ProCAST User Manual

PROCAST USER MANUAL

VERSION 2006.0

Revised version (April 2006) - CL/PRCA/06/02/00/A

Version 2006.0

Page 1

ProCAST User Manual

PROCAST USER MANUAL

INTRODUCTION SOFTWARE CAPABILITIES SOFTWARE ORGANIZATION USER MANUAL PRESENTATION WHAT'S NEW VERSION 2006.0 VERSION 2005.0 VERSION 2004.1 VERSION 2004.0 IDENTIFIED BUGS, PROBLEMS AND LIMITATIONS GETTING STARTED SOFTWARE LAUNCH PROBLEM SET-UP CALCULATION RESULTS DISPLAY SOFTWARE MANAGER FILE MANAGER MODULE CALLS ADVANCED MODULE CALLS RUN LIST SOFTWARE CONFIGURATION CUSTOMIZED INSTALLATION FLEXLM PRE-PROCESSING INTRODUCTION GEOMETRY IMPORT THERMAL GEOMETRY ASSIGNMENTS MATERIALS ASSIGNMENT INTERFACES ASSIGNMENT BOUNDARY CONDITIONS ASSIGNMENT PROCESS CONDITIONS ASSIGNMENT INITIAL CONDITIONS ASSIGNMENT RUN PARAMETERS ASSIGNMENT

7 7 9 11 12 12 19 24 29 33 36 36 38 46 48 51 51 55 63 66 69 73 74 78 78 82 84 85 85 90 95 100 102 104

Version 2006.0

Page 2

ProCAST User Manual

FLUID FLOW & FILLING RADIATION STRESS DATABASES MATERIAL DATABASE MATERIAL PROPERTIES THERMODYNAMIC DATABASES Calculation of Stress Properties Databases limitations Influence of alloying elements INTERFACE DATABASE BOUNDARY CONDITIONS DATABASE PROCESS DATABASE STRESS DATABASE STRESS MODELS AND PROPERTIES Digitized Hardening Plastic and Viscoplastic properties determination RUN PARAMETERS GENERAL RUN PARAMETERS THERMAL RUN PARAMETERS CYCLING RUN PARAMETERS RADIATION RUN PARAMETERS FLOW RUN PARAMETERS TURBULENCE RUN PARAMETERS STRESS RUN PARAMETERS MICRO RUN PARAMETERS PRE-DEFINED RUN PARAMETERS RUN PARAMETERS RECOMMENDATIONS POROSITY MODELS POROS=1 POROS=4 POROS=8 DENSITY DEFINITION ACTIVE FEEDING SGI POROSITY MODEL VIRTUAL MOLD FILTERS EXOTHERMIC CYCLING LOST FOAM THIXO CASTING CENTRIFUGAL CASTING MULTIPLE MESHES AND NON-COINCIDENT MESHES GEOMETRY MANIPULATION MESH OPTIMIZATION USER FUNCTIONS USER FUNCTIONS TEMPLATES External heat transfer coefficient Function External temperature Function Emissivity Function Heat flux Function Interface heat transfer coefficient Function Mass Source Flow Rate Function

106 110 122 131 131 135 141 150 153 156 164 167 182 187 191 205 206 220 221 227 233 234 238 246 248 252 255 257 261 262 266 266 267 268 269 276 282 285 287 290 295 298 299 303 307 308 313 313 314 315 316 317 318

Version 2006.0

Page 3

ProCAST User Manual

Mass Source Vector Function Translation Vector Function Imposed Velocity Vector Function Solid Transport Velocity Vector Function External Function RUN OF THE CALCULATION SOLVER TROUBLESHOOTING RESULTS VIEWING INTRODUCTION FIELD SELECTION DISPLAY TYPES DISPLAY PARAMETERS TAPE PLAYER CURVES GEOMETRY MANIPULATION RESULTS ANALYSIS CRITERION FUNCTIONS POROSITY FATIGUE LIFE INDICATOR HOT TEARING INDICATOR CRACKING INDICATOR FRECKLES INDICATOR RESULTS EXPORTS INTRODUCTION GEOMETRY RADIATION FACES TEMPERATURE HEAT FLUX DISPLACEMENTS STRESS PARALLEL SOLVER INTRODUCTION HOW DOES PROCAST PARALLEL WORKS ? USE OF THE PARALLEL SOLVER REPEATABILITY LIMITATIONS MACHINE CONFIGURATION

319 320 321 322 322 324 324 327 331 331 337 341 351 369 370 376 385 385 394 395 396 397 398 399 399 400 401 402 405 406 407 408 408 411 416 421 422 423

Version 2006.0

Page 4

ProCAST User Manual

PARALLEL VERSION INSTALLATION HARDWARE AND OS LAM/MPI AND MPICH COPYRIGHTS ADVANCED POROSITY CALCULATIONS INTRODUCTION ADVANCED POROSITY PRE-PROCESSING GENERAL INFORMATION MATERIAL PROPERTIES GAS AND BUBBLE PROPERTIES PROCESS INFORMATION CALCULATION SETTINGS EXAMPLES PREFIX_PORO.D INPUT FILE ADVANCED POROSITY SOLVER ADVANCED POROSITY POST-PROCESSING MICROSTRUCTURES INTRODUCTION CASE SET-UP AND RESULTS EXAMPLES IRON AND STEEL INTRODUCTION TO IRON AND STEEL IRON AND STEEL MODELS CASE STUDIES "2-D" INTRODUCTION MESHING CASE SETTING CAFE-3D

425 428 430 433 433 435 436 436 438 440 444 447 452 454 458 458 462 467 474 474 480 484 505 505 507 514 517

INVERSE MODELING INTRODUCTION MODEL SET-UP INVERSE RUN FILE FORMATS INVERSE APPENDIX INPUT-OUTPUT FILES

524 524 527 534 538 542 552

Version 2006.0

Page 5

ProCAST User Manual

TIPS & TRAPS RESTART CONVERGENCE PROBLEMS STRESS CALCULATIONS GAPS IN STRESS MODELS STRESS VISUALIZATION TUTORIALS GENERAL INTRODUCTION FLOW CHART MESHCAST PRECAST PROCAST VIEWCAST RESTART BOUNDARY CONDITIONS SYMMETRY TIME STEPS MATERIAL PROPERTIES CASTING MATERIAL EXOTHERMIC MATERIAL FILTER MATERIAL MOLD MATERIAL THERMODYNAMIC DATABASES CYCLING MODELLING VIRTUAL MOLD PROCESS TEMPLATES HPDC-CYCLING HPDC-FLOW LPDC GRAVITY-SAND INVESTMENT FILTER STRESS NON-COINCIDENT MODEL

557 557 558 559 560 562 565 565 565 567 568 569 569 571 575 581 583 585 586 589 591 591 592 598 604 609 610 636 647 666 687 710 734 751

Version 2006.0

Page 6

ProCAST User Manual

INTRODUCTION
SOFTWARE CAPABILITIES
ProCAST is a software using the Finite Elements Method (FEM). It allows the modeling of Thermal heat transfer (Heat flow), including Radiation with view factors, Fluid flow, including mold filling, Stresses fully coupled with the thermal solution (Thermomechanics). Beside that, it includes also microstructure modeling and porosity modeling. Special models are included in order to account for thixo casting and lost foam. Specific features are included to account for processes such as high pressure die casting, centrifugal, tilt. Finally, customized models for foundry processes, such filters, sleeves are included.

Thermal calculation

Version 2006.0

Page 7

ProCAST User Manual

Mold filling and Fluid flow calculation

Thermomechanical calculation

Version 2006.0

Page 8

ProCAST User Manual

SOFTWARE ORGANIZATION
The software is organized around a Manager, which calls the different modules : MeshCAST : the mesh generator PreCAST : the pre-processor, coupled with databases DataCAST / ProCAST : the solvers ViewCAST : the post-processor and data export unit

The following figure is presenting the structure of the software. First, the geometry, in the form of a CAD model is loaded into MeshCAST, to generate a FEM mesh. Then, the calculation is configured in PreCAST, the Pre-processor. PreCAST is linked to Thermodynamic Databases for the automatic determination of the material properties from thermodynamic databases. Before the solver ProCAST is launched, a "data conditioner" named DataCAST is run. Finally, the results can be viewed or exported (for further processing) in the Postprocessor ViewCAST.

The ProCAST solvers are divided in "Physical modules" with the following capabilities :

Version 2006.0

Page 9

ProCAST User Manual

Thermal module
Heat conduction (Fourier equation) Latent heat release during solidification Cycling in die casting Sleeves (insulating and exothermic) Non-coincident meshes Solidification time Secondary Dendrite Arm spacing Porosity indicator

Radiation module
Net radiation method Full view factors capabilities Mirror and rotational symmetries Relative motion of materials Solid or surface enclosures

Fluid flow module

Navier-Stokes equation Penalization of the flow in the mushy zone and in the solid Mold filling algorithm, with free surface Filter model Newtonian and Non-Newtonian flow Thixo casting models Lost Foam model Tilt pouring Centrifugal casting Turbulent models

Stress module
Elastic, Elastic-plastic, Elasto-visco-plastic Rigid or vacant materials Automatic calculation of the air gap heat transfer Contact algorithm between the different materials Contact pressure Die Fatigue prediction Hot Tearing indicator

The features linked to specific processes (e.g. cycling, tilt, ...) are embedded in the corresponding physical modules.

Version 2006.0

Page 10

ProCAST User Manual

USER MANUAL PRESENTATION

After a "Getting started" chapter, which presents briefly the set-up of a simple thermal case, the different modules of the software, starting by the "Software Manager" are presented. The "Pre-processing" chapter is probably the most important as it describes the setting up of a case, from the FEM mesh to the run of the calculation. After an introduction on common features, this chapter is divided according to the "Physical modules", i.e. Thermal, Fluid Flow & Filling, Radiation and Stress. Then, the Databases, Run Parameters and Advanced features are presented. After the Run of the calculation chapter, three chapters are dedicated to the Results viewing, the Results analysis and the Results exports. Finally, Tips & Traps and Tutorials will illustrate the use of the software. MeshCAST, the mesh generator, is presented in a separate manual.

Version 2006.0

Page 11

ProCAST User Manual

WHAT'S NEW
VERSION 2006.0
This section is describing the main news of ProCAST v2006.0. The links are referring to the corresponding section of the manual for more details.

General / Manager
In the "Copy" panel, the possibility to access directly to the "Template" directory of the installation was added (see "Software Manager/File Manager" section for more details). The access to the CAFE pre-processor and post-processor were introduced in the Manager, as well as to the Advanced Porosity post-processor (see the "Software Manager/Module calls" and "Software Manager/Advanced module calls" sections for more details). A bug in the "Copy" functionnality was corrected. The possibility to launch DMP calculations on Linux in Batch was introduced (see the "Parallel Solver/Use of the Parallel solver" section for more details). The FlexLM libraries were upgraded from 9.2 CRO to 10.1.3 STL. A bug in the status window (wrong update of the number of cycles) was corrected.

MeshCAST
A link between GEOMESH and MeshCAST was created in order to speed-up the repair and surface meshing process (especially to account for the more efficient readers of GEOMESH) (see the MeshCAST manual for more details). Surface mesh Assembly (see the MeshCAST manual for more details). Boolean Assembly (see the MeshCAST manual for more details). The Surface mesh algorithm was improved in order to have a better quality surface mesh. Especially, cylinders defined by only one surface can now be meshed automatically. Thin cylinders (e.g. cooling chanels) are not collapsed anymore when the mesh size is coarse. The Autofix is now fully automatic (no more need to specify a length) A functionnality to remove automatically filets was introduced (see the MeshCAST manual for more details). The Layered shell mesher was re-introduced in MeshCAST (see the MeshCAST manual for more details). The format of the volume mesh has been adapted to allow the handling of files with more than 10 millions elements. When a Volume mesh is loaded, new buttons "Select All" and "Deselect All" were introduced in order to select/deselect automatically all the domains.

Version 2006.0

Page 12

ProCAST User Manual

The "Partial layer" option was corrected (it was giving the same results are the Full layer option). The "*.bstl" files can now be read on Linux.

PreCAST
Time-dependant Inlet boundary conditions were introduced (see the "Boundary Conditions Database" section for more details). In the case of shot piston, it is possible now to define the position of the piston in three different ways : position vs time, velocity vs time, velocity vs position (see the "Process Database" section for more details). Wall BC and Velocity BC are now exclusive (i.e. if a Wall BC with a "Select All" is set, the previously defined Velocity BC's will be preserved). Now, it is not anymore possible to set BC's on Periodic BC faces (the Periodic BC faces are "protected", like Symmetry faces). A new "Pick" function has been introduced, to pick the coordinates and the node number of a given point selected interactively on the model (see the "Geometry Manipulation" section for more details). Stress properties (Youngs modulus, Poisson Ratio, Expansion coefficient) are now automatically calculated, based upon the Thermodynamic databases (see "Calculation of Stress Properties" section for more details). When defining the interfaces, the "Wireframe" mode has been reactivated (in order to see a domain which is inside an other one). The Plastic stress properties (i.e. the Hardening) can now also be defined as a set of tabulated tensile tests curves in an ASCII file (see "Digitized Hardening" section for more details). In the case of Microstructure modelling, the default values of the Run parameters were removed (see the "Micro Run Parameters" section for more details). The mesh can be now optimized also in the case of Periodic BC. If a non-coincident mesh is used, a warning prevents the definition of a virtual mould. New User functions have been introduced (see "User Functions" section for more details). A problem in the initialization of CORE materials during cycling has been corrected. A problem of Extract of temperatures with the prefix_t.unf file larger then 2 GB has been corrected. Enclosures can now again be defined by QUAD elements A new Run Parameter (GATEFS) was introduced for the handling of the third stage pressure in HPDC (see the "Thermal Run Parameters" and "Active Feeding" sections for more details). The set-up of Centrifugal cases was changed for more accuracy and the RELVEL Run Parameter was introduced (see the "Centrifugal casting" and the "Flow Run Parameters" sections for more details). Due to the changes in PreCAST and DataCAST, it is advised to reload the d.dat and p.dat files generated on previous versions and save them before rerunning the cases with the version 2006.0.

Version 2006.0

Page 13

ProCAST User Manual

The Run Parameters Recommendations have been updated in order to account for the changes in the algorithm (see the "Run Parameters Recommendations" section for more details).

Thermodynamic databases
Stress properties (Youngs modulus, Poisson Ratio, Expansion coefficient) are now automatically calculated, based upon the Thermodynamic databases (see "Calculation of Stress Properties" section for more details). The latest available Computherm databases are included in this version. The material properties calculated with these databases will be more accurate than the previous ones. The Computherm manual (from Computherm LCC) is added in the Software installation (in the dat/manuals/PDF direcrtory). This manual describes for each alloying system the phases which are calculated, the limitations as well as the validations which have been made.

ViewCAST
A new "Pick" function has been introduced, to pick the coordinates, node number and values (e.g. Temperature, Pressure, ...) of a given point selected interactively on the model (see the "Geometry Manipulation" section for more details). The liquidus and solidus temperature are now shown on the Temperature scale (see the "Results Viewing/Display Parameters" section for more details). The previously defined scales (i.e. the scales defined manually during a previous session) are now automatically stored and loaded (see the "Results Viewing/Display Parameters" section for more details). Stored views were introduced (see the "Results Viewing/Geometry Manipulation" section for more details). When a model is loaded for the first time (or when no view was stored), the model is displayed in isometric view, with the gravity pointing downwards. In case of a Stress calculation, it is now possible to calculate all the Stress results (e.g. Effective Stress, Principal Stress, SigmaX, ...) all at once before viewing them (see the "Results Viewing/Field selection" section for more details). In the case of stress calculation, the possibility to view the total displacement was introduced (see the "Results Viewing/Field selection" section for more details). The displacements relative to a plane were introduced for the visualization of deformed geometries (stress calculation) (see the "Results Viewing/Field selection" section for more details). When a stress calculation is performed, it is possible to display the underformed geometry in wireframe together with the deformed contours (see the "Results Viewing/Display Parameters" section for more details). The calculation of the solidification time was improved in order to account for remelting (if any). The calculation of the Filling time was changed in order to ignore the piping area.

Version 2006.0

Page 14

ProCAST User Manual

The pressures calculated by the "Advanced Porosity module" can now be visualized in ViewCAST (see the "Advanced Porosity Post-processing" section for more details). In order to better visualize pockets of air, a new display mode of the free surface ("Foreground") was added (see the "Results Viewing/Display Parameters" section for more details). Stress results can be viewed also when the directory is in Read-only mode (on Windows only). This is especially useful to view stress cases stored on a CD or a DVD. The stress results are computed and stored in a temporary file on the local disk of the computer (see the "Results Viewing/Field selection" section for more details). If more than one symmetry is defined in the model, it can be automatically retrieved (it was previously working only with one symmetry). In the case of a Tilt casting, it was mandatory to have the p.dat file present in the directory to view the case. This obligation was removed. In the XYPlot window, when nodes were selected interactively, the selection window (with the cursors) was disappearing each time the model had to be rotated or moved. This limitation is now removed.

Solver
The filling algorithm was improved. A new Run Parameter FREESFOPT was introduced (see the "Flow Run Parameters" section for more details). The filling solver was improved in order to better handle Tilt pouring models. A new Run parameter "TILT=1" was introduced in the case of Tilt models (see the "Flow Run Parameters" section for more details). For centrifugal casting, the algorithm was improved by the introduction of the RELVEL Run Parameter (see the "Centrifugal casting" and the "Flow Run Parameters" sections for more details). The convergence, the accuracy and the CPU time of the Stress module was significantly improved. Two new models (Norton law and Strain Hardening Creep) were introduced in the Stress module to account for viscoplasticity and creep (see "Stress models and Properties" section for more details). The Plastic stress properties (i.e. the Hardening) can now also be defined as a set of tabulated tensile tests curves in an ASCII file (see "Digitized Hardening" section for more details). In the case of Stress calculations with Vacant or Rigid materials, the memory management was changed in order to reduce the amount of memory needed (in the scalar version only). Stress calculations can now be run with piping (i.e. it is now allowed to set PIPEFS<1 in the case of stress calculations). The POROS=1 model was extended to allow the treatment of the Graphite expansion in the case of SGI (see "SGI Porosity model" section for more details). The POROS=1 model was improved for the treatment of the third phase in HPDC. A new Run Parameter (GATEFS) was introduced (see the "Thermal Run Parameters"and "Active Feeding" sections for more details).

Version 2006.0

Page 15

ProCAST User Manual

A bug in the POROS=1 model was corrected (in the identification of the independant liquid pockets). A problem in the POROS=1 model with FREESF=2 filling was corrected. Microstructure models were improved (see "Microstructures" section for more details). The memory requirement was significantly reduced when running a case with many materials. Time-dependant Inlet boundary conditions were introduced (see the "Boundary Conditions Database" section for more details). New User functions have been introduced (see "User Functions" section for more details). A problem of Restart in the case of gravity casting was corrected. The problem of heat transfer between a shot piston and the casting (there was no heat transfer) has been solved. The logic of the flow Switch Off was changed in order to avoid some unexpected situations. If LVSURF < 1, the value of PIPEFS can not be set to 1. A problem in the FLOW=9 option was corrected. It is possible now to run Stress models with materials which have no phase change. A bug in the case of Stress calculation in the presence of filters has been corrected. It is possible now to run a Microstructure calculations together with the Turbulent model. A problem in Cycling with FREESF=1 was corrected. In the case of moving solids, the translation is also called at step 0. A problem in the case of DIE COMBO was corrected. A problem in the calculation of the piping when using Periodic Boundary conditions was fixed. A problem in the calculation of the "Freckles predictor" has been corrected, as well as a problem of Restart with Freckles prediction. If the "statpro" directory (where the status file is written) is momentarily not available, the calculation will not crash anymore. In the case of delayed filling (i.e. if the filling does not start since the beginning of the analysis), the DTMAXFILL timestep will not be active before the filling starts. A bug in the treatment of the TOPEN has been corrected. A problem in the Units of the user function "externalcompute.c" was corrected. A problem of "datacast -u" in the case of stress was fixed (the stresses of vacant domains were lost). A problem of convergence with non-coincident meshes was fixed. A problem of memory allocation was corrected in the case of Radiation problems. The Run Parameters Recommendations have been updated in order to account for the changes in the algorithm (see the "Run Parameters Recommendations" section for more details).

Version 2006.0

Page 16

ProCAST User Manual

Parallel processing (DMP)

The Stress module has been parallelized. Please note that no specific set-up should be done for parallel processing of Stress calculations. The set-up exactly the same as the scalar version. The "DataCAST" module is now the same between Scalar and DMP (e.g. one can run DataCAST on a scalar machine and then launch the calculation on N CPU's with the DMP version or vice-versa). ProCAST 2006.0 contains the first release of the Windows DMP version (see the "Parallel solver" chapter for more details). The Lost Foam and the Non-Newtonian module has been parallelized (same set-up as the scalar version). The TRI2QUAD=1 functionnality (for radiation) has been parallelized. Some problems encountered in the DMP versions were fixed.

CAFE
The CAFE pre-processor and post-processor were introduced in the Manager (it is no longer needed to have the calcosoft-3D installation to set-up and visualize CAFE calculations) (see the "CAFE-3D" chapter for more details). The CAFE solver (in post-processing mode) was introduced in the ProCAST 2006.0 distribution (it is no longer needed to have the calcosoft-3D installation to run CAFE calculations) (see the "CAFE-3D" chapter for more details).

Advanced Porosity module

The quality of the solver was significantly improved in order to have a better convergence (robustness) and a better accuracy of the results. The defalult value for the "convergence.tolerance" was modified (from 1e-12 to 1e-10) in order to be compatible with the modified algorithm. A bug in the storage frequency was corrected. The pressures calculated by the "Advanced Porosity module" can now be visualized in ViewCAST (see the "Advanced Porosity Post-processing" section for more details). It is possible to view Advanced Porosity results (with the calcososft postprocessor) across plateforms (automatic swap option). The calcosoft post-processor allows now to view results calculated using Hex or Wedges mesh. A bug in the viewing of Advanced Porosity results in the presence of enclosures has been corrected. Template files of the poro.d input file were added in the software installation (see the Template button in the "Copy" of the Manager in the end of the "Software Manager/File Manager" section for more details)

Manual
(The following sections of the manual were added or changed) The Stress properties recommendations were modified. One should note that it is very important to set stress properties according to the "rules" described in

Version 2006.0

Page 17

ProCAST User Manual

the manual. Othewise, one could have convergence problems (see the "Stress Models and Properties" section for more details - end of the section). The Computherm manual (from Computherm LCC) is added in the Software installation (in the dat/manuals/PDF direcrtory). This manual describes for each alloying system the phases which are calculated, the limitations as well as the validations which have been made. The Run Parameters Recommendations have been updated in order to account for the changes in the algorithm (see the "Run Parameters Recommendations" section for more details). A new section was added in the "Microstructure Module" chapter (see the "Iron and Steel" section for more details).

Version 2006.0

Page 18

ProCAST User Manual

VERSION 2005.0
This section is describing the main news of ProCAST v2005.0. The links are referring to the corresponding section of the manual for more details.

General / Manager
The manager was modified in order to allow the launch of Parallel calculations, of the Advanced Porosity module and of the Inverse module (see the "Software Manager/Advanced module calls" section for more details). A "Run list" functionality was added in order to launch automatically a serie of calculation (batch processing) (see the "Software Manager/Run list" section for more details). A Graphical status was introduced in the Manager (see the "Software Manager/Module calls" section for more details). The F2 key can be used to scroll between the different case names which are present in the same directory (see the "Software Manager/File manager" section for more details). The case name was added in the top bar of all the modules (MeshCAST, PreCAST, ViewCAST). The suffix ".mesh" has been added to the Filter of the "File Manger". The FlexLM locking system was upgraded to FlexLM 9.2 CRO, for a better security. This version will work with 2005 CRO license keys. LMtools can be accessed directly from the Manager (see the "Software Manager/Software configuration" section for more details).

MeshCAST
The call of the Volume mesher of MeshCAST in batch was re-introduced (see the end of the "Software Manager/Module calls" section for more details). The possibility to read Parasolid files with the .x_t suffix was added. The automatic "zoom-out" was suppressed when the ADDLINE operation was performed. A units problem in MeshCAST-2D was corrected.

PreCAST
A GATENODE Run parameter was introduced in order to handle "Gate feeding" with a shot piston (see the "Active Feeding" and "Thermal Run Parameters" sections for more details). The computation algorithm of the Virtual mold was modified (to avoid odd behavior in some cases). The virtual mold size can be defined automatically (see the "Virtual mold" section for more details).

Version 2006.0

Page 19

ProCAST User Manual

The possibility to activate User functions was introduced in PreCAST (see the "User Functions" section for more details). A bug in the units of the "Tilt" axis was corrected. A bug in the storage of the "K-factor" (non-Newtonian model) was corrected. The "Extract" was extended in order to allow it in the presence of filters. The "automatic" setting of FLOW=3 was reviewed. A "Erase line" was added in the "Chemical Composition" definition of the material properties. The TENDFILL Run parameters was introduced in order to automatically stop a calculation at the end of the filling (see the "General Run Parameters" section for more details). New alloys were introduced in the Material database. Some properties of existing alloys were improved. The Stress material properties database was complemented. A bug in the treatment of non-coincident interfaces was corrected (it was creating a double heat transfer on some nodes in some specific cases). The COARSEC and COARSEP Run parameters were removed from the interface as they are no longer used. When PreCAST is opened, the model is oriented according to the ISO orientation.

Thermodynamic databases
The automatic calculation of the "Thermal conductivity" from the thermodynamic databases was introduced for all databases (see the "Thermodynamic Databases" section for more details) The content of the Computherm databases was improved. This will lead to more accurate properties predictions. Moreover, the computation of some alloys (which was not possible with previous versions) was improved. This means that the properties calculated with this version will be different (but more accurate) than the previous ones (e.g. the solidus temperature of Nibased alloys is now much more realistic than before). A problem in the Mg database was fixed, giving now the right results. The liquid density of Pure Aluminium was changed in order to give more accurate density predictions of Al-alloys (e.g. in the previous version, the solidification shrinkage of an A356 alloy was about 9.15%, whereas it is now 5.45%). This will affect the density calculations for all Al alloys.

ViewCAST
ViewCAST was adapted in order to allow the viewing of result files coming from a different platform (e.g. viewing on Windows results computed on UNIX) (see the "Software Manager/Software configuration" section for more details). Automatic creation of AVI animations was introduced (see the "Resultsviewing/Introduction" section for more details). The GIF/AVI capture action can also be started from the "File" menu.

Version 2006.0

Page 20

ProCAST User Manual

Export of the "Shrinkage porosity" and of the "Advanced porosity" fields has been introduced. An automatic pre-defined setting for "Shrinkage porosity" was introduced (see the end of the "Results viewing/Display types" section for more details). Short cuts for the different display modes were introduced (see the "Results viewing/Display types" section for more details). The export of "Reversed displacements" was corrected, as well as the X-Y plots of displacements. The maximum number of timesteps which could be exported was increased from 200 to 5000. In Tilt casting, the particles are now "tilting" with the part. In addition, a new "Tilt" option in the "Parameters" menu was introduced in order to deactivate the tilt during the viewing (see the "Display Parameters" section for more details). The "Fill time" plot was introduced. A problem with the calculation of RGL on Linux was fixed.

Solver
User functions were introduced for "interface heat transfer coefficients" and "Heat boundary conditions" (see the "User Functions" chapter for more details). A bug with the TCLOSE (DIE COMBO cycling) was corrected. A problem in the cementite formation of cast iron (microstructure module) was fixed. The interpolation of the linear thermal expansion coefficient (stress module) was changed from linear by intervals to piecewise linear. A problem with the calculation of stress in the presence of piping (stress module) was corrected. In such case, one should set PIPEFS = 0 (see the "Thermal Run Parameters" section, under PIPEFS for more details). A new possibility was included in order to handle the "Gate feeding" in hpdc in the presence of a shot piston (see the "Active Feeding" and "Thermal Run Parameters" sections for more details about the GATENODE Run Parameter). A bug in the POROS 1 model was corrected (to avoid unexpected behaviors in some specific cases). Moreover, the display of the macroshrinkage (i.e. holes inside the casting) has been changed (one will have shrinkage porosity values of 1, but no internal "Voids" will be shown. "Voids" will be used only for pipe shrinkage at the surface of the model). Please note that the value of FVOL has no meaning for the viewing of porosity results. A problem in the Surface load (stress module) which was appearing in some cases was corrected. The influence of the gap width on the interface heat transfer coefficient, with non-coincident meshes, was introduced. The viscoplastic model (stress module) was improved, for a better accuracy. The stress module was extended in order to allow the use of filters. DIE COMBO was extended for the use with non-coincident meshes. A problem in the handling of the "RESERVOIR" was corrected in order to allow to fully empty the reservoir.

Version 2006.0

Page 21

ProCAST User Manual

A problem with the Freckles indicator was fixed. A new stop criterion (TENDFILL) was introduced. It allows to stop the calculation N seconds after the end of filling (see the "General Run Parameters" section for more details). The CAFE-3D coupled solver was introduced in the standard ProCAST solver executable (see the "CAFE-3D" section for more details). Undesired "Sticking" in the pouring cup during Tilt pouring was improved. For that purpose a new TILT Run Parameter was introduced (see the "Flow Run Parameters" section for more details). A "loading increment" Run Parameter LOADSCL was introduced for more accuracy in structure analysis type of problems (see the "Stress Run Parameters" section for more details). A problem with RDEBUG=2 was fixed.

Parallel processing
The Parallel processing solver was introduced in this version for selected platforms (see the "Parallel Solver" chapter for more details).

Advanced Porosity module

The Advanced Porosity module (originally developed in CALCOSOFT-3D) was introduced in the ProCAST distribution (see the "Advanced Porosity Calculations" chapter for more details). The module was improved (with respect to one of the CALCOSOFT-3D 2004.0 distribution). Extension of the Advanced Porosity module in order to allow the calculation of the porosity in more than one casting domain. Direct reading of the ProCAST files was introduced (no more result files conversion needed). Storage of the porosity results at FEM nodes, allowing the visualization of the results in ViewCAST. Removal of the enthalpy conversion, allowing for a direct compatibility of the data between the ProCAST and the Advanced Porosity module. Possibility to define directly a density curve (instead of expansion coefficients of the different phases) and thus to use the same density curve as in the ProCAST calculation (the density curve of ProCAST is automatically used and thus, the definition of this curve is not anymore necessary in the porosity module). Direct reading of the fraction of solid and density curves in the ProCAST data files (the definition of these curves is not anymore necessary for the porosity module).

Manual
(The following sections of the manual were added or changed) A new section was added to list the limits in the chemical compositions for the Thermodynamic databases (see the "Databases limitations" section for more details).

Version 2006.0

Page 22

ProCAST User Manual

A section on the "Influence of alloying elements" on thermophysical properties obtained with thermodynamic databases was added. A new "Software Manager/Advanced module calls" section was added. A new "Software Manager/Run list" section was added. The "Software Manager/Software configuration" section was complemented. A new "Parallel Solver" chapter was added. A new "Advanced Porosity Calculations" chapter was added. A new "User Functions" section was added in the "Pre-processing" chapter. A new "CAFE-3D" chapter was added. The "Virtual mold" section was modified in order to account for the automatic definition of the box size. The description of the new Run Parameter TENDFILL was added in the "General Run Parameters" section. The description of the new Run Parameter GATENODE was added in the "Active Feeding" and "Thermal Run Parameters" sections. Updates in the "Results viewing/Display types" section were made. Details about non-coincident tolerances were added at the end of the "Multiple Meshes and non-coincident meshes" section. The description of how to disable both the porosity and the piping calculation (by setting POROS = 0 and PIPEFS = 0) is now included in the "Thermal Run Parameters" section (under PIPEFS). The set-up of a cycling calculation including a Porosity calculation is now described at the end of the "Cycling" section. The transition between plastic and viscoplastic models was revised in the "Stress Models and Properties" section. Advices about the setting of Yield stress values in the liquid have been added at the end of the "Stress Models and Properties" section. The principles of plastic and viscoplastic properties determination from experimental measurements are now described.

Version 2006.0

Page 23

ProCAST User Manual

VERSION 2004.1
This section is describing the main news of ProCAST v2004.1. The links are referring to the corresponding section of the manual for more details.

General
On some platforms, the graphics performances were very poor (in terms of display speed). This was corrected. However, it may still happen that on some graphics cards, the graphical performances are not optimum (in terms of display refresh). In such cases, the configuration of the X-Server may need to be adjusted. The installation structure of the software was complemented in order to allow the installation in a Read only location. Each user has now the possibility to set a customized directory for his Preferences (including the MeshCAST user customized icons) and databases. (Customized installation).

MeshCAST
On some platforms (mainly Windows), MeshCAST was not always able to use all the available memory (e.g. it was crashing at 1.1 GB, whereas the computer had 2 GB of RAM). This problem was due to the way "reallocation" of memory was done in the code and how the operating system was managing the memory allocation. The memory "reallocation" technique was modified in order to prevent such problems, allowing to use all (or most) of the available memory. In such situation, temporary files will be created in the working directory (see the MeshCAST manual for more details about this option and the "Module calls" section for the activation of this option). In the case of problems with Shell meshing with symmetry, a new functionality was added in order to allow the curing of the problem (see the MeshCAST manual for more details about this option). The User defined icons can now be stored in a central location (or in a customized user defined location) (Customized installation). Some corrections when reading Parasolid files were made. ANSYS surface mesh input is now available.

PreCAST
The viscosity and density can be automatically calculated from the thermodynamic databases, based upon the chemical composition (Thermodynamic Databases). The thermal conductivity for Aluminium alloys can be determined automatically from the thermodynamic databases, based upon the chemical composition (Thermodynamic Databases).

Version 2006.0

Page 24

ProCAST User Manual

A new Run Parameter was introduced (DTMAXFILL) in order to define the maximum timestep to be applied during the filling stage. This run parameter is optional (General Run Parameters). A "Die close" was introduced in the Die Combo definition. This allows to account for a closed period of the die before the next filling starts (Interface Database, Cycling Run Parameters) A special effort in the "Standard" Run Parameters definition was made (in PreCAST and in the Solver). Thus, it is now possible to set "Standard" sets of Run parameters for each process (see the "Run parameters Recommendations" section for more details). The automatic Mesh Optimization was introduced in PreCAST. This means that it will not be necessary anymore to optimize meshes in MeshCAST and that the optimization will automatically take into account the interface set-up (Warning : the Optimization should not be used on a model which contains Extracted initial conditions coming from a non-optimized model). A more sophisticated model (but simple to use) to define filters (with pressure drop information) was introduced, aside from the current filter model (see the definition of Filters in the "Material Properties" section). The inverse set-up was re-introduced in PreCAST. This was the opportunity to simplify it to make it more user friendly. The Microstructure Run Parameters were introduced in PreCAST. The size of the model (and not only the min-max) is now displayed when the case is loaded. A correction in the "Extract menu" was done. A correction of the Virtual mold scale display was done. A bug in the definition of time-dependent displacements was corrected. The "Select Remain" was selecting both outside faces and faces at interfaces. This was corrected. Only the outside surfaces are selected. The direct access in databases with a Right Mouse click was extended to all databases. Units in the "Thermal Expansion coefficient" (Stress database) were corrected.

ViewCAST
A new "Particle tracing" has been introduced. The streamlines are automatically computed during the calculation and thus the display in ViewCAST is fast (see "Display parameters" for the viewing and "Flow Run Parameters" for the set-up). When a case is viewed during the calculation, the timesteps can be automatically updated during the ViewCAST session (see the Display parameters section for more details). In XY-Plots, the Import/Export of the node numbers (as well as the scale) was not working well. This was corrected. See the "Curves" section for the format of the Import/Export. The GIF capture was not always working with screen resolutions different from 1280x1024. This was corrected. When capturing a GIF picture/animation, if the ".gif" extension is not specified, it will be automatically added.

Version 2006.0

Page 25

ProCAST User Manual

The units of the SDAS display were wrong. It was corrected. The node numbers and elements numbers can now be displayed (see the "Geometry manipulation" section). During the filling calculation, a "smoothing" algorithm of the free surface was suppressed (as it was causing some problems in some filling situations). Thus, in the case of coarse mesh or strong zoom, one could see a more "rough" free surface than in previous versions (this does not affect the quality of the filling pattern). A bug in the stress solver, in the presence of non-coincident meshes was corrected.

