROUTINE NLTRD2 Avoidance: 1) Rerun the model using SOL 101 with APP HEAT and insert PARAM,SPCGEN,1 in the bulk data section. The PARAM,SPCGEN,1 entry will cause constraint entries for the unattached extra grid points to be written to the punch file and these constraint entries can be subsequently used in solution 159. 2) Rerun the transient model by using the the older and significantly slower unsymmetric solver by setting the NASTRAN SYSTEM(209)=0. @ 40839; 4252 ; Documentation/QRG - RESTART, DMIG - Inefficient Restart or SFM Restarts with DMIG entries may unneccessarily re-process the entire model even though no changes to the input have occurred. The way to verify that this is happening is to run with DIAG 56 and to look in the .f04 file for "changed" and "DMIG" on the same line. If these words occur on a line output by the RESTART module, the program will re-run the entire model. Avoidance: This is currently a limitation of the sort routines used in Nastran. Other potentially problem entries include BNDFL, DMIAX, ELIST, DMIAX, ELIST, MFLUID, RADLIST, and RADMTX. If there are no changes to any BNDFL, DMIAX, DMIG, ELIST, MFLUID, RADLIST, and RADMTX entries, then insert the following alter to disable the checking for changes in these entries. compile ifpl alter 'restart.*matpool','' message //'restart checking on MATPOOL disabled' $ 7816 IERROR

@ 40843;

; Beam Library - PBARL, PBEAML - SFM 6498 on Large Models A very large model with PBARL or PBEAML entries may fail with the following error messages: *** SYSTEM FATAL MESSAGE 6498 (Beam Lib, allocate_Master_Buffer) API MESSAGE FOLLOWS. USER INFORMATION: Memory Limit exceeded. Could not allocate Memory for the Main Buffer Pool! MSC.Software has not been able to duplicate this error and the model that causes this error is confidential. The client model contains over 1.3 million entries, but only a few PBARLl entries. Avoidance: There are three possible avoidances. 1. Replace the pbarl/pbeaml entries with equivalent pbar and pbeam entries. 2. Allocate more memory. 3. Run the job on a non-Cray computer since Cray computers require more memory.

@ 40874; )

; Heat Transfer Analysis - NLCOMB - Poor Message, UFM 4016 (GETEMP In Version 70.7 Heat Transfer analysis the following message indicating that there is missing TEMP entry data will be output for any point with an SPC constraint that does not have a corresponding TEMP entry. USER FATAL MESSAGE 4016 (GETEMP) THERE IS NO TEMPERATURE DATA FOR ELEMENT FATAL ERROR 8127187 IN SET 1

Avoidance: Check all the grids with SPC constraints and ensure that there is a corresponding TEMP entry. @ 40892; 2001; Documentation/QRG - PBEAM The following remark should be added to the Quick Reference Guide description of the PBEAM entry: Remark 13: In nonlinear analysis the location of the 8 plastic rods is the same on the PBEAM as on the PBCOMP SYMOPT=0. @ 40909; 2001; Documentation/QRG - SEUPPLOT, SEPLOT, PLOT - Missing Plot Output Starting with Parts superelements in Version 69 each PLOT command requires an SEUPPLOT or a SEPLOT command. This requirement and an example should be noted under the Nastran Quick Reference Guide PLOT command description.

An example follows: some general plot commands such as VIEW etc. SEPLOT 10 FIND SCALE ORIGIN 1 SET 1 PLOT SET 1 ORIGIN 1 $ SYMBOL 1 LABEL GRID SEUPPLOT 0 FIND SCALE ORIGIN 1 SET 1 PLOT SET 1 ORIGIN 1 $ SYMBOL 1 LABEL ELEMENT @ 40910; t ; Nonlinear Transient Analysis - RESTART/CGAP - Missing CGAP Outpu If a nonlinear transient analysis restart run adds a request for CGAP element force/stress output the output may be missing. Avoidances: Request the output in a cold start run. @ 40934; 2001; Optimization/Design Sensitivity - PBARL, PBEAML - Illegal input to DPNCH6 When both PBARL and PBEAML entries are included in an optimization task, the job may fail in routine dpnch6 with a SYSTEM FATAL MESSAGE indicating illegal input to subroutine dpnch6. Avoidance: Set PARAM,DESPCH,-1 @ 40936; 2001.0.1; Documentation - Missing Output, UFM 316 The Nastran Quick Reference Guide does not specify a limit for element identification numbers. With the large field format 16 digits may be specified but id's greater than 8 digits may result in missing output or USER FATAL MESSAGE 316. Avoidance: Do not specify element ids that are greater than 8 digits.