Solver
Although many improvements were made in version 2004.0 in the filling algorithm, in order to reduce numerical diffusion (to have "less viscous" liquid), this did create some detrimental behaviors in some situations. Thus, the filling algorithm was further refined and tests have shown much better results in all the geometries and processes which were calculated. A special effort was put in the Run Parameters settings for filling simulation. Now, the default values can be used for all processes (only WALLF should be customized in order to account for the "friction" of the liquid at the mold wall) (Flow Run Parameters, Run Parameters Recommendations). In LPDC, the stability of the filling with a pressure boundary condition was totally reviewed and now good results are obtained also with pressure filling (Flow Run Parameters, Run Parameters Recommendations) The POROS=1 model was improved and a few bugs were corrected. The description of the POROS=1 model was also improved (see in particular the description of the effect of FEEDLEN) The visco-plastic model was slightly changed in order to allow a more straightforward definition of the input parameters from measured data and the possibility to set a Norton law (with threshold) for creep (Stress models and properties). One should be careful that viscoplastic models which have been set in previous version should be modified in PreCAST in order to add the normalization stress. Otherwise, the solver will not work. Porosity calculations can now be done together with cycling. In the Lost Foam model, there was situations where small bits of foam were not burning. This situation was corrected. The stress solver was extended in order to account for moving solids. A bug in the Restart of Stress calculation was corrected, leading to a much improved convergence of the restarted step. A bug in the GATEFEED=1 option was corrected in the case of timedependant pressure or inlet velocity (GATEFEED was not activated in this cases) A check was introduced at the beginning of the calculation in order to prevent a missing pressure BC when GATEFEED is set to 1 (to prevent piping in pressure die casting, in thermal only calculations). A bug was corrected in the "inlet BC" algorithm when nodes at the interface between the casting and the mold were selected. One should be careful in such

Version 2006.0

Page 26

ProCAST User Manual

cases that the filling time may not be exactly correct (if interface nodes are selected). A bug in the micro module for hyper-eutectic cast iron was corrected. When the microstructure module is used, the material properties definition has slightly changed (as some data can be automatically obtained from the thermodynamic databases) - see the "Microstructures/Case set-up and Results" section for more details. A print-out bug was corrected in the d.out file. On some platforms, it was still not possible to generate files larger than 2 GB. This is now corrected. One should be careful however that the file system (kernel) should be well configured in order to accept such large files. However, as it was not possible to generate a 64 bits version of the User interfaces (and especially ViewCAST) on IBM (due to Tcl/Tk problems), it is not possible to visualize such large models on IBM at this stage. Work around : visualize these results on a different platform. A bug in the calculation of the permeability of a mushy zone at low fraction of solid was corrected. This has an effect on cold shots or early solidification during filling (in case of very large interface heat transfer coefficients). DataCAST execution time was reduced for large models.

Manual
(The following sections of the manual were added or changed) New Microstructures module chapter New Inverse module chapter New Thixo casting chapter New Lost Foam chapter New section on Thermodynamic databases for automatic computation of density, viscosity and thermal conductivity New section on Mesh Optimization in PreCAST New section on Particle tracing (see "Display parameters" for the viewing and "Flow Run Parameters" for the set-up) New section on "customized installation" with a Read/Write zone New section on the Input-Output files New section and changes in the Filter definition (see the "Material Properties" section for more details) New "Run parameters Recommendations" section New Run parameter DTMAXFILL (General Run Parameters) New Run parameter TCLOSE for Die Combo (Interface Database, Cycling Run Parameters) New Run parameter VACUUM for stress calculations (Stress Run Parameters) Changes in the material properties definition in the case of Microstructure calculations have been made ("Microstructures/Case set-up and Results") Correction of the GAPMOD Run parameter definition (Stress Run Parameters)

Version 2006.0

Page 27

ProCAST User Manual

Correction of the definition of the modified interface heat transfer coefficient during stress calculations, when air gaps are formed or when there is a contact pressure (Gaps in stress models) Changes in the definition of the visco-plastic and creep model (Stress models and properties) Correction in the description of the FEEDLEN effect in the POROS=1 porosity model. New tutorial for LPDC.

Version 2006.0

Page 28

ProCAST User Manual

VERSION 2004.0
This section is describing the main news of ProCAST v2004.0. The links are referring to the corresponding section of the manual for more details.

General
All the user interfaces have been re-developed, using the Tcl/Tk and OpenGL technology. Thus, NutCracker is not anymore necessary to run the software on Windows machines The software is licensed using the FlexLM tool (FlexLM). This gives the possibility to have floating licenses Direct access to the on-line help

Manager
A new Software Manager was introduced (to replace "PCS") Directories can be created and managed (File Manager) Cases can be automatically copied, with prefix changes (File Manager) The different modules can be called interactively (Module calls) The calculation can be monitored with a direct status access (Module calls) The software configuration can be tailored in the Manager (Software configuration)

MeshCAST
The volume meshing algorithm is significantly improved in order to generate less elements and nodes with a better quality The 2-D mesh generator is replaced by a new 2-D mesh generator based upon the one of calcosoft-2D. It is extended in order to automatically generate 3-D meshes of one slice for both cartesian and axisymmetric modeling The maximum number of surfaces which can be allocated to an enclosure was increased to 5000 The new user interface is available for UNIX/Linux users

PreCAST
PreCAST was totally re-developed in a new user interface, based upon the same fundamental principles as the previous version (Pre-processing) Interactive tools to manipulate the geometry (rotate, zoom, center, drag) (Geometry manipulation) Different display types (wireframe, hidden, hidden mesh, shade) (Geometry manipulation) New symmetry definition, with interactive pick of the coordinates (see the Radiation section)

Version 2006.0

Page 29

ProCAST User Manual

Enhanced definition of the Virtual mold, with possibility to view the thermal depth and the mold box (Virtual Mold) Display of the size of the model (Geometry Assignment) New Material database design (Material Database) Coherency about material properties definition (enthalpy, Cp, fs and L) (Material Properties) Direct visualization of tables (Material Database) Direct access to material and boundary condition databases (Materials assignment, Boundary conditions assignment) Search and Sort tools for each database (Material Database) Exothermic material definition (Exothermic) More explicit definition of the material domain types (casting, mold, foam, filter, exothermic, ...) (Materials assignment) Definition whether a material domain is empty or not in the "Materials assignment" menu Possibility to remove interfaces which have been previously created (Interfaces assignment) More flexible way to define non-coincident interfaces (Interfaces assignment) Possibility to change interactively the non-coincident interface tolerance (Interfaces assignment) In all databases, possibility to Import or Export the tables (Material Database) Enhanced selection of boundary conditions, automatic propagation, selection of remainder, selection of interfaces (Boundary conditions assignment) Copy, Paste and Clip capabilities in the Boundary conditions selection (Boundary conditions assignment) A "velocity calculator" was introduced in order to calculate the desired inlet velocity boundary condition (Boundary Conditions Database) For radiation problem with an enclosure, the alignment of the surface vectors is simplified (Radiation) For all databases, the database management (add, read, copy, delete) can be done in the same window as the assignments (Material Database) New design of the Run Parameters layout (Run Parameters) Automatic settings of up to nine Run Parameters Templates (Pre-defined Run Parameters)

ViewCAST
ViewCAST was totally re-developed in a new user interface, based upon the same fundamental principles as the previous version (Results viewing, Results analysis, Results export) Better rendering of the models and of the free surface, including shading and transparency Interactive tools to manipulate the geometry (rotate, zoom, center, drag) (Geometry manipulation) Full capability "Tape player" to step forward and backwards in single steps or animated mode (Tape Player)

Version 2006.0

Page 30

ProCAST User Manual

More comprehensive way to select the different view modes of the material domains (solid, invisible, wireframe, hidden mesh, shaded, transparent) (Geometry manipulation) Direct access to the different view modes (snapshot, slice, scan, cut-off) (Display types) Interactive slicing is available at any time (Display types) Enhanced timestep display definition, with the possibility to define time intervals for the display (in addition to step intervals) (Display parameters) Full export of stress results (in a Neutral format which can be imported in any stress software) (Results Export/Stress) Full export of the displacements (in a case of stress calculation), with the possibility to "reverse" the displacements in order to perform reverse engineering of distorted molds (Results Export/Displacements) Automatic calculation of the solidification time (Field selection) Enhanced symmetries settings (same panel as in PreCAST) (Display parameters) Merge of the former PostCAST capabilities into ViewCAST. This allows to have only one post-processor for the snapshots and the time evolutions curves (Results viewing/Curves), as well as for the generation of outputs, such as Niyama, cooling rates, SDAS, ... (Results analysis/Criterion functions) The XYPlots settings have been totally reviewed (Results viewing/Curves) The node selection for XYPlots can now be stored, as well as the scale of the graph for further use (Results viewing/Curves) Three different modes for the viewing of the free surface have been introduced (On, Off, Only) (Display parameters) The definition of the scale is totally new and interactive. The scale is kept in memory for each field (Display parameters) The materials can be "exploded" for a better viewing (Geometry manipulation) The display of vectors is improved. It is possible to define the length of the arrows, as well as to have a unique length for all the arrows (Display parameters). The unit vector length is now shown on the screen during snapshots and slices An automatic scan of slices (scanning along the X-, Y- or Z-direction) can be done at a given step (Display types) The definition of slices (cuts) is improved. It is now possible to view or not the geometry behind (or ahead) of a slice (Display types) The "cut-off" capabilities were improved. It is now possible to specify two cut-off values (Display types) The display in cut-off mode in the presence of a free surface is now well treated (Display types) The information about the case status (prefix, displayed field), time, timestep has been reviewed and is now shown on the bottom of the screen (Results viewing/Introduction) The free faces information is now generated in DataCAST and thus the loading of a case is much faster Full viewing capabilities are now available when symmetries are activated (cut-off, free surface, ...)

Version 2006.0

Page 31

ProCAST User Manual

Solver
The flow solver for filling was extensively reviewed. In particular the flow along walls was significantly enhanced (WSHEAR=2). The stability of the solver was thus improved for filling (Flow Run Parameters) A new Run Parameter (WALLF) was introduced in order to better account for the surface quality of the mold (Flow Run Parameters) The EDGE Run parameter was suppressed (due to the WALLF introduction) The end of the filling can be shortcut with the new Run parameter ENDFILL (Flow Run Parameters) A new "Pool" model was introduced (RESERVOIR type) in order to facilitate the modeling of cases with large horizontal free surfaces, such as tundishes (Materials assignment) Improved lost foam model The inlet boundary condition was improved in order to preserve the right amount of incoming mass The Porosity model (POROS=1) was significantly improved in order to give more accurate predictions (Porosity models) The density definition in the case of porosity calculations was extended (Density definition) The gate feeding (in the case of injection - hpdc - lpdc) was changed in order to be fully automatic (Active feeding) Enhanced enthalpy formulation for more accuracy of the thermal solver Exothermic materials can be modeled (Exothermic) Large files (over 2 GB) can now be handled The visco-plastic model for stress was adapted to account for creep (Stress models and properties) Fatigue prediction is available (Fatigue life indicator) Hot tearing calculations are available (Hot tearing indicator) Cracking indicator is introduced (Cracking indicator)

Version 2006.0

Page 32

ProCAST User Manual

IDENTIFIED BUGS, PROBLEMS AND LIMITATIONS

This section contains the identified bugs and problems which could not be corrected in this version.

General
On UNIX, the command window which is used to launch the manager is not available for any further operations (as long as the manager is open). On some platforms, there may be an error message when exiting the different modules. This message can be ignored.

Manager
Problems may be encountered if the working directory name contains dots ("."). If the Status is refreshed too often (by clicking on the "Status" button), a Tcl error may be encountered. This happens if the refresh is done exactly in the same time than the writing of the stat file by the solver. This can also happen if two refresh are done in a very short time interval.

PreCAST
On some platform, the "Get coords" functionality to define symmetries is not working properly. On Sun, windows are opened behind the application, which can not be accessed. The display of the Virtual mold thermal depth has some problems. If GATEFEED is set to 1 and POROS to 0, the calculation does not start if no pressure BC is set (as there is a mistake in the logic of such testing). Work around : one should set GAFEFEED = 0 when POROS = 0. When a case is opened directly from PreCAST (and not from the Manager), if this case is located in a different directory (from the directory where PreCAST was launched) and if this directory contains a ".", the case can not be opened. Work around : open the case directly from the Manager. The material names should not contain blank spaces (this will generate an error in DataCAST : "An enthalpy function is referenced in 5 0 data set 1 which has not been given". Work around : if an old "d.dat" file is containing such a material name, one can edit the d.dat file and change the material name in the "card 5.0" to suppress the blanks ( 5 0 1
NICKEL_IN_718 1 )

When a translation velocity as a function of position, v(x), is defined by a table, the velocity values should not contain any zero values.

ViewCAST
Some problems were observed in the visualization of stresses on SGI. When displaying a slice with vectors, some problems may occur.

Version 2006.0

Page 33

ProCAST User Manual

When displaying vectors in a slice, the contour of the slice does not appear. It is not possible to superimpose a contour and vectors (in slices). Sometimes, the particle tracing is giving wrong results at the last steps of the filling. In same cases, the automatic timestep update is freezing ViewCAST on SGI. Work around : do not activate the automatic timestep update on SGI. The visualization of result files larger then 2 GB is not possible on IBM (as the executable of ViewCAST is 32 bits). Due to Tcl/Tk problems on IBM 64 bits, it was not possible to create an executable of ViewCAST 64 bits. Work around : use an other platform to visualize such results. On UNIX, depending on the configuration of the X-Server, it was observed that the performances of the graphics may be not optimum. When a case is opened directly from ViewCAST (and not from the Manager), if this case is located in a different directory (from the directory where ViewCAST was launched) and if this directory contains a ".", the case can not be opened. Work around : open the case directly from the Manager. Particle tracing is not available in Tilt and Reservoir cases (i.e. when there is no inlet velocity, inlet pressure on inlet BC). In Tilt casting the velocity vectors will not "rotate" with the part and thus they will appear in the wrong direction. Work around : deactivate the tilt rotation in ViewCAST when viewing at velocity vectors. It is not possible to visualize only a filter. Work around : select also an other material. In Lost foam calculations, one see non-zero velocities in the mould at step zero. Just ignore these velocities which are meaningless. It is not possible to view the Particle tracing with symmetries (i.e. the traces are not replicated).

Solver
It is not possible to fill simultaneously two separated cavities. One should fill only one cavity at a time or set (by hand in the p.dat file) the Run parameter COLDSHUT = 0 (this will deactivate the cold shut detection algorithm). In models where there are more than one RESERVOIR domain, they should all touch each other and they should be all full at the beginning. Otherwise, problems may be encountered. PIPEFS must be set to zero if it is to be followed by an Advanced Porosity calculation. It is not possible to model temperatures below -273C (below absolute zero) with the THERMAL=2 model. Work around : use THERMAL = 1. When a domain defined as RESERVOIR is remaining full during the whole filling, the Pool model will be activated at the wrong fill fraction (i.e. at LVSURF minus the volume of the reservoir). Work around : set this domain to CASTING instead of RESERVOIR, as if this domain remains full, it is useless to define it as RESERVOIR.

MeshCAST

Version 2006.0

Page 34

ProCAST User Manual

On some Windows OS, the Open in MeshCAST is not done in the current working directory.

Miscellaneous
Some error messages may appear during the use of CONVBIN, however, one can ignore these messages.

Version 2006.0

Page 35

ProCAST User Manual

GETTING STARTED
SOFTWARE LAUNCH
Launch - Windows platform
When the software is installed on Windows, a directory named PAM-SYSTEM is automatically installed on the Desktop of the computer. Inside this directory, a "ProCAST 2006.0 Manager" icon is present.

To launch the software, double click on this icon and the ProCAST Manager will start.

Launch - UNIX / Linux Platforms

On UNIX / Linux platforms, the ProCAST Manager is launched by typing the command "cast" in a Console window.

ProCAST 2006.0 Manager

Both on Windows and on UNIX-Linux, the ProCAST 2006.0 Manager has the following appearance.

Version 2006.0

Page 36

ProCAST User Manual

The "on-line help" can be opened with the "Help" button (1). The working directory should be set with the browse tools (2) . The "Case" name (or prefix) should be defined (3). The different modules can be called with the upper buttons (4). The software is ready to be used.

Version 2006.0

Page 37

ProCAST User Manual

PROBLEM SET-UP
In order to get started, a very simple thermal problem will be set-up, step by step. To do so, a very simple mesh located in the installation directory will be used. Please note that the meshing, described in the MeshCAST manual will be addressed in this section (please refer to the MeshCAST manual for more details). The mesh can be found in the following directory : On Windows : %ProCAST20050%\Test\Geometry\Test.mesh On Unix / Linux : $ProCAST20050/Test/Geometry/Test.mesh

Software launch
This "Test.mesh" files should be copied in the desired working directory (C:\Temp\Start in the following example). Then, the software manager should be launched (see previous section) and one should browse to the working directory and set the Case name (prefix) to "Test".

Pre-processing launch
When the "PreCAST" button is pressed (in the upper black band), the preprocessing module opens and the mesh ("Test.mesh" file) is automatically loaded.

Version 2006.0

Page 38

ProCAST User Manual

A window with some information about the loaded model (number of domains, of nodes and elements, as well as the model size) appears. Any click on the menus will automatically close this window. The geometry can be rotated interactively with the mouse. The icon can be pressed in order to view the geometry in "Solid" mode.

Materials menu
The first operation to perform is to assign material properties to the two domains. The first domain should be selected (1) and the desired material properties should be selected in the database (2). An "H13-steel" is selected in this case. Then the "Assign" button should be pressed in order to assign the selected material properties to the given material domain (3). The material property assignment should be repeated for the second material domain (let's select the "AL-7%Si0.3%Mg-A356" alloy. Finally, the material type ("Mold" for the first domain and "Casting" for the second domain) should be specified (4). This is done with a right click in the "Type" column.

Version 2006.0

Page 39

ProCAST User Manual

Interface menu
In the interface menu, one should create the interfaces between the different material domains and assign the desired interface heat transfer coefficients. The type of interface should be specified. To do so, one should click once on "EQUIV" (1), which will change into "COINC" (for coincident interface). The "Apply" button (2), which is turning orange should be pressed to accept the coincident interface type. A message will appear on the screen, which should be acknowledged. The desired interface heat transfer coefficient should be selected in the database (3) and assigned to the corresponding interface (4).

Version 2006.0

Page 40

ProCAST User Manual

Boundary Conditions menu

The cooling of the outside of the mold with the air, as well as the top surface of the casting should be defined in the "Boundary Conditions" menu. The type of boundary condition should be selected with the "Add" button (1). The "Heat" type should be selected in the list which is appearing. Then, this "Heat" boundary condition appears in the BC-Type column (2). A "Heat" boundary condition allows to define a heat transfer coefficient between the model and the ambient air, as well as the air temperature. The corresponding values are selected in the database (3) and the "Assign" button is pressed (4) to link the Heat boundary condition to the database values. The database entry number is then indicated in the "DB entry" column (5). This boundary condition should be then applied to the model. To do so, the "Select all" icon should be pressed (6) and all the model is colored in red (7), showing that all outside surfaces were selected. Finally, the "Store" button, which was turned to yellow, should be pressed in order to save the selection (8).

Version 2006.0

Page 41

ProCAST User Manual

Process menu
The gravity should be defined in the "Process menu". To do so, click on the menu "Process" and select the "Gravity" sub-menu. This will open the "Gravity" panel. A gravity in the -Z direction is set and the panel is closed with the "Close" button.

Version 2006.0

Page 42

ProCAST User Manual

Initial Conditions menu

The initial temperature of both material domains should be specified in the "Initial Conditions" menu. Each domain should be selected (1) and the initial temperature should be entered in the field in (2).

Version 2006.0

Page 43

ProCAST User Manual

Run Parameters Menu

Finally, the calculations parameters should be specified in the "Run Parameters" menu. In this example, all the default values can be used. One could just modify the value of TSTOP (in the "General/Standard" tab) to 500C (in order to stop the calculation when all the temperature are below 500C).

Version 2006.0

Page 44

ProCAST User Manual

Save and Exit

As the set-up of the model is finished, one can save and exit. In the "File" menu, one can select the "Exit" sub-menu. The software will ask whether one would like to save the model before exit. Select "Yes" and the software will first save the model and then exit PreCAST. The pre-processing stage is finished and one is ready to launch the calculation.

Version 2006.0

Page 45

ProCAST User Manual

CALCULATION
The calculation can now be launched in the Manager. To do so, the "ProCAST" button should be pressed (1), which will open the confirmation window for the run. The "Execute DataCAST first" should be activated (2) and then one can confirm the launch of the solver by pressing the "Run" button (3).

On Windows, a Command window will open and the DataCAST and ProCAST are automatically launched. A PAUSE message is displayed when the calculation is finished.

Version 2006.0

Page 46

ProCAST User Manual

The progress of the calculation can be monitored by pressing on the "Status" button of the Manager (top right button). The following window displays all the information about the current situation of the calculation.

Once the calculation is finished, or anytime during the calculation, one can view the results in the post-processor ViewCAST.

Version 2006.0

Page 47

ProCAST User Manual

RESULTS DISPLAY
To launch the post-processor, the button "ViewCAST" should be activated in the Manager. Once the ViewCAST window is opened, one can interactively rotate the mouse (with a left click in the graphics area (1)). The material domains to be displayed can be selected with the icon (2). The result to be displayed (i.e. Temperature, Fraction of Solid, ...) should be selected in the "Contour" menu (3). In order to display a snapshot, the icon (4) should be active. Then, one can display the different timesteps with the tape player (5), either as an animation or as a step-bystep display. The scale can be changed by a click on the scale values (6).

Slices through the model can be obtained with the Slice icon (1). When no slices are defined, it will automatically open the slice definition window (2). Then, one should select the "Add ->" button and the "XYZ Plane" sub-menu, which will open the slice selection panel (3). When the desired slice is selected, the Apply button should be pressed and the slice will be displayed. One can select more than one slice by repeating the "Add ->" operation. Finally, one can show the different timesteps with the Tape player (4).

Version 2006.0

Page 48

ProCAST User Manual

Time evolutions can also be shown with the "XY Plot" menu. The desired field (e.g. Temperature) should be selected and the desired locations to be plotted should be defined (e.g. with the "Interval" sub-menu (a value of 50 nodes interval was selected in the figure hereafter).

Version 2006.0

Page 49

ProCAST User Manual

To Quit ViewCAST, the File/Exit menu should be used. The software is asking whether you want to store the last view. If it is stored, the current position and orientation of the model on the screen, as well as the material selection, will be retrieved at the next launch of the post-processor.

Version 2006.0

Page 50

ProCAST User Manual

SOFTWARE MANAGER
FILE MANAGER
When the ProCAST software is started, the Manager is launched, as seen in the figure below. The Manager allows to browse in the desired working directory, to create new directories and cases, to copy files from one directory to an other and to launch the different modules of the software.

The default starting directory can be set in the "Installation settings" (see the "Software configuration" section). Firstly, the working directory should be defined (with the "browse" "create directory" should be defined. and

icons) and secondly, the name of the case (prefix name)

Version 2006.0

Page 51

ProCAST User Manual

When several cases are present in the same working directory, the prefix of the first one (in alphabetical order) is selected. In order to scroll through the other cases, one can use the F2 key. The File Manager allows to perform some actions on the files, such as "Refresh" (to refresh the file list), "create new text file" "Duplicate file" and "Delete file" . , "Rename file" ,

A set of files can be copied from one directory to the other, using the button. Then, two windows are opened (see figure below) and files can be copied from the left window to the right one. If the "Keep Prefix" check box is activated, the prefix remains unchanged, otherwise, the prefix is changed to the one defined under "Case" in the lower orange section.

Version 2006.0

Page 52

ProCAST User Manual

The extensions "d.dat", "p.dat", "d.out" and "p.out" are well handled by the copy. However, if it is desired to copy "unf" files, the user is prompted for the "old prefix" and the "new prefix", as shown below :

In the "Copy" panel, the icon gives a direct access to Template files (for the user functions and for the porosity input file).

Version 2006.0

Page 53

ProCAST User Manual

Version 2006.0

Page 54

ProCAST User Manual

MODULE CALLS
The different modules of ProCAST can be called from the top bar buttons :

Firstly, the FEM mesh should be generated with MeshCAST. Then the case should be set-up with the pre-processor (PreCAST). Once all the parameters and conditions are defined, the calculation files should be prepared with DataCAST. The calculation can be run at this time by calling the ProCAST solver. Finally, the results can be visualized with the post-processor ViewCAST. At any time during and after the modeling, the status of the calculation can be monitored with Status. MeshCAST, the meshing module of ProCAST, is launched with the "MeshCAST" button. Then the desired case should be opened into MeshCAST (there is no automatic launch as the software does not know whether the .gmrst, the .sm or the .mesh file should be opened). Please refer to the MeshCAST User Manual for further details. A Command window is opened in the background in order to allow the viewing of error messages if necessary. This Command window can be closed at the end by a simple <RETURN>.

The Manager allows also to launch the "2-D mesh generator" or MeshCAST with the -M option. To do so, a "right click" on the "MeshCAST" button should be

Version 2006.0

Page 55

ProCAST User Manual

done. This opens a sub-menu, where one can select between MeshCAST-2D or MeshCAST-3D and MeshCAST-3D with the -M option.

When "MeshCAST-3D -M" is selected, the following window opens :

This special mode of MeshCAST should be used only in the case of large memory requirements (see the MeshCAST manuals for more details). When PreCAST is launched, if a "d.dat" file is present, it will automatically be loaded. Otherwise, the browser will prompt in order to select a ".mesh" file.

Version 2006.0

Page 56

ProCAST User Manual

When a right click in made on the "PreCAST" button, it gives access to a submenu in order to activate the "CAFE pre-processor"

Interfaces assignmentThe DataCAST module is translating one input file generated by the pre-processor PreCAST (i.e. the d.dat file) into binary files (*.unf) which will be used by the calculation. Some error checking is also performed at this stage. When the DataCAST button is activated, a Command Window appears. This allows to view potential warnings or error messages. The operation is finished when the PAUSE appears on the window. Then, one should just hit any key (e.g. RETURN) in order to close this Command window and proceed with the calculation itself.

Version 2006.0

Page 57

ProCAST User Manual

When the DataCAST module is launched, a confirmation window is prompted to the user in order to confirm the launch of DataCAST (as DataCAST is erasing existing result files, this confirmation allows to prevent any undesired results loss).

Moreover, the "Update option" of DataCAST (see the "Run the Calculation/Solver" section for more details) can be activated from this window. Once the data are ready (after PreCAST and DataCAST), the ProCAST solver can be launched, with the "ProCAST" button. The following window is then opened in order to "Confirm" the start of the calculation. Moreover, it is possible to automatically launch "DataCAST" before the ProCAST solver, by checking the "Execute DataCAST first" box.

Version 2006.0

Page 58

ProCAST User Manual

Upon "Run", a Command Window will be automatically opened and the solver will be started. As this window is independent from the Manager, it is possible to use the Manager (and the other modules, such as the post-processing for instance) while the calculation is running. When the calculation is finished, a PAUSE will be prompted and one has to hit any key (e.g. RETURN) to make the window disappear. If there is a problem during the execution, a warning or an error message may be printed in this window.

The inverse solver can also be called from the "ProCAST" button (see the Software configuration section for more details about how to activate the Inverse option):

Version 2006.0

Page 59

ProCAST User Manual

Results can be viewed at any time (even during the run of the calculation), with the ViewCAST button. If the case name is set, as well as the current working directory, the case is automatically loaded in the post-processor, as shown below.

A right click on the "ViewCAST" button gives access to the "Advanced Porosity" and the "CAFE" post-processors (i.e. the calcosoft post-processor) :

Version 2006.0

Page 60

ProCAST User Manual

To monitor the calculation, the "Status" button can be used at any time. A graphical status of the case will be shown with the relevant information concerning the calculation (i.e. percent filled, solid fraction, timestep, current time, cpu time, ...) corresponding to the case name indicated in the lower orange window. The status is refreshed every 15 seconds, unless the "Status" button is pressed again. To exit the Status window, press the "Return to File Manager" button. Please note that the Graphical status is working only when activated in the working directory of the case (as the prefixp.dat file is read).

Finally, all these modules can be called manually from a Command Window. The "Command Window" button opens automatically a Window in the working directory.

Version 2006.0

Page 61

ProCAST User Manual

Then, the modules can be called in the following way (or via ad-hoc aliases) : For Windows : Manager MeshCAST PreCAST DataCAST ProCAST ViewCAST %ProCAST20060%\bin\manager.exe %ProCAST20060%\bin\meshcast %ProCAST20060%\bin\precast prefix %ProCAST20060%\bin\datacast prefix %ProCAST20060%\bin\procast prefix %ProCAST20060%\bin\viewcast prefix

For Unix (tcsh, csh) : Manager MeshCAST PreCAST DataCAST ProCAST ViewCAST $ProCAST20060/bin/manager.exe $ProCAST20060/bin/meshcast $ProCAST20060/bin/precast prefix $ProCAST20060/bin/datacast prefix $ProCAST20060/bin/procast prefix $ProCAST20060/bin/viewcast prefix

The Volume mesher of MeshCAST can be called in batch mode with the following command : On Windows : %ProCAST20060%\bin\meshcast -batch prefix.sm Flag On Unix (tcsh, csh) : $ProCAST20060/bin/meshcast -batch prefix.sm Flag "Flag" is an integer value defining : 0 : no layer 1 : partial layer 2 : full layer

Version 2006.0

Page 62

ProCAST User Manual

ADVANCED MODULE CALLS

Advanced modules, such as the Parallel solver of ProCAST, the Advanced Porosity module, the CAFE module or the Inverse solver, can now be called from the manager. To do so, firstly, the Manager should be configured in the following way (in the "Installation Settings" window.

In the "Preferences" tab, one can select which "modules" should be activated (under "Modules display") in the Manager :

Version 2006.0

Page 63

ProCAST User Manual

By default, none is activated. "ProCAST Parallel" will activate the possibility to launch the parallel solver (see the "Parallel Solver" chapter for more details). "Advanced Porosity module" will activate the possibility to launch and visualize Advanced Porosity calculations (see the "Advanced Porosity Calculations" chapter for more details). "Inverse module" will activate the possibility to launch inverse calculations form the Manager (see the "Inverse modeling" chapter for more details). "CAFE Module" will activate the possibility to launch an visualize CAFE calculations (see the "CAFE-3D" chapter for more details). When these modules are activated in the "Preferences", the "ProCAST" window has the following appearance (more details about the functionnalities in the corresponding chapters) :

Version 2006.0

Page 64

ProCAST User Manual

Version 2006.0

Page 65

ProCAST User Manual

RUN LIST
The "Run List" allows to launch automatically several calculations in a row, as well as to "chain" different calculations (such as a ProCAST thermal calculation followed by an Advanced Porosity calculation). This allows also to launch calculations in batch mode.

When the "Run List" menu is selected the Manager window appears as follows :

If it is create a "Run List" from cases in the same directory, one should check that the desired case name is specified in the bottom window, then the desired solver should be selected in the left window (Available Solvers) and the "Add button"

Version 2006.0

Page 66

ProCAST User Manual

should be pressed. The corresponding run will appear in the right window (Solvers to be run). Then, one can add as many runs as desired (see example hereafter where the Case1 and Case2 are launch sequentially).

In the following example, the ProCAST thermal calculation (including DataCAST) is automatically followed by the "Advanced Porosity" run.

If the cases to be run are not all in the same directory, one should first select the Solver to be run (in the left window) and then press the "Folder" button (above the "Add" button). This is opening a browser, where one can select the desired directory and the desired case to be run. Once the "Open" button of the browser is pressed, the corresponding line in the right window is added. With this method, one can run sequentially a serie of calculation which are located in different directories. It is also possible to remove an entry from the right list, by selecting it and pressing the "Remove" button. Once the list is ready, the "Run Computations" button can be pressed. This will automatically launch the calculations. Otherwise, it is also possible to save the "Batch file", with the "Create Batch File" button. This file will automatically store the batch file in the working directory.

Version 2006.0

Page 67

ProCAST User Manual

One can also edit the batch file or run it with the corresponding buttons. In these two cases a browser will open automatically, to select the desired file.

Version 2006.0

Page 68

ProCAST User Manual

SOFTWARE CONFIGURATION
The software can be configured in the "Installation Settings" window :

Preferences tab

The Preferences can defined, such as the Default directory which will appear every time the Manager is launched, the preferred Text editor can be configured.

The access to the "Advanced modules" (Parallel, Advanced Porosity and Inverse) can be configured in this window (see the "Advanced module calls" section for more details).

Version 2006.0

Page 69

ProCAST User Manual

ViewCAST File format compatibility

ViewCAST allows now a full compatibility between UNIX and Windows results. This means that it is possible to visualize on a Windows machine, results which have been calculated on a Unix machine, or vice-versa. To do so, one should "Activate the swap mode", as shown below.

Then, the user can select the desired default mode between the two above choices. This default mode will be used when the "ViewCAST" button will be pressed. However, it is still possible to use the other mode (i.e. Swap format), by making a right click on the "ViewCAST" button. Then, one can chose between the "No swap format" (which means that ViewCAST and the ProCAST run were performed on the same plateform) or the "Swap format" (which means that ViewCAST and the ProCAST run were performed on a different plateform).

FlexLM manager : LMtools

It is possible to open the FlexLM Manager (LMtools), by clicking on the corresponding button in the "Preferences" tab of the "Installation settings".

Version 2006.0

Page 70

ProCAST User Manual

This will work only if FlexLM is installed in the "standard" directory.

Installation tab

In the "Installation" tab, one can check that the environment variable ProCAST20060 is well configured. It should point towards the v2006.0 directory, above the "bin", "dat" and "lib" directories.

The "Command window Launch Option" check box (in the bottom of this Installation Settings window" allows to activate the Command windows when PreCAST and ViewCAST are launched. This mode should be activated in case of problems in order to see the possible error messages which may occur. In normal situation, this mode should be deactivated. The "DOS Path Converter" (on Windows only) can be useful to convert "long directory names" into short "DOS" ones. In ViewCAST, the time is displayed at the bottom of the screen. By default, the time is shown in seconds. However, in the case of large casting modeling, it may be useful to see the time in hours:min:sec, whereas for high pressure die casting, milliseconds are preferred. The default settings can be set in the file "manager.tcl", which is located in the "bin" directory of the installation. At about 30 lines from the top of the file, there is the following lines :
### Viewcast ### global viewcastOption set viewcastOption "-T1"

The viewcastOption variable is controlling the time display as follows :

Version 2006.0

Page 71

ProCAST User Manual

set viewcastOption "-T1" set viewcastOption "-T2" set viewcastOption "-T3"

(display in seconds) (display in hours:min:sec) (display in milliseconds)

Version 2006.0

Page 72

ProCAST User Manual

CUSTOMIZED INSTALLATION
In some cases, the software is installed in a centralized location of the company. In previous versions, it was necessary the set the installation directory of the software with Read/Write access rights, in order to allow the access to the databases and to the status files. From version 2004.1, ProCAST allows to install the software in a Read only location and to customize the location of the files which have to be modified by the user (databases, status files, preferences). To do so, two optional Environment variables are available : ProCAST20060_DB for the location of the databases. ProCAST20060_USER for the status files and preferences. If a user, or a group of users, would like to have the databases in a different location then the software installation, one should copy the "db" directory in the desired location (e.g. /my_db_location/db). Then, the ProCAST20060_DB environment variable should be set (manually) to the /my_db_location directory. In the same way, if a user would like to set it own preferences and status files, one should copy the "pref" and "statpro" directories in the desired location (e.g. /my_pref_location/pref and /my_pref_location/statpro). Then, the ProCAST20060_USER environment variable should be set (manually) to the /my_pref_location directory. Please note that the "Preferences" directory is containing preferences of the Manager, as well as the User customized icons of MeshCAST.