@ 40937;

2001; Miscellaneous - SPOINTs, SEP1 - SWM 4203, UWM 4203 If a model has superelements and contains more than 32,000 SPOINTs the run may fail after issuing SYSTEM WARNING MESSAGE 4203 and USER WARNING MESSAGE 4203, indicating that the SEMAP datablock has been marked as non-executable in SEP1. Avoidance: If possible use GRID points in place of SPOINTs.

@ 40953; 2001; Optimization/Design Sensitivtiy - DRESPi - No Message, Traceback in dopr3m If a DRESP2 entry references a DESVAR, DRESP1, DNODE, or DVPRELx with ID = 11000. The job will fail in the DOPR3 module. The traceback from the log file follows:

Signal received: SIGSEGV - Segmentation violation Traceback: Offset 0x000002e4 in procedure Offset 0x000004e8 in procedure Offset 0x00000284 in procedure Offset 0x00000110 in procedure Offset 0x0000005c in procedure Offset 0x000009ac in procedure --- End of call chain --dopr3m_ dopr3_ xsem03_ xsemdr_ nastmon_ mainal

Avoidance: Specify another DRESP id. @ 40974; 2001.0.1; Superelements - part superelements, SNORM - Wrong answers If PARAM,SNORM or the SNORM Bulk Data entry is specified with part superelements or BEGIN SUPER and main Bulk Data superelements, then wrong answers will result. Avoidance: Insert PARAM,SNORMPRT,2 or 3. @ 40990; 2001.0.1; Superelements - PARAM,SESDAMP - Incorrect Damping If param,sesdamp,yes is used, then superelement component modes will have modal damping applied based on the SDAMP command in each superelement's SUBCASE. The residual structure will use the SDAMP command in the first residual structure SUBCASE, unless there is none, in which case it will use the last SDAMP called in a superelement subcase, which may result in unexpected damping in the solution.. Avoidance: To be safe, if using PARAM,SESDAMP,YES, always explicitly specify SDAMP in the residual subcase. If zero damping for the residual is desired, SDAMP should refer to a TABDMP1 with 0.0 damping. @ 41009; 2001.0.1; P-Elements - PBEAM - Non-structural mass is ignored For beam p-elements, the non-structural mass (NSM) on the PBEAM bulk data entry is ignored. Avoidance: An avoidance is to adjust the density to account for the non-structural mass. @ 41020; ; Optimization/Design Sensitivity - CSUPER - SFM 3007 (MTRXIN) If external superelements are used in optimization, and DESMAX has a value greater than the number of cycles required to converge, and the run is not saving the database (i.e. scr=yes), the run may fail with the following message:. SYSTEM FATAL MESSAGE 3007 (MTRXIN) ILLEGAL INPUT TO SUBROUTINE MTRXIN FATAL ERROR Avoidance: Either set DESMAX to the number of cycles required to converge or run with SCR=NO on the submittal. @ 41021; 2001.0.1; Aeroelastic Analysis - APD - SFM 4276

If an AEFACT entry has a duplicate station defined, then the APD module fails with a floating point exception. Avoidance: The avoidance is to remove or correct the duplicate station specification. @ 41034; 2001.0.1; P-Elements - OUTPUT - Wrong Answers If the output requests specified by the DISPLACEMENT, VELOCITY, and ACCELERATION case control commands are different from the corresponding output requests on the OUTPUT bulk data entry, then the output may be incorrect. Avoidance: The only avoidance is to make sure the output requests are consistent. @ 41036; 2001.0.1; Miscellaneous - CORDi - UFM 6915 (GP1C) When defining a coordinate system in MSC.Nastran, three non-colinear points are required. If the dimensions of the three points are such that there are more than 15 orders of magnitude difference between them, the program may incorrectly issue the message: *** USER FATAL MESSAGE 6915 (GP1C) indicating that the points are colinear. Avoidance: Use 3 points which have locations within 10 orders of magnitude of each other when defining coordinate systems. @ 41039; ; DMAP - DRMH1 - SFM 4276 EC 7777 on DEC ALPHA Use of the DRMH1module to generate transfer matrices for certain element types may cause an underflow to occur. Subsequent use of the matrices will generate the following error on Alpha machines: *** SYSTEM FATAL MESSAGE 4276 (tbk) ERROR CODE 7777 PID= 0 Avoidance: This error appears to occur only on Compaq Alpha Unix systems. If possible, try running on a different type of computer. @ 41050; P 2001; Distributed Memory Parallel - RBEi, DMP, SEQP - Traceback in SEQ If a Distributed Memory Parallel, DMP, run contains rigid elements or mpc equations with many dependent degrees of freedom, then the job may fail in the SEQP module with the following traceback: Signal received: SIGSEGV - Segmentation violation Traceback: Offset 0x000006d4 Offset 0x000013b0 Offset 0x00000f0c Offset 0x000002c8 Offset 0x000001b8 Offset 0x0000005c in in in in in in procedure procedure procedure procedure procedure procedure seqcon_ seqmpc_ seqp_ xsem02_ xsemdr_ nastmon_

Offset 0x00000c30 in procedure mainal --- End of call chain --Avoidance: Increasing still be a performance undesirable for models freedom. Version 2001 memory required. @ 41057; memory may allow the job to execute but there may penalty that makes Distributed Memory Parallel with a large number of dependent degrees of has been enhanced to provide an estimate of the

; Nonlinear Analysis - PCOMP, TEMP(LOAD) - Wrong Answers A nonlinear analysis with temperature loading applied to composite elements where the PCOMP specifies an unsymmetric thermal coefficient expansion in the 1 and 2 direction results in wrong answers.

@ 41059; 2001.0.1; Aeroelastic Analysis - AMG - Floating Point Exceptions The above error can result when using Strip Theory aerodynamics with a Prandtl-Glauert correction to the lift curve slope (CLA < 0) and a supersonic Mach number is chosen on a MKAEROi entry. This is because the correction involves computing a term of the form sqrt(1.0 - M**2); i.e., taking the square root of a negative number Avoidance: If strip theory is to be used with lift curve slope corrections, use CLA > 0 @ 41061; ; Nonlinear Analysis - PLOADX1, PARAM,FOLLOWK - UFM The follower stiffness for PLOADX1 is incorrect. This could cause some nonlinear static problems in SOL 106 to diverge. Avoidance: Turn off the follower stiffness by specifying param, followk, no. @ 41065; swers ; Heat Transfer Analysis - VIEW Factor using CQUAD8 yields poor an The calculation of the VIEW factors for CQUAD8 elements for curve surfaces such as cylinder and sphere in heat transfer analysis yields poor results when compared to the CQUAD4 element. Avoidance: Use the CQUAD4 element. @ 41066; 2001; Documentation - PARAM FOLLOWK - Missing Solution Sequences The Version 70.5 Quick Reference Guide description of PARAM,FOLLOWK does not include all applicable solution sequences. The description should read: In Sols 103, 105, 106, 107, 108, 109, 110, 111, 112, 115, 116, FOLLOWK=YES(default) requests...... and on the fourth line: For FOLLOWK=YES in Sols 103, 105, 107, 108, 109, 110, 111, 112, 115, and 116 a separate static subcase is required ....... @

41069; PYAD)