Version 2006.0

Page 73

ProCAST User Manual

FLEXLM
The software is locked using the FlexLM tool. For "Node locked" licenses, the following procedure can be used (for floating licenses, please refer to the general FlexLM documentation). The following is valid for Windows only. For Unix, please refer to the FlexLM documentation. In order to activate your license, a license file should be requested to ESI. The first step is to identify the machine on which the software has to run. To do so, run the "FlexLM License Manager", by a double click on the corresponding icon in the "PAM-SYSTEM" directory which is on your desktop.

This will open the following panel. Go to the second tab "System Settings" (1) and send this panel to your local distributor. You can also transform this information in a text file, using the "Save HOSTID Info to a File" button (2).

Version 2006.0

Page 74

ProCAST User Manual

Then, the licence file which is received from ESI should be called "pam_lmd.lic" (usually, it is called "*****.cry" and you should rename it to "pam_lmd.lic" previously it was called "license.dat"). Then, the "pam_lmd.lic" file should be placed in the C:\flexlm directory. When one of the following modules (PreCAST, ProCAST or ViewCAST) is called for the first time, the following screen will be prompted :

Please select "Specify the License File" and press "Next>".

Version 2006.0

Page 75

ProCAST User Manual

The location of the "pam_lmd.lic" file should be specified. The "Browse" button could be used to find the "C:\flexlm\pam_lmd.lic" file :

Version 2006.0

Page 76

ProCAST User Manual

Once the "C:\flexlm\pam_lmd.lic" file is specified, the "Next>" button can be pressed in order to finish the FlexLM installation.

When the "Finish" button is pressed, the software is unlocked and all the modules can be used.

Version 2006.0

Page 77

ProCAST User Manual

PRE-PROCESSING
INTRODUCTION
To start the Pre-processor, the "PreCAST" button should be used. If a mesh file (case.mesh) or a "d.dat" file (cased.dat) is present, the case will be automatically opened.

If there is no case present in the working directory, then the browse window will open so that the user can select the desired input files (see the Geometry import section). When the case is loaded, a window appears with some information about the model, such as the number of materials, the number of nodes and elements, as well as the model size. Then, the pre-processor is ready to set-up a case.

Version 2006.0

Page 78

ProCAST User Manual

The top bar menu is divided in 9 menus which allow to perform all the operations to set-up a case :
File Geometry Materials Interface Boundary Conditions Process Initial Conditions Run Parameters Inverse Help

First the model should be opened or saved in the File menu. It allows also to quit the Pre-processor.

Then, in the Geometry Menu, symmetries can be defined, as well as the virtual mold characteristics. Moreover, some features of the FEM mesh can be checked, such as negative Jacobians, or the volume of each domain.

Version 2006.0

Page 79

ProCAST User Manual

In the Material menu, the characteristics of each domain (or each material) can be defined. In addition to the material properties, one can specify the type of the domain (casting, mold, filter, foam, ...), as well as if it will be empty or not at the beginning of the calculation (for mold filling).

The Interface menu has no sub-menus. It opens the window which allows to define the interactions between the different materials, such as heat transfers. The Boundary condition menu allows to define all the interactions between the different materials and the outside world (i.e. on the outside surfaces of the model), such as external cooling, velocities at the surface of the model for flow calculations, displacements or constraints for stress calculations, etc...

The Process menu gives access to the definition of the gravity, as well as the definition of the motion of the different domains or enclosures.

Version 2006.0

Page 80

ProCAST User Manual

The initial temperatures of each materials are defined in the Initial Conditions menu.

The Run Parameters menu, as well as the Help menu have no sub-menus. All the calculation parameters are defined in the Run Parameters window. The on-line Help can be access from the Help menu. Below the menus, icons allows to perform a number of operations linked to the display of the model on the screen. These icons are described in the Geometry manipulation section.

The next sections are presenting the set-up of a case, according to the following flow chart.

Version 2006.0

Page 81

ProCAST User Manual

GEOMETRY IMPORT
When PreCAST is started from the Manager, it is automatically reading a mesh or a d.dat file (if they are present in the working directory, with the selected prefix). The priority is first a "d.dat" and then a ".mesh" file. If there is no mesh file or d.dat file present in the working directory with the corresponding prefix, the pre-processor opens with the browser window.

Then the user has the choice of the input format, through the following filter :

The Pre-processor is able to read PreCAST input files (*d.dat), also called "Restart" files, or meshes coming from MeshCAST (*.mesh), from PATRAN (*.out) or I-DEAS (*.unv).

Version 2006.0

Page 82

ProCAST User Manual

The reading of multiple meshes (for non-coincident meshes or for radiation calculation) is described in the "Advanced features" section of the Pre-processor.

Version 2006.0

Page 83

ProCAST User Manual

THERMAL
Thermal model
The Thermal module allows to perform a heat flow calculation, by solving the Fourier heat conduction equation, including the latent heat release during solidification. The typical results which can be obtained are the following : Temperature distribution Fraction of solid evolution Heat flux and thermal gradients Solidification time Hot spots Porosity prediction

Flow chart
This section describes the set-up of a thermal case. It is also the opportunity to introduce the general work flow of the pre-processor, as well as some aspects which may be used by different modules (e.g. symmetry).

Each step described in the above flow chart is described in the following sections.

Version 2006.0

Page 84

ProCAST User Manual

Geometry assignments
Once the geometry is loaded, the following operations can be performed on the geometry :

Symmetries can be defined at this stage (see the "Thermal/Radiation" section for more details). The definition of the Virtual mold is also done at this level. (see the "Virtual Mold" section for more details). In the "Check Geom" menu, the following features are accessible :

Neg-Jac (negative Jacobian) and Neg-Area correspond to problems in the mesh. These buttons allow to locate where these problems are in the mesh in order to give indication where to modify the mesh in MeshCAST. Volumes gives access to the volume of each material domain, whereas Min-Max indicates the dimensions of the model.

Materials assignment
Once the model is loaded (see the "Geometry import" section), the first operation is to define the different materials with their properties and attributes. This is performed in the Material/Assign menu.

On the right of the window, two frames are shown. The top one contains the material list (or domain list), whereas the bottom one corresponds to the material database.

Version 2006.0

Page 85

ProCAST User Manual

When one clicks on the different materials in the material list, the corresponding domains are highlighted (in the picture below, the hidden mode with mesh was selected - see the "Geometry manipulation" section for the other display modes).

Version 2006.0

Page 86

ProCAST User Manual

In the lower frame ("Material database" list), the list of all available material properties in the material database is displayed. To manage the database entries, please refer the "Databases" sections. The {T} or {F}which are indicated before the material name are telling wheter material properties are present in this material for Thermal only calculations (T) or for Thermal and Fluid flow calculations {F}. If a {*} appears, it means that the material properties definition is uncomplete and that this material entry can not be used for a calculation at this stage.

In the top frame ("Domain list"), all the domains (or materials) present in the mesh are listed. When a mesh is loaded, it appears as follows :

Version 2006.0

Page 87

ProCAST User Manual

Then, one should assign a Material to each domain, to define the type of each domain and to specify whether the domain is empty or not at the beginning of the calculation. To assign Materials, (1) one should select the desired domain in the upper list, (2) select the desired material in the material database list, and (3) click on the Assign button. This should be repeated for each domain.

Then, the "Type" of each material should be defined. To do so, make a right click on the "CASTING" word and the available list of possible selection will appear :

Version 2006.0

Page 88

ProCAST User Manual

Mold : the mold material should be set to "Mold". This will be used for cycling calculation (in die casting) in order to allow the calculation of the heating of a die during cycling (i.e. the temperature of the mold domains will not be reset to the initial temperature at the beginning of each cycle). Casting : the casting material should be set to "Casting". This setting is necessary in particular for all the domains where fluid flow will occur. For a cycling calculation, the casting domains initial temperatures will be reset at the beginning of each cycle. Filter : filter domains should be set with the "Filter" type (see the "Filters" section for more details). Foam : for lost foam calculations, the domains where the foam is present at the beginning of the calculation should be set to "Foam". Of course, during the filling, the casting material will replace the foam, as it burns. Insulation : this type has no specific effect on the solver. It will correspond to a "Mold" type of material. At this moment, this is for information purposes. Exothermic : this will activate the Exothermic properties of the sleeve (if they are defined in the corresponding material properties). If the material properties are containing the exothermic information, but the "Exothermic" type is not activated, the exothermic model will not be activated (see the "Exothermic" section for more details). Core : a core type material should be defined in the case of cycling, where cores are placed into the mold at each cycle. This means that unlike mold materials, the initial temperature of the cores will be reset at the beginning of each cycle. Reservoir : a Reservoir type material is a domain where the free surface will always be perpendicular to the gravity. This allows to simplify the free surface computation and it is especially useful in the case of tundish modeling. The "RESERVOIR" domains should have an "EQUIV" interface with the other CASTING materials.

Version 2006.0

Page 89

ProCAST User Manual

Please note that if there is more than one RESERVOIR domain, they should all touch each other and they should be all full at the beginning. Otherwise, problems may be encountered. If one want to empty the reservoir, no special BC should be specified. If one would like to keep the reservoir full, a pressure BC should be set at the surface of the reservoir. Finally the user has to specify which domains are empty at the beginning of the calculation (for mold filling calculations). One should make a left click on the "No" to turn it to "Yes" (which means that Yes the domain is empty). On additional click returns to No. Of course, more than one domain may be empty (if the casting is made out of several mesh domains).

When one makes a right click on the material name in the upper list, the lower list is pointing on this particular material. This is very useful if one wants to see the material properties of this material.

Interfaces assignment
Once the Materials are defined, one should define the Interfaces, with the Interface menu.

As for the Material window, two frames appear on the right of the window. The top one contains the list of all the possible interfaces, whereas the lower one shows the Interface database. To manage the database entries, please refer the "Databases" sections.

Version 2006.0

Page 90

ProCAST User Manual

Firstly, one should define the type of interfaces in the upper right window :

On the left, the "Material Pair" are shown. "1 and 3" means that there is an interface between material 1 and material 3. By default, the Type of the interface is set to "COINC". By clicking on the "COINC" text, one can toggle between "COINC", "NCOINC" and "EQUIV". When a mesh is generated with MeshCAST (or with most common mesh generator), the elements which are on either side of an interface (i.e. adjacent elements which belongs to two different domains) are sharing the same nodes. This is called a coincident mesh.

Version 2006.0

Page 91

ProCAST User Manual

EQUIV option When two domains are part of the same entity (i.e. they both belong to the casting with the same material properties, but they were meshed separately for technical reasons), one will set an "equivalenced" interface between them (EQUIV). It means that there will be a continuum between the two domains, with a continuous temperature profile across the interface, as well as continuous velocity field. In such a case, the nodes at the interface (shown in orange in the figure below) are shared by the elements on both sides. This EQUIV option can also be used if one has different materials in the two domains, but the materials are welded together (i.e. with a total bounding between the two materials).

COINC option At an interface between two different materials, such as the casting and the mold, there is usually a temperature drop. In this case, the nodes at the interface should be doubled (for a coincident interface), in order to distinct temperature on each side of the interface. As during the mesh generation, there is one node at the interface, it is necessary at this stage to duplicate all the interface nodes (as shown in green in the figure below). This duplication operation is performed when "COINC" is selected (for "coincident nodes"). The interface, which is shown in yellow in the figure below has in fact a zero thickness.

NCOINC option It is also possible to generate a non-coincident mesh (i.e. where the elements on both sides of the interface are not matching, which means that they are not sharing the same nodes), by adding different meshes together (see the "Advanced

Version 2006.0

Page 92

ProCAST User Manual

features" section for more details on non-coincident meshes). In this case, one has to specify that the interface is non-coincident, with the "NCOINC" option.

When one toggles between the different options, the interface appears in red and green. It is thus possible to well identify whether it corresponds well to the desired interface (see figure below, which was obtained in hidden mesh mode )

An other way to view the desired interfaces is to click on the "Material Pair" and the material on both sides will be highlighted in red and green respectively - the first material in the list is in red and the second one in green (see figure below, which was obtained in hidden mesh mode )

Version 2006.0

Page 93

ProCAST User Manual

Once the desired selections (between COINC, EQUIV and NCOINC) are done for each possible interface, the STORE button (which is highlighted in orange) should be pressed.

Then the pre-processor will automatically create the double nodes and a message will appear to confirm that the number of nodes of the model has increased.

Version 2006.0

Page 94

ProCAST User Manual

As it is now possible in version 2006.0 to go from EQUIV to COINC and viceversa, the user should be careful that as nodes were duplicated or removed, some boundary condition assignments may be corrupted, as well as the extracted initial conditions. Thus, in this case, they should be re-assigned. Once the types of interfaces are defined, one has to apply the corresponding heat transfer coefficients (for COINC and NCOINC only, as nothing as to be specified for EQUIV). To assign Interface heat transfer coefficients, (1) one should select the desired Material Pair in the upper list, (2) select the desired interface heat transfer coefficient in the interface database list, and (3) click on the Assign button. This should be repeated for each coincident or non-coincident interface.

For non-coincident meshes, it is possible to have access to the non-coincident tolerances with a right click on the NCOINC label. A window will open with the two tolerances.

Boundary conditions assignment

Version 2006.0

Page 95

ProCAST User Manual

After the definition of the interfaces, the Boundary conditions should be specified. This is done in the "Boundary Conditions" menu.

Three types of boundary conditions can be applied : Surface boundary conditions ("Assign Surface"), which correspond to all the conditions applied to the outside of the model or the outside of a given material domain. This is the most commonly used type of boundary conditions. Volume conditions ("Assign Volume"), which corresponds to conditions which are applied in a whole volume (e.g. volumetric heat or mass source). Boundary conditions assigned to Enclosures ("Assign Enclosure") in case of radiation problem (see the "Radiation" section for more details).

Assign Surface
The principles of Boundary conditions definition are described in the following figure.

Version 2006.0

Page 96

ProCAST User Manual

Firstly, the desired boundary conditions should be "Added" (1) in the upper list (2). With the "Add ->" button, one has the following choices :

The boundary conditions corresponding to Thermal problems are "Symmetry", "Periodic", "Temperature" and "Heat". Then the "location" where the boundary condition should be applied on the geometry should be specified. The selection tools (3) allow to "paint" the desired area on the geometry (4). The values to be assigned to the boundary conditions should be selected in the database (5) and then they are assigned to the corresponding boundary condition (6). See the "boundary conditions database" section for more details about the different type of boundary conditions, as well as the database management. It is possible to have a quick access to the database entries in the following way. When one makes a right click on the boundary condition entry in the upper list (2), the lower list is pointing on this particular boundary condition entry in the lower list (5). The "Selection tools" allow to "Select", "Deselect", "Propagate", "Clip", "Copy" and "Paste" faces or nodes on the geometry.

Version 2006.0

Page 97

ProCAST User Manual

For some boundary conditions, such as "Heat" or "Symmetry", faces of the Finite Element Mesh are selected. For other boundary conditions, such as "Temperature", "Velocity", ... the boundary conditions are applied on nodes. The choice between faces or nodes is automatic.

Selection of individual faces or nodes.

Deselection of individual faces or nodes.

Select all.

Deselect all.

Select and propagate. All the faces or nodes which have an angle with the neighbors smaller than the specified propagation angle will be selected.

Deselect and propagate. All the faces or nodes which have an angle with the neighbors smaller than the specified propagation angle will be deselected.

Definition of the propagation angle.

Select remainder. The remaining faces or nodes which have not yet been selected are selected.

Select interface. For "Heat" boundary conditions, the selections are applied on external faces only. However, for cycling, one would like sometimes to apply a Heat BC on faces which lie at interfaces. This allows to select those interfaces automatically. In this case, the following panel is opened. It is proposing the list

Version 2006.0

Page 98

ProCAST User Manual

off all active materials. One could select one or more material and all the faces of the selected materials which are lying on an interface will be selected.

Button to clip a model. This allows to perform a selection inside a model, where is it not accessible from the outside.

Button to "Backtrack" the clip.

Copy of selection. All the nodes or faces which are applied on the geometry for the active boundary condition will be copied in the memory.

Paste of selection. The copied selections (above button) can be pasted on a different boundary condition (i.e. on the active boundary condition when the Paste button is pressed). Please note that one can not copy the node selection of a Temperature BC to a Heat BC (as faces are expected).

Assign Volume
In the "Assign Volume" menu, the following entities can be assigned :

Version 2006.0

Page 99

ProCAST User Manual

Volumetric Heat corresponds to a "heat source" in the entire volume (see the "Boundary Conditions Database" section for more details). Surface Heat allows to specify how the free surface inside a given material domain is cooling down. It is possible to assign a "Heat" boundary condition to the free surface inside the domain. With this, it is possible to assign a convective heat transfer coefficient and an external temperature which will drive the cooling condition of the free surface itself. For instance, this could be important to take into account the cooling of the free surface of a swimming pool. Please note that it is not possible to activate the radiation model with view factors on a free surface. Momentum Source corresponds to a force term on the liquid (momentum force) in the entire volume (see the "Boundary Conditions Database" section for more details). Mass Source corresponds to a different way of applying an inlet of metal in the case of filling (see the "Boundary Conditions Database" section for more details). Filter Heat allows to define the interface heat transfer coefficient between the liquid metal and the filter material. Thus, the cooling of the hot liquid, when passing through an initially cold filter can be modeled using a "Filter heat". The interface heat transfer coefficient (which is the same as the one defined in the "Interfaces assignment" section) is applied to the entire filter domain.

Assign Enclosure
In the case of radiation problems with View Factors, the ambiance or the furnace can be modeled with an enclosure. See the "Pre-Processing/Radiation" secton for more details. This menu is not used for Thermal problems without "View Factor radiation".

Process conditions assignment

For Thermal problems (as well as for flow), the gravity vector should be defined in the "Process" menu.

Gravity

Version 2006.0

Page 100

ProCAST User Manual

For Thermal only problems, it is important to define the gravity direction for the calculation o of the porosity (using the POROS=1 model). The gravity is defined in the following panel. For standard problem, a constant gravity (1) is defined. The three components of the gravity are defined in (2). If one clicks on the X, or the Y, or the Z letter, automatically, the gravity vector is set in the X-, Y- or Z- direction. With two clicks, the negative direction will be set.

The "Rotate" is used only for Fluid flow problem (see the "Fluid Flow & Filling" section for more details).

Assign Volume
This option is used to specify a translation, a rotation or a revolution to the material domains. As this is mainly used in the case of Radiation problems, please look for more details in the "Pre-processing/Radiation" section.

Assign Enclosure
This option is used only for radiation problems (see the "Preprocessing/Radiation" section for more details.

Version 2006.0

Page 101

ProCAST User Manual

Initial conditions assignment

The initial temperature of each material should be defined in the initial condition menu.

The initial temperature can be defined either as a Constant value throughout the material domain, or as an Extracted temperature field coming from a previous calculation. In the case of "Constant", the list of all the Material domains is displayed and one can enter the initial temperature (in the lower white field) for each one.

In order to apply an extracted temperature field (e.g. temperature at the end of the cycle N 10, as initial condition for a filling calculation), one should perform a first calculation on the same mesh. Then, one could specify the case, as well as the step to be considered for the initial conditions. In the case of an "Extract", the following screen appears.

Version 2006.0

Page 102

ProCAST User Manual

All available material domains are listed in (1). The desired materials are selected in this list and the case name from which the thermal field must be extracted (2) should be selected with the Browse button (3). The directory (4) and the prefix should be selected (5). Then, the timestep (6) should be defined in the field (7). When the prefix (and the directory) of the desired case is selected, if one clicks on it in the "prefix" column (2), the extracted temperatures are shows (see picture below).

Version 2006.0

Page 103

ProCAST User Manual

Run parameters assignment

The following Run Parameters should be specified :

Version 2006.0

Page 104

ProCAST User Manual

The thermal module should be activated with THERMAL = 1 or 2. The storage frequency of the thermal results (TFREQ) should be specified, as well as the activation of the porosity model (POROS) . Please refer to the "Thermal Run Parameters" section for the full description of all the Run Parameters.

Version 2006.0

Page 105

ProCAST User Manual

FLUID FLOW & FILLING

Fluid flow model
The fluid flow module allows to perform mold filling calculation (free surface) as well as fluid flow computation, by the resolution of the Navier-Stokes equation. The typical results which can be obtained are the following : Filling behavior Free surface evolution Natural and forced convection currents Dynamic pressure of the liquid Entrapped gas Filter behavior

Flow chart
This section describes the additional set-up necessary for fluid flow and mold filling. For the set-up of a thermal case, please refer to the "Thermal" section of the Pre-processor.

Materials menu

Version 2006.0

Page 106

ProCAST User Manual

Concerning the material assignments, the only requirement is that the domains in which fluid flow calculation will be performed are defined as "Casting" domains (or "Filter" or "Foam" - see the "Advanced features" section for more details). Then, one should define whether the corresponding domain is Empty or not (in the case of mold filling). One can change from No to Yes and vice-versa by clicking on the text directly.

Moreover, the material properties of the fluid material should have the flow properties (e.g. viscosity). This can be checked in the Material database list. The Material name should be preceded by an {F} (for Fluid properties).

Interface menu
Concerning the interfaces, nothing special should be done concerning the fluid flow. One can however notice that if the casting (i.e. the flow domain) is noncoincident with the mold domain(s), one will need to set a zero velocity boundary condition all around the casting domain (in order to prevent "leaks"). See the "Advanced features" section for more details on non-coincident meshes.

Version 2006.0

Page 107

ProCAST User Manual

Boundary conditions menu

Concerning the boundary condition assignments, one has to specify the velocity, pressure, wall and/or inlet BC. See the "boundary conditions database" section for more details about the different type of boundary conditions, as well as the database management. See the "Thermal/Boundary Conditions Assignment" section for more details about how to assign boundary conditions. Please note that when a filling-flow calculation is performed on a model where there is no mold, it is necessary to set a zero-velocity boundary condition around the whole mesh (i.e. at all the location where a mold would be in contact with the casting). A zero velocity boundary condition can also be replaced by a WALL boundary condition, which has the same effect.

Process menu
For fluid flow problems, the gravity has to be defined (see the "Thermal/Process Conditions Assignment" section for more details).

Initial conditions menu

Nothing should be specified in this menu for fluid flow problems.

Run parameters menu

The following Run Parameters should be specified :

Version 2006.0

Page 108

ProCAST User Manual

The fluid flow module should be activated with FLOW = 1 or 3, as well as the free surface model (FREESF) for mold filling and the gas model (GAS). The storage frequency of the fluid flow results (VFREQ) should be specified. Then, information about the reference pressure (PREF) and pressure driven inlet (PINLET), about the final fill fraction (LVSURF) and the filling parameter (WSHEAR, WALLF) shall be defined. Please refer to the "Flow Run Parameters" section for the full description of all the Run Parameters.

Version 2006.0

Page 109

ProCAST User Manual

RADIATION
Radiation model
The radiation module allows to perform complex radiative problems (e.g. investment casting), with the calculation of the shadowing effects (view factor calculations). The typical results which can be obtained are the following : Effect of radiative heat transfer Shadowing effect

Radiation can be either treated as a simple radiative flux (described in the "Thermal/Boundary conditions" section), or with a complex radiation algorithm which takes into account reflexions, obstructions and shadowing effects. The setup of a case with such complex radiation (called hereafter "Radiation with view factors") is explained in this section. As Radiation with view factors involves the calculation of the interaction of the components (casting and mold) with the environment (furnace, castshop, ...), it is necessary to include the environment into the model. This is done with an "Enclosure". If the casting is put into a furnace, the Enclosure is the furnace itself (or the inner skin of the furnace). However, it the casting (and mold) is sitting on the floor of the castshop, one should set an "artificial" enclosure which will surround the casting and which will have the same effect as the environment. An enclosure can be either a solid (represented by a solid 3-D FEM mesh, as the casting, mold, etc...), or by a closed surface (represented by a closed FEM surface mesh). The figure below shows a casting within a solid enclosure (for symmetry reasons, only a sector is modeled).

Version 2006.0

Page 110

ProCAST User Manual

The figure below presents a casting within a surface enclosure shown in grey (for symmetry reasons, only half of the geometry is modeled)

Flow chart
This section describes the additional set-up necessary for radiation calculations. For the set-up of a thermal case, please refer to the "Thermal" section of the Preprocessor.

Version 2006.0

Page 111

ProCAST User Manual

File/Multiple meshes menu

When a model contains an enclosure (i.e. a surface mesh which will act as an enclosure), there are two ways to load the corresponding meshes : a) the enclosure is built into MeshCAST, together with the solid mesh. In this case, the enclosure should be "tagged" as an enclosure. b) the enclosure is built separately in MeshCAST, as a conventional surface mesh (without and "enclosure tag"). Then, the solid mesh and surfaces meshes should be loaded into PreCAST using the "File/Multiple meshes" menu. Please note that in previous versions, the surface meshes of the enclosures had to be "tagged" as enclosures. This is not anymore needed.

Geometry/Symmetry menu
Firstly the symmetries (if any) should be defined. ProCAST is able to deal with symmetry implying one mirror, two orthogonal mirrors, a single rotation of n sectors and a combination of them. In order to illustrate the different possibility of symmetry, consider the simple but explicit example of two concentric cylinders (the inside cylinder is the casting and the outside cylinder is a solid enclosure). The full geometry is shown in the figure hereafter.

Version 2006.0

Page 112

ProCAST User Manual

The full geometry of the two concentric cylinders.

The following figures illustrate the different possibility of simplification by symmetry. One mirror (M1)

Two examples of simplified geometry by one mirror.

Two orthogonal mirrors (M1 and M2)

Version 2006.0

Page 113

ProCAST User Manual

Two examples of simplified geometry by two orthogonal mirrors.

A simple rotation of n sectors (R)

A simplified geometry by a rotation of 8 sectors.

A rotation of n sectors (R) associated with a mirror (M1)

Version 2006.0

Page 114

ProCAST User Manual

Two examples of simplified geometry by a rotation (left: 8 sectors, right: 4 sectors) and one mirror.

A rotation of n sectors (R) associated with two orthogonal mirrors (M1 and M2)

A simplified geometry by a rotation with 4 sectors and two orthogonal mirrors.

The symmetries are defined in the Geometry/Symmetry menu :

Version 2006.0

Page 115

ProCAST User Manual

In case of a rotation, the axis of rotation should be defined by the coordinates of two points and the number of sectors. Then, the "Rotational" check box should be checked. One should be careful that the mesh should be defined so that the selected number of sectors will not create overlaps of the mesh (i.e. to specify 7 sectors for a mesh which corresponds to a sector of 60). For mirror symmetries, each plane of symmetry should be defined by the coordinates of three points (which should not be co-linear). Then the "Mirror-1" and "Mirror-2" (if applicable) check box should be checked. If two mirror symmetry are used, one should be careful to make sure that the two planes are orthogonal. In the case of mirror symmetry with a rotation, the axis of rotation should be either perpendicular or parallel to the mirror plane(s). In case of parallel plane(s), the axis of rotation should be within the mirror plane(s). Then, the "Apply" button should be used to validate the symmetry definition. To disable a symmetry, just uncheck the corresponding check box. The "Get Co-ord" button allows to pick nodes of the FEM mesh for an interactive definition of the mirror planes or rotational axis. To use it, first click in the X coordinate box of the point which should be defined interactively, then click on "Get Co-ord" and finally, click on the desired node on the geometry. The corresponding coordinates will fill automatically the corresponding fields. Repeat it for the other points. Please note that one should click very close to the desired node (otherwise, it may be possible that a node "behind" is selected).

Materials and Interface menu

Version 2006.0

Page 116

ProCAST User Manual

Nothing specific should be defined at the level of the Materials and Interface menus. The complex radiation method should be set in the "Boundary conditions" and "Process" menus.

Boundary conditions menu

When the ambiance or the furnace is modeled with an enclosure, one should specify the temperature and the emissivity of the enclosure in the "Assign Enclosure" menu.

The enclosure can be divided in different sets (1) (in order to apply different temperatures and/or emissivities). One can Add new sets (2). Then, using the selections tools (3) (see the "Pre-processing/Thermal" section for the full description of the selection tools), the desired surfaces of the enclosure can be "painted" (4) and stored (5). Finally, the desired temperature and emissivity can be selected in the database (6) and assigned (7). Please note that one should assign to each set both a Temperature entry and an Emissivity entry. In order to compute well the View Factors of the enclosure, one should make sure that the surfaces of the enclosure (i.e the triangles or the quadrangles) are well oriented (i.e. the surfaces are pointing inwards). To do so, the following icons are available next to the selection tools :

Version 2006.0

Page 117

ProCAST User Manual

View icon : Viewing of the face orientation

Reverse icon : Reversal of the face orientations

Align icon : Automatic alignment of the face orientations To view the face orientations, one should first select the desired "Enclosure set" and then press the "View" icon. The arrows are drawn for the selected set as shown in the figure hereafter.

If all the arrows are pointing outwards, the "Reverse" icon allows to reverse the orientation of all arrows. If only a few arrows are pointing in the wrong direction, one should use the "Align" icon in order to have all the arrows pointing in the same direction. Then, one may need to use the "Reverse" icon to point the arrows inwards.

Process menu
The process menu allows to define the motion (if any) of the enclosure (Assign Enclosure) with respect to the casting or of material domains (Assign Volume).

Version 2006.0

Page 118

ProCAST User Manual

Moreover, it allows to define the gravity vector (see the "Thermal/Process Conditions Assignment" section for more details). When "Assign Enclosure" is opened, the following panel appears on the upper right corner of the window. Each enclosure set is displayed and three database entries are possible. The first column corresponds to a "Translation", the second one to "Rotation" and the third one to "Revolution". If a "*" appears, it means that no motion is defined for this set. Otherwise, the number indicates the corresponding database entry.

In the case of "Assign Volume", the same type of panel appears, but instead of the enclosure sets, the different material domains are listed. Again in this case, it is possible to specify a Translation, a Rotation or a Revolution. Any combination of the three is possible, however, please note that the user should check that there is no conflict between these motions and prevent any inter-penetration of the different materials and/or enclosures.

On the bottom right, the motion database (Process database) appear (see the "Process Database" section for more details).

Version 2006.0

Page 119

ProCAST User Manual

Run parameters menu

The following Run Parameters should be specified :

Version 2006.0

Page 120

ProCAST User Manual

The radiation module is automatically activated when Heat Boundary conditions with View Factors ON are set. Thus, no specific Run Parameter should be activated at this stage. There is however three Run parameters which are important to set if the enclosure is moving. In this case, the update of the view factors will be triggered by the VFTIME or VFDISP parameters. If VFDISP is activated, the Enclosure ID on which the motion will be recorded should be specified. Please refer to the "Radiation Run Parameters" section for the full description of all the Run Parameters.

Version 2006.0

Page 121

ProCAST User Manual

STRESS
Stress models
The stress module allows to perform a thermo-mechanical calculation. The typical results which can be obtained are the following : Stress distribution Deformations (elastic and plastic) Displacements Gap formation Elastic springback Die fatigue Hot tears Cracks

In order to address these different aspects, five different stress models are available. Linear Elastic Elasto-plastic Elasto-viscoplastic Rigid Vacant

These different models are presented in the Databases section.

Flow chart
This section describes the additional set-up necessary for stress calculations. For the set-up of a thermal case, please refer to the "Thermal" section of the Preprocessor.

Version 2006.0

Page 122

ProCAST User Manual

Materials menu
The stress properties should be assigned in the Materials/Stress menu.

This is opening the following window. Firstly, each domain should be selected (1). Then, the desired properties should be chosen in the database list (2) and then the "Assign" button (3) should be pressed in order to link the material with the corresponding stress properties. The stress properties, as well as the different stress models are described in the "Stress database" section.

Version 2006.0

Page 123

ProCAST User Manual

Interface menu
Nothing specific should be defined at the level of the interfaces.

Boundary conditions menu

The user can specify three types of stress boundary conditions, as shown in the figure below.

Version 2006.0

Page 124

ProCAST User Manual

The specification of the displacements is very important in order to guarantee that the model will not rotate or move in an unexpected fashion, under the effect of stresses. The figure below is summarizing typical displacement constraints.

One should be careful not to over constrain the model with too many displacements boundary conditions. If this is the case, this may induce artificial stresses locally, as shown in the figure hereafter.

Version 2006.0

Page 125

ProCAST User Manual

One should note that a symmetry is also inducing displacement constraints as shown in the figure below.

It is necessary to specify displacements on the casting domain when the mold is set to "Vacant". However, when the mold is set to "Rigid", it is not necessary to specify displacements on the mold, except if one wants to model the Elastic spring-back when the mold is opened. In the case of die casting, it is recommended to set a zero vertical displacement at all the top and bottom points of the upper and lower dies (black triangles on the figure below), as well as one lateral zero displacement on each side of each die (blue triangle on the figure below).

Version 2006.0

Page 126

ProCAST User Manual

One should be aware that the application of only a "Surface load" on the top die (as shown below) is not enough.

This can be explained with the following sketch.. In the left case (blue), only two loads are applied to the part. This part is not constraint and it can move in any direction. In order to make sure that the part will not move, one has to apply in addition the appropriate zero displacements (right case in green). In the center

Version 2006.0

Page 127

ProCAST User Manual

case (orange), the displacement constraints are not sufficient and the part could rotate around the bottom right point.

Finally, "Point load" can be applied locally to model the effect of a local force.

Process Initial conditions menu

Nothing should be specified in these menus for stress problems.

Run parameters menu

The following Run Parameters should be specified :

Version 2006.0

Page 128

ProCAST User Manual

The stress module should be activated with STRESS = 1. The storage frequency of the stress results (SFREQ), as well as the calculation frequency of the stresses (SCALC) should also be specified. Please refer to the "Stress Run Parameters" section for the full description of all the Run Parameters. In addition, when a Stress calculation is performed, one should set PIPEFS = 0. This allows to prevent an unexpected effect of the piping on the stress calculation.

Advanced stress features

Elastic Springback

The stress module of ProCAST allows to simulate the Elastic Springback occurring when a mold is opened. In the figure below (left picture), the mold in blue is closed and it induces some stresses on the "T-bar". When the mold is

Version 2006.0

Page 129

ProCAST User Manual

opened (right picture), the corresponding mold resistance is removed and the stresses are relaxed. As the elastic strain is released, the shape of the part is changing (on the right picture, both the part and the mold are represented, although the mold is not anymore present).

To model the Elastic Springback effect, the user just needs to define a function for the interface heat transfer coefficient (at all the interfaces which will not be anymore in contact when the mold opens). One should define a zero interface heat transfer coefficient when the mold opens and automatically, the mechanical effect of the mold will be removed in the model. However, if one want to continue to simulate the cooling of the casting, while in the air, the user should not forget to set a Heat boundary condition on the casting surface (and eventually on the inside mold surface) with a non-zero value from the opening time (the technique is the same as the one used in cycling calculations).

Version 2006.0

Page 130

ProCAST User Manual

DATABASES
Material Database
All the material properties are stored in Databases. This section describes the database containing the material properties for Thermal and flow calculations. The Database containing stress properties is described in the "Databases / Stress Database" section. Moreover, the "Thermodynamic" databases are described in the "Databases / Thermodynamic databases" section. In the Material Assignment menu (see "Material assignments"), the content of the material database is shown (see below). Above the material list, the database management buttons are present :
Read : the database entry can be read and modified (if the user has the appropriate rights) Add : a new entry can be created Copy : an existing entry can be copied in order to create a new entry Del : an exisiting entry can be deleted Sort : the list of materials is sorted by alphabetical order Search : a search on the material name can be done

When an existing material is opened (with the READ button), a window appears with the following content :

Version 2006.0

Page 131

ProCAST User Manual

Firstly, the material name, the user, as well as the creation date of the material is shown. Then, the material properties are organized in different tabs, with a hierarchical structure :
Composition Thermal Conductivity Density Specific Heat Enthalpy Fraction Solid Latent heat Liquidus-Solidus Exothermic Fluid Viscosity Surface Tension Permeability Filter Comments

The yellow tabs indicate that values are defined for the corresponding properties. The white tab shows the active one. Once the "Thermal" or "Fluid" tab is selected, a second level of tabs becomes active (and so on). The figure below shows the definition of the thermal conductivity :

Version 2006.0

Page 132

ProCAST User Manual

The above figure shows the standard panel for material properties definition. Firstly, the user has to define whether the material property is defined by a

Version 2006.0

Page 133

ProCAST User Manual

Constant or by a Table (function of temperature), by clicking on the corresponding radio button. When a Constant is selected, the user can choose the units and enter the constant material property in the white field below "Enter". When a Table is selected, the user can input the temperature-dependant property in the white field on the bottom left of the screen. Each time a new line is filled, the graph on the right is updated. In this case, the units for the temperature and the property can be selected. Above the table, the buttons allow to erase the whole table, erase only the selected line, import or export the table. For imports, the table should be in the form of a text file, with X and Y values on the same line, separated by at least a blank. The number of lines should not exceed 100. The Export format is the same. If a "Table" is defined and the user decides to switch to a "Constant" (or viceversa), a warning will be prompted saying that the table will be lost. A confirmation will then be asked to the user. Of course, the data will be totally lost only when the material is "Stored". All the properties which can be defined as temperature-dependant are organized in the same way. The only thing which changes is the hierarchy of tabs, as shown below for the Fluid/Viscosity/Carreau-Yasuda/Zero Viscosity definition.

Version 2006.0

Page 134

ProCAST User Manual

Beside these properties, some of them are defined only by constants, such as the latent heat, the liquidus and solidus temperature or the filter properties. In this case, only the corresponding constant(s) should be entered (see figure below).

The following section "Material Properties" describes the different properties and when it should be defined.

Material Properties
Good material properties are the best base for a good simulation. Properties could be found in several locations, such as litterature, material suppliers, universities, web, ... www.matweb.com is material properties website which contains many useful data. www.matdata.net is a search engine for material properties. An other way to obtain material properties is through the Thermodynamic Databases which are embedded into ProCAST (see the "Thermodynamic Databases" section for more details).

Thermal problems

Version 2006.0

Page 135

ProCAST User Manual

No phase change

For thermal problems (with or without solidification), the minimum data which are required are the following (typically for mold materials) :
Thermal conductivity Specific heat Density

These properties can be either constant or temperature dependant.

With phase change (solidification)

When solidification is present (i.e. for casting materials), one should define in addition the following properties :
Fraction of solid Latent heat Liquidus and Solidus temperatures

The fraction of solid curve must be temperature-dependant. It should start at 0.0 at high temperature and increase to 1.0 towards the low temperatures. The fraction of solid should be a strictly descending curve and it should be strictly defined between 0.0 and 1.0. If it is not the case, a warning will be issued. If there is a isothermal transformation (e.g. eutectic plateau), it should be "spread" over an interval of one degree. The latent heat, liquidus and solidus temperatures are defined by constants. Please note that the liquidus and solidus temperatures should be consistent with the fraction of solid curve (no consistency checks are performed). The liquidus and solidus temperatures are used for the porosity models and for the calculation of the permeability of the mushy zone in the case of flow calculations. ProCAST offers an alternative in the definition of the phase change. Instead of defining the specific heat, and the latent heat, one can define the corresponding enthalpy curve. The enthalpy as a function of temperature, H(T), is defined as follows :

where cp(T) is the specific heat as a function of temperature, L is the latent heat and fs is the fraction of solid. As there are two ways of defining the phase change, the software is automatically detecting if there is a conflict in order to have either :

Version 2006.0

Page 136

ProCAST User Manual

specific heat Latent heat

or
enthalpy

If this is the case, the following message is displayed :

and the user has to select which definition is preferred. In previous versions (v4.x.x and v3.x.x), there was no check to prevent both definitions. Thus, if a model which was created in a previous versions is loaded into PreCAST v2006.0, the following warning will be displayed (during the load in PreCAST) :

The user will need to resolve the conflict, by selecting which data are to be kept (i.e. either enthalpy or specific heat/latent heat) for each material which has this duplicate definition.
Density

The density is used in thermal calculation (it multiplies the specific heat, the enthalpy and the latent heat), as well as in fluid flow calculations and in porosity calculations. Please refer to the "Porosity models" section for more details about the density definition.

Fluid flow problems

Version 2006.0

Page 137

ProCAST User Manual

For fluid flow problems, it is mandatory to define the viscosity. Then, optionnal definitions are available, such as "Surface Tension", "Permeability" and "Filter".
Viscosity

Several viscosity models are available in ProCAST :

Newtonian Carreau-Yasuda Power-cutoff

The Carreau-Yasuda model corresponds to Non-Newtonian flow, where the viscosity depends upon the shear rate (see the equation below) :

with :

strain rate zero strain rate viscosity infinite strain rate viscosity n a phase shift Power law coefficient Yasuda coefficient

The above parameters can be defined in the database (as constants or as function of temperature) as follows :

The Power-cutoff is used in the case of Thixocasting.

Surface tension

This option is not described as it has not been validated at this stage. One can say that in conventional casting processes, the surface tension effects are certainly negligible in comparison to the simplifications made in the free surface algorithms.

Version 2006.0

Page 138

ProCAST User Manual

Permeability

For a casting material, the permeability is defined by a Karman-Cozeny model, modified by Beckermann at low fraction of solid. The user has also the ability to define its own permeability table, as a function of temperature. In this case, a high permeability corresponds to a "free flow", whereas a low value corresponds to "no flow". For "casting" materials, the permeability is applied only in between the solidus and the liquidus temperatures. For mold materials (in the case of lost foam), a permeability should be defined. In this case, one can define a constant or a temperature dependant permeability. For Filter materials, if the Permeability is defined, it will override the default permeability calculated from the Filter tab.
Filter

Filters are characterized by the following properties :

Void fraction Surface area Pressure Drop

The void fraction (Fv) corresponds to the amount of "porosity" or void inside the filter. This value is dimensionless [-]. The definition of this value is mandatory in all cases. The Surface area (Sa) corresponds to the amount of "interface" between the filter material and the air (when the filter is empty) per unit volume (see example below). This value is used for the calculation of the thermal exchange between the filter and the liquid metal going through, as well as for the automatic permeability calculation. The units are the reversed of a distance (e.g. [1/m]). The definition of this value is mandatory in all cases.

Version 2006.0

Page 139

ProCAST User Manual

The Permeability of the filter (i.e. its resistance to the flow) can be calculated in three different ways.
a) Automatic permeability calculation

From the Void Fraction (Fv) and the Surface Area (Sa) definitions, the permeability can be automatically computed (based upon Karman-Cozeny), according to the following relationship :

This mode is activated if the "Pressure Drop" and "Permeability" tabs are not defined.
b) Pressure drop calculation

If the "Pressure Drop" tab is defined, then, the permeability is calculated, using the following values (coming from simple experiments) :

Version 2006.0

Page 140

ProCAST User Manual

and the following equation :

where v, P and x are measured values which can be made in a simple experiment (or provided by the Filter supplier). Please note that the Flow rate, v, corresponds to the velocity used in the experiment and not the velocity of your casting model. If both the "Pressure Drop" and "Permeability" tabs are defined, the "Permeability" values are ignored (and replaced by the ones obtained from the above equation).
c) Specified permeability

It is also possible to define a given permeability value in the "Permeability tab". In this case, the "Void fraction" is not used for the automatic Karman-Cozeny relationship. See the "Filters" section for more details about the settings of cases with Filters.

Thermodynamic Databases

Version 2006.0

Page 141

ProCAST User Manual

Material properties, such as the enthalpy curve and the solidification path (i.e. the fraction of solid curve versus temperature), density, viscosity and thermal conductivity can be computed automatically from thermodynamic databases. ProCAST has an automatic link with thermodynamic databases to calculate these properties. It is thus possible to compute the enthalpy curve, the fraction of solid curve, the density, the viscosity and the thermal conductivity, based upon the chemical composition, for the following systems and the following alloying elements (the elements shown in blue were added in version 2006.0) :

CompuTherm LLC databases

Al database: Ag Al B C Cr Cu Fe Ge Hf Mg Mn Ni Sc Si Sn Sr Ti V Zn Zr Fe database: Al C Co Cr Cu Fe Mg Mn Mo N Nb Ni P S Si Ti V W Ni database: Al B C Co Cr Fe Hf Mo N Nb Ni Re Si Ta Ti W Zr Ti database: Al B C Cr Cu Fe H Mo N Nb Ni O Si Sn Ta Ti V Zr Mg database: Ag Al Ca Ce Cu Fe Gd Li Mg Mn Nd Sc Si Sr Y Zn Zr

The other alloying elements which are not present in this list are not available in the database and will have no effect on the computed material properties. More details about composition limitations are given in the next section. To activate the thermodynamic database, one should go in the "Composition" tab of the material properties window.

Version 2006.0

Page 142

ProCAST User Manual

Then the Base alloy (i.e. Al, Fe, Ni, Ti or Mg) should be entered, as well as each alloying element with its concentration (in weight percent). Once the chemical composition is entered, the "Apply->" button is pressed and the "Scheil" or "Lever" option is selected to start the computation of the fraction of solid and of the enthalpy.

Version 2006.0

Page 143

ProCAST User Manual

"Scheil" and "Lever" correspond to two different microsegregation models. In the case of "Lever", the Lever Rule is applied, which corresponds to a complete mixing of the solute in the solid (i.e. very good diffusion in the solid). On the other hand, the Scheil model corresponds to no diffusion at all in the solid phase (both model consider complete mixing or infinite diffusion in the liquid). The main difference between the two models is the shape of the fraction of solid curve at the end of solidification, as well as the solidus temperature (see figure below).

For most alloys, it is recommended to use the Scheil model, except for low alloy steels where the diffusion in the solid is very fast.

Version 2006.0

Page 144

ProCAST User Manual

In the example above, when the Scheil button is selected, the following curves appear :

One could see the fraction of solid curve, as well as the fractions of the different phases as a function of temperature. In the same time, automatically, the fraction of solid curve, the liquidus and solidus temperatures, the enthalpy curve, the density, the viscosity and the thermal conductivity are stored in the database, as shown hereafter (of course the value are finally stored only when the "Store" button is pressed, before exiting the database).

Thermal conductivity

Version 2006.0

Page 145

ProCAST User Manual

Density

Enthalpy

Version 2006.0

Page 146

ProCAST User Manual

Fraction of solid

Version 2006.0

Page 147

ProCAST User Manual

Liquidus-Solidus

Viscosity

Version 2006.0

Page 148

ProCAST User Manual

Please note that when a thermodynamic database is used, as the enthalpy is calculated, the specific heat and the latent heat should not be defined (as they are contained in the enthalpy). During the Thermodynamic database calculation, a file named "prefix.phs" is created. It contains for each temperature the phase fractions, as well as the composition of each phase. This information is not needed for a ProCAST calculation, but it can be interesting for other purposes (e.g. growth kinetics calculations).

For some chemical composition, it may happen that the software which extracts the data from the Thermodynamic database is not able to find the right set of stable phases at low temperature.

Usually, this does not affect the determination of properties of interest (which are more near the solidification range). Thus, it is possible to use the calculated values

Version 2006.0

Page 149

ProCAST User Manual

as such. If such situation occurs, one should check the calculated data in order to be sure that it covers at least the temperature range of interest. In very few cases, it is possible that the density calculation does not give relevant results, as shown in the figure hereafter. In this case, these result should not be used (or the wrong values should be erased).

As a general rule, if the results are not realistic, it is advised to suppress (i.e. ignore) the elements which are present in very low concentrations. This is especially true for traces of Sulfur (S) and Phosphorus (P) in steels, which are sometimes "corrupting" the results.

Calculation of Stress Properties

Beside the thermal properties, it is possible to calculate automatically some Stress properties. At this stage, the Young's modulus, the Poisson's ratio and the Thermal expansion coefficient can be calculated based upon the phases obtained from the thermodynamic databases. When the Properties calculation is started in PreCAST with either the Scheil or the Lever model, the following window appears :

Version 2006.0

Page 150

ProCAST User Manual

The user has the choice of either not calculate the Stress properties, to create a new entry in the Stress database or to substitute/Add the data to an existing entry. If a new entry is created the user has to specify its name (without spaces). If the user would like to substitute or add the calculated data to an existing entry of the stress database, one should select the desired entry with the Browse button. Once these choices are made, the computation can be started (of both the thermal and stress properties) with the "Compute" button. Please note that the other Stress properties (i.e. Yield stress, hardening, viscoplastic,...) can not be calculated at this stage. Thus these properties will remain empty. The following figures are showing examples of computed Stress data from an A356 alloy.

Version 2006.0

Page 151

ProCAST User Manual

Young's modulus

Version 2006.0

Page 152

ProCAST User Manual

Poisson's ratio

Thermal expansion coefficient

Databases limitations
The Computherm databases can be used for the following elements and in the following ranges. More information can be obtained on the www.computherm.com web site. The recommended composition ranges mentioned hereafter are not strict limits. These are ranges which were extensively tested. The Computherm manual (from Computherm LCC) is added in the Software installation (in the dat/manuals/PDF directory). This manual describes for each alloying system the phases which are calculated, the limitations as well as the validations which have been made.

Al database

Version 2006.0

Page 153

ProCAST User Manual

Developed for Al-rich alloys such as commercial casting and wrought alloys. Tested with more than 40 commercial Al alloys.
20 Components : Major alloy elements: Al, Cu, Fe, Mg, Mn, Si, Zn Minor alloy elements: Ag, B, C, Cr, Ge, Hf, Ni, Sc, Sn, Sr, Ti, V, Zr Major phases: Liquid, Fcc_A1(Al), Diamond_A4(Si), Al5Cu2Mg8Si6, Al8FeMg3Si6, Eps, Sigma-(Al,Cu,Zn)2Mg, T(Al,Cu,Zn)49Mg32, Al20Cu2Mn3, Al23CuFe4, Al7Cu2Fe, SAl2CuMg, TAO(t), a-AlFeSi, b-AlFeSi, AL15_FeMn3Si2(aAlMnSi), AlMnSi-Beta, AlCu_Theta(q), Al13Fe4, AlMg_Beta, Al11Mn4, Al12Mn, Al4Mn, AL6_FeMn, Al3Ni1, AlSr4, Mg2Si,Al3Zr, Al3Sc_x Recommended composition range (in wt%): Al 80 ~ 100 Cu 0 ~ 5.5 Fe 0 ~ 1.0 Mg 0 ~ 7.6 Mn 0 ~ 1.2 Si 0 ~ 17.5 Zn 0 ~ 8.1 other 0 ~ 0.5

Fe database
Developed for Fe-rich alloys.
18 Components: Al, C, Co, Cr, Cu, Fe, Mg, Mn, Mo, N, Nb, Ni, P, S, Si, Ti, V, W. 59 Phases: Liquid, BCC_A2 (ferrite), HCP_A3, FCC_A1(austenite), TCP phases, Carbides, and so on. Recommended Composition Limits (wt%): Fe > 50 Ni < 31 Cr < 27 Co, Mo < 10 V, W < 7 C, Cu, Mn, Nb, Si, Ti < 4 Al, Mg, N < 0.5 P, S < 0.05

Version 2006.0

Page 154

ProCAST User Manual

It was observed that alloy elements which are present is very small quantities (such as P and S) may cause problems in the phase determination. As these elements do not affect significantly the material properties (although it may have important effects in other fields), it is recommended to remove these elements for the computation.

Mg database
Developed for commercial Mg-rich alloys 16 components: Mg, Ag, Al, Ca, Ce, Cu, Gd, Li, Mn, Nd, Sc, Si, Sr, Y, Zn, Zr Contains more than 200 phases. Recommended composition range (in wt%): Not yet available

Ni database
Developed for commercial Ni-rich alloys.
17 Components: Al, B, C, Co, Cr, Fe, Hf, Mo, N, Nb, Ni, Re, Si, Ta, Ti, W, and Zr. 63 Phases: Liquid, Fcc_A1(g), L12_Fcc(g), TCP phases, Carbides, and so on. Recommended Composition Limits (wt%): Ni > 50 Al, Co, Cr, Fe < 22 Mo, Re, W < 12 Hf, Nb, Ta, Ti < 5 B, C, N, Si, Zr < 0.5

Ti database
Developed for commercial Ti-rich alloys such as alpha, alpha+beta, and beta alloys.
18 Components: Al, B, C, Cr, Cu, Fe, H, Mo, N, Nb, Ni, O, Si, Sn, Ta, Ti, V and Zr 108 Phases: Liquid, BCC_A2(b), HCP_A3(a), DO19_Ti3Al(a2), Laves, and so on. Recommended Composition Limits (wt%):

Version 2006.0

Page 155

ProCAST User Manual

Ti > 75 Al, V < 11 Mo, Nb, Ta, Zr < 8 Cr, Sn < 5 Cu, Fe, Ni < 3 B, C, H, N, O, Si < 0.5

Influence of alloying elements

The goal of this section is to illustrate the influence of alloying elements on properties and to show why properties obtain with thermodynamic databases may differ from experimental data. An AlSi9Cu3Fe will be considered to illustrate this example. The usual average chemical composition of such alloy is the following :

When the Scheil model is used, with the above composition, 10 phases are found (in addition to the liquid phase) :

Version 2006.0

Page 156

ProCAST User Manual

Starting from AlSi9Cu3Mg0.3, the other alloying elements are added progressively. The effect on the solid fraction curve is shown in the following figures (please note that the Temperature scale is changing from one graph to the next one):

Version 2006.0

Page 157

ProCAST User Manual

AlSi9Cu3Mg0.3

Version 2006.0

Page 158

ProCAST User Manual

AlSi9Cu3Mg0.3Fe1.3 (the effect of Fe is mainly visible at the liquidus)

AlSi9Cu3Mg0.3Fe1.3Mn0.55Ni0.55Zn1.2 (Mn, Ni and Zn are mainly affecting the second half of the curve)

Version 2006.0

Page 159

ProCAST User Manual

AlSi9Cu3Mg0.3Fe1.3Mn0.55Ni0.55Zn1.2Cr0.15 (Cr is rising the liquidus temperature from 612 to 640C)

Version 2006.0

Page 160

ProCAST User Manual

AlSi9Cu3Mg0.3Fe1.3Mn0.55Ni0.55Zn1.2Cr0.15Ti0.15 (When Ti is added, the Al3Ti phase appears, with a very high liquidus temperature above 760C) The following figure are showing the solid fraction curves for all the alloys together.

Version 2006.0

Page 161

ProCAST User Manual

In order to see the effect near the liquidus, the same figure is shown, with a different vertical scale. Only the first 5% are shown.

Version 2006.0

Page 162

ProCAST User Manual

One can see very well on this latter figure the effect of Ti. Ti is added in order to create this Al3Ti phase which is stable at very high temperature, which is acting as inocculant. The amount of this phase is very small (around 0.5%). The above example is showing that one should be careful with the use of Thermodynamic databases. In this case, for instance, it would be advisable to ignore the Ti for the thermodynamic computation, in order to avoid this "artificially" high liquidus temperature. This explains also why there are differences observed between literature values (measurements) and computed values for liquidus and solidus temperature. This is due to the fact that such values are measured usually by Thermoanalysis and that small amount of solid (like the few percents due to Fe, Cr and Ti near the liquidus temperature) can not be detected. One should note that the usual literature value of the liquidus for this alloy is 588C which corresponds to a computed value of about 5% of fraction of solid. The measurement of the solidus temperature is even more difficult and thus, it is normal to observe differences.

Version 2006.0

Page 163

ProCAST User Manual

Interface Database
The "Interface" database contains two types of entries :
Interface Die Combo

"Interface" corresponds to "standard" interface heat transfer coefficients between two different domains, where a "coincident" or "non-coincident" interface has been defined (not applicable in case of perfect contact - EQUIV). "Die Combo" is a specific entry to define a "composite" type of heat transfer in the case of die casting. It allows to define automatically the sequence of closed and opened dies.

Interface
For the definition of the Interface heat transfer coefficient, the following panel is opened.

The structure of the above panel is described hereafter (see the numbers in the panel below). Firstly, the type of the database entry is shown on the top of the panel (1). Then, the entry can be labeled with a Keyword (2) which will help to recognize it's content for later use. If a constant heat transfer coefficient is to be defined, it's value should be entered in the corresponding field (3). Then the units can be defined (4). The available choice of units is proposed when one clicks on the unit box. If it is desired to define a "Temperature-" and/or a "Time-"dependant heat transfer coefficient, the corresponding table can be defined in the "Temperature" or "Time" buttons (5) (for more details about the definition of

Version 2006.0

Page 164

ProCAST User Manual

tables, please refer to the "Boundary Conditions Database" section). It is also possible to specify that the interface heat transfer coefficient will be defined by a User Function (6). Finally, the data are saved with the "Store" button (7).

If the interface heat transfer coefficient is non-constant, the value defined in (3) and the table(s) defined in (5) are all multiplied to get the final value (at a given temperature and a given time). It is allowed to define only a table without a constant (in this case, it is like if the constant is equal to 1). If a User function is specified, neither a constant value, nor a table can be defined.

Die Combo
The "Die Combo" definition is used only in the case of die casting, to handle automatically the sequence of closed and opened dies (see the "Cycling" section for more details). The "Die Combo" database entry corresponds to the following panel.

Version 2006.0

Page 165

ProCAST User Manual

It allows to define : a) the interface heat transfer coefficient (when the die is closed) in the "Constant" field and/or as a Temperature-dependant table ("Temperature" button). b) when the die opens, the interface heat transfer coefficient between the die and the air is defined by a heat transfer coefficient (Air Coeff) and by an ambient temperature (Air Temp). c) if there is a Spray cooling stage, it can be defined by the corresponding heat transfer coefficient (Spray Coeff) and the Spray temperature (Spray Temp). The check-box "Attached until Ejection" will tell whether the switch between the interface heat transfer and the air cooling will be done when the dies opens or upon ejection of the part. These times are given in the "Run parameters", in the "Cycles" tab, as shown below.

Version 2006.0

Page 166

ProCAST User Manual

Boundary Conditions Database

The "Boundary Conditions" database contains the following types of entries in the "Assign Surface" menu :
Temperature Heat Velocity Pressure Inlet Turbulence Vent Inject Displacement Point load Surface load

in the "Assign Volume" menu :

Volumetric Heat Momentum Source Mass Source

and in the "Assign Enclosure" menu :

Emissivity

Temperature
"Temperature" allows to specify (or impose) a temperature at a given point. It is especially used to specify inlet temperatures in filling calculations.

Version 2006.0

Page 167

ProCAST User Manual

The following panel corresponds to the "Temperature" database entry.

As most of the panels of the Boundary Conditions database have the same structure, the principles will be described for this one only. Firstly, the type of the database entry is shown on the top of the panel (1). Then, the entry can be labeled with a Keyword (2) which will help to recognize it's content for later use. If a constant value of temperature is to be defined, it's value should be entered in the corresponding field (3). Then the units can be defined (4). The available choice of units is proposed when one clicks on the unit box. If it is desired to define a "Time-"dependant Temperature, the corresponding table can be defined in the"Time" button (5). Finally, the data are saved with the "Store" button (6).

The following figure shows the choice of available units (when one clicks on the unit box).

Version 2006.0

Page 168

ProCAST User Manual

In order to define a Time-dependant function, one should click on the "Time" button (see 2 in the figure below). This automatically opens a table, as well as a graph. The values of the table can be entered in (3). It is possible to manage the table (4), with "Erase" operations, as well as "Import" or "Export". For the import, the data should be stored in a text file, with the X and Y values on the same line. One can have up to 100 lines (i.e. 100 points). The Export will also create an text file. For both the Import and the Export, it is possible to Browse in order to find the right location of the imported or exported file. Finally, when the table is well defined, one should SAVE it (5), before Storing (6) the database entry.

Heat
The "Heat" boundary condition allows to define the heat transfer between the outside faces of a given domain and the outside world (air, water, ...). The following panel is available for Heat database entry definition.

Version 2006.0

Page 169

ProCAST User Manual

The following equation shows the three possible contributions to the "Heat" boundary condition (all the contributions are added, although they may not be all active in the same time or in the same cases). The first term corresponds to a specified flux. It can be used if a given heat flux was measured for instance. The second term corresponds to Convective cooling. This is the most common definition of the cooling of an external face. It is defined by a heat transfer coefficient with the ambiance and by an external temperature (of the ambiance). The third term is used at high temperature, when radiation becomes important. In this case, the transfer is proportional to the StefanBolzmann constant and the emissivity and the difference of the fourth power of the temperatures (surface temperature and ambiance temperature). The above equation contains the four values which are shown in the panel below (1), such as the "Film Coeff", the "Emissivity", the "Ambient Temp" and the "Heat Flux". These values can be also "Time" and/or "Temperature" dependant (2). In this case, the value of the constant (if any) is multiplied by the corresponding table(s). It is also possible to define the different parameters by the corresponding "User Functions" (3).

Version 2006.0

Page 170

ProCAST User Manual

In the case of "Complex" radiation calculations, one should include "View Factors" (see the "Radiation" section for more details). In order to active the "Radiation" model, the "View Factor" button (see item 4 in the figure above) should be turned ON.

Velocity
The definition of the velocity is made in the following panel. In this case, the three components of the velocity vector should be defined. If a Time- or a Pressuredependant velocity is set, each component of the vector will be multiplied by the corresponding table(s). If a gate should be active until a given filling fraction is reached, the "Fill Limit" slider can be used for that purpose. This Fill limit has not the same effect as the Run parameter LVSURF (see the "Run parameters" section for more details). LVSURF allows to stop the filling, but more important it allows to switch off the Fluid flow solver when the filling is finished. On the other hand, the Fill limit defined in this boundary condition should be used if multiple gates are present. It allows to switch off automatically one gate when a given fill fraction is reached an the other gate will continue until LVSURF is reached. Although it is possible to define a Pressure-dependant velocity, one should be very careful in the use of such capability, as the pressure solution is very sensitive.

Version 2006.0

Page 171

ProCAST User Manual

When an inlet velocity is set to fill a casting, there are situations where the user would like to know which value of velocity should be set in order to achieve a given filling time. To do so, a "Velocity calculator" is available.

Version 2006.0

Page 172

ProCAST User Manual

To use the velocity calculator, the following operations should be done (see the numbers in the upper figure). Firstly, one should "Add" a new velocity boundary condition (1) and assign the corresponding nodes on the geometry. Then, open the "Velocity" boundary condition entry (2) and click on "Velocity calculator" (3). The "Calculator" window will appear (4) and one should select all the domains which correspond to the casting (5). In the above example, the casting is made of one domain only. Then, the expected filling time should be entered in (6) and the velocity is automatically calculated by pressing the button in (7). The velocity magnitude appears then in red (8). The value should be reported in (9), in the appropriate component of the vector. Finally, the values should be stored (10). One should be aware that the velocity calculator will give accurate results if all the nodes on the inlet surface are selected. If only a few nodes are selected, the inlet surface will "overflow" to the neighboring elements and the filling time will be shorter than the one specified. In order to prevent that, it is advised to design a small additional volume, corresponding to the jet of liquid entering the casting.

Pressure
The pressure is defined in the following panel, which has the same principles as the "Temperature" panel described above.

Version 2006.0

Page 173

ProCAST User Manual

Inlet
Inlet boundary conditions are used to specify in the same time an "equivalent" inlet velocity together with an inlet temperature. The inlet Flow rate can be time dependant.

Turbulence
In the case of a Turbulent model, one can specify the "Turbulent" state at the entry of the liquid in the calculation domain.

Version 2006.0

Page 174

ProCAST User Manual

Vent
If the "Gas" model is activated, the air can escape through vents. Vents are applied on nodes of the casting domains. Each vent is represented by an "equivalent tube" by which the air can escape. Thus, a vent is described by the diameter and the length of the "equivalent tube", as well as a roughness and the exit pressure.

Inject

Version 2006.0

Page 175

ProCAST User Manual

In the case of low pressure die casting (lpdc), it is possible to model the liquid bath of metal and to apply a pressure on it (in order to drive the metal in the mold) via injected air. The inject BC allows to specify the amount of injected air. Please note that this method may be delicate to converge.

Displacement
In the case of stress calculations, the displacement of the model should be constraint. The Displacement BC allows to define these constraints. Please note that if a field is not filled, the displacement is not constraint in this direction (e.g. it is possible to specify only a "0" in the x-field which will mean that the node can not move in the X-direction, but it is free to move in the Y and Z-directions.

Version 2006.0

Page 176

ProCAST User Manual

Point Load
For stress calculations, it is possible to apply a load (or a force) at a given location. The three components of the load should be defined.

Version 2006.0

Page 177

ProCAST User Manual

Surface Load
In the case of stress calculations, one can define a load applied on a given surface. In this case, it corresponds to a pressure. The direction of the load is specified by the three components of the load vector.

Version 2006.0

Page 178

ProCAST User Manual

Volumetric Heat
In some thermal problem, it may be necessary to define the generation of heat inside a given material (e.g. to model the effect of an electrical resistance). In this case, the thermal power which is generated per unit volume should be defined. This heat source will be applied on the entire specified material.

Version 2006.0

Page 179

ProCAST User Manual

Momentum Source
In the case of flow problems, one would like to model a source of momentum in a given material. This could be the case to take into account electromagnetic forces or the effect of a propeller. In this case, the momentum force vector should be defined. Please note that this force will be applied in the whole specified material.

Mass Source
In the case of filling calculations, one would like to define the amount of metal which enters into the material with a Mass source (instead of a velocity or an inlet BC). In this case, one can specify that a given amount of metal (Flow rate) is "appearing" at the given location (X, Y and Z coordinates) at the specified temperature. The metal will "appear" with a zero velocity.

Version 2006.0

Page 180

ProCAST User Manual

Emissivity
For radiation calculations, the emissivity of the mold, the casting top surface, the furnace or the enclosure should be specified. The emissivity should be defined between zero and one.

Version 2006.0

Page 181

ProCAST User Manual

Process Database
It is possible to define Translations, rotations and revolutions of domains. This is used especially for radiation calculations with view factors (where the furnace is moving with respect to the part, or vice-versa. The following panels allow to define these data.

Translation
The translation of a given material domain or of an enclosure is defined with the following panel. The translation can be defined in three different ways : x(t) - Translation vector or position (i.e. position vs time) v(t) - Translation velocity as a function of time (i.e. velocity vs time) v(x) - Translation velocity as a function of position (i.e. velocity vs position)
x(t) - Position vs Time

The position of the domain is defined as a function of time. This position is relative to the original position at step 0. The translation vector is defined by X, Y and Z which is multiplied by the Time function (in order to have the domain moving, it is necessary to define a time function. Otherwise, the domain will remain fixed - no motion). When x(t) is defined, it is possible to define the translation vector with user functions.

Version 2006.0

Page 182

ProCAST User Manual

v(t) - Position vs Time

The velocity of the domain vs time can be defined with the panel hereafter. The direction of the velocity vector is defined by the U, V and W components. This velocity vector can be either constant or it can be multiplied by a Time function. When v(t) is used, the software is transforming it in x(t) automatically in order to have a translation position which is independant from the timestep.

Version 2006.0

Page 183

ProCAST User Manual

v(x) - Position vs Position

The velocity of the domain vs Position can be defined with the panel hereafter. The direction of the velocity vector is defined by the U, V and W components (it should be a "unit vector"). This velocity vector can be either constant or it can be multiplied by a "Distance" function. The "Distance" is the relative distance with respect to the initial location of the domain which moves. When v(x) is used, it is mandatory to have a non-zero velocity for the initial position (i.e. the zero position). Please note that it is also mandatory to have always non-zero velocities defined (i.e. the velocity should never be zero).

Version 2006.0

Page 184

ProCAST User Manual

Rotation
The following panel allows to define the rotation of a material domain or an enclosure. The axis of rotation should be defined by two points.

Version 2006.0

Page 185

ProCAST User Manual

The rotation is defined in the "Time" button, which opens the following panel, for the definition of the angle as a function of time.

Revolution
The "Rotation" definition is used when less than one turn is done during the entire process. When one has several turns, the "Revolution" definition is used. In this case, the axis of revolution should be defined by two points.

Version 2006.0

Page 186

ProCAST User Manual

The "speed" of revolution can be defined either as a constant or as a function of time, with the following panel.

Stress Database
To access the Stress database, one should to in the Material/Stress menu

Version 2006.0

Page 187

ProCAST User Manual

Then, the list of available materials (stress properties) is shown in a window in the lower right corner (as for the standard material database. The database can be managed in the same way with the "Read", "Add", "Copy" and "Del" capabilities.

When a new stress database entry is created, the following window appears.

Version 2006.0

Page 188

ProCAST User Manual

The first thing to do is to select the desired model among the five following options :
Vacant Rigid Linear-Elastic Elasto-Plastic Elasto-ViscoPlastic

"Vacant" is used to specify that the domain will not participate to the stress calculation. Thus, no stress and no strain will be calculated and the domain will not participate to the contact algorithm (i.e. the domain will not create any resistance to the neighboring domains). No properties should be defined for a Vacant domain. No stress calculation will be done in a "Rigid" domain, however, the domain will participate to the contact algorithm (i.e. the neighboring domains will not be allowed to penetrate in the Rigid ones). No properties should be defined for a Rigid domain. For the three other models, the following data are required :

Version 2006.0

Page 189

ProCAST User Manual

All the properties can be defined as Constants or as Temperature dependant tables. The standard input panel is used as for the material database (see the "Material Database" section for more details) :

The three models, as well as the meaning of the corresponding properties, are described in the "Stress Models and Properties" section.
Warning

For stress calculations, the thermal material properties must include the phase change (i.e. fraction of solid curve). Otherwise, no stresses will be calculated in

Version 2006.0

Page 190

ProCAST User Manual

this material (as the fraction of solid will not be defined, it will be considered as "zero" and no stress will be calculated).

Stress Models and Properties

Warning

Before defining the stress properties, one should be aware of the following. For stress calculations, the thermal material properties must include the phase change (i.e. fraction of solid curve). Otherwise, no stresses will be calculated in this material (as the fraction of solid will not be defined, it will be considered as "zero" and no stress will be calculated). The three stress models available in ProCAST can be summarized in the following figure.

Linear Elastic model

The Elastic model is mainly characterized by the Young's modulus. It corresponds to the slope of the initial part of the stress-strain curve.

Version 2006.0

Page 191

ProCAST User Manual

Beside the Young's modulus, one should define the Poisson's Ratio and the Thermal Expansion coefficients. The value of the Poisson's ratio is usually around 0.3 for metals.

The thermal expansion can be defined by two different ways :

"Thermal Strain" "Secant" thermal expansion coefficient

In the case a Strain curve as a function of Temperature is measured, one can directly input such a curve in the "Thermal Expansion/Strain" tab.

Version 2006.0

Page 192

ProCAST User Manual

However, in the equation (see above), it is the "Secant" thermal expansion coefficient which is used. Thus, one can define "Secant" thermal expansion coefficient directly.

The figure above shows how to transform a measured "Strain curve" (shown in blue) in a Secant thermal expansion coefficient. To do so, one needs to define a Reference Temperature (Tref), shown in the above figure as To, which corresponds to a zero strain. The "Secant" thermal expansion coefficient at temperature T2 corresponds to the slope S2 of the line between a zero strain (To) and the strain at the temperature T2. In the case of a constant Thermal Expansion coefficient, the strain curve is a straight line and the coefficient corresponds to the slope of this line. One does NOT need to define a reference temperature in the case of a Constant coefficient.

Version 2006.0

Page 193

ProCAST User Manual

The Reference Temperature is defined in the lower part of the "Secant" panel (see above).

Elasto-Plastic model
For the Elasto-Plastic model, the properties described above for the Elastic model (i.e. the Young's modulus, the Poisson's ratio and the Thermal Expansion coefficients) should also be defined. In addition, one should define the Yield stress and the Hardening coefficient.

Version 2006.0

Page 194

ProCAST User Manual

The Yield stress corresponds to the stress at which plastic deformation starts. It can be temperature dependant.

The hardening coefficient corresponds to the slope of the stress-strain curve in the plastic range. Four different models of hardening are available in ProCAST :

Version 2006.0

Page 195

ProCAST User Manual

Linear hardening is defined as follows :

whereas Non-linear hardening is defined as :

Version 2006.0

Page 196

ProCAST User Manual

The Hardening can also be defined freely in an ASCII file where one can enter any "digitized" curve. More information about this possibility is in the "Digitized Hardening" section. In order to take into account for the "Kinematic" non-isotropic hardening behavior (Bauschinger effect), the Amstrong-Frederick model is available :

x is called the back stress. It corresponds to the "movement" of the center of the Yield surface. Isotropic and Kinematic models can be used either individually or together.