2001; Distributed Memory Parallel - DMP, PARAM,RESVINER - SFM 5423 (DM MSC.Nastran Version 70.7 with DMP with PARAM,RESVINER,YES fails in SOL 111 with: *** SYSTEM FATAL MESSAGE 5423 (DMPYAD) ATTEMPT TO MULTIPLY INCOMPATIBLE MATRICES Avoidance: Insert the following DMAP alter: compile resvec alter '\/paa','' dbstatus vafs//s,n,novafs $ if ( novafs>0 ) then $ merge mxx,,,,vafs,/mxx1 $ else $ equivx mxx/mxx1/-1 $ endif $ mpyad mxx1,dga,/paa $ endalter

@ 41071; 2001.0.1; Inertia Relief - Enhancement to PARAM,INREL,-2 An inertia relief option based on measuring deflections from a body-based coordinate system was introduced in Version 70.5. It is requested with param, inrel, -2. This method uses constraint equations that cause deflections to be measured from a coordinate system tied to the model's center of gravity. In Version 70.7 the rigid body modes used to generate the constraint equations are based on shapes made by moving a point at the basic origin through 6 unit motions. Revision 1 of the ir2a.v707 sssalter package for improved efficiency (dated 2 Jan 2001) now has the modes calculated around the grid point specified on the param, grdpnt entry. The default action of this parameter is to use the basic origin, as before. This change is in the revised sssalter package on the MSC web site, and is being considered for Version 2001. The alter to achieve this action is: compile phase0 list $ use grdpnt as reference for rb modes alter 'IF ( GRDPNT.*0 OR INREL=-2 ) REFPNT=0','' $ V70.7 IF ( GRDPNT<0 ) REFPNT=0 $ @ 41109;

; 6 by 6 gets 7 roots with LAN Some small eigensolutions compute extra roots by repeating one of the roots and its eigenvector. This tends to cause fatal exits in the dynamics solvers such as frrd1 and trd1. this error is caused by making two or more rows of the modal matrices linearly dependent.

An avoidance is to use the AHOU metod instead. @ 41133; ; Nonlinear Analysis - PCOMP, PARAM.LGDISP - Missing Composite Str ain Output Homogenous composite strain output will be missing in nonlinear analysis when param,lgdisp,1 is specified.

Avoidance: Request STRAIN output in case control. Also, a request for STRESS for at least one element and appy the following alter: compile nlstatic alter 'sdr2.*oesxc' SDR2 CASESX2,CSTMS,MPTS,DIT,EQEXINS,, ETT,EDTS,BGPDTH,,,UGNI,EST,,,,,GPSNTS, DEQATN,DEQIND,/ ,,,OESXd,,/'GNST'//3 $ ofp oesxd/ $ @ 41137; ; Eigenvalue Analysis - EIGRL - Lanczos unreliable for diagonal in put matrices. The Lanczos method gets eigenvectors with poor orthogonality when there are repeated roots, and both the mass and stiffness matrix are diagonal. On some occasions these matrices can also cause overflow exits. Diagonal matrices occur most commonly on research efforts when using dmap. Avoidance: Use the AHOU method instead. @ 41142; ; Parts Superelements - SELOC, Exterior SPOINTs - UFM 4156 (SEMA) If SELOC is used to locate a parts superelement that has of freedom that include generalized coordinates that are spoints and qset1 entries, then the run may fail in the the following message when the superelement is attempted the residual: exterior degrees defined by SEMA1 module with to be attached to

*** USER FATAL MESSAGE 4156 (SEMA) DIMENSIONS OR TYPE OF DATA BLOCK SCRATCH , SUPERELEMENT 10 TRAILER=( 72 72 1 2 18 662) ARE INCONSISTENT WITH SUPERELEMENT MAP. FATAL ERROR Avoidance: The error occurs because the program assumes every scalar point has six degrees of freedom. The error may be avoided by using the internally generated q-set DOFs with the SENQSET entry. Remove the SECONCT bulk data entry. @ 41146; al exits ; Eigenvalue Analysis - EIGRL - maxset=1 or =2 on eigrl causes fat