Elasto-ViscoPlastic model
For the Elasto-ViscoPlastic model, the properties described above for both the Elastic model (i.e. the Young's modulus, the Poisson's ratio and the Thermal

Version 2006.0

Page 197

ProCAST User Manual

Expansion coefficients) and the Elasto-Plastic model (i.e. the Young's modulus and the Hardening coefficient) should also be defined.

In order to account for the visco-plasticity, three models are available : a) Perzyna b) Norton c) Strain Hardening Creep
Perzyna model

This model allows to describe the secondary (steady-state) creep with threshold.

Version 2006.0

Page 198

ProCAST User Manual

Norton model

This model allows to describe the secondary (steady-state) creep with no threshold.

Version 2006.0

Page 199

ProCAST User Manual

As there is no threshold value, the data for the Yield Stress and the Hardening will be ignored.

Strain Hardening Creep model

This model allows to describe both the primary (strain hardening) creep and secondary (steady-state) creep regimes with a possible threshold.

Version 2006.0

Page 200

ProCAST User Manual

For the three models, one have (for the plastic model) :

or

Moreover, the value of the normalization stress ( * ) is defined by the user (usually a value of 1 is recommended, in the same unit system as the measured stresses). One should note that the value of and n will depend upon the selected value for *. One should be careful that viscoplastic models which have been set in versions prior to 2005.0 should be modified in PreCAST in order to add the normalization stress (*). Otherwise, the solver will not work. Further details on the Perzyna law can be found in : "Numerical Modelling in Materials Science and Engineering", M. Bellet, M. Rappaz and M. Deville, Spinger, 2003, pp. 306-310.

Summary

Version 2006.0

Page 201

ProCAST User Manual

With the three available models (elastic / elasto-plastic / elasto-viscoplastic), it is possible to cover the whole range of behavior occurring in a solidifying alloy. The figure below shows how one can "go" from one model to the other, by setting the appropriate temperature-dependant material properties.

At high temperature, one can have a viscoplastic behavior. In this case, setting a hardening parameter (H in case of linear hardening) to zero will "cancel" the plastic model, whereas setting the "Viscous parameter", will activate the viscoplastic model. On the other hand, at low temperature, if the hardening parameter will be set to a "high" value, the plastic model will be active, whereas one can "deactivate" the visco-plastic model by setting a "Viscous parameter" to 0 (it is allowed to set it to zero, because the "0" is considered as a "deactivation flag" and not as a "zero" value). Please note that it is advised to set the transition of the "Viscous parameter" between 0 (at low temperature) to a given value (at intermediate temperature), within a very small range of 0.001C. One should however notice that in most cases, the use of an elastic-plastic model gives very realistic results, without the burden of finding appropriate data for the visco-plastic behavior. The principles of the determination of plastic and visco-plastic properties from experimental measurements are described in the "Viscoplastic properties determination" section.

Temperature-dependant stress data

The three graphs hereafter are showing how it is recommended to define the Elastic Young's modulus, the Yield stress and the Hardening coefficient (in the case of Linear Hardening), over the whole temperature range. These

Version 2006.0

Page 202

ProCAST User Manual

recommendations will allow the best compromise between the physics and the convergence of the stress solver. Please note that the definition of these parameters in the mushy zone will influence the hot tearing indicator prediction.

The value of the Young's modulus in the mushy zone (below a fraction of solid of 20%) should be set as a constant value corresponding to the value of fs = 20%. Usually, this corresponds to values between 50 and 500 MPa. Please note that the Youngs's modulus can be automatically calculated in PreCAST as function of the chemical composition, using the thermodynamic databases.

Version 2006.0

Page 203

ProCAST User Manual

The value of the Yield stress in the mushy zone (below a fraction of solid of 50%) should be set as a constant value corresponding to 5-10 MPa. One should not set values below 5 MPa.

If no data are available, the Hardening (in case of linear hardening) can be set to about 1/20 of the Young's modulus. In all cases, the Hardening should be set to 0 MPa for fraction of solid smaller than 50%. Remark : If a tensile test is made on a sample which will be fully mushy at a given time, a non-zero value of the Hardening should be set for fraction of solid below 50%. However, this case is never occurring in usual casting processes. Concerning the Poisson's ratio, if a value of 0.5 is set, it is automatically changed in the software to 0.48. Together with these data, the corresponding Run parameters are recommended (see the "Stress Run Parameters" section for more details) : CRITFS = 0.5 CONVS = 0.01 PENALTY = 0.01 AVEPEN = 0.1 mm (for large casting, this value can be increased) SCALC = 5

Version 2006.0

Page 204

ProCAST User Manual

Digitized Hardening
Instead of defining the Hardening either as Linear or with a Powerlaw, it is possible to enter in an ASCII file digitized hardening curves (i.e. the plastic part of tensile test curves at different temperatures). To activate this mode, one should select the "Table" tab in the "Hardening" tab :

Then, the number of tables specified in the ASCII file should be set. The corresponding ASCII file should be called : *ss1.dat (e.g. : prefixss1.dat) and it should have the following structure :
STRESS_UNIT 4 CURVE 1 POINTS 2 TEMPERATURE 1 293. 0. 8.1159e+01 0.02 1.7684e+02 0.03 2.2671e+02 0.04 2.8462e+02 CURVE 2 POINTS 2 TEMPERATURE 1 823. 0. 9.838 0.01 3.89105e+01 0.05 4.73871e+01

Version 2006.0

Page 205

ProCAST User Manual

The stress units are defined first (the unit code corresponds to the one of the ProCAST d.dat file - e.g. 4 corresponds to MPa). Then, the number of the curve should be specified, followed by the number of points in the curve. The temperature at which the curve is defined (with the unit code before it - e.g. 1 corresponds to degrees Kelvin, 2 to Centigrade and 3 to Fahrenheit), followed by the curve itself (strain - stress). Please note that only the plastic part of the curve (i.e. the hardening) is defined by the tables. This means that each curve should start by a zero strain (this is mandatory) and that the corresponding stress value is the Yield stress at this temperature. Thus, the Yield stress defined in the "Yield Stress" tab in PreCAST will not be used in the case of Tables (the values will be ignored). Each curve can have a different number of points. Above the last point, the Stress is extrapolated as a constant (i.e. perfect plasticity).

Plastic and Viscoplastic properties determination

In order to determine the plastic and viscoplastic properties (o, H, , n) at different temperatures, including the transition, two methods are suggested (for the Perzyna model with linear hardening) : Tensile test method Creep test method

Version 2006.0

Page 206

ProCAST User Manual

Tensile test method

In the "tensile test method", one should perform controlled tensile tests. These tensile tests should be performed for different temperatures using various strain rates. To do so, the tensile test should be done at constant RAM speed (in order to pull at specified strain rates). The stress and strain should be recorded during these tests. This is usually obtained with a Gleeble machine. The figure below shows the results of such measurements : at low temperature (T1 - green curve), the curve is independent from the strain rate. Thus, the behavior is elastic-plastic. At high temperature (T3 - red curves), one can see that the stress level is depending upon the strain rate and that there is no hardening (viscoplasticity only), as the curve is horizontal at a given stress level. At "intermediate" temperature (T2 - blue curves), one has both plasticity (with hardening) and viscoplasticity (as the curve depends upon the strain rate).

Version 2006.0

Page 207

ProCAST User Manual

Low Temperature

At low temperature (usually T1 < Tm/3), the tensile test curve is independent from the strain rate. This means that different tests made at different RAM speeds are giving the same curve (see the green curve in the figure below). From this curve, one can get the Yield stress (o) and the Hardening coefficient (H). The value of the viscous parameter () should be set to zero in order to "disable" the viscoplasticity. The Power (n) can be set to any value as it will not be used.

Version 2006.0

Page 208

ProCAST User Manual

High Temperature

At high temperature (usually T3 > Tm/2), the tensile test curve is dependent from the strain rate and the stress level is constant (no hardening) after the transition stage (see the red curves in the figure below). As there is no hardening, the value of H should be set to zero. Therefore the threshold (y) is equal to the Yield stress o.

Version 2006.0

Page 209

ProCAST User Manual

In order to obtain the viscoplastic parameters ( and n), one should first determine the value of the Yield stress (o), which can be equal to zero. To do so, one should perform a tensile test at the lowest possible strain rate (possibly zero !) and see the value of the stress on the plateau. If this value is very low, one can consider that the Yield stress (o) is zero. In order to determine this value, one can also make loading-unloading tests, with increasing loads until plastic deformation is obtained. Then, for each measured tensile test curve, the value of (- Y)/* (which is the stress level of the plateau minus the Flow stress, Y, which is equal to o in this case, as H = 0) and d/dt should be plotted in a log-log graph. The value of and n can be deduced from this graph (see below).

Version 2006.0

Page 210

ProCAST User Manual

Intermediate Temperature

At intermediate temperature (usually Tm/3 < T3 < Tm/2), the tensile test curve is dependent from the strain rate and the stress level is not constant (hardening) after the transition stage (see the blue curves in the figure below). The hardening (H) corresponds to the slope of the curves (after the transition stage). In this model, the hardening is the same for all strain rates. In order to obtain the viscoplastic parameters ( and n), one should first determine the value of the Yield stress (o), which can be equal to zero. To do so, one should perform a tensile test at the lowest possible strain rate (possibly zero !) and see the value of the stress zero plastic strain. If this value is very low, one can consider that the Yield stress (o) is zero. In order to determine this value, one can also make loading-unloading tests, with increasing loads until plastic deformation is obtained. Then, for each measured tensile test curve, the value of (- Y)/* (which is the extrapolated stress level at zero strain minus the Flow stress, Y. In this case, Y is equal to o, as pl = 0 - because the extrapolation is performed a zero strain) and d/dt should be plotted in a log-log graph. The value of and n can be deduced from this graph (see previous figure).

Version 2006.0

Page 211

ProCAST User Manual

Creep test method

In the "creep test method", one should perform controlled creep tests (plastic deformation as a function of time for a given fixed load). The creep tests should be performed for different temperatures using various loads (stress). The figure below shows the results of such measurements : at low temperature (T1 - green curves), the curves are independent from time (plateau). Thus, the behavior is elastic-plastic. At high temperature (T3 - red curves), one can see that the plastic strain is increasing linearly with time (viscoplasticity only). At "intermediate" temperature (T2 - blue curves), one has both plasticity (with hardening) and viscoplasticity.

Version 2006.0

Page 212

ProCAST User Manual

Low Temperature

At low temperature (usually T1 < Tm/3), the following curves are measured at different loads.

Version 2006.0

Page 213

ProCAST User Manual

From these curves, one can get the Yield stress (o) and the Hardening coefficient (H). To do so, the values (shown in the graph below) of and (at the different loads) should be plotted in a - graph. A linear or Power Law Hardening can be fitted on those points.

Version 2006.0

Page 214

ProCAST User Manual

Beside the above, the value of the viscous parameter () should be set to zero in order to "disable" the viscoplasticity. The Power (n) can be set to any value as it will not be used.
High Temperature

At high temperature (usually T3 > Tm/2), the slope of each curve represents the strain rate, d/dt (see the red curves in the figure below).

Version 2006.0

Page 215

ProCAST User Manual

In order to obtain the viscoplastic parameters ( and n), one should first determine the value of the Yield stress (o), which can be equal to zero. To do so, one should perform a creep test at decreasing loads. The Yield stress is reached when there is no more plastic strain. If this value is very low, one can consider that the Yield stress (o) is zero. Then, for each measured creep curve (shown above), the value of strain rate (d/dt) should be plotted as a function of (- Y)/* (which is the stress level of the plateau minus the Flow stress, Y, which is equal to o in this case, as H = 0) in a log-log graph (see below). The value of and n can be deduced from this graph (see below).

Version 2006.0

Page 216

ProCAST User Manual

Beside the above values, as there is no hardening, the value of H should be set to zero. Therefore the threshold (y) is equal to the Yield stress o.
Intermediate Temperature

At intermediate temperature (usually Tm/3 < T3 < Tm/2), the creep test curves are dependent from both the load (stress level) and time (see the blue curves in the figure below).

Version 2006.0

Page 217

ProCAST User Manual

From these curves, one can get the Yield stress (o) and the Hardening coefficient (H) as described above in the "Low Temperature" section. To do so, the values (shown in the graph below) of and (at the different loads) should be plotted in a - graph. A linear or Power Law Hardening can be fitted on those points. In order to obtain the viscoplastic parameters ( and n), one should first determine the value of the Yield stress (o), which can be equal to zero. To do so, one should perform a creep test at decreasing loads. The Yield stress is reached when there is no more plastic strain. If this value is very low, one can consider that the Yield stress (o) is zero. Then, for each measured creep curve (see blue curves in the zoom below), the values of the INITIAL strain rate (d/dt) should be plotted as a function of (Y)/* (which is the load, i, minus the Flow stress, Y. In this case, Y is equal to o, as pl = 0 - because the extrapolation is performed at time=0, where the strain is zero) in a log-log graph (as described in the "High Temperature" section).

Version 2006.0

Page 218

ProCAST User Manual

Version 2006.0

Page 219

ProCAST User Manual

RUN PARAMETERS
This section contains the comprehensive description of all Run parameters. The ones shown in blue are either new or have been modified or further described, with respect to previous versions. The different categories of Run Parameters are organized in Tabs. The first Tab layer (1) corresponds to the different modules of ProCAST. For each modules, the Run Parameters are divided in a "Standard" tab (2), with the most often used parameters and one or two "Advanced" tabs (2) with less often used Run Parameters. The values are then defined in (3).

The values of the Run Parameters are stored in a file named "prefixp.dat" (often called the "p.dat" file). The Run Parameters can thus be changed either in the PreCAST interface, or directly "by hand" in the "p.dat" file (see example below).

Version 2006.0

Page 220

ProCAST User Manual

The default values are not anymore specified in the manual, as they are set in the "default_p.dat" file (see the "Pre-defined Run Parameters" section for more details).

General Run Parameters

The Run Parameters which are used by any module are grouped in the "General" tab. It is also possible to defined the units which will be used for the display of the results (it will not affect the results themselves).

Version 2006.0

Page 221

ProCAST User Manual

The Run Parameters of the "General/Advanced" tab are usually not changed from the default values, except in very specific cases. It corresponds mainly to the parameters of the different general solver algorithms.

Version 2006.0

Page 222

ProCAST User Manual

General
AVEPROP
specifies the method to be used in calculating the properties for each element. ProCAST will calculate the properties at each Gauss point or you may specify that the properties be calculated only at the element center and that this value will be used as an average for the whole element. This averaging reduces, somewhat, the finite element integration time required. This averaging does not apply to the specific heat or enthalpy calculations. Choose from: 0 to calculate at each point, or 1 to use the average

CGSQ
specifies the Conjugate Gradient Squared solver flag. The values specified in this parameter may be added together. This allows you to "build" a customized solver approach for your simulation.

Version 2006.0

Page 223

ProCAST User Manual

Choose from: 0 = Use the default iterative solver ( TDMA ), 1 = Use the CGSQ solver on the U momentum equation, 2 = Use the CGSQ solver on the V momentum equation, 4 = Use the CGSQ solver on the W momentum equation, 16 = Use the CGSQ solver on the energy equation, 64 = Use the CGSQ solver on the turbulence intensity equation, 128 = Use the CGSQ solver on the turbulence dissipation equation, or 512 = Use the CGSQ solver on the density equation for compressible flow

CONVTOL
specifies the convergence tolerance which will be used in conjunction with the default non-symmetric iterative solver. Enter a floating (real) value.

DIAG
specifies the diagonal preconditioning flag for the symmetric solver. Choose from: 0 = use partial Cholesky preconditioning for everything, 8 = use diagonal preconditioning for pressure, 16 = use diagonal preconditioning for energy, and 16384 = use diagonal preconditioning for radiosity

DT
specifies the initial time step size. Setting DT to zero when INILEV > 0 will cause ProCAST to use the DT at step INILEV. Enter a floating (real) value. The default is 1.0000e-03. Select the units of time from: {sec | min}

DTMAX
specifies the maximum time step size. Enter a floating (real) value. Select the units of time from: {sec | min}

DTMAXFILL
specifies the maximum time step size which will be used during the filling stage only. Once the filling is finished, the DTMAX value will be used. If DTMAXFILL is not defined (i.e. it is set to zero), the value of DTMAX will be used for the whole calculation. Enter a floating (real) value. Select the units of time from: {sec | min}

INILEV
specifies the initial time level. When an analysis is first started, INILEV should be equal to zero. When you are resuming an analysis, INILEV should be set to the time step from which you would like to continue. Note: You must have results for that time step.

Version 2006.0

Page 224

ProCAST User Manual

Enter an integer value.

LUFAC
specifies the preconditioning parameter for the CGSQ solver. This parameter may speed-up calculations of large models Choose from: 0 to use diagonal preconditioning, or 1 to use partial LU factorization preconditioning

NEWTONR
turns on the NEWTON Raphson technique for the energy equation. Choose from: 0 to turn off the Newton Raphson technique, 1 to turn the Newton Raphson technique on, or 2 to turn on the Newton Raphson technique and use bsplines The default is 0. Option 2 results in using b-splines instead of linear line segments in the representation of the thermal properties. It is suggested that all thermal input data be smoothed before attempting to use b-splines. Enter an integer value.

NPRFR
specifies the printout frequency. This controls the time step interval at which results are output to the prefixp.out file. Enter an integer value.

NRSTAR
specifies the number of allowable restarts before the entire run is abandoned. A restart occurs when the maximum number of corrections is reached. If too many restarts are taking place, it could indicate problems with the model setup. Enter an integer value.

NSTEP
specifies the number of time steps to take in the current run and is used in conjunction with TFINAL. ProCAST will terminate the run when it reaches this limit or the TFINAL value, whichever occurs first. Enter an integer value.

PRNLEV
specifies the level of nodal results to be printed out. The values specified in this parameter may be added together. This allows you to collect combinations of nodal information in a single run. Choose from: 0 = no printout, 1 = nodal velocities, 8 = nodal pressures, 16 = nodal temperatures,

Version 2006.0

Page 225

ProCAST User Manual

64 = nodal turbulence intensities, 128 = nodal turbulence dissipation rates, 1024 = nodal displacements, 8192 = surface heat fluxes, and 32768 = nodal magnetic potentials

SDEBUG
specifies the level of solution debugging messages to be captured. These messages are written to the p.out file. Choose from: 0 to capture no solution debugging messages, or 1 to obtain information concerning, solver performance, time step control, and the free surface model

TENDFILL
specifies the delay after the end of the filling at which to terminate a ProCAST analysis. If this parameter is zero, this parameter will not be active. If one wants to stop the calculation right after the end of the filling, one should set a small value, different from 0. Enter a floating (real) value. Select the units of time from: {sec | min}

TFINAL
specifies the simulated time at which to terminate a ProCAST analysis. If this parameter is zero, the run will be stopped by the TSTOP or NSTEP parameter. If NSTEP and TSTOP and TFINAL parameters are set, the simulation will be terminated based upon which parameter is reached first. Enter a floating (real) value. Select the units of time from: {sec | min}

TSTOP
specifies the temperature at which to terminate a ProCAST analysis (i.e. when all the temperatures at all nodes are below the TSTOP temperature). If this parameter is zero, the run will be stopped by the TFINAL or NSTEP parameter. If NSTEP and TSTOP and TFINAL parameters are set, the simulation will be terminated based upon which parameter is reached first. Enter a floating (real) value. Select the units of Temperature from: {C | K | F}

TMODR
specifies the time step modification factor for restarts. If MAXCOR correction steps are taken without convergence, the time step is multiplied by TMODR. Therefore, this number should be less than 1. Enter a floating (real) value.

TMODS

Version 2006.0

Page 226

ProCAST User Manual

specifies the time step modification factor for normal stepping. If the number of correction steps is less than or equal to NCORL, the subsequent time step is multiplied by TMODS. If the number of correction steps is greater than or equal to NCORU, the subsequent time step is divided by TMODS. Enter a floating (real) value.

Units
PUNITS
specifies the pressure units to be used in the outputs. Choose from: {N/m**2 | Pa | KPa | MPa | bar | dyne/cm**2 | atm | psia | Ksi | lb/ft**2}

QUNITS
specifies the heat flux units to be used in the outputs. Choose from: { W/m**2 | cal/cm**2/sec | cal/mm**2/sec | Btu/ft**2/sec | Btu/in**2/sec | cal/cm**2/min | cal/mm**2/min | Btu/ft**2/min | Btu/in**2/min}

TCUNITS
specifies the thermocouple units to be used in the outputs and is only used for inverse modeling. Choose from: {C | F | R | K}

TUNITS
specifies the temperature units to be used in the outputs. Choose from: {C | F | R | K}

VUNITS
specifies the velocity units to be used in the outputs. Choose from: {m/sec | cm/sec | mm/sec | ft/sec | in /sec | m/min | cm/min | mm/min | ft/min | in/min} The choice of the units does not have any effect on the calculated results.

Thermal Run Parameters

The Thermal model activation, the Temperature results storage frequency, as well as the Porosity model and parameters are the main parameters to be defined in this tab.

Version 2006.0

Page 227

ProCAST User Manual

Version 2006.0

Page 228

ProCAST User Manual

Thermal
CINIT
Reserved for future use.

CLUMP
specifies the capacitance matrix lumping factor. Enter: 0 to use consistent matrix, or 1 to use diagonal matrix

CONVT
specifies the convergence criterion for temperature. A value of around one degree is generally appropriate. Values larger than the mushy (liquidus--solidus) zone range are not recommended. Enter a floating (real) value. Select the units of temperature from: {C | F | R | K}

Version 2006.0

Page 229

ProCAST User Manual

CRELAX
specifies the heat capacity relaxation parameter. Enter a floating (real) value.

FEEDLEN
"Feeding length" (distance), at which macroporosity can occur, beyond the MACROFS isosurface. Enter a floating (real) value.

GATEFEED
allows to specify whether liquid can be fed at the ingate or not. In the case of injection (i.e. hpdc or lpdc), the shrinkage at the gate is compensated by the liquid pushed by the piston (for hpdc). On the other hand, in gravity casting, there is no feeding at the top of the risers and piping will occur. GATEFEED=1 tells the software to activate the feeding at the ingate, thus leading to no piping at this location. GATEFEED=1 will automatically activate the "Active feeding" where inlet velocities or pressures are set (see the "Active Feeding" section for more details). Enter an integer value.

GATEFS
GATEFS allows to control the critical fraction of solid above which there is no more feeding during the third stage pressure in HPDC (see the "Active Feeding" section for more details). This Run Parameter has to be added or modified manually in the p.dat file (it does not appear in PreCAST). A higher value of GATEFS will involve less porosity. Enter a value between zero and one. The default value is 0.95.

GATENODE
GATENODE is an alternative to GATEFEED. When a gate feeding should be applied, but there is no external surface of the model where an inlet surface or a pressure can be applied. This is the case for instance when a shot piston is modeled. In such case, one can define a location in the volume where this gate feeding is applied (typically in the middle of the biscuit). In order to define this location, the corresponding node number should be specified. In such case, no pressure or inlet velocity BC needs to be specified. To find the desired node number, it is advised to use the "pick" selection in ViewCAST (Run DataCAST first) (see the "Active Feeding" section for more details). Enter an integer value.

LINSRC
specifies the source term linearization parameter for micromodels. This parameter may be used in conjunction with micromodels that control the evolution of the solid fraction and thus the release of latent heat. The default value of zero indicates that the heat generation will only appear in the right hand side source

Version 2006.0

Page 230

ProCAST User Manual

term. A value of one will give some contribution to the diagonal terms of the left hand side matrix. This improves numerical stability, but does require that the LHS be factored, which would normally happen anyway. Enter: 0 = no linearization, or 1 = for linearization of the source term

MACROFS
Parameter for the macroporosity calculation. It sets the limiting fraction of solid between the macroporosity and the microporosity formation. The value should be set between 0 and 1. Enter a floating (real) value.

MOBILE
Enter a floating (real) value.

MOLDRIG
The rigidity of the mold has an influence on the amount of porosity in the case of expanding alloys. If the mold is totally rigid, the casting can not expand and thus the alloy expansion will be "available" for the "refill" of the existing porosity. On the other hand, if the mold is very soft (or weak), the casting will expand and thus there will be no "refill" of the porosity (of course, the reality is more complex as the solid shell is thick enough it will act as a "rigid" mold, even if the sand mold is weak. In order to take the mold rigidity into account, the Run Parameter MOLDRIG is introduced (see the "SGI Porosity model" section for more details). MOLDRIG should be defined by a value between 0 and 1. All the net expansion is multiplied by MOLDRIG. Thus, with MOLDRIG=1, corresponding to a rigid mole, the expansion will be fully accounted. On the other hand, no expansion will be taken into account if MOLDRIG=0. The expansion will be compensated by the mold movement because the mold is too weak to hold the expansion in this case. For real situations, the value of MOLDRIG should be set somewhere between 0 and 1 depending upon the casting processes. MOLDRIG should be added "manually" in the p.dat file by the user. The default value is 1.

PIPEFS
Parameter for the piping calculation (during a porosity calculation). It sets the limiting fraction of solid for piping formation. The value should be set between 0 and 1. In order to disable the porosity calculation, one should set POROS = 0. However, this may still lead to some piping calculation. In order to disable also the piping, one should set in addition PIPEFS = 0. For stress calculations, PIPEFS must be set to 0. This is needed in order to prevent the unexepected influence of the pipe on the stress calculation. One should note that if in the physical sitution, piping must occur, this will be taken into account in the calculation as "macropores" when PIPEFS = 0.

Version 2006.0

Page 231

ProCAST User Manual

Enter a floating (real) value.

POROS
specifies the porosity calculations to be performed (see the porosity models description). Choose from: 0 for no porosity calculation, 1 - most advanced porosity model 4 - evolution of the POROS=8 model, which allows to handle multiple free surfaces 8 - old porosity model from version 3.2.0 In order to disable the porosity calculation, one should set POROS = 0. However, this may still lead to some piping calculation. In order to disable also the piping, one should set in addition PIPEFS = 0.

QFREQ
specifies the time step interval for writing heat flux data to the unformatted results file. This parameter can be used to reduce the size of the prefixq.unf file. Heat flux results may not be of interest to everyone, so it may be desirable to minimize the size of this file. Enter an integer value.

TFREQ
specifies the time step interval for writing temperature data to the unformatted results file. This parameter can be used to reduce the size of the prefixt.unf file, which can become quite large for problems with many nodes and time steps. Note that it is only possible to restart a run from one of the time steps that was written out. Enter an integer value.

THERMAL
specifies the thermal analysis to be performed. Choose from: 0 for no thermal analysis. Solve flow equations alone, 1 to perform thermal analysis, using temperature as the primary variable, or 2 to perform thermal analysis, using enthalpy as the primary variable

TRELAX
specifies the temperature relaxation parameter. This is used for computing the initial guess for the temperature field in the predictor step. TRELAX should be greater than or equal to zero and less than or equal to one. Enter a floating (real) value.

USERHO
specifies the how the density in the mushy zone is calculated. Choose from:

Version 2006.0

Page 232

ProCAST User Manual

0 - automatic density calculation, by extrapolation of the densities at liquidus and solidus, weighted by the fraction of solid. 1 - the density table, which is defined in the database is used.

Cycling Run Parameters

When cycling is modeled, one should define the corresponding Run Parameters in this tab (number of cycle and cycling time). In addition, when "Die Combo" interface definitions are used, additional parameters should be set.

NCYCLE
specifies the number of casting cycles to be simulated in a continuous mode. This parameter is used along with TCYCLE. Both NCYCLE and TCYCLE must be set. This parameter is typically used in die casting and permanent mold problems. Enter an integer value.

Version 2006.0

Page 233

ProCAST User Manual

TCYCLE
specifies the time of casting cycle to be simulated in a continuous mode (duration of a cycle). This parameter is used along with NCYCLE. Both NCYCLE and TCYCLE must be set. Enter a floating (real) value. Select the units of time from: {sec | min}

TOPEN
specifies the time at which the mold opens, during one casting cycle. This is used in conjunction with the "Die Combo" interface definition. Enter a floating (real) value. Select the units of time from: {sec | min}

TEJECT
specifies the time at which the part is ejected from the mold, during one casting cycle. This is used in conjunction with the "Die Combo" interface definition. Enter a floating (real) value. Select the units of time from: {sec | min}

TBSPRAY
specifies the time of the beginning of the spray sequence, during one casting cycle. This is used in conjunction with the "Die Combo" interface definition. Enter a floating (real) value. Select the units of time from: {sec | min}

TESPRAY
specifies the time of the end of the spray sequence, during one casting cycle. This is used in conjunction with the "Die Combo" interface definition. Enter a floating (real) value. Select the units of time from: {sec | min}

TCLOSE
specifies the time at which the mold closes (before the start of the next filling), during one casting cycle. This is used in conjunction with the "Die Combo" interface definition. Enter a floating (real) value. Select the units of time from: {sec | min}

Radiation Run Parameters

When a calculation with Radiation, including View Factors is run, the following parameters should be defined. If the "View factor OFF" option is selected, it is not needed to define these Run parameters.

Version 2006.0

Page 234

ProCAST User Manual

Version 2006.0

Page 235

ProCAST User Manual

ANGTOL
specifies the angle tolerance to be used with VFLIM. Radiation faces which are grouped using VFLIM tolerance are further differentiated by their solid angle. Enter a floating (real) value.

ENCLID
specifies an enclosure identification number. This parameter is used in combination with VFDISP for updating view factors by a displacement interval. ENCLID indicates which enclosure set is to be tracked, in case all the enclosure elements are not moving at the same rate. Enter an integer value.

EPTOL
specifies the emissive power tolerance to be used with VFLIM. Radiation faces which are grouped using VFLIM tolerance are further differentiated by their solid angle. Enter a floating (real) value.

Version 2006.0

Page 236

ProCAST User Manual

RDEBUG
specifies the user debug parameter for printing detailed view factor information. Various combinations of these files may be obtained by adding together these numbers. For example, RDEBUG = 7 gives all three files. Note that these files can be quite large, especially the prefix.vf. Enter an integer value based upon the following: 1 for face to face view factors after symmetrization, in the prefix.vf file, 2 for face to group view factors after symmetrization, in the prepfix.view file (necessary to see FACE TO GROUP in ViewCAST), or 4 for row sum errors before symmetrization, in the prefix.serr file (necessary to see ROW SUM ERRORS in ViewCAST

RFREQ
specifies the radiation update frequency. This provides a mechanism for recomputing the radiosities at some time step interval other than one. This is particularly useful if you are performing a filling transient along with the view factor radiation model. In this case, the time step size may be small due to the filling whereas the mold temperature may not be changing very rapidly. You can save some computational time by recomputing the radiosities at every tenth step, for example. Enter an integer value.

TRI2QUAD
specifies the option to group or not triangles into quadrangles for the radiation view factor calculation. When TRI2QUAD is set to 1, adjacent triangles are grouped in order to obtain quadrangles (providing the angle between the triangles is not too large). This has the effect of reducing the number of radiative face (by about 50%) which is cutting down the CPU time significantly (by about 75%).

VFDISP
specifies the displacement interval for updating view factors in the radiation model if there are moving relative surfaces. This is used in conjunction with ENCLID and will be used in preference to VFTIME if both are specified. Enter a floating (real) value. Choose the units of length from: {m | cm | mm | ft | in}. The default is m.

VFLIM
specifies the view factor limit. This parameter is used to agglomerate faces in the view factor calculations. This reduces the size of the radiosity matrix and speeds up the radiation calculations.

VFLIM
can be set to a fraction between zero and one. If one face occupies less than this fraction of the total view space, as seen from another face, the first face is combined with some others. A value of 0.01 is a good starting point. Enter a floating (real) value.

Version 2006.0

Page 237

ProCAST User Manual

VFTIME
specifies the time interval for updating view factors in the radiation model if there are moving relative surfaces. Enter a floating (real) value. Choose the units of time from: {sec | min}. The default is sec.

Flow Run Parameters

When fluid flow and/or filling models are activated, the "Flow" Run Parameters should be defined. The main parameters are present in the "Standard" tab. In the "Advanced 1" tab, one should mainly define the WSHEAR and the WALLF parameters. In most cases, all the other parameters can remain as the default values.

Version 2006.0

Page 238

ProCAST User Manual

Version 2006.0

Page 239

ProCAST User Manual

ADVECTW
specifies the weighting of advection velocities and controls the degree of nonlinearity of the momentum equations. ADVECTW can take on values between zero and one. Velocities at the last time step are used as the advecting velocities if a value of zero is used. Velocities at the current time step are used as the advecting velocities if a value of one is used. Numerical experience has shown that the accuracy of natural circulation flows can be enhanced by using a factor of 0.5. For most filling analyses, a value of zero works fine and requires much less computational time. Enter a floating (real) value.

COMPRES
specifies whether this is an incompressible flow problem or a compressible flow problem. Choose from: 0 to specify an incompressible flow problem, or 1 to specify a compressible flow problem

Version 2006.0

Page 240

ProCAST User Manual

CONVV
specifies the convergence criterion for velocity. The value given here is a fraction of the maximum velocity calculated at each step. Generally, .05 or 5% is appropriate. Enter a floating (real) value.

COUPLED
specifies whether the energy and fluid solutions should be coupled or decoupled within a time step. When the analysis is decoupled, the momentum and pressure equations are solved repeatedly until convergence. Subsequently, the energy equation is solved until convergence, assuming the flow field is fixed. With a coupled analysis, the energy equation is solved in the same loop with momentum and pressure. Both the momentum and temperature convergence criteria have to be met to terminate the loop. This method is more accurate, but usually takes more computational time. Choose from: 0 to decouple energy and fluid solutions within a time step, or 1 to fully couple energy and fluid solutions within a timestep

COURANT
specifies the courant limit on time step size. This parameter is only used for filling problems. If COURANT is set to 1.0, the time step will be adjusted so that the fluid will advance no more than one element length. This is a fairly severe limit on time step size, but will give the most accurate results for filling transients. Acceptable results can usually be obtained with values around 100. Enter a floating (real) value.

DETACHTOP
allows to have a better treatment of the detachment of liquid which is in contact with horizontal top walls (like "roofs"). A value of 1 allows a better detachment of such liquid regions, however it takes some more CPU time. This Run parameter should be added/changed manually in the p.dat file.

EDGE
This Run parameter is not anymore available.

ENDFILL
Sometimes, the user may not be interested by the filling of the last percents (e.g. the end of the filling of a riser or an overflow). If ENDFILL = 0.98, once 98% will be reached, the remaining 2% will be filled in one timestep.

FFREQ
specifies the flow update frequency. This provides a mechanism for re-computing the velocities at some time step interval other than one. This might come into play if you were solving a conjugate heat transfer problem where the velocity field is

Version 2006.0

Page 241

ProCAST User Manual

changing on a longer time scale than the temperatures. This option is not appropriate for free surface problems. Enter an integer value.

FLOW
controls the use of fluid equations. Choose from: 0 do not solve fluid equations, 1 to solve fluid equations, 3 to solve fluid equations during filling, but switch over to thermal only analysis when the LVSURF fill limit is reached and NCYCLE = 1, 9 to solve fluid equations during filling, but switch over to thermal only analysis when the LVSURF fill limit is reached and NCYCLE > 1 The default is 0 if there are no "Casting" materials. If "Casting" materials exist, the default is 1.

FLOWDEL
specifies the delay time between the end of fill and a switch to a thermal only, in the case of FLOW = 3 simulation. This option is used in conjunction with velocity boundary conditions with active fill limits. The time delay corresponds to the time for the fluid to completely settle down in the casting before the thermal only phase begins. Enter a floating (real) value. Select the units of time from: {sec | min} The default is sec.

FREESF
specifies the free surface model number to be used. Choose from: 0 = no free surface model activated 1 = use the momentum dominated movement of free surface, rapid filling model, 2 = use the gravity dominated movement of free surface, slow filling model.