The number of vectors in a block for the LAN method is controlled by the maxset input (field 7) on the eigrl entry. When set to any explicit value, several classes of small problems have been observed to fail, with diagnostics including floating point overflow, lnherr subroutine errors., and other obscure diagnostics. Examples include those with Lagrange Multiplier variables and those with diagonal input matrices. Avoidance: Use the default value for maxset (leave it blank), or use the AHOU method. @ 41164; 2001.0.1; File Management Section - EXPAND, INIT - SFM 1012 Allocating too much space to scratch files may lead to poor messages such

as; "SPECIFIED RECORD IS BEYOND MAXIMUM FOR THIS FILE SYSTEM" or "DBSET SCRATCH IS FULL AND NEEDS TO BE EXPANDED. USER ACTION: USE FILE MANAGEMENT STATEMENT "EXPAND" FOR THIS DBSET." Avoidance: The diagnostic stating to use EXPAND FMS statement is applicable only to restarts. The user is referred to the Version 2001 Installation and Operations Guide or tan 4893 on the msc web site for methods to make larger databases. @ 41165; ; Eigenvalue Analysis - EIGRL - Possible Wrong Answers If f1 is specified as 0.0 on the EIGRL entry and the model contains a "large mass" for enforced motion the run may produce poor answers. Avoidance: @ 41178; Leave f1 blank.

; Parts Superelements - SEP1X - WM 6795, SFM 6790 If a PART superelement does not have any attachment points between it and the rest of the model, the program will correctly issue SWM 6795, but may incorrectly issue SYSTEM FATAL MESSAGE 6790 (SEP021) followed by an incorrect list of superelements. Avoidance: Be sure that PART superelements have points which connect to the rest of the model.

@ 41192; 2001.0.1; Nonlinear Analysis - PARAM,FOLLOWK, PLOADi - Poor Efficiency In most cases the formulation used for follower stiffness results in an unsymmetrical matrix. Since the default value for parameter FOLLOWK is yes in SOL106 most problems with pressure loads will have an unsymmetrical stiffness unless param,followk,no is specified. This may lead to an inefficient solution. Avoidance: If the follower force effects are not great, a more efficient solution may be achieved by setting PARAM,FOLLOWK,NO. @ 41198; SB) ; Modal Transient Response - SUPORT1, PARAM,RESVEC - SFM 6424 (DFN If a modal transient response analysis has a static preload, a SUPORT1 entry, and PARAM,RESVEC,YES is specified the run may fail with : SYSTEM FATAL MESSAGE 6424 (DFNSB) THE NUMBER OF ROWS IN THE LOAD VECTOR AND THE FACTOR MATRIX ARE INCOMPATIBLE followed by: *** USER FATAL MESSAGE 5423 (SMPYAD) ATTEMPT TO MULTIPLY OR ADD INCOMPATIBLE MATRICES Avoidance: Insert the following alter into your file: compile resvec, list alter 'SSG3.*LXX,,KXX,PX3,,,,,/UX1'(,1) $ 144 PARAML USET//'USET'//////'r'/S,N,NOxSET $

@ 41215; 2001.0.1; DMAP - TRNSP, APPEND - Segmentation Violation in TRNSP The APPEND module will append incompatible matrices and this may cause problems in subsequent modules including a segmentation violation in TRNSP module. Two steps have been taken to address this problem in MSC.Nastran 2001. First, the following warning message is issued: *** USER WARNING MESSAGE 5423 (APPEND) ATTEMPT TO APPEND INCOMPATIBLE MATRICES ID NZWD DENS TRAILER FOR X 60 10000 TRAILER FOR Y 20 10000 Second, the program will select the greater number of rows for the output matrix. For example if APPEND is requested to append a 10X40 matrix with a 30X20 matrix instead of creating a 10X60 matrix a 30x60 matrix will be output and the unspecified terms will be set to 0. @ 41221; ; Distributed Memory Parallel - DOMAIN, dmp - Job Hangs if Insuffi cient Space Distributed Memory Parallel jobs, dmp=n, executed on Sun systems may hang if insufficient space exists. Currently, this is a limitation of the Sun MPI implementation. Avoidance: The only avoidance is to provide more space. @ 41237; ption 2001; Aeroelastic Analysis - GI, SPLINE1, CAERO1 - Floating Point Exce If a SPLINE1 with the FPS option is used to connect a CAERO1 that only has one spanwise box, then a floating point exception may occur in the GI module. This error will also occurs if the spline connects to a chordwise strip of elements on a CAERO1 that is 1 spanwise box wide. Avoidance: Split the CAERO1 entry so that it has two span wise boxes, or use the IPS option on the SPLINE1. @ 41241; ly ; Miscellaneous - SMPYAD - Segmentation Fault, Signal 11, Alpha On On Alpha platforms, a run may fail during a SMPYAD operation with a message in the log file: Receiving signal number 11: Segmentation fault Job aborting.. Avoidance: There are 2 possible avoidances either replace the SMPYAD in the dmap with 2 equivalent MPYAD's, or set NASTRAN SYSTEM(129)=1. @ 41257; 2001.0.1; Complex Eigenvalue Analysis - OFREQ - Extra Output OFREQ feature does not work for solution output requests in complex IS 102 40 10 2 2 IS 101 20 30 2 2 COLS ROWS FORM TYPE

eigenvalue analysis @ 41260; 2001.0.1; Frequency Response - FRRD1 - SFM 3001 If the model in frequency response does not have any stiffness, the program will fail in FRRD1 with SFM 3001 indicating that executive file 0 does not exist. @ 41269; 2001.0.1; Elements - SYSTEM(218) - SFM 4276 (tbk) , Arithmetic Trap in g mtstd Setting SYSTEM(218) to a real number may cause an arithmetic trap in routine gmtstd on some machines. Avoidance: Use the keyword "T3SKEW" instead. @ 41276;

; Lagrange Multiplier Technique - Wrong Answers When using enforced velocity in transient analysis with the Lagrange Multiplier approch, the LMT variables may show an instability. This appears as a high frequency signal at the forcing variable only. If this is encountered, contact MSC, and ask for details on residual mass. Mention this Client Service Request Number (41276).

@ 41280;

; DMAP - DIAGONAL - Module DIAGONAL may give a bad matrix trailer If a rectangular matrix (form=2) is processed by the DIAGONAL module using the 'SQUARE' option, the output matrix will have an incorrect trailer (form=6) indicating that it is symmetric, when it should be rectangular(form=2). Avoidance: If the input matrix to DIAGONAL is rectangular and the 'SQUARE' option is specified, use MODTRL to update the trailer.

@ 41301;

2001; Transient Analysis - GPSTRESS- Missing Output If GPSTRESS output is requested in SOL 112 in Version 70.7, there will be no corresponding output created. Avoidance: Insert the following DMAP alter: compile sedrcvr alter 'call gpstress'(2),'' call gpstress postcdb,BGPDTN,ECTS,est,CSTMS,estnl,casedr, eqexins,oes1x,oesnlx,ostr1/ egpsf,ogs1,egpstr,oeds1,ogds1,eldct,gpdct, ogstr1,egpstra/ app/otape2/statics/app $ endalter