FREESFOPT
The filling algorithm (in the case of FREESF = 1) was significantly improved in version 2006.0. From now on, three filling algorithms are available : FREESFOPT = 0 : it corresponds to the filling algorithm of version 2005.0. It is less precise than the two other algorithms, however it is the most robust algorithm. FREESFOPT = 1 and 2 corresponds to two improved filling algorithms. The difference between these two algorithm is the "numerical balance" between the mass conservation contribution to the free surface and the momentum contribution. With FREESFOPT = 1, the mass conservation contribution is more important than the momentum contribution. This algorithm is more robust, but may lead to slightly less precise results. With FREESFOPT = 2, the momentum contribution is predominant over the mass conservation contribution. This model is supposed to be the most precise, however it is more sensitive to the quality of the mesh. In order to have good results with FREESFOPT = 2, the mesh should be well adapted to the local free surface (e.g. one should have enough nodes in the

Version 2006.0

Page 242

ProCAST User Manual

stream of the liquid in the case of a thin jet of liquid). This means that the FREESFOPT = 2 algorithm is less robust as the 1 (or the 0). As a conclusion, it is strongly advised to use FREESFOPT = 1 in most cases. If more precision is required, the mesh should be refined and well tuned so that FREESFOPT = 2 could be used. The FREESFOPT = 0 option is provided for backwards compatibility purposes, however, it is not advised to be used. The default value is 1.

GAS
specifies whether or not to consider the trapped gas effects. If the option to consider trapped gas effects is chosen, trapped gas effects will be considered even when the model contains no vents, gas injection, or gas diffusion through the mold. When features normally found in a gas problem (vents, injection, or gas diffusion through the mold ) are present in a model, GAS will be set automatically. Choose from: 0 to not consider trapped gas effects, or 1 to consider trapped gas effects

HEAD_ON
specifies the approach to be used when calculating gravitational term in the momentum equation for flow problems without free surfaces. Choose from: 0 = calculate as rho - rho_ref, or 1 = calculate as rho * g

HIVISC
specifies different solution methods for viscosity in the flow problem. Choose from: 0 = normal flow problem, 1= high viscous flow problem. To be used when the Reynolds number is less the one. This method only works for viscosity less than 104 poise. In this case, the advection terms are neglected, symmetric solvers are employed on the momentum equations, and large degrees of pressure relaxation are utilized, or 2 = very high viscous flow problem. To be used when the Reynolds number is less the one. This method is always preferred. In this case, the advection terms are neglected and momentum effect on implicitly included within a Poisson pressure equation. This option usually allows for much larger time steps than HIVISC = 1

LVSURF
provides a way to switch from the filling transient to a mode where advection is due to buoyancy and shrinkage. LVSURF turns all inlets off. It is assumed thereafter that the free surface is perpendicular to the gravity vector. This allows the time step to increase significantly. The number represents the fraction of the total casting volume which is to be filled before changing modes. Enter a floating (real) value.

Version 2006.0

Page 243

ProCAST User Manual

MLUMP
specifies the mass matrix lumping factor. Choose from: 0.0 to use a consistent matrix, or 1.0 to use a diagonal matrix

NNEWTON
specifies whether the flow is newtonian or non-newtonian. Choose from: 0 to indicate Newtonian flow, or 1 to indicate non-newtonian flow, where viscosity is a function of shear rate

PENETRATE
Flag to activate the algorithm of inter-penetrating meshes (with a value of 1). This is used especially for the modeling of a shot piston in hpdc. Enter an integer value.

PFREQ
specifies the "Particle tracing" launch frequency in the solver. Particles are launched at each node of the inlet (defined by a velocity BC, an inlet pressure BC or an Inlet BC), every PFREQ steps. A value of 50 is recommended (see "Display parameters" for the description of Particle tracing) Enter an integer value.

PINLET
specifies a pressure driven inflow. Setting PINLET to 1 indicates that all the pressure boundary conditions are also inflow boundary conditions. Use of this option allows one to avoid using thin filled regions at the inlets of pressure driven problems. It allows for filling of metal without having an initial layer of fluid. Enter an integer value of 0 (off) or 1 (on).

PLIMIT
specifies the pressure cutoff limit. You can use this parameter to turn off an inlet velocity when the back pressure exceeds the given value. This is useful particularly in cases where cold shuts are occurring. Otherwise, the program will keep trying to force more mass into the fluid region, even though there is no place for it to go, and the pressure will continue to rise. Enter a floating (real) value. Choose the pressure units from: {N/m**2 | Pa | KPa | MPa | bar | dyne/cm**2 | atm | psia | Ksi | lb/ft**2}

PREF
specifies the pressure which is to be subtracted from any boundary condition pressure in order to convert an absolute pressure into a gauge pressure. This parameter comes into play when: (1) there is trapped gas, (2) a pressure boundary condition drives the flow, (3) there are vents, and/or (4) there is gas injected. For

Version 2006.0

Page 244

ProCAST User Manual

example, if the pressure boundary condition drives the flow at a gauge of 1 atmosphere, the boundary condition is set to 2 atm. PREF should be set to 1 atm. Enter a floating (real) value. Choose the pressure units from: {N/m**2 | Pa | KPa | MPa | bar | dyne/cm**2 | atm | psia | Ksi | lb/ft**2}

PRELAX
specifies the pressure relaxation factor. PRELAX, to have an effect, should be greater than zero and less than one. If it is left to the default value of one, ProCAST will automatically compute an appropriate relaxation factor. Enter a floating (real) value.

RELVEL
For centrifugal casting, the fluid flow should be solved in a Relative velocity reference frame (i.e. in a rotating velocity reference frame). Thus, for centrifugal casting cases, one should set RELVEL to 1. This Run parameter should be added manually in the p.dat file (it does not appear in PreCAST). 0 = standard case - no centrifugal (default value) 1 = activation of the relative velocity reference frame - to be used for centrifugal casting. Enter an integer value.

TILT
For Tilt pouring problems, one should activate the "Tilt" mode, by setting TILT to 1. This improves the appearance of undesired sticking in the pouring cup. This Run parameter should be added manually in the p.dat file (it does not appear in PreCAST). 0 = standard filling mode (default value) 1 = activation of the Tilt mode Enter an integer value.

TPROF
This parameter indicates that a thermal boundary layer profile is used at the wall for the energy equation with advection. This has been found to reduce false diffusion errors. Choose from: 0 = do not use boundary layer profile, or 1 = use boundary layer profile Enter an integer value.

TSOFF
This parameter specifies the time at which to switch off the flow solution. For example, TSOFF 1 42, indicates that the flow solution will be turned off 42 seconds into the simulation. If a cyclic analysis is being performed, then the flow solution will be turned off 42 seconds into each cycle. Choose from: 0 = turns this option off, or

Version 2006.0

Page 245

ProCAST User Manual

a real value sets the time Enter a floating (real) value. Choose the time units from: {sec | min}.

VFREQ
specifies the time step interval for writing velocity and pressure results to the unformatted files. This parameter can be used to reduce the size of these results files, which can become quite large for problems with many nodes and time steps. Note that it is only possible to restart a run from one of the time steps that was written. Only the steps that are written can be viewed with post-processing. Enter an integer value.

VPROF
This parameter indicates that a flow boundary layer profile is used at the wall for the momentum equation with advection. This has been found to reduce false diffusion errors, although it is not very useful with WSHEAR=2. Choose from: 0 = do not use boundary layer profile, or 1 = use boundary layer profile Enter an integer value.

WALLF
specifies the wall slip behavior. The default value is 0.9. A value of 0.98 correspond to more slip along the wall, whereas a value of 0.8 will act as if the mold surface is rougher (more friction). It is advised to use a value of 0.8 for sand gravity casting, a value of 0.9 for gravity die casting and a value of 0.98 for high pressure die casting (HPDC). For LPDC, this value is not used, as WSHEAR=0 must be used (in this case the WALLF algorithm is not activated).

WSHEAR
specifies whether or not the wall shear formulation will be used. The wall shear formulation will convert no-slip boundary conditions into wall traction conditions. Choose from: 0 to indicate that wall shear formulation will not be used, or 2 to indicate wall shear formulation will be used. The use of the Wall shear formulation allows to have non-zero velocities at the mold wall, which is more representative of the reality (slip of the liquid along walls). For HPDC and Gravity casting, a value of WSHEAR = 2 should be used. For LPDC, a value of WSHEAR = 0 must be used. The WSHEAR = 1 option is not anymore supported.

Turbulence Run Parameters

Version 2006.0

Page 246

ProCAST User Manual

CMU
specifies the proportionality constant used in the turbulent viscosity equation. Enter a floating (real) value.

CONE
specifies the proportionality constant used in the production of turbulent energy dissipation. Enter a floating (real) value.

CTWO
specifies the proportionality constant used in the destruction of turbulent energy dissipation. Enter a floating (real) value.

KAPPA
specifies the Von Karman's constant, usually taken as 0.4 Enter a floating (real) value.

Version 2006.0

Page 247

ProCAST User Manual

SIGMAE
specifies the diffusivity modifier used in the turbulent energy dissipation transport equation. Enter a floating (real) value.

SIGMAK
specifies the diffusivity modifier used in the turbulent kinetic energy transport equation. Enter a floating (real) value.

TBRELAX
specifies the turbulence relaxation parameter. Enter a floating (real) value.

TURB
specifies whether the turbulent flow model is turned on or off. A model started with TURB = 1 can be restarted at a later time with TURB = 0. This allows laminar conditions to be considered during mushy or natural circulation flows. Once TURB has been set to zero, the turbulence model can not be restarted at a later time. Setting TURB to one for a flow problem which has no turbulence boundary conditions assigned is OK; the software will automatically define them. Enter: 0 to turn the turbulent flow model off, or 1 to turn the turbulent flow model on

Stress Run Parameters

Version 2006.0

Page 248

ProCAST User Manual

Version 2006.0

Page 249

ProCAST User Manual

AVEPEN
AVEPEN corresponds to the "Average Penetration". This corresponds to the maximum average penetration which is allowed during the calculation. During the calculation, the PENALTY is automatically changed in order that the penetration is not larger than AVEPEN. The goal is to have the lowest possible PENALTY number to speed-up convergence, and AVEPEN allows to set the upper limit. The default value of AVEPEN is 0.1 mm. For large casting, it is advised to increase the value of AVEPEN for faster convergence.

CRACK
The cracking indicator model is activated with CRACK = 1 or 3. By default, the value of CRACK is set to zero. This Run parameter is not available in the Run Parameters menu and it should be added manually in the p.dat file. CRACK = 1 activates the cracking model, without feedback on the stress properties

Version 2006.0

Page 250

ProCAST User Manual

CRACK = 3 activates the cracking model, with a feedback on the stress properties. In this case, the hardening properties are changed, according to the cracking model.

CRITFS
This corresponds to the critical fraction of solid where the stress calculation starts. By default, it is set to 0.5 (i.e. 50% fraction of solid). This critical value is also used for the computation of the hot tearing.

CONVS
specifies the convergence criterion for the stress calculation. Enter a floating (real) value.

GAPMOD
specifies the treatment of the interface heat transfer coefficient. With a value of 1, the interface heat transfer coefficient is automatically modified to account for the air gap formation (an additional thermal resistance is computed as a function of the gap width, taking into account conduction and radiation through the air or vacuum). With a value of 0, the interface heat transfer coefficient which is defined in PreCAST will not be modified during the calculation.

LOADSCL
When a load is applied in non-linear problem, it has to be applied incrementaly (within a timestep). LOADSCL is the number of increments for this loading. The higher the value of LOADSCL, the more accurate is the result (and the higher is the CPU time). This has to be used mainly for structure analysis type of problems (e.g. tensile test). For usual casting, as the loading is progressive (due to gradual temperature changes), it is not necessary to apply such increments. The default value of LOADSCL is 1. This value has to be added manually in the p.dat file.

PENALTY
PENALTY controls the level of "penetration" allowed by the "contact algorithm". As the displacements are computed numerically at interfaces, there is always some penetration between two bodies (as long as they are touching). High values of PENALTY means that the allowed penetration is very small and this leads to a more difficult convergence of the algorithm. Small values of the PENALTY allows to have more penetration (which means a more easy convergence). The default value for Penalty is 1, but for thin sections, it could be advised to set it to 0.01. During the calculation, the PENALTY is automatically changed in order to optimize the CPU time and in order to be within the AVEPEN limit. When a calculation is restarted, it is advised to set the PENALTY to zero. In this case, the last PENALTY which was used in the calculation will be automatically set. If a non-zero value is defined, it will be used for the restart step. One should note that if the PENALTY was decreased during the calculation, to leave the default value may lead to a very long convergence at the restart step. Thus, it is

Version 2006.0

Page 251

ProCAST User Manual

strongly advised to set it to zero for a restart. The PENALTY at each stress timestep is indicated in the p.out file.

SCALC
specifies the time step interval the stress calculation is performed. It is thus possible to perform the stress calculation only every 10 thermal steps. SCALC and SFREQ are independent values and one will not affect the other. Enter an integer value.

SFREQ
specifies the time step interval for writing stress results to the unformatted files. This parameter can be used to reduce the size of these files, which can become quite large for problems with many nodes and time steps. Note that it is only possible to restart a run from one of the time steps that was written. SCALC and SFREQ are independent values and one will not affect the other. One should however be careful that SFREQ is a multiple of SCALC. Enter an integer value.

STRESS
specifies whether the stress calculation is turned on or off. Enter: 0 to turn the stress calculation off, or 1 to turn the stress calculation on

VACUUM
When GAPMOD = 1, the interface heat transfer coefficient depends upon the gap width. With VACUUM = 0, air conduction is taken into account (i.e. a heat resistance corresponding to heat conduction through air is taken into account). With VACUUM = 1, no heat conduction is taken into account. In both cases, radiative transfer through the interface is taken into account (when there is a gap). This Run parameter should be added manually in the p.dat file.

Micro Run Parameters

Version 2006.0

Page 252

ProCAST User Manual

EQNMAX
First nucleation parameter of the dendritic primary phase. Maximum density of nuclei of the Gaussian distribution.

EQSTD
Second nucleation parameter of the dendritic primary phase. Standard deviation of the Gaussian distribution.

EQUNDER
Third nucleation parameter of the dendritic primary phase. Average undercooling of the Gaussian distribution.

EUNUCL
First nucleation parameter of the eutectic phase. Nucleation factor.

EUPOWER
Second nucleation parameter of the eutectic phase. Nucleation exponent.

Version 2006.0

Page 253

ProCAST User Manual

EUGROW
Eutectic growth kinetics constant.

MICRO
Activation of the microstructure module. The module is activated with a value of 1. See the "Microstructures" chapter for more details about the models. It is possible to select the default values for a given system with the "Select Default Values" button.

The default values correspond to the following table :

Version 2006.0

Page 254

ProCAST User Manual

Pre-defined Run Parameters

ProCAST allows to define "Pre-defined sets" of Run Parameters in the Preferences.

Version 2006.0

Page 255

ProCAST User Manual

When one clicks on the "Select Pre-defined Set" button, a list of the available sets are displayed. Then, the user has to select the desired one. In the case below, six sets are available, among ten possible sets.

Version 2006.0

Page 256

ProCAST User Manual

The values contained in the Pre-defined sets correspond to the Recommended Run Parameters for each process. It is thus strongly advised to use these pre-defined sets or to customize them according to your specific processes (see below). The configuration of the Pre-defined sets can be easily performed by the user. One should just save the a "d.dat" file, with the desired Run Parameters in the "dat/pref" directory of the installation, under the name "predefined_x_p.dat", where x is a number between 1 and 10 (see below).

The label which appears in the Run Parameter window above should be included in the predefined file as follows (TITLE line) :

Moreover, in the "dat/pref" directory, there is also a file named "default_p.dat". This file contains the "default" values of the Run Parameters which will appear when a new case is created. The user can change the content of this file (which has exactly the same structure as a normal p.dat file) at his convenience.

Run Parameters Recommendations

ProCAST provides the access to many Run parameters, in order to allow the treatment of all kind of situations.

Version 2006.0

Page 257

ProCAST User Manual

However, for an everyday use, only a few Run parameters have to be set or modified. This section is presenting the most "popular" Run Parameters that should be set, with proposed values, for each main family of processes. These parameters recommendations may be slightly different from previous versions, as the solver algorithms have been modified. These recommended Run parameters correspond to the one which are pre-defined in the "Pre-defined Run Parameters" window. It is thus advised to activate the "pre-defined" set corresponding to the process and then to set the appropriate stopping criteria.

For all processes

Stopping criteria

(it is advisable to set a stopping criterion in order to limit the CPU time and avoid unnecessary storage of results) TFINAL TSTOP
Porosity

POROS = 1 (this model is now recommended for all processes) MACROFS = 0.7 FEEDLEN = X (The value FEEDLEN depends upon the size of the mushy zone and thus, the size of the casting. A value ranging from a few millimeters to a few centimeters is recommended. This should be calibrated with experiments. A value of 0 is not advised as this will produce a uniform microporosity throughout the part, beside the macroshrinkage)

Gravity casting
Timestep handling

DT = 1e-3 DTMAXFILL = 1e-1 DTMAX = 0.5 - 5 (depending upon the size of the model and thus the solidification time)
Porosity

PIPEFS = 0.3 GATEFEED = 0

Filling

WSHEAR = 2 FREESFOPT = 1 WALLF = 0.8 LVSURF = 0.98

High pressure die casting (HPDC)

Timestep handling

Version 2006.0

Page 258

ProCAST User Manual

DT = 1e-6 to 1e-4 (it depends upon the initial velocity of the first stage) DTMAXFILL = 1e-2 DTMAX = 0.2 - 1 (depending upon the size of the model and thus the solidification time)
Porosity

PIPEFS = 0.0 GATEFEED = 1

Filling

WSHEAR = 2 FREESFOPT = 1 WALLF = 0.99 LVSURF = 1.0 PINLET = 1 for a pressure filling or PINLET = 0 for a velocity/inlet filling

Low pressure die casting (LPDC)

Timestep handling

DT = 1e-3 DTMAXFILL = 1e-2 (it is important to limit the timestep during the filling of an LPDC part. A value of 1e-2 is recommended for filling time of about 5-20 s.). DTMAX = 0.2 - 1 (depending upon the size of the model and thus the solidification time)
Porosity

PIPEFS = 0.0 GATEFEED = 1

Filling

WSHEAR = 0 (never use WSHEAR = 2 for LPDC) FREESFOPT = 1 WALLF = 0.8 LVSURF = 1.0 PINLET = 1 for a pressure filling or PINLET = 0 for a velocity/inlet filling

Tilt casting
Timestep handling

DT = 1e-3 DTMAXFILL = 1e-1 DTMAX = 0.5 - 5 (depending upon the size of the model and thus the solidification time)
Porosity

PIPEFS = 0.3 GATEFEED = 0

Filling

WSHEAR = 2 FREESFOPT = 1 WALLF = 0.8 TILT = 1 DETACHTOP = 1

Version 2006.0

Page 259

ProCAST User Manual

Centrifugal casting
Timestep handling

DT = 1e-3 DTMAXFILL = 1e-1 DTMAX = 0.5 - 5 (depending upon the size of the model and thus the solidification time)
Porosity

PIPEFS = 0.3 GATEFEED = 0

Filling

WSHEAR = 2 FREESFOPT = 1 WALLF = 0.8 RELVEL = 1

Version 2006.0

Page 260

ProCAST User Manual

POROSITY MODELS
The different porosity models available in ProCAST are summarized hereafter. Then, a full description of each model is provided.

POROS = 1
The POROS = 1 model corresponds to the latest porosity model of ProCAST. It accounts for coupled micro and macroporosity, as well as pipe shrinkage. It can be applied to both gravity casting and injection (either hpdc or lpdc). This model can be used with or without flow calculations.

POROS = 4
The POROS = 4 model corresponds to the same model as POROS=8 (see below), with the additional treatment of multiple free surfaces for piping. This model can be used with or without flow calculations.

POROS = 8
The POROS = 8 model corresponds to the porosity model which was available in version 3.2.0 of ProCAST (using the POROS=1 Run Parameter of v3.2.0). Although this model is less sophisticated than the current POROS = 1 model, it was kept in this version for the users who have calibrated the model to their casting and who obtained good results in the past. The only difference between the current POROS=8 model and the one available in version 3.2.0 is the method used to compute liquid pockets. Thus, this may lead to slight differences between the versions. This model can be used with or without flow calculations. One should notice that even if POROS = 0 is set, piping will be still calculated and displayed. In order to disable the piping calculation (in the case of a THERMAL only calculation), one should set FREESF to zero.

To disable the porosity and piping calculations

In order to disable the porosity calculation, one should set POROS = 0. However, this may still lead to some piping calculation. In order to disable also the piping, one should set in addition PIPEFS = 0.

Version 2006.0

Page 261

ProCAST User Manual

POROS=1
Parameters of the model
POROS MACROFS PIPEFS USERHO FEEDLEN 1 (porosity model) 0.7 (critical fs for porosity formation) 0.3 (critical fs for piping formation) 1 (flag to set the density model) 1 5.0e-3 (critical feeding length)

It is necessary to define the gravity vector in order to have meaningful results with POROS=1.

Model
When a casting solidifies, pockets of liquid are created, surrounded by a mushy zone and then a solid shell. Automatically, the casting is divided into "regions" within which the fraction of solid is lower than one or that are bounded by walls (or symmetry planes). As solidification proceeds and depending upon the complexity of the geometry, the number of "regions" may increase with time. A region can thus be split in more regions. A region can disappear when all nodes have completely solidified.

When a "region" is cooling down, if the density is increasing with decreasing temperature (the usual case for most alloys), some shrinkage occurs. At each timestep, the accumulated shrinkage occurring at all the nodes which have a solid fraction equal to or lower than MACROFS, plus those nodes between the

Version 2006.0

Page 262

ProCAST User Manual

MACROFS and MACROFS + FEEDLEN isosurfaces, is computed. This shrinkage is then distributed according to the following scenarios: a) Find the highest point of the region that is on a free surface and has a fraction of solid lower than PIPEFS. In this case, piping occurs and the free surface of the casting (usually the riser) goes down by the amount corresponding to the shrinkage (For display purposes, in the pipe, the shrinkage porosity value is set to 1 and FVOL is set to a value below 0.5, so that it will exhibit piping, i.e. "empty nodes").

b) Find the highest point of the region that is on a free surface and has a fraction of solid higher than PIPEFS. In this case, macroshrinkage occurs at that point. This will also correspond to piping (same result as a) above). However, instead of showing an empty volume, it will have a "Shrinkage porosity" value of 1. (For display purposes, in the macroshrinkage, the shrinkage porosity value is set to 1 and FVOL is set to a value above 0.5, so that it will not exhibit piping, i.e. "empty nodes", but shrinkage).

Version 2006.0

Page 263

ProCAST User Manual

c) No nodes of the "region" are on a free surface with a fraction solid lower than PIPEFS. In this case, no more piping can occur and macroshrinkage in the bulk of the casting will appear. The macroporosity will appear at the highest most liquid point of the region (e.g. if there is a pocket of liquid which is surrounded by a mushy zone, itself surrounded by a solid shell, the macroporosity will start at the highest point of the liquid pocket). (For display purposes, in the macroporosity region, the shrinkage porosity value is set to any value and FVOL is set to a value above 0.5, so that it will not exhibit piping, i.e. "empty nodes", but shrinkage).

Version 2006.0

Page 264

ProCAST User Manual

During the same time, microporosity is computed in the following way: a) microporosity can appear only in the zone where the fraction of solid is in between MACROFS and 1. b) within this zone, two situations may occur : b1) there is still some mushy zone (or liquid) below MACROFS. In this case, microporosity can occur only at a distance higher than the value of FEEDLEN from the MACROFS isosurface (Zone A in the figure below). This means that, if high gradients are present, the distance between the MACROFS and solidus isosurface is smaller than FEEDLEN, no microporosity occurs (Zone B in the figure below). The amount of microporosity is equal to the density change between the local fraction of solid and 1. b2) there is no more mushy zone below MACROFS. In this case, the FEEDLEN parameter is not active anymore and there could be microporosity in the whole region between MACROFS and 1. This is due to the fact that as there is no more "open liquid" to feed the shrinkage, local microporosity has to occur in order to compensate the local shrinkage. The amount of microporosity is equal to the density change between the local fraction of solid and 1.

c) on the other hand, if FEEDLEN=0, the amount of microporosity is the same throughout the part (except where there is macroporosity). In this case, the amount of microporosity is everywhere equal to the density change between Fs=MACROFS and Fs=1.

Version 2006.0

Page 265

ProCAST User Manual

d) if FEEDLEN is set to a very large value (larger than the casting side), the region between MACROFS+FEEDLEN is not existing anymore. Thus, no microporosity is created until the whole pocket is above MACROFS (i.e. no microporosity created according to b1) above). Then, microporosity can start to occur, according to b2) above. The amount of porosity is displayed in ViewCAST under the "Shrinkage porosity" Contour. The unit is volume fraction [-]. In general, values which are higher than 0.01 can be considered as macroporosity, whereas regions with a value lower than 0.01 correspond to dispersed microporosity. Symmetry planes are taken into account in the liquid pocket calculation (i.e. symmetry planes "close" the pockets, even if there is liquid at the symmetry wall. See the "Density definition" section for more details about how to define the density. See the "Active feeding" section for more details about feeding from the piston or the ingate in the case of hpdc and lpdc.

POROS=4
Parameters of the model
POROS MACROFS 4 0.7 (porosity model) (critical fs for porosity formation)

Model
This model is based upon the same algorithm as the POROS=8 model (see below). In addition, multiple piping can be considered. This means that piping occurs at the highest free surface of each region. Thus, one can have piping at the top of risers which are at different levels (this was not the case in version 3.2.0 and thus does not occur with POROS=8) See the "Density definition" section for more details about how to define the density.

POROS=8
Parameters of the model
POROS MACROFS 8 0.7 (porosity model) (critical fs for porosity formation)

Version 2006.0

Page 266

ProCAST User Manual

In version 3.2.0, this Run parameters was called MOBILE (MOBILE still exists in version 2004.0 for other purposes. Thus, it should not be used anymore for porosity settings).

Model
When the casting solidifies, pockets of liquid are created, surrounded by a mushy zone and then a solid shell. As soon as a pocket of liquid is surrounded by a zone which has a solid fraction higher than MACROFS, the density of each node inside the pocket (from fs=1 to fs=MACROFS) is recorded (as a "critical density"). Then, the amount of porosity of each of these nodes is equal to the density variation between this "critical density" and the density of the solid. Piping is occurring at the highest free surface of the model. Thus, if there are two separate regions with risers at different heights, only the highest one will exhibit piping, even if the lower one is in an isolated region. POROS=4 corrects this situation (see above). Symmetry planes are taken into account in the liquid pocket calculation (i.e. symmetry planes are "closing" the pockets, even if there is liquid at the symmetry wall. The amount of porosity is displayed in ViewCAST under the "Shrinkage porosity" Contour. The unit is volume fraction [-]. See the "Density definition" section for more details about how to define the density.

Density definition
For most alloys, the density at the liquidus is lower than the density at the solidus, thus leading to porosity. By default, the porosity module of ProCAST is using the density curve in the mushy zone which is defined in the material database (see the green curve in the Graph A below). However, if this density change in the mushy zone is not well known, it is possible to automatically calculate the density in the mushy zone, as an average of the liquid and solid densities, weighted by the fraction of solid. The density of the liquid and of the solid is calculated as a function of temperature, by extrapolating the density slopes at the liquidus and solidus respectively.

Version 2006.0

Page 267

ProCAST User Manual

The user has the possibility to activate this automatic density computation with the Run parameter USERHO. If USERHO is set to 0 (instead of 1 which is the default value), the density curve which is defined in the material properties, between the solidus and liquidus, will be ignored and the automatic density computation will be activated. (Please note that in version 4.x.x, the default value was 0). This is valid for all the Porosity models (POROS = 1, 4, 8).

Active Feeding
In the case of injection (e.g. high or low pressure die casting), for a while, the shrinkage is compensated by the piston in the case of hpdc and by the liquid bath for lpdc, thus leading to no piping. ProCAST can account for such "active feeding", by setting the Run parameter GATEFEED=1. In this case, no piping will occur, but liquid will feed the ingate, as long as the fraction of solid is lower than GATEFS (i.e. there is feeding in all regions which are within the same GAFEFS isosurface as the gate). The gate is defined as being the region where in inlet velocity is applied, or where a pressure is applied (or where GATENODE is applied). One should note that the porosity level will NOT depend upon the value of the pressure. The Active feeding is there only to compensate for the shrinkage of the region in contact with the piston, as long as the fraction of solid is lower then

Version 2006.0

Page 268

ProCAST User Manual

GATEFS (i.e. within the same GATEFS isosurface as the gate). This does not correspond fully to a real third stage pressure, however, the value of GATEFS could be adjusted with pressure if desired). For HPDC and LPDC, it is advised to set PIPEFS = 0.0 in order to prevent piping at the top of the casting, if the gate is closing too early. The active feeding is valid only with the POROS = 1 model. In the case of a thermal only calculation (for HPDC or LPDC), one should set a Pressure boundary condition at the ingate (in order to activate the active feeding) and one should disable the FLOW run parameter (as normally a Pressure BC would automatically switch ON the flow solver). Of course, GATEFEED should also be set to 1 in this case. In order to activate GATEFEED, one should apply a pressure of inlet velocity BC on external faces. There are cases where there are no external faces on which to apply these boundary conditions. This happens in the case of a filling with a shot piston. To account with such situation, one can apply instead a "Gate feeding" condition on an inside location of the casting (defined by its node number). In general, it is advised to select a node inside the final biscuit, which will remain liquid during most of the casting process. To define this node number, the GATENODE Run Parameter should be specified, followed by the node number. In order to find the node number corresponding to the desired location, it is advised to use the "pick" capability of ViewCAST. This can be done before the run of the case, just after DataCAST. Warning : In the case of a thermal only calculation, a Pressure BC should be set at the ingate in order to switch ON the active feeding. In such a case, the Pre-processor automatically is activating the flow solver (FLOW=3). As a consequence, one should set FLOW=0 manually in the d.dat file in order to deactivate the flow solver (but the porosity calculation will be performed with active feeding).

SGI Porosity model

Instead of shrinking during solidification, some alloys do also exhibit some expansion. The most well known material which exhibits this behavior is the Nodular Cast Iron, also called Spheroidal Graphite Iron (SGI). Grey Iron may also exhibit such behavior. The volume change as a function of temperature is shown in the figure hereafter for different cast irons.

Version 2006.0

Page 269

ProCAST User Manual

POROS model in case of expansion

In order to account for such expansion, the POROS=1 model was adapted in the way describe hereafter. In the "pockets" of liquid and mushy zone which have a solid fraction lower than MACROFS (+FEEDLEN), the integral of the density change at each timestep is performed. In the case of expanding material, some location will expand and some will shrink. If the total density change corresponds to a net shrinkage, the "regular" model applies (i.e. this macroshrinkage will occur as the highest most liquid point). If it corresponds to a net expansion, two scenario may occur : a) Find the highest point of the region that is on a free surface and has a fraction of solid lower than PIPEFS. In this case, the free surface of the casting (usually in the riser) will go up by the amount corresponding to the net expansion. b) Find the highest point of the region that is on a free surface and has a fraction of solid higher than PIPEFS. The amount of expansion is applied proportionally to all of the nodes in the region that have pre-existing porosity (thus, the macroporosity which appeared earlier, when there was a net shrinkage, will be "refilled"). For example, if the amount of expansion is enough to "refill" 50% of the total porosity in the region, then the porosity of each node is reduced by 50%. c) No nodes of the "region" are on a free surface with a fraction solid lower than PIPEFS. The expansion is distributed to all the nodes in the region as described in b).

Version 2006.0

Page 270

ProCAST User Manual

At the same time, microporosity is computed in the following way: a) Microporosity can appear or disappear (partially or fully) only in the zone where the fraction of solid is in between MACROFS (+FEEDLEN) and 1. b) Within this zone, two situations may occur: b1) There is still some mushy zone (or liquid) below MACROFS. In this case, microporosity can occur or disappear (partially or fully) only at a distance greater than the value of FEEDLEN from the MACROFS isosurface (Zone A). This means that, if high gradients are present, the distance between the MACROFS and solidus isosurface is smaller than FEEDLEN, no microporosity occurs or disappears (Zone B in the figure below). The amount of microporosity formation or disappearance is equal to the density change between the local fraction of solid and 1. b2) There is no more mushy zone below MACROFS. In this case, the FEEDLEN parameter is not active anymore and there could be microporosity in the whole region between MACROFS and 1 for shrinkage. This is due to the fact that as there is no more "open liquid" to feed the shrinkage, local microporosity has to occur in order to compensate the local shrinkage. The amount of microporosity is equal to the density change between the local fraction of solid and 1. If the density change between the local fraction of solid and 1 is positive (expansion), there is no microporosity formation. The already formed micro porosity can be refilled partially or fully during expansion depending on the degree of the density change.

c) On the other hand, if FEEDLEN=0, the amount of microporosity is the same throughout the part (except where there is macroporosity). In this case, the amount of microporosity is everywhere equal to the density change between Fs=MACROFS and Fs=1. Again if the density change between the fs=MACROFS and fs=1 is positive (expansion), there is no microporosity formation. The already formed micro porosity can be refilled partially or fully during expansion depending on the degree of the density change. d) If FEEDLEN is set to a very large value (larger than the casting size), the region between MACROFS+FEEDLEN is not existing anymore. Thus, no microporosity is created or refilled until the whole pocket is above MACROFS (i.e. no microporosity created or refilled according to b1) above). Then, microporosity can start to occur or disappear, according to b2) above.
Mold rigidity

The rigidity of the mold has an influence on the amount of porosity in the case of expanding alloys. If the mold is totally rigid, the casting can not expand and thus the alloy expansion will be "available" for the "refill" of the existing porosity. On the other hand, if the mold is very soft (or weak), the casting will expand and thus

Version 2006.0

Page 271

ProCAST User Manual

there will be no "refill" of the porosity (of course, the reality is more complex as the solid shell is thick enough it will act as a "rigid" mold, even if the sand mold is weak. In order to take the mold rigidity into account, the Run Parameter MOLDRIG is introduced. MOLDRIG should be defined by a value between 0 and 1. All the net expansion is multiplied by MOLDRIG. Thus, with MOLDRIG=1, corresponding to a rigid mole, the expansion will be fully accounted. On the other hand, no expansion will be taken into account if MOLDRIG=0. The expansion will be compensated by the mold movement because the mold is too weak to hold the expansion in this case. For real situations, the value of MOLDRIG should be set somewhere between 0 and 1 depending upon the casting processes. MOLDRIG should be added "manually" in the p.dat file by the user. The default value is 1.
Density curve

For expanding materials, the density defined in PreCAST should be not anymore monotonic. The density can increase (locally) with increasing temperature.

One should note that such density curve will not be obtained when the material properties are computed with the Computherm database. This is because the expansion is depending upon the microstructure which itself depends upon the cooling rate (see next section). Thus, such curve should be defined manually by the user, based upon experiments.

Version 2006.0

Page 272

ProCAST User Manual

Coupled Microstructure - Porosity calculation

When a microstructure calculation is performed in the case of SGI, the density is automatically calculated at each location (i.e. each node) of the casting and this local density is used "on-line" in the porosity calculation (POROS = 1 model only). The following figures are showing the temperature at different locations of a casting and the corresponding density curves (with the same colors). As one can see in the bottom graph, the fast cooling zones exhibit only shrinkage (red curve), whereas the low cooling rate zones exhibit a very important expansion (blue curve). In between, one have mixed behaviors.

Cooling curves calculated with the Microstructure module of ProCAST (top) with the corresponding computed density curves (bottom) (SGI with 3.5% C, 2.39% Si, 0.1% Mn, 0.023% Cr, 0.045% Mn, 0.03% Ni, 0.01% Cu, 0.0055% V, 0.0056% Ti, 0.117% Co)

Version 2006.0

Page 273

ProCAST User Manual

The above example is showing the advantage of using a coupled Microstructure Porosity calculation. If the Microstructure module is not available, one should use the density curve which corresponds to the "closest" average cooling rate.
Example

The figures hereafter are showing an example of an expanding alloy. One can see very well that the level of the metal in the left riser (which has an insulating sleeve) is first going down (overall shrinkage) and then it is going back up (overall expansion).

Version 2006.0

Page 274

ProCAST User Manual

The following figure is showing an enlargement on the riser where the liquid level is going down and then up.