@ 41308; 2001.0.1; P-Elements - DDRMM - Incorrect Principal Strains

For solid p-element dynamics in SOL 112 (modal transient), the principal strains calculated with DDRMM are incorrect. Avoidance: A workaround is to turn DDRMM off by setting param,ddrmm,-1. @ 41335; ed 2001; Nonlinear Analysis - Complex Eigenvalue - Material Damping Ignor The calculation of complex eigenvalues in SOL 106 does not include the effect of structural damping defined by Ge on the material entries. Avoidance: Insert the alter: compile nlstatic alter 'call semr2' equivx kllrh/kaa/-1 $ type parm,,i,n,nobggnl=-1,nok4ggnl=-1 $ EMG MESTNL,CSTMS,MPTS,DIT,,MUGNI,ett,,DEQATN,DEQIND,BGPDTS, GPSNTS,,,,ditid/ x1,x2,,,belmnl,bdictnl/ -1/0/s,n,nobggnl/s,n,nok4ggnl/NSKIP/ /////////tabs/sigma/K6ROT/LANGLES/NOC//////unsymf $ if ( nobggnl>-1 ) EMA GPECT,bDICTNL,bELMNL,BGPDTS,SILS,CSTMS,,/ bggNL, $ call xadd5 bgg,bggnl,,,/bggt/cu/cu/cz/cz/cz $ dbstatus bggt//s,n,nobggt $ if ( nobggt>0 ) CALL SEMRB USET,GM,GOA,BGGt,casesx2,cmlama,bk4zz,mphvz,goq,dit/ BAA,BMG,BSF/ NOA/NOMSET/norc/norset/nocset/nobset/noqset/ notset/false/0 $ if ( nok4ggnl>-1 ) EMA GPECT,KDICTNL,KELMNL,BGPDTS,SILS,CSTMS,,/ K4ggNL,/1 $ call xadd5 k4gg,k4ggnl,,,/k4ggt/cu/cu/cz/cz/cz $ dbstatus k4ggt//s,n,nok4ggt $ if ( nok4ggt>0 ) CALL SEMRB, USET,GM,GOA,K4GGt,casesx2,cmlama,bk4zz,mphvz,goq,dit/ K4AA,K4MG,K4SF/ NOA/NOMSET/norc/norset/nocset/nobset/noqset/ notset/false/0 $ endalter or, if possible, use PARAM,G to define the structural dampin @ 41359; rs 2001; Documentation/Aeroelastic Analysis - SPLINEi - Unexepected Answe The USAGE field of the SPLINEi entries is not clearly documented that the "BOTH" option is the ONLY one in which a single SPLINEi entry will transform both loads and displacements (wherein the virtual work transpose of forces to get displacements is used). Since the set of structural degrees of freedom appropriate for load application may not be the same set that is appropriate to represent the important deflections for the aeroelastic correction, the FORCE and the DISP options allow the the splines to be defined separately. However, FORCE or BOTH spline must

be defined for each aerodynamic box to yield a "complete" mapping of forces. It is unusual (but not unheard of) to have aerodynamics for which forces are not desired (e.g., to impose 2D-like boundary conditions or "infinite aspect ratio" boundary conditions one typically uses extraneous aerodynamics). Because there are these legitimate cases in which the forces are not desired, no error messages or warnings are issued. Please be careful that you have correctly and completely modelled the desired usage of the splines. BOTH, the default USAGE, is a good default unless you are particularly modeling around or for a particular purpose for which a separation of force and displacement mapping is desired. @ 41365; M 9050

GL; Acoustic Analysis - STATSUB, PARAM, RESVEC, RESVINER, INRLM - UF If the STATSUB Case Control or PARAMs, RESVEC, RESVINER, or INRLM is requested in a fluid-structure model then the job may fail with UFM 9050. Avoidance: Specify a SUPORTentry for any grid point in the fluid model.

@ 41379; EC 7777

; Nonlinear Transient Analysis - TSTEPNL - Poor Message, SFM 4276,

If NDT on the TSTEPNL entry is incorrectly set to a value less than 3, the program will fail with a Floating Point Exception. Avoidance: Specify a value greater than 3 for NDT. @ 41387; ; Parallel Processing - DMP, GPWG - Wrong Answers for GPWG The GRID POINT WEIGHT GENERATOR output is incorrect in Distributed Memory Parallel solutions. Avoidance: The only avoidance is to execute a checkout run without dmp. @ 41389; 2001; Optimization/Design Sensitivity - DRESP2 - Incorrect DRESP2 Resp onse Output The print of the DRESP2 response values that are controlled by the P2 parameter on the DOPTPRM entry may be in error. The error occurs when there are more DRESP2 values than the length of the maximum record in the RESP12 datablock (this value is the second word of the trailer of the RESP12 datablock). The incorrect value that is printed is not used for subsequent calculations with the exception that it may get used in calculating the sensitivities of DRESP2's that reference other DRESP2's. Optimization is not affected. @ 41408; ; DMAP - MATGPR - Illegal Input to MATGPR The MATGPR module fails with "Illegal input" error messages when attempting to print matrices that contain the k or ks sets in aeroelastic analysis.