Version 2006.0

Page 275

ProCAST User Manual

VIRTUAL MOLD
ProCAST offers the capability of modeling a mold without meshing it, with the Virtual Mold option. This is especially useful in the case of large sand casting. It can also be used in permanent mold casting, if one is mainly interested in the filling behavior. When a Virtual mold is used, one should define the dimension of the mold (which is an orthogonal box, aligned with X, Y and Z, the material properties of the mold and the interface heat transfer coefficient between the different part of the casting and the mold. The Virtual mold model considers that the thermal diffusion is occurring in the mold, according to "half diffusion distances". Thus, the model calculates for each face of the casting what is the half diffusion distance. This distance is either the one between the face and the limit of the mold or half the distance between the face and an other face in front. At the limit of the mold, an adiabatic boundary condition (i.e. no flux) is considered. Thus, it is advised to defined a "box" which is large enough, in order not to "saturate" the virtual mold. A good start for the Virtual Mold definition is to look at the minimum and maximum size of the model in the "Geometry/Check Geom/Min-Max" menu.

Then, one should open the "Geometry/Virtual Mold" menu, to get the following :

Version 2006.0

Page 276

ProCAST User Manual

The size of the Virtual Mold box can be defined in the following fields. One has the possibility to define automatically the virtual mold box size, using the "Default Size" button. One should first select on the right, whether the virtual mold box should be 1, 2, 3, 4 or 5 times larger than the part, in the X, Y and Z directions respectively. For small components, it is advised to use a virtual mold 5 times the component size. For large castings, a virtual mold with a size of 2 o r 3 times the component size will be enough. In case there is any doubt, it is advisable to take a large virtual mold rather than the reverse, in order to avoid saturation.

Version 2006.0

Page 277

ProCAST User Manual

Please note that one should NOT use the real size of the mold for the Virtual mold computation. As the model considers that there is no cooling of the sides of the virtual mold, the box should be large enough in order to avoid satuaration of the mold (and thus too slow cooling). The size of the Mold box can be visualized with the "Show Mold" button (in the case of the figure hereafter, the virtual mold of a size 1 time larger than the component was selected, for viewing purposes).

Version 2006.0

Page 278

ProCAST User Manual

When the Mold size is good, one can "compute" the Virtual Mold, using the "Compute Mold" button. Please note that the computation may take sometime, depending upon the size (i.e. the number of surface elements) of the model. The computation time is independent upon the size of the mold box itself (in the case of the figure hereafter, the virtual mold of a size 1 time larger than the component was selected, for viewing purposes).

Once the Virtual Mold is computed, the "Thermal depth" can be visualized with the "Show Depth" button. The color scale can be changed with the "Set Scale" button, which opens the following panel :

Version 2006.0

Page 279

ProCAST User Manual

Finally, if needed, it is possible to erase the Virtual mold, using the "Remove Mold" button. Before setting a Virtual mold, a few precautions should be taken : When a virtual mold is used in conjunction with a symmetry, it is mandatory to set first the symmetry and then to generate the virtual mold. Otherwise, the symmetry of the virtual mold will not be applied. If the mesh is made out of several material domains, it is mandatory to define first the interfaces between the different domains and then to generate the Virtual mold. If interfaces are changes after the Virtual mold creation, it will be destroyed. Once the Virtual Mold is generated, it appears in the Material list in the Material properties assignment. One should assign the desired material properties to this Virtual mold domain.

In the same way, the interfaces between the Virtual Mold and the different material domains are automatically generated. These are labeled "Virtual". One should assign an interface heat transfer coefficient to each "Virtual" interface.

Version 2006.0

Page 280

ProCAST User Manual

Version 2006.0

Page 281

ProCAST User Manual

FILTERS
If Filters are present in the process, they can be modeled. Firstly, the filter should be meshed as a separate material domain, as shown hereafter.

Once the model is loaded, a Filter material should be assigned to the corresponding material domain (see the "Databases/Material Properties" section for the description of the filter properties definition).

Version 2006.0

Page 282

ProCAST User Manual

Then, the "FILTER" type should be assigned to this material domain. In the Interface menu, the interfaces between the casting and the filter material (on both sides) should be kept as "EQUIV" (i.e. no interface should be created between the casting material and the filter).

Version 2006.0

Page 283

ProCAST User Manual

Finally, the interface heat transfer coefficient between the liquid metal (casting material) and the filter material can be defined in the "Boundary Conditions/Assign Volume/Filter Heat" menu. The list of the Filter materials is shown, as well as the interface heat transfer coefficient database, ready for an assignment.

Version 2006.0

Page 284

ProCAST User Manual

EXOTHERMIC
Exothermic sleeved can be modeled in ProCAST, with the appropriate heat generation. Firstly the exothermic sleeve should be meshed as a separate material domain. Then, the corresponding material properties should be assigned and the "EXOTHERMIC" type should be defined (if the EXOTHERMIC type is not defined, the exothermic energy, as defined hereafter will not be released) :

The material properties of the Exothermic sleeve must be defined as follows :

Version 2006.0

Page 285

ProCAST User Manual

The "standard" properties (thermal conductivity, density and specific heat) should be defined as usual (1) for the sleeve material. In addition, the exothermic properties should be defined in the "Exothermic" tab (2). The Exothermic energy (3) corresponds to the amount of energy which is generated during the burning of the sleeve. The Ignition Temperature (3) corresponds to the temperature at which the exothermic reaction is initiated. The "burning kinetics" is defined in the table (4-5), as a fraction of burning, which is a function of time. Once the exothermic reaction is started (i.e. when the temperature is going above the ignition temperature), the exothermic energy will be released according to the burnt fraction.

Version 2006.0

Page 286

ProCAST User Manual

CYCLING
ProCAST allows to model the cycling sequence in the case of die casting (i.e. to account for the heating of the die mold during the first cycles of casting). The figure below shows the principles of cycling, as well as the different sequences.

In order to model cycling, one should be able to take into account the fact that when the mold is closed, there is an interface heat transfer between the casting and the mold and when the mold is opened, one should consider a cooling of the die and of the casting with the ambiance. One should thus switch between an interface heat transfer coefficient and two "Heat" boundary conditions.

Version 2006.0

Page 287

ProCAST User Manual

The above diagrams are showing in green the interface heat transfer coefficients as a function of time for both the interface between the casting and the fixed die (solid green line) and between the casting and the mobile die (dashed green line). One could see that the interface heat transfer coefficient goes to zero when the casting and the mold are not anymore in contact. On the reverse, as soon as the mold opens, a "Heat" boundary condition should be set (the red curve on the above figure corresponds to the heat transfer coefficient and the dashed blue curve corresponds to the ambient (or external) temperature. One could see on the above figure that during the spray sequence, the heat transfer coefficient increases and the spray temperature is also taken into account. When the die is closed, before the next filling, the interface heat transfer coefficient between the dies is again activated and the interfaces between the casting and the die(s) are set to zero. One should note that the Pre-defined Run parameters for HPDC-cycling are not designed in order to peform a Porosity calculation during the cycling calculation. If this is desired, one should set to :
POROS 1

Version 2006.0

Page 288

ProCAST User Manual

PIPEFS 0. GATEFEED 1

Then, a pressure BC should be set at the ingate in order to trigger Gate feeding.

Version 2006.0

Page 289

ProCAST User Manual

LOST FOAM
ProCAST provides the capability to model "lost foam" casting or "evaporative pattern" casting. The cavity is filled with foam and is surrounded by a sand mold. During the filling, the hot liquid metal is heating up the foam which is burning, leaving the space for the liquid metal. In the Lost Foam process, the filling is controlled by the rate of burning of the foam, as well as the escape of the gas through the ingate or through the permeable sand mold. The set-up of a Lost Foam case should be done as follows. Firstly, the geometry should contain at least three components : a part of the downsprue which is empty, the cavity filled with the foam and a sand mold (see figure below).

The following figure shows how the materials should be set. The downsprue should be assigned with the casting material and it should be set as EMPTY=YES. The cavity, which is filled with foam should be assigned with a "Foam" material (EMPTY=NO) and the mold should be set with a sand material (EMPTY=NO).

Version 2006.0

Page 290

ProCAST User Manual

There are no special requirements for the material properties definition of the casting material. For the Foam material, the thermal conductivity, the density, the specific heat and the latent heat (of burning) of the foam should be specified. Moreover, under Tsolidus and Tliquidus, one should specify the temperature at which the foam starts and ends to burn respectively. One should note that the burning kinetics (and thus the filling rate) is influenced by the density, the specific heat, the latent heat and the burning temperature. The larger these values, the slower the burning kinetics will be. This is due to the fact that the foam should be heated up by the liquid metal and that if these quantities are larger the amount of heat required to burn the foam will be larger and thus it will take more time. For the sand, in addition to the usual thermal properties, it is necessary to define it's permeability (in the "Fluid/Permeability" tab). Typical values range from 1e-6 to 1e-7 cm**2. The interface between the casting material and the foam material should be set as "EQUIV" (i.e. no interface), whereas the other interfaces should be set as "COINC".

In addition the usual "Heat" boundary condition (to cool down the outside of the mold) and the inlet Temperature, one should set two Pressure BC. One for the top of the down sprue (on the whole surface) and the other for the outside of the mold. It is recommended to set a value of 1 atm. on the outside of the mold and a value slightly larger (e.g. 1.05 atm) on the top of the downsprue.

Version 2006.0

Page 291

ProCAST User Manual

Version 2006.0

Page 292

ProCAST User Manual

The Run Parameters should be configured as a regular Gravity filling problem with an inlet pressure. Additional Run parameters should be set in the p.dat file for Lost Foam :

Version 2006.0

Page 293

ProCAST User Manual

FOAMHTC 0.02 FOAMHTCMAX 0.25 BURNZONE 1.0 GASFRAC 0.1 DIAG 262144 These parameters are regulating the rate of transfer of heat between the liquid metal and the foam. When the liquid metal front is at a distance of "BURNZONE" from the foam, the "interface heat transfer" between them is equal to the value of FOAMHTC/BURNZONE. This heat transfer is increasing as the liquid get closer from the foam. When the liquid is "touching" the foam, a maximum heat transfer coefficient (equal to FOAMHTCMAX) is set. FOAMHTC and FOAMHTCMAX are defined in the p.dat file in CGS units and BURNZONE is in "cm". BURNZONE corresponds to the "usual" distance between the foam and the liquid metal. One should be careful that the mesh size must be finer than the value of BURNZONE. One should note that the burning kinetics will change when the value of BURNZONE is changed (as the interface heat transfer coefficient is equal to FOAMHTC divided by BURNZONE). As the rate of heat transfer between the foam and the liquid metal is not well known, it may be needed to calibrate the filling time with the values of FOAMHTC and FOAMHTCMAX. To shorten the filling time, one should increase these values. The value of GASFRAC corresponds to the fraction of the foam which is transformed into gas (during the burning). The rest is mainly transformed in liquid traces. A value of 0.1 (10%) is recommended. In order to allow for smooth Restarts, the Run Parameter DIAG should be set to 262144. This setting is not necessary if no Restart is performed.

Version 2006.0

Page 294

ProCAST User Manual

THIXO CASTING
ProCAST has a dedicated model for thixo casting (or semi-solid casting). When a semi-solid material is injected in a mold cavity, it's viscosity is depending upon the shear rate (as well as the shear rate which was encountered by the metal previously during the injection). When the shear rate is high, the solidifying dendrites are broken and the fluidity is increasing (i.e. the viscosity is decreasing). In order to account for such a behavior, a specific "Power cut-off" model was designed. In order to set-up a Thixo casting case, one should define the appropriate material properties of the casting material and define an additional simple input file (prefixg0.dat). In the Material properties definition, all the thermal properties should be defined as usual. In the Fluid tab, the Viscosity should be defined with the "Power-Cutoff" tab.

The values of the "Zero viscosity", the "K Factor" and the "Power", which can be Temperature dependant, should be calibrated with experiments. Please note that the value of n should be negative (in order to have a decreasing viscosity with an increasing shear rate).

Version 2006.0

Page 295

ProCAST User Manual

The principle of the Thixo casting model is to divide the mesh in different regions. For each region, it is possible to define the critical "cut-off" value of the shear rate . This definition is done in a small input file named prefixg0.dat. The format of this ASCII file is the following (one line per domain of the mesh) :
Domain_number critical_shear_rate

The following example shows a g0.dat file for a case with 4 domains. One should note that domain 2 is the mould and a dummy value (of 1) should be set. The units of this critical cut-off shear rate is [s].

The principle for the definition of the critical cut-off shear rates is described hereafter. It is based upon the fact that dendrites are broken in locations where the shear rate is high and then, even if the shear rate is decreasing afterwards (in an opened cavity), the viscosity will not increase anymore (as the dendrites have been broken). This means that the viscosity in a cavity which is following a gate will remain at a rather low level, corresponding to the shear rate of that gate. One can illustrate this principle with the following figure.

In domain 3, the viscosity of the semi-solid material is about the one which corresponds to the shear rate that was encountered in the gate (i.e. domain 2). Thus, one should set a "cut-off" value in domain 3 which corresponds to the mean shear rate of domain 2. In the same way, the "cut-off" value of domain 5 should

Version 2006.0

Page 296

ProCAST User Manual

correspond to the shear rate encountered in domain 4 (which should be a higher value than domain 2 as the section is smaller). In domains 1, 2 and 4, it is not necessary to set a cut-off value (the default value of 1 can be used), as the effective shear rates will correspond well to the reality. To determine the shear rate in a gate, one can view the "Non-Newtonian Shear Rate" in ViewCAST.

Finally, in order to activate the Power cutoff model, the two following Run parameters should be set :
HIVISC = 2 NNEWTON = 2

Version 2006.0

Page 297

ProCAST User Manual

CENTRIFUGAL CASTING
Starting from version 2006.0, centrifugal casting cases should be set-up in the following way. Only the CASTING domains should be set with a revolution velocity (i.e. not the mold materials). This must be defined in the "Process/Assign Volume" menu, with a "Revolution". The Run Paremeter RELVEL should be set to 1. This Run parameter is activating the resolution of the fluid equations in a "Relative velocity reference frame" to handle the rotation. The rest of the set-up should be done as for a standard gravity casting (i.e. WSHEAR = 2, WALLF = 0.8, FREESFOPT = 1 (or 2)).

Version 2006.0

Page 298

ProCAST User Manual

MULTIPLE MESHES AND NON-COINCIDENT MESHES

ProCAST has the capability to handle non-coincident meshes (see figure below).

If the different domains are meshed separately, they can be loaded in PreCAST using the "File/Multiple Meshes" menu.

Then, a window opens, which allows to load the different meshes.

Version 2006.0

Page 299

ProCAST User Manual

Once the meshes are loaded, the set-up of the case in PreCAST should be done as for conventional meshes, except for the definition of the interfaces (see figure below). Firstly, all the non-coincident interfaces, where at least one element is coincident on each side, are automatically listed in the interface list. One should change the "Type" from EQUIV to NOCOIN (1). Then, for the non-coincident interfaces which are not listed, one should "create" manually these interfaces, with the Add button (2). This will open the"Add Interface Pair" window, where the material numbers of each side can be defined (3). The "Master-Slave" concept is used in the calculation of the heat transfer across the non-coincident interface. It is advised to set the casting as the Master and the mold as the Slave (although it does have only a minor importance on the accuracy of the results). Then, one should check the "Non-coincident Interface" button (4). Once all the interfaces are

Version 2006.0

Page 300

ProCAST User Manual

defined, one can "Apply" the whole selection (5), which will create the appropriate interfaces.

When a non-coincident interface is defined, it is possible to change the default tolerances used for the detection of the nodes on the opposite side. To do so, one should make a right click on the NCOINC label and the following window will appear :

The "In-plane Tolerance" corresponds to the maximum distance between the two surfaces in order to have a contact (distance normal to the plane of the interface). The "Perimeter Tolerance" corresponds to the maximum distance around an element of the master surface where a node of the slave surface can be found.

Version 2006.0

Page 301

ProCAST User Manual

By default, the In-plane and Perimeter tolerances are taken as a fraction of the smallest edge of the whole mesh. Thus, if the mesh has a quite heterogeneous mesh size, these tolerances may be too small (and thus, there will be "no contact" and thus no heat transfer at these non-coincident interfaces). If such a case occurs, one can change (i.e. increase) these tolerances. One should however be careful not to use too large tolerances so that nodes beyond the opposite surface will be taken into account. A good tolerance value should be about half of the mesh size of the corresponding surface.

Version 2006.0

Page 302

ProCAST User Manual

GEOMETRY MANIPULATION
The geometry can be manipulated and can be displayed with different modes, using the following icons.

Manual rotation of the model (it opens the Rotate window). Otherwise, the model can be rotated interactively with the mouse at any time (as long as the Center or Drag icon is not activated). If the model is rotated interactively with the mouse, while the "Shift" key is pressed in the same time, the model is rotating along the axis perpendicular to the screen (the mouse should be rotated horizontally).

Restore the X-Y orientation of the model (Z-axis perpendicular to the screen)

Interactive zoom (the model is enlarged when the cursor is moved towards the bottom of the screen and it is reduced when the cursor is moved towards the top of the screen)

Auto-scale (automatic scaling of the model so that it fits into the graphics window)

Center of the model (the location of the model which is selected will move to the center of the screen)

Drag of the model (interactive move of the model on the screen)

Version 2006.0

Page 303

ProCAST User Manual

Material (or domain) selection

Wire frame display mode

Hidden wire frame display mode

Version 2006.0

Page 304

ProCAST User Manual

Hidden display mode with the mesh

Version 2006.0

Page 305

ProCAST User Manual

Shaded display mode

Activation of the enclosure viewing (for radiation models)

Version 2006.0

Page 306

ProCAST User Manual

MESH OPTIMIZATION
The speed of a FEM calculation depends upon the structure of the mathematical matrices which have to be solved. This depends upon the way the FEM mesh is numbered (i.e. which nodes belong to which element and vice-versa). It is possible to optimize the FEM mesh numbering in order to minimize the CPU time. This was done previously in MeshCAST. However, as the addition of interfaces is creating new nodes, it was necessary to know already in MeshCAST which interfaces had to be created. In order to simplify this operation, it is now possible to optimize the mesh in PreCAST, just before SAVING a case. This will take automatically into account the created interfaces and will thus guarantee that the mesh is optimum. As the optimization operation may be rather long for large meshes, it is not done by default at every save of the model. In order to activate the optimization, one should set it in the "File" menu :

This optimization will be done before exiting PreCAST (it is not performed with a "Save" or "Save as" operation, but in this case a warning is displayed upon Exit). One should note that if an optimized model is loaded again in PreCAST, it is not necessary to optimize it again, as long as the interface settings are not changed. In the case of initial conditions EXTRACTion, it is not possible to optimize the mesh when the extracted temperatures were coming from a non-optimized mesh.

Version 2006.0

Page 307

ProCAST User Manual

USER FUNCTIONS
ProCAST allows the use of User Functions. User Functions can be used to define boundary conditions in a very flexible way, such as combined time, temperature and/or space dependant heat transfer coefficients. Three categories of Functions are available in ProCAST : User Functions to define a given condition, instead of a constant or a table External Functions which can be called from the User Functions, in order to get the values of some fields at any node. External Computation Function which can be called during the execution of the calculation (e.g. for coupling with external softwares).

User Functions description

Currently, the following User Functions are available : Interface BC : interface heat transfer coefficient Heat BC : external heat transfer coefficient Heat BC : external temperature Heat BC : emissivity Heat BC : heat flux Imposed velocity (X, Y and Z components) Solid transport velocity (X, Y and Z components) Translation vector (X, Y and Z components) Mass source position (X, Y and Z components) Mass source flow rate

Each of these functions have the following arguments : Time Local temperature Local fraction of solid Local coordinates Material number Boundary condition ID

This allows to define conditions as a function of any of the above arguments (e.g. a time- and space-dependant interface heat transfer coefficient). In addition, "external functions" allow to get the value of the main fields (e.g. temperature, fraction of solid, velocities, ...) at any location of the model. This

Version 2006.0

Page 308

ProCAST User Manual

allows to define for instance a heat transfer coefficient as a function of a "control temperature", corresponding to a given location, somewhere else ("remote control"). User Function Templates are available in the next section.

External Functions description

From any User Function, it is possible to get the value of some fields (e.g. Temperature, fraction of solid, velocity) at any node, specified by its node number. To do so, the following functions are available : usertemp1(node#) for temperature userfs1(node#) for the fraction of solid uservx1(node#) for the X-component of the velocity uservy1(node#) for the Y-component of the velocity uservz1(node#) for the Z-component of the velocity

In addition, it is possible to get the node number by giving the coordinates, using the following function : nodNum (xin, yin, zin, domain#, xout, yout, zout)

where xin, yin and zin are the coordinates specified by the user, xout, yout and zout the coordinates of the closest nodes which corresponds to the returned node number. domain# is the domain number in which the search node should be found.

External Computation Function description

Beside User Functions for the definition of flexible conditions, ProCAST is providing an "External User Function", which is called at the following times : at the beginning of the calculation at the beginning of each timestep at the end of each timestep at the end of the calculation

This allows to perform a number of operations, as defined by the user, at these different moments. As for the User Functions, the "internal functions" allow to retrieve the values of the main fields (e.g. temperature, fraction of solid, velocities, ...) at any location of the model. Such External calls can be used for instance to couple ProCAST with other softwares. The Template of the External Function is available in the next section.

Version 2006.0

Page 309

ProCAST User Manual

Language and Compiler

All the User Functions and External Functions should be written in C language, based upon the provided Templates. In order to be able to use User Functions, a C compiler and linker should be available on the machine. Specific compiler and linker are needed (only the compiler/linker which have been used for the creation of the executables are guaranteed to work with the software). There is no guarantee that the software will work with other compilers/linkers.
Windows

On Windows, the compiler and linker of the Microsoft Visual C++ 6.0 package has been used to create the executables. An "reduced" version of this compiler and linker, called "Microsoft Visual C++ 2005 Express Edition" is available for free on the web (information valid in April 2006). It can be downloaded from the following link : http://msdn.microsoft.com/vstudio/express/visualc/download/ If one would like to use ProCAST with "Microsoft Visual C++ 2005 Express Edition", one should "manually" set the following (if the software is installed in the "C:\Program Files\Microsoft Visual Studio 8\" directory (default location)) : Add to the PATH environment variable, the location of the "bin" directory of the Visual tool kit (e.g. "C:\Program Files\Microsoft Visual Studio 8\VC\bin") Add to (or create) the LIB environment variable, the location of the "lib" directory of the Visual tool kit (e.g. "C:\Program Files\Microsoft Visual Studio 8\VC\lib") Add to (or create) the INCLUDE environment variable, the location of the "include" directory of the Visual tool kit (e.g. "C:\Program Files\Microsoft Visual Studio 8\VC\include") The version of April 2006 is containing a bug which requires to copy the "mspdb80.dll" from the "C:\Program Files\Microsoft Visual Studio 8\Common7\IDE" directory to the "C:\Program Files\Microsoft Visual Studio 8\VC\bin" directory. Otherwise the compiler and linker will not work.
Linux

Intel C compiler (icc), version 8.1

SGI

C compiler version 7.3.1

IBM

C compiler version 6

Version 2006.0

Page 310

ProCAST User Manual

SUN

C compiler version 5.4

HP (HP-UX 11.11)

C compiler version B11.11.06

HP (HP-UX 11.22)

C compiler version A.05.50

Use of User Functions

Once the User Functions are defined (i.e. programmed), they should be placed in the current working directory (i.e. next to the standard input files). Then, when the "procast" executable is launched (either from the Manager, or "manually" in a Command window), the User Functions are automatically compiled and linked. A local DLL is created, as well as a local executable. Then, this executable is automatically launched and the calculation starts. This DLL and local executables are automatically deleted at the end of the execution (if the calculation is stopped manually or if it crashes, these files will remain in the local directory).

Units
ProCAST allows to define the inputs with almost any kind of units. In order to define which units should be taken into account into the software, two possibilities are provided to the user. 1. Default units can be specified in an installation file (either the main central installation or in the local user preference file (see the "customized installation" section for more details). These units will be used in all the cases run by the user. This ASCII file is called "UserFunctions_units.dat" and is located in the "dat/pref" directory. 2. Specific units can be used for a given case. These specific units should be specified in a text file in the local execution directory. This ASCII file is called "prefix_units.dat". In both cases, the same units should be used in all the User Functions of the same case. The files "UserFunctions_units.dat" or "prefix_units.dat" have exactly the same format as follows :
time 1 length 1

Version 2006.0

Page 311

ProCAST User Manual

temperature 2 velocity 1 heatflux 1 heattransfercoefficient 1 massflowrate 1

Each type of variable available in user functions are mentioned in the above list. Then, a unit code is following (as an integer value). The above values correspond to the default units (set at the installation) used by the user routines of the ProCAST solver (SI Units and degree Celsius). One can change these units, with a text editor, using the following nomenclature (corresponding to the standard unit codes of ProCAST, used in the d.dat file) :
temperature 1 = Kelvin 2 = Celsius 3 = Fahrenheit length 1 2 3 4 5 velocity 1 2 3 4 5 6 7 8 8 time 1 = sec 2 = min heatflux 1 2 3 4 5 6 7 8 = = = = = = = = W/m2 cal/cm2/sec cal/mm2/sec Btu/ft2/sec Btu/in2/sec cal/cm2/min Btu/ft2/min Btu/in2/min = = = = = = = = = m/s cm/s mm/s ft/s in/s m/min cm/min ft/min in/min = = = = = m cm mm ft in

heattransfercoefficient

Version 2006.0

Page 312

ProCAST User Manual

1 2 3 4 5 6 7 8

= = = = = = = =

W/m2/K cal/cm2/C/sec cal/mm2/C/sec Btu/ft2/F/sec Btu/in2/F/sec cal/cm2/C/min Btu/ft2/F/min Btu/in2/F/min Kg/sec g/sec lb/sec Kg/min g/min lb/min

massflowrate 1 = 2 = 3 = 4 = 5 = 6 =

User Functions Templates

The following section is presenting the templates of all the User and External Functions. The meaning of the arguments are described in the comments of the functions. External heat transfer coefficient Function (convehtransfer.c) External temperature Function (texternal.c) Emissivity Function (emissivity.c) Heat flux Function (heatflux.c) Interface heat transfer coefficient Function (interhtransfer.c) Mass Source Flow Rate Function (masssourceflowrate.c) X-component Mass Source Vector Function (xmasssource.c) - same for Y and Z X-component Translation Vector Function (xtranslation.c) - same for Y and Z X-component Imposed Velocity Vector Function (vximposed.c) - same for Y and Z X-component Solid Transport Velocity Vector Function (vxsolidtransport.c) same for Y and Z External Function (externalcompute.c)

External heat transfer coefficient Function

This function is called : convehtransfer.c

It is used to define a convective heat transfer coefficient on the outside of a domain, to be used in conjunction with an external temperature.

Version 2006.0

Page 313

ProCAST User Manual

#include <stdio.h> #include <stdlib.h> #define real double #ifdef WIN32 #define EXPORT _declspec(dllexport) EXPORT real func_convehtransfer(char*, int, real, real, real, real, real, real, int); #else real func_convehtransfer(char*, int, real, real, real, real, real, real, int); #endif extern extern extern extern extern extern /* * */ real usertemp1(int); real userfs1(int); real uservx1(int); real uservy1(int); real uservz1(int); int nodNum (real,real,real,int,real*,real*,real* );

convective heat transfer coefficient (applied on external surfaces)

real func_convehtransfer( char prefix[], /* case name */ int dimension, /* 2 = 2D ; 3 = 3D */ real temp, /* current temperature */ real fs, /* current fraction of solid */ real time, /* current time */ real x_coor, /* local coordinates: x */ real y_coor, /* local coordinates: y */ real z_coor, /* local coordinates: z */ int numBC) /* boundary condition ID number */ { /* ------------- Do not change anything above this line ------------- * * ------------- Program your function below this line ------------- */

/* ------------ Do not forget to remove the call to exit ------------ * * ------------ hereafter before running the calculation ------------ */ printf("---> exit in C user function convehtransfer <---\n"); exit(1);

External temperature Function

This function is called : texternal.c

It is used to define an external temperature on the outside of a domain, to be used in conjunction with a convective heat transfer.
#include <stdio.h> #include <stdlib.h>

Version 2006.0

Page 314

ProCAST User Manual

#define real double #ifdef WIN32 #define EXPORT _declspec(dllexport) EXPORT real func_texternal(char*, int, real, real, real, real, real, real, int); #else real func_texternal(char*, int, real, real, real, real, real, real, int); #endif extern extern extern extern extern extern /* * */ real usertemp1(int); real userfs1(int); real uservx1(int); real uservy1(int); real uservz1(int); int nodNum (real,real,real,int,real*,real*,real* );

ambient temperature (applied on external surfaces)

real func_texternal( char prefix[], /* case name */ int dimension, /* 2 = 2D ; 3 = 3D */ real temp, /* current temperature */ real fs, /* current fraction of solid */ real time, /* current time */ real x_coor, /* local coordinates: x */ real y_coor, /* local coordinates: y */ real z_coor, /* local coordinates: z */ int numBC) /* boundary condition ID number */ { /* ------------- Do not change anything above this line ------------- * * ------------- Program your function below this line ------------- */

/* ------------ Do not forget to remove the call to exit ------------ * * ------------ hereafter before running the calculation ------------ */ printf("---> exit in C user function texternal <---\n"); exit(1); return 0; }

Emissivity Function
This function is called : emissivity.c

It is used to define an emissivity on the outside of a domain.

#include <stdio.h> #include <stdlib.h> #define real double #ifdef WIN32 #define EXPORT _declspec(dllexport)

Version 2006.0

Page 315

ProCAST User Manual

EXPORT real func_emissivity(char*, int, real, real, real, real, real, real, int); #else real func_emissivity(char*, int, real, real, real, real, real, real, int); #endif extern extern extern extern extern extern /* * */ real usertemp1(int); real userfs1(int); real uservx1(int); real uservy1(int); real uservz1(int); int nodNum (real,real,real,int,real*,real*,real* );

emissivity coefficient (applied on external surfaces)

real func_emissivity( char prefix[], /* case name */ int dimension, /* 2 = 2D ; 3 = 3D */ real temp, /* current temperature */ real fs, /* current fraction of solid */ real time, /* current time */ real x_coor, /* local coordinates: x */ real y_coor, /* local coordinates: y */ real z_coor, /* local coordinates: z */ int numBC) /* boundary condition ID number */ { /* ------------- Do not change anything above this line ------------- * * ------------- Program your function below this line ------------- */

/* ------------ Do not forget to remove the call to exit ------------ * * ------------ hereafter before running the calculation ------------ */ printf("---> exit in C user function emissivity <---\n"); exit(1); return 0; }

Heat flux Function

This function is called : heatflux.c

It is used to define a heat flux on the outside of a domain.

#include <stdio.h> #include <stdlib.h> #define real double #ifdef WIN32 #define EXPORT _declspec(dllexport) EXPORT real func_heatflux(char*, int, real, real, real, real, real, real, int); #else real func_heatflux(char*, int, real, real, real, real, real, real, int); #endif

Version 2006.0

Page 316

ProCAST User Manual

extern extern extern extern extern extern /* * */

real usertemp1(int); real userfs1(int); real uservx1(int); real uservy1(int); real uservz1(int); int nodNum (real,real,real,int,real*,real*,real* );

heat flux coefficient (applied on external surfaces)

real func_heatflux( char prefix[], /* case name */ int dimension, /* 2 = 2D ; 3 = 3D */ real temp, /* current temperature */ real fs, /* current fraction of solid */ real time, /* current time */ real x_coor, /* local coordinates: x */ real y_coor, /* local coordinates: y */ real z_coor, /* local coordinates: z */ int numBC) /* boundary condition ID number */ { /* ------------- Do not change anything above this line ------------- * * ------------- Program your function below this line ------------- */

/* ------------ Do not forget to remove the call to exit ------------ * * ------------ hereafter before running the calculation ------------ */ printf("---> exit in C user function heatflux <---\n"); exit(1); return 0; }

Interface heat transfer coefficient Function

This function is called : interhtransfer.c

It is used to define an interface heat transfer coefficient between two domains.

#include <stdio.h> #include <stdlib.h> #define real double #ifdef WIN32 #define EXPORT _declspec(dllexport) EXPORT real func_interhtransfer(char*, int, real, real, real, real, real, real, int); #else real func_interhtransfer(char*, int, real, real, real, real, real, real, int); #endif extern extern extern extern real real real real usertemp1(int); userfs1(int); uservx1(int); uservy1(int);

Version 2006.0

Page 317

ProCAST User Manual

extern real uservz1(int); extern int nodNum (real,real,real,int,real*,real*,real* ); /* * */ interface heat transfer coefficient

real func_interhtransfer( char prefix[], /* case name */ int dimension, /* 2 = 2D ; 3 = 3D */ real temp, /* current temperature */ real fs, /* current fraction of solid */ real time, /* current time */ real x_coor, /* local coordinates: x */ real y_coor, /* local coordinates: y */ real z_coor, /* local coordinates: z */ int numBC) /* boundary ID number */ { /* ------------- Do not change anything above this line ------------- * * ------------- Program your function below this line ------------- */

/* ------------ Do not forget to remove the call to exit ------------ * * ------------ hereafter before running the calculation ------------ */ printf("---> exit in C user function interhtransfer <---\n"); exit(1); } return 0;

Mass Source Flow Rate Function

This function is called : masssourceflowrate.c

It is used to define a variable flow rate of a mass source.

#include <stdio.h> #include <stdlib.h> #define real double #ifdef WIN32 #define EXPORT _declspec(dllexport) EXPORT real func_masssourceflowrate(char*, int, real, real, real, real, int); #else real func_masssourceflowrate(char*, int, real, real, real, real, int); #endif extern extern extern extern extern extern /* * */ real usertemp1(int); real userfs1(int); real uservx1(int); real uservy1(int); real uservz1(int); int nodNum (real,real,real,int,real* ,real* , real* );

mass source flow rate coefficient

Version 2006.0

Page 318

ProCAST User Manual

real func_masssourceflowrate( char prefix[], /* case name */ int dimension, /* 2 = 2D ; 3 = 3D */ real time, /* current time */ real x_coor, /* local coordinates: x */ real y_coor, /* local coordinates: y */ real z_coor, /* local coordinates: z */ int numMat) /* number of the domain */ { /* ------------- Do not change anything above this line ------------- * * ------------- Program your function below this line ------------- */

/* ------------ Do not forget to remove the call to exit ------------ * * ------------ hereafter before running the calculation ------------ */ printf("---> exit in C user function masssourceflowrate <---\n"); exit(1); return 0; }

Mass Source Vector Function

This function is called : xmasssource.c (same for ymasssource.c and zmasssource.c)

It is used to define the X-component of the location of a mass source.

#include <stdio.h> #include <stdlib.h> #define real double #ifdef WIN32 #define EXPORT _declspec(dllexport) EXPORT real func_xmasssource(char*, int, real, int); #else real func_xmasssource(char*, int, real, int); #endif extern extern extern extern extern extern /* * */ real usertemp1(int); real userfs1(int); real uservx1(int); real uservy1(int); real uservz1(int); int nodNum (real,real,real,int,real* ,real* , real* );

masssource vector of a domain : x - component

real func_xmasssource( char prefix[], /* case name */ int dimension, /* 2 = 2D ; 3 = 3D */ real time, /* current time */ int numMat) /* number of the domain */ { /* ------------- Do not change anything above this line ------------- *

Version 2006.0

Page 319

ProCAST User Manual

* ------------- Program your function below this line

------------- */

/* ------------ Do not forget to remove the call to exit ------------ * * ------------ hereafter before running the calculation ------------ */ printf("---> exit in C user function xmasssource <---\n"); exit(1); return 0; }

Translation Vector Function

This function is called : xtranslation.c (same for ytranslation.c and ztranslation.c)

It is used to define the X-component of the translation of a domain.

#include <stdio.h> #include <stdlib.h> #define real double #ifdef WIN32 #define EXPORT _declspec(dllexport) EXPORT real func_xtranslation(char*, int, real, int); #else real func_xtranslation(char*, int, real, int); #endif extern extern extern extern extern extern /* * */ real usertemp1(int); real userfs1(int); real uservx1(int); real uservy1(int); real uservz1(int); int nodNum (real,real,real,int,real* ,real* , real* );

translation vector of a domain : x - component

real func_xtranslation( char prefix[], /* case name */ int dimension, /* 2 = 2D ; 3 = 3D */ real time, /* current time */ int numMat) /* number of the domain */ { /* ------------- Do not change anything above this line ------------- * * ------------- Program your function below this line ------------- */

/* ------------ Do not forget to remove the call to exit ------------ * * ------------ hereafter before running the calculation ------------ */ printf("---> exit in C user function xtranslation <---\n"); exit(1); return 0;

Version 2006.0

Page 320

ProCAST User Manual

Imposed Velocity Vector Function

This function is called : vximposed.c (same for vyimposed.c and vzimposed.c)

It is used to define the X-component of an imposed velocity BC.

#include <stdio.h> #include <stdlib.h> #define real double #ifdef WIN32 #define EXPORT _declspec(dllexport) EXPORT real func_vximposed(char*, int, real, real, real, real, real, real, int); #else real func_vximposed(char*, int, real, real, real, real, real, real, int); #endif extern extern extern extern extern extern /* * */ real usertemp1(int); real userfs1(int); real uservx1(int); real uservy1(int); real uservz1(int); int nodNum (real,real,real,int,real* ,real* , real* );

ambient temperature (applied on external surfaces)

real func_vximposed( char prefix[], /* case name */ int dimension, /* 2 = 2D ; 3 = 3D */ real temp, /* current temperature */ real fs, /* current fraction of solid */ real time, /* current time */ real x_coor, /* local coordinates: x */ real y_coor, /* local coordinates: y */ real z_coor, /* local coordinates: z */ int numBC) /* number of boundary condition */ { /* ------------- Do not change anything above this line ------------- * * ------------- Program your function below this line ------------- */

/* ------------ Do not forget to remove the call to exit ------------ * * ------------ hereafter before running the calculation ------------ */ printf("---> exit in C user function vximposed <---\n"); exit(1); return 0; }

Version 2006.0

Page 321

ProCAST User Manual

Solid Transport Velocity Vector Function

This function is called : vxsolidtransport.c (same for vysolidtransport.c and vzsolidtransport.c)

It is used to define the X-component of transport velocity of a solid.

#include <stdio.h> #include <stdlib.h> #define real double #ifdef WIN32 #define EXPORT _declspec(dllexport) EXPORT real func_vxsolidtransport(char*, int, real, real, real, real, real, real, int); #else real func_vxsolidtransport(char*, int, real, real, real, real, real, real, int); #endif extern extern extern extern extern extern /* * */ real usertemp1(int); real userfs1(int); real uservx1(int); real uservy1(int); real uservz1(int); int nodNum (real,real,real,int,real* ,real* , real* );

ambient temperature (applied on external surfaces)

real func_vxsolidtransport( char prefix[], /* case name */ int dimension, /* 2 = 2D ; 3 = 3D */ real temp, /* current temperature */ real fs, /* current fraction of solid */ real time, /* current time */ real x_coor, /* local coordinates: x */ real y_coor, /* local coordinates: y */ real z_coor, /* local coordinates: z */ int numBC) /* number of boundary condition */ { /* ------------- Do not change anything above this line ------------- * * ------------- Program your function below this line ------------- */

/* ------------ Do not forget to remove the call to exit ------------ * * ------------ hereafter before running the calculation ------------ */ printf("---> exit in C user function vxsolidtransport <---\n"); exit(1); return 0; }

External Function
This function is called : externalcompute.c

Version 2006.0

Page 322

ProCAST User Manual

This function is called at the start of the calculation, at the start of each timestep, at the end of each timestep and at the end of the calculation.
#include <stdio.h> #include <stdlib.h> #define real double #ifdef WIN32 #define EXPORT _declspec(dllexport) EXPORT void func_externalcompute(char*,int,int,real); #else void func_externalcompute(char*,int,int,real); #endif extern extern extern extern extern extern /* * * * * * * * */ real usertemp1(int); real userfs1(int); real uservx1(int); real uservy1(int); real uservz1(int); int nodNum (real,real,real,int,real*,real*,real* );

function called at the beginning, the end of the calculation and at the beginning of each timestep loop = 0 : function is called at the start of the computation loop = 1 : function is called at the start of the current timestep loop = 2 : function is called at the end of the current timestep loop = 3 : function is called at the end of the computation

void func_externalcompute( char prefix[], /* case name */ int loop, /* loop value : 0/1/2/3 */ int timestep, /* current timestep */ real time) /* current time */ { /* ------------- Do not change anything above this line ------------- * * ------------- Program your function below this line ------------- */ }

Version 2006.0

Page 323

ProCAST User Manual

RUN OF THE CALCULATION

SOLVER
Once the problem is set-up in PreCAST, two operations should be performed in order to run the solver : Run DataCAST Run ProCAST

DataCAST
DataCAST converts the input data stored in ASCII in the prefixd.dat file into binary data ready for the solver. DataCAST also cleans any previously existing result files and prepare the new ones. This means that when DataCAST is run, all results will be erased. Some error checking is also performed in DataCAST. DataCAST is called with the DataCAST button of the Software manager. When DataCAST is finished, one should press Return to close the corresponding command window (see the Module calls section of the Software Manager chapter for more details). See the "DataCAST -u" option at the end of this section.

ProCAST
ProCAST is running the solver. No message is displayed by ProCAST, except if the solver is stopping for an unexpected reason. ProCAST is called with the ProCAST button of the Software manager. When ProCAST is finished, one should press Return to close the corresponding command window (see the Module calls section of the Software Manager chapter for more details).

Status

Version 2006.0

Page 324

ProCAST User Manual

During the calculation, the status of the model can be monitored. To do so, use the "Status" button of the Manager (see the Module calls section of the Software Manager chapter for more details). The prefix of the case to be monitored should be present in the "Case" field and when the Status button is pressed, a window with the calculation status is displayed :

The status information is stored in the software installation directory (in the dat/stat directory), under the name prefix.stat. This means that the status of any calculation is available at any time, by any user who is using the same software installation. As a consequence, if two calculations with the same prefix are run in the same time, there will be a conflict and the status will be once the one of calculation A and once of calculation B. Such a conflict however has no influence at all on the calculations themselves.

Batch calculations
It is possible to run several calculations sequentially in batch mode. To do so, a simple script should be written and launch in a Command window, as shown in the example hereafter (for Windows) :
cd calculations_A cd case_1 %ProCAST20060%\bin\procast case_1 cd .. cd case_2 %ProCAST20060%\bin\procast case_2 cd .. cd .. cd calculations_B cd case_3 %ProCAST20060%\bin\procast case_3 cd .. cd ..

Version 2006.0

Page 325

ProCAST User Manual

DataCAST -u option
It is possible to run a calculation until a given step (e.g. step 500) with given settings and then to continue the calculation with different settings (e.g. if one would like to test different heat transfer coefficients from step 500). To do so, run the calculation as usual until step 500. Then, in order to prevent any error, copy the results in a safe place (as if a wrong operation is performed with DataCAST, all the results files will be erased). Load the d.dat file in PreCAST and make the desired changes (e.g. change a heat transfer coefficient). Change the Run parameter INILEV to 500 and set DT = 0 (so that the last stored timestep will be used as the first one of the restart) and exit PreCAST. Then, in order to update the calculation binary files, DataCAST has to be run. However, in order to prevent DataCAST to erase the existing result files, it is necessary to activate the -u option.

To use the "-u" option of DataCAST, the corresponding check box should be activated (see figure above). One could also launch it from a Command window with the following syntax :
datacast prefix -u

Then, the solver ProCAST can be launched and the calculation will be continued from the step 500, with the new settings. Please note that it is not necessary to run DataCAST if only the Run Parameters are changed (i.e. only the content of the p.dat file is changed).

Version 2006.0

Page 326

ProCAST User Manual

TROUBLESHOOTING
During the execution of the solver, information are stored in the prefixp.out file. In case of problem, it may be useful to check the convergence of the solver in this file. At each timestep, a number of information are printed in the p.out file. The most important ones (hereafter in the case of a thermal only calculation) are the step number (1), the timestep (2), the fraction which has already solidified (3) and the iterations (4). If a calculation seems not to run well, it can be useful to look at the evolution of the timesteps, during the course of the calculation. If the timestep is significantly reduced, it means that the convergence of the calculation is poor. In addition, the number of iterations per timesteps (i.e. the number of line shown in (4) in the figure below, "T" indicating the resolution of the Thermal model) is a good indication of the quality of the convergence. The least number of iterations, the better the convergence.

In the case of a mold filling calculation, the prefixp.out file is looking like the following figure. One can check that the "free surface model" is well activated (1), that the model is filling with the filling percentage (2). The iterations (3) for the Thermal model ("T"), for the pressure solver ("P") and for the three components of the velocity ("U", "V" and "W") are shown. When more than 5-10 iterations are necessary, one can anticipate a convergence problem. Finally, the memory used by the calculation (for the current timestep) is shown (4).

Version 2006.0

Page 327

ProCAST User Manual

One of the Run Parameter is COURANT. This parameter is limiting the timestep in order to prevent the free surface to advance more than N elements during one timestep. When the mesh has a large number of elements and when the elements are small, this limitation may be quite drastic for small COURANT values. When this limit is reached, the following message appears in the prefixp.out file :

Version 2006.0

Page 328

ProCAST User Manual

The Run Parameter LVSURF allows to switch off the filling calculation when the casting fill fraction has reached the LVSURF value. A message is printed in the prefixp.out file to acknowledge this switch.

The fluid flow model will still be solved until the next stored timestep and then it will be switched off (see figure below). The calculation will then continue with a thermal only calculation (and stress if activated).

Version 2006.0

Page 329

ProCAST User Manual

When the calculation is not converging, the following message is printed in the prefixp.out file and the calculation of the given timestep is restarted with a reduced timestep. This can happen a few times during the calculation and one should not worry about it (typically, if the inlet velocity is suddenly increased, the timestep may be too large and it should be reduced through this mechanism). However, if the calculation does not converge several consecutive times, it may indicate a problem in the problem set-up or in the input data (non relevant material properties for instance).

Version 2006.0

Page 330

ProCAST User Manual

RESULTS VIEWING
INTRODUCTION
Once the calculation is performed, the results can be viewed with the ProCAST post-processor, called ViewCAST. ViewCAST can be called from the Manager as follows :

The following results can be visualized in the post-processor : snapshot (or maps) slices (or cuts) scan of slices x-ray views evolutions

Depending upon the type of calculation which was performed, the following entities (or variables) can be viewed (raw results) : Scalar fields : Temperature Fraction of solid

Version 2006.0

Page 331

ProCAST User Manual

Heat flux magnitude Velocity magnitude Pressure Stresses Strains and displacements ...

Vector fields : Velocities Heat fluxes

This chapter is describing how these results can be viewed. In the Result analysis chapter, the display of more entities, such as porosity, solidification time, secondary arm spacings, etc... will be described. The export of the calculated results, to be used in other softwares are described in the Results Exports chapter. Once ViewCAST is started, if the prefix of the case is well defined, the case is automatically loaded in the post-processor and the initial temperature field is displayed :

The background of the screen can be switched from black to white in the "Parameters/Reverse Video" menu :

Version 2006.0

Page 332

ProCAST User Manual

The screen is divided in different zones : 1. Display zone 2. Scale zone 3. Information zone 4. Menus 5. Icons 6. Tape player The post-processor is "driven" by the Menus, the icons and the Tape player :

Firstly, the case should be opened in the File menu (if it is not opened automatically when the ViewCAST is launched from the Manager). Then, the field (scalar) to be displayed (e.g. Temperature) is selected in the Contour menu. In the case of vectors, the desired field is selected in the Vector menu - see the Field selection section for more details. The times (or steps) to be displayed are selected in the Steps menu, whereas all the parameters to configure the display (e.g. symmetry, cut-off values, type of display, ...) are defined in the Parameters menu - see the Display parameters section for more details. Then, the Picture menu allows to select the type of

Version 2006.0

Page 333

ProCAST User Manual

display (i.e. snapshot, slice, scan slices or cut-off view) - see the Display types section for more details. Once these selections are made, the display is activated with the Tape player, either as individual pictures or frame, or as animations (see the Tape Player section for more details). The X-Y Plots menu allows to draw curves (time evolutions) of the different fields (see the Curves section). The Actions menu allows to calculate some parameters from the raw results (such as SDAS, isochrons, Niyama, etc...) - see the Results analysis chapter for more details. Please note that the Actions and X-Y Plots menus are replacing the former module PostCAST (version 400 and before), as well as the File/Export menu. The Geometry manipulation, as well as the different viewing modes is described in the Geometry manipulation section. To File/Exit menu allows to quit the post-processing session. Upon Exit, the user is asked whether he would like to keep the information about the current orientation of the model as well as the definition of the material display (i.e. which materials are active and in which mode) for the next launch of the post-processor.

GIF/AVI capture
All the pictures which are displayed in ViewCAST can be captured as GIF/AVI files or GIF/AVI animations. To do so, the opens the following window icon should be used, which

Version 2006.0

Page 334

ProCAST User Manual

The file name can be specified and one can browse in the desired directory (by default it is stored in the working directory and the file name corresponds to the case prefix) - the suffix (.gif or .mpg) is optional. One has to select between a Still picture or an Animated picture and to select between the GIF or the AVI format. It is also possible to activate the Reverse Video (with the "Reverse Video" check box) in order to have a white background. Then the picture is stored when the "Apply" button is pressed. For an animation, it will start to store the animation when the Forward button of the Tape player is activated. The "Cancel" button allows to exit without any action. If the filename is already existing, the following prompt window is shown :

One has the choice between "Append" (i.e. to append or add the picture to the previous one(s)), to "Overwrite" (i.e. to replace the existing file by this one) or "Cancel" to abort the capture operation. The AVI files are compressed using a "Motion jpeg" compressor. It has been observed that with old versions of Powerpoint, such "Motion jpeg" files were not recognized, unless they are renamed as prefix.mpg (instead of prefix.avi).

Version 2006.0

Page 335

ProCAST User Manual

When an animation is stored in the GIF format, it can be interrupted at any time (and the GIF animation will be stored until the interruption). However, it is not possible to interrupt an animation when an AVI file is created. If the animation is interrupted before the end, the AVI file will not be created. Thus, one has to specify the starting and ending steps before the creation of an AVI animation.

Version 2006.0

Page 336

ProCAST User Manual

FIELD SELECTION
The entities to be displayed have to be selected in either the Contour or the Vector menu. Contour corresponds to scalar variables, such as Temperature, Fraction of solid, Heat flux intensity, Pressure, Velocity magnitude, Principal stress, Xdisplacement, etc... Vector corresponds to vector variables, such as Velocities or Heat fluxes. It is possible to display either one Contour or one Vector field at a time, or to superimpose one Contour field, with one Vector field.

Contour
The Contour menu is divided according to the different models which can be activated in the ProCAST solvers. The different "Contour" fields which are available are shown hereafter

One should note that the "Solidification Time" corresponds to elapsed time between the liquidus and the solidus temperatures. The Shrinkage Porosity is only calculated in the Run Parameter POROS is set to a non-zero value.

Version 2006.0

Page 337

ProCAST User Manual

As the stresses (i.e. the stress tensor) are calculated at Gauss points in the solver, it is necessary to compute in the post-processing the different stress entities that can be viewed (e.g. Effective stress, Principal Stress, Sigma X, ...). As these calculations can be rather long, it is possible to compute all the desired one in one shot, with the "Prepare Stress Results" option. In this case, the following panel appears :

Version 2006.0

Page 338

ProCAST User Manual

The stress result files which are present and up-to-date are shown in Green. The ones which are not present are shown in Red. If there are files which are present but not up-to-date (i.e. these files are older than the latest computed stress results by the solver), they appear in Orange. In this panel, it is possible to select all the desired stress results to be computed and to press "Apply" to start the computation. The results which are in green will not be recomputed (unless the "Overwrite Existing Files" is checked).
Remarks on the Stress results

The "Average Normal Stress" corresponds to the average of the normal stresses along the X, Y and Z planes. This does not correspond to a stress normal to an external surface (thus it is possible that the Average Normal Stress is not zero on a casting surface). The "Displacement -> Plane" option corresponds to the difference (between the initial timestep and the selected timestep) of distance between a User specified plane and a given point (the distance is taken normal to the plane). When Stress results are to be viewed from a CD-ROM (on Windows only), the calculated stress results (e.g. Principal Stress 1) can not be stored on the CD (as it is Read-only). Thus, when the directory where the case is located is in Readonly mode, automatically, the stress results are created and stored in a local file

Version 2006.0

Page 339

ProCAST User Manual

on the PC and are automatically used (even in a latter post-processing session). These results are stored in a directory named : C:\Documents and Settings\USER\Local Settings\Temp\Prefix . One should just be careful that if two cases having the same prefix are visualized from a CD, the data will be overwritten. Please note that the fields which are shown in the menus correspond only to the available result files in the working directory. For instance, if results are computed in the "Action" menu - such as Niyama (see the Result analysis chapter for more details), the corresponding item will be added in the menu.

Vector
Vector results, such as Heat Flux and Fluid velocity can be selected in the Vector menu.

It is possible to view both vectors and contours, if desired. In this case, the vectors will be displayed in white (respectively black). If vectors only are shown, the color of the vector can be set in the Parameters menu (see the Display Parameters section for more details) to the vector magnitude, to the temperature or the pressure. In order to shown the vectors only, the Contour/None option should be selected.

Version 2006.0

Page 340

ProCAST User Manual

DISPLAY TYPES
The type of display can be selected either with the following icons

or with the Picture Menu

Snapshot - 3D view of the selected field (e.g. Temperature), displayed on the surface of the selected domains. For filling the free surface is viewed in 3D. This mode can be accessed directly with the F5 key. Slice - 2D cuts (shown within the 3D model). One or several slices can be selected, either along the X, Y and/or Z planes or along any plane. This mode can be accessed directly with the F6 key. Scan - Scan of 2D slices along either the X, Y or Z axis. This mode can be accessed directly with the F7 key. CutOff - X-Ray view of the inside of a model. The model is made partially transparent (according to a criterion such as above a given temperature) in order to see features which are inside the model, such as pockets of liquid, surrounded by a solid shell (which is made transparent). This mode can be accessed directly with the F8 key. The selected mode can be viewed by a red square around the corresponding icon.

Version 2006.0

Page 341

ProCAST User Manual

Examples of Snapshots (during filling and after), of slice and CutOff (X-Ray) view

Example of Scan slices

Each Display type, except Snapshot, has ad-hoc settings (available in the Parameters menu), as described hereafter (Slice Data, Scan Data and Scan Data).

Version 2006.0

Page 342

ProCAST User Manual

Slice definition
For Slice Data definition (see figure below) : 1. Select "Slice Data" in the Parameters menu 2. Select "Add" and then XYZ Plane (for orthogonal slices). This will open the slice selection window (3) 3. Select the X, Y or Z plane and then define the location of the plane with the slider 4. The plane is shown on the model

Version 2006.0

Page 343

ProCAST User Manual

5. Press "Apply", the corresponding selection will appear in the list (6) and the slice will be displayed. It is possible to use the Tape Player at this stage to see this slice over the timesteps, either as single pictures or as animations.

As shown in the figure below, several slices can be selected. Moreover, the slice characteristics can be stored in a file (file named "prefix.clip" in the local directory) with the "Store" button and can be retrieved in a later session with the "Read" button. The "Delete" button allows to delete the selected slice. The Show "Yes"/"No" toggle allows to activate or deactivate slices, without deleting them.

Version 2006.0

Page 344

ProCAST User Manual

In the following case, two slices are activated, as shown in the frame above.

When one slice only is activated, one can draw this slice with or without the Background (using the "Display Background" button.

Version 2006.0

Page 345

ProCAST User Manual

With "Display Background : None", only the slice is shown (upper right). With "Display Background -", the slice is shown together with the background (upper left). With the "Display Background +", the foreground is shown, up to the slice (lower left). When the model is rotated (for instance around a vertical axis in the above example), the background becomes the foreground and vice-versa. This option is not available with more than one slice. With the "Add->" button, it is also possible to select "Any Plane" :

In this case, the following window opens. One should define the slice plane by the coordinates of three points. To select interactively the points on the mesh, the "Get Co-ord" button can be used (to select the point which is highlighted in red). Once the three points are selected, it is possible to move this plane along its normal, with a given Offset.

Version 2006.0

Page 346

ProCAST User Manual

It is also possible to rotate this plane with the "Rotate" button. In this case, the following window allows to define the rotation axis, as well as the angle of rotation.

Finally, the "Copy Plane" is used to copy a plane which was already defined in order to modify its characteristics. This is especially useful when one would like to define several parallel planes which are not in the X, Y or Z directions. One will define one plane with the standard Anyplane definition procedure and then, one will copy this plane and offset it to the appropriate value. This can be repeated for the other parallel planes.

Scan definition
The definition of the Scan parameters is done in the Parameters/Scan Data menu, which opens the following window :

Version 2006.0

Page 347

ProCAST User Manual

The Scan direction is selected, as well as the number of planes (or slices) which will be shown. The largest dimension of the model in the selected direction is divided by the number of planes. Then, the display and the scan through the planes is activated with the Tape Player. Beside this slice mode, it is possible at any time to slice interactively the model, using the "Scan" slider (see figure below).

Cut-off definition

Version 2006.0

Page 348

ProCAST User Manual

The Cut-off (or X-Ray) parameters are defined in the Parameters/CutOff Data menu, which opens the following window :

The Cut-off parameters are defined by two values (e.g. two temperatures). Then the user has to select among five different display possibilities : Above Red : only the zones which are above the upper value (in red) are shown Below Blue : only the zones which are below the lower value (in blue) are shown Inbetween Bounds : only the zones which are in between the upper and the lower values are shown Outside Bounds : only the zones which are below the lower value and above the upper value are shown Isosurfaces : the isosurfaces corresponding to the two values are shown

The figure hereafter is illustrating these different possibilities.

Version 2006.0

Page 349

ProCAST User Manual

Shrinkage and Advanced Porosity display

When the Shrinkage Porosity or Advanced Porosity contours are selected for the first time, automatically, the "cut-off" mode is activated, with the definition "above 0.01" (which corresponds to a porosity value higher than 1%). Of course, it is always possible to go back to the normal snapshot mode, by clicking on the snapshot icon .

Version 2006.0

Page 350

ProCAST User Manual

DISPLAY PARAMETERS
Firstly, the timesteps to be viewed should be defined in the "Steps" menu, which opens the following window :

The user can choose between Steps and Time increments. In blue, the available minimum and maximum values are shown. The time values are always indicated in seconds. If the selected Step increment is smaller than the available stored one, it will take the first available step for the display. If ViewCAST is launched during a calculation, the "Update" button allows to "refresh" the available steps and to view what has been calculated between the launch of ViewCAST and the current instant. One can also set "ViewCAST" such that the update is made in a "continuous" mode during the calculation (this is useful when ViewCAST is launched while the ProCAST calculation is still running). To do so, the "Setup" button (of the above screen) should be pressed and the Update interval can be specified (please note that if it is intended to leave the ViewCAST session opened for a few hours, it is advisable to set an Update interval not too small, as this will load the CPU) :

Version 2006.0

Page 351

ProCAST User Manual

The Display parameters are defined in the Parameters menu :

Reverse Video
By default, the background of the screen is black. It can be reversed to white with the Reverse Video option.

Free Surface
To visualize a mold filling four options are available to view the free surface (i.e. the surface between the liquid metal and the air) :

Version 2006.0

Page 352

ProCAST User Manual

The option "Free Surface (Foreground)" is used to view the free surface even when it is hidden by the liquid or the rest of the casting. One can see in the upper figure hereafter (Back face) the free surface in the ingate. If we rotate the casting (lower figures), the free surface is behind. If the Free Surface is ON, we do not see it. If the Free Surface is set as "Foreground", one can see it although it is behind. Of course, one can also see ONLY the free surface.

Version 2006.0

Page 353

ProCAST User Manual

The "Foreground" option can be quite useful to view entrapped gas pockets (which are fully surrounded by liquid).

Enclosure
In the case of a radiation problem with an enclosure (this is valid only for surface enclosures as solid ones are considered as materials) the display mode of the enclosure can be selected among six choices : Invisible, Wireframe, Hidden, Solid, Translucent or Shade :

The following figure shows an example of a shaded enclosure.

Version 2006.0

Page 354

ProCAST User Manual

Feature Angle
When the geometry is displayed in Wireframe, one can tune the "density" of lines. In the following example, a Feature Angle of 10 and 60 degrees has been used.

Version 2006.0

Page 355

ProCAST User Manual

Vector settings
When vectors are selected, it is possible to color them according to the vector magnitude or according to Temperature or Pressure. By default, the color is White. When the "Parameters/Vector Settings" menu is called, the following panels opens.

Version 2006.0

Page 356

ProCAST User Manual

One can choose between an automatic vector length or a manual vector length. It is also possible to define a Unique vector length (i.e. all the vectors will be displayed with the same length, regardless of the vector magnitude (except if it is zero). When vectors are shown together with a Contour Plot (i.e. a scalar variable such as Temperature), it is not possible to color the vectors. The vectors can be plotted with or without the arrow at the tip of the vector.

Displacement Mag.
In the case of Stress calculations, the display of the deformations can be artificially increased in order to better view them. The real displacements are multiplied by the "Displacement Magnitude" value.

Version 2006.0

Page 357

ProCAST User Manual

Slice Data, Scan Data, CutOff Data

These options are described in the "Display types" section.

Titles
Titles (text), arrows and cirlces can be added to the graphics, using the Parameters/Titles menu, which opens the following window :

Then, one can "Add ->" either a String (text), an Arrow or a Circle. When a String is selected, the text has to be typed in the corresponding field and then the text should be placed with the mouse. When an Arrow is selected, it should be located with the mouse. The first click should be at the starting point of the line and the second click should correspond to the end of the arrow. For circle, the two opposite corners of the square enclosing the desired circle should be defined. To delete an item, click on the desired one (in the left grey zone), which will highlight it in red and then press the Delete button. All the information of this panel can be Stored in an ASCII file and retrieved (with the Read button) for later use.

Version 2006.0

Page 358

ProCAST User Manual

Symmetry
For cases which have been calculated with symmetry, it is possible to "reconstruct" the full geometry, using the symmetry menu :

The figure below illustrate a mirror symmetry. For more details about the definition of the symmetries, please refer to the Radiation section of the Preprocessing chapter.

Version 2006.0

Page 359

ProCAST User Manual

Display Foam
In the case of a lost foam calculation, one can display or not the foam with this toggle menu.

Display Pipe
In case of porosity, if piping or macroshrinkage occurs, the corresponding volumes will be shown as empty. However, in the visualization of the "Shrinkage porosity", it may be useful to view together the porosity and the pipe shrinkage. To do so, the "Display Pipe" OFF mode will allow to see the empty regions as "Shrinkage porosity" with a value of 1 (i.e. 100% void). This mode is only applicable for the Contour "Shrinkage porosity" and will not affect the viewing of the other fields, such as the Temperature, Fraction of solid, ...

Particle tracing
The particle tracing allows to follow streamlines during the filling. The particle traces are calculated during the filling calculation (as it is needed to have access to all the timesteps, which is not the case during the post-processing). To do so,

Version 2006.0

Page 360

ProCAST User Manual

"virtual particles" are launched at each inlet node (defined by a velocity BC, an inlet pressure BC or an Inlet BC) at given intervals specified by PFREQ (see "Flow Run Parameters"). This means that it is not possible to compute particle traces in the case of tilt casting or casting with a filled reservoir of material (as there are no inlet boundary conditions where particles could be launched). Moreover, if the inlet is applied to a few nodes only, the number of traces is not very significant. This will be improved in future versions with more sophisticated means to launch particles. To view the Particle traces, activate "Particle Tracking" in the "Parameters" menu. The traces can be viewed either alone or together with any contour or vectors. One should note that the traces are always displayed in the foreground of a contour plot (no hidden view of the traces). In order to allow the viewing of successive batches of particles, the "tail" of the traces are erased, as the traces are progressing, leaving the place to new traces.

Version 2006.0

Page 361

ProCAST User Manual

Depending upon the geometry and the mesh size, it is possible that one or two particles are going out of the model. Please note that this does not harm the other results.

Tilt
In case of a tilt pouring calculation, it is possible to disable the viewing of the model rotation. By default, the display of the rotation is active.

Display Undeformed

Version 2006.0

Page 362

ProCAST User Manual

In a Stress calculation, it is possible to superimpose the Undeformed geometry (in wireframe mode) to the deformed geometry. The extent of the part deformation can be adjusted with the "Displacement Mag." option. If the wireframe has "too many lines", it is possible to adjust that with the "Feature Angle" option.

Version 2006.0

Page 363

ProCAST User Manual

Scale
The scale can be modified by a direct click on the scale itself (1). Then, the Scale settings panel opens (2).

Version 2006.0

Page 364

ProCAST User Manual

The scale can be defined either automatically, or semi-automatically, via the "Min-Max" or "Semi-Auto" settings. It is also possible to change each value independently (with the Manual field), by clicking on each value to be changed.

Version 2006.0

Page 365

ProCAST User Manual

For the temperature scale, if there is a material where the solidification range is defined (i.e. a casting material), there is also the possibility to select automatically the Solidus-Liquidus interval (Tsol-Tliq). In the same line, when the Temperature is displayed, the Solidus and Liquidus temperatures are indicated on the color scale :

Version 2006.0

Page 366

ProCAST User Manual

It is possible to deactivate this display by unchecking the "Display Tsol-Tliq" in the above window. When a scale was defined other than "Automatically", the minimum and maximum values are stored for each field (the storage is done upon exit of the model if the "Store the last view and Exit" option is selected. When the model is loaded the next time, these settings will be automatically retrieved. The values are indicated in the "Stored" field (the values can not be modified interactively, use the Min-Max fields above). The values are stored in a local file called prefix.scale (this file can be copied in an other case if one wants to use the same values for this new model). Finally, the color scale can be change, by clicking on the color map itself. The following panel opens and the colors can be changed. It is also possible to store and load a different color scale.

Version 2006.0

Page 367

ProCAST User Manual

One should note that the scale and the color map is stored in memory for each field (e.g. temperature, fraction of solid, velocity magnitude, ...). Thus, if the scale is changed for the display of temperature and then fraction of solid is shown, when the temperature field is displayed again, the modified scale is kept.

Version 2006.0

Page 368

ProCAST User Manual

TAPE PLAYER
To activate the display, the Tape Player must be used. This allows to display either individual pictures or animations.

As for a Tape recorder, the Tape Player has the following functions. Displays the first define time or time-step (Goto first) Plays the animation backwards (Rewind) Displays the previous step (Step back) Pause (stops the animation) Display the next step (Step forward) Plays the animation forwards (forward) Display the last define time or time-step (Goto end) The definition of the first and last times or time-steps, as well as the time or timestep increment between two frames is made in the Steps menu (see the Display parameters section for more details).

Version 2006.0

Page 369

ProCAST User Manual

CURVES
Curves can be viewed in ViewCAST, using the X-Y Plots menu. This allows to visualize time evolutions of the calculated fields, such as Temperature, pressure, velocity, etc... Please note that these capabilities are replacing the former PostCAST module.

Three types of selections are available : Interval - selection of the curves every N nodes Nodes - selection node by node External - loading of a set of external curves (e.g. measurements)

Interval
When Interval is selected, the user has to specify the Nodal interval. This means that curves will be selected every N nodes (1, 501, 1001, 1501, ...). Only the nodes of the materials which are active (as Solid) will be selected.

In addition, the user can select the Units which will be used for the display of the curves. When the Apply button is pressed, the curves are read in the files and displayed.

Version 2006.0

Page 370

ProCAST User Manual

Nodes
When "Nodes" is selected, the following node selection window appears :

Version 2006.0

Page 371

ProCAST User Manual

The node numbers can be entered manually, or an interactive selection can be made with the button.

In addition, the user can select the Units which will be used for the display of the curves. When the Apply button is pressed, the curves are read in the files and displayed. The node numbers (as well as the defined scale) can be stored (Export) and retrieved (Import) in a file. To do so, the "File" button should be pressed and the following window is opening :

In the case of an Export, the file name should be specified and it will be stored in the working directory. For Import, one can "Browse" the desired file. The format of the Export file is the following :
-1 0 3.56 10 30 10. 1550. 0

Version 2006.0

Page 372

ProCAST User Manual

100 234

The first line contains the scale. As the scale is optional, the line should start by a "-1" and should be followed by xmin, xmax, ymin, ymax. Then, there is a "0" or a "1" to specify whether the Y-scale is defined automatically of not. A value of "0" means that the specified ymin and ymax are taken into account, whereas a value of "1" means that the Y-scale is set automatically. Then, the list of the nodes should be specified. Finally, a file named "prefix.tt" is created when the curves are displayed, with the x and y coordinates of the displayed curves. These data can be then either viewed together with other results (see the External section below) or exported to spreadsheets (e.g. Excel).

External
When curves are selected with the Nodes mode (see above), the displayed curves are stored in the same time in a text file named "prefix.tt". The fist line of the file (see below) contains the number of time-steps (132), the number of stored curved (3) and the corresponding node numbers (567, 9587 and 23129). Then, each line contains the time and the field values for each curve.

Any file which as the above format can be loaded and displayed in ViewCAST with this External option :

Version 2006.0

Page 373

ProCAST User Manual

The desired files should be selected with the Browse. Please note that if the format should be as described above, the name of the file is free. Once this selection is accepted with the Apply button, the selected curves will be displayed in the same time as the ones selected with the Nodes option.

X-Y Plot settings

The display of the X-Y plots can be tuned in the X-Y Plot settings window. For the settings of the time scale, the calculated minimum and maximum values are shown in blue (1). The vertical scale can be adjusted automatically with the "Auto" check box (2).

Version 2006.0

Page 374

ProCAST User Manual

Version 2006.0

Page 375

ProCAST User Manual

GEOMETRY MANIPULATION
The geometry can be manipulated with the following icons :

Opens the Rotate definition window (see below)

Restores the original orientation of the model (with the Z-axis perpendicular to the screen)

Interactive zoom (zoom up when the mouse is move towards the bottom of the screen)

Auto-scale (or zoom out)

Interactive center of the model (the location which is clicked is moved towards the center of the screen)

Interactive drag of the model

Rotation
When the Rotate icons is activated, the following Rotate window is opened :

Version 2006.0

Page 376

ProCAST User Manual

With the X, Y and Z sliders, the rotation can be prepared. Exact angles can be manually entered in the fields on the right. The rotation will be applied only when the "Rotate" button is pressed. If the "Rotate" button is pressed again, a new rotation (with the same values) is applied again. The "Reset" button resets all the sliders to zero. By default, rotation are with positive angles. If one clicks on the "+" sign, it becomes negative (for negative rotation).

Automatic Rotation in the X-direction

, in the Y-direction

and n the

Z-direction can be activated, as well as a standard isometric view . When the left mouse key is used, the model is oriented with the X, respectively Y and Z, axis perpendicular to the screen, pointing outwards, whereas the opposite orientation (pointing inwards) can be obtained with a right mouse key click. Finally, the X, Y and Z keys can be used to rotate the model by 10 around the x, y and z axis respectively. When the <ctrl> key is pressed in the same time, the rotation direction is reversed. When the <shift> key is pressed in the same time, the rotation is by 30 instead of 10.

Interactive Rotation
At any time, the model can be rotated interactively with the mouse. The left mouse key should be pressed and the model is rotated when the mouse is moved. The center of rotation corresponds to the center of the screen (green arrow). If the "Shift" key is pressed during the rotation, the model will rotate around an axis which is perpendicular to the screen and located in the center of the screen (green cross). The rotation is activated upon the horizontal motion of the mouse (while the left mouse key is pressed).

Zoom

Version 2006.0

Page 377

ProCAST User Manual

Beside the zoom icon (which allows to increase or decrease the size of the model), the F2 and F3 keys can be used to increase, respectively decrease the size of the model (zoom in and zoom out). Moreover, with the center mouse key, it is possible to zoom interactively a specific zone of the model with a rubber band box. Once the center key is pressed, it defines one corner of the rubber band box and the opposite corner is defined by the movement of the mouse until the key is released. Then, the zoomed model is automatically displayed. It is possible to zoom the model several time using this technique. To zoom out, use the Auto-scale icon .

Stored Views
It is possible to store 6 different views in the "Parameters/Store View" menu. The views are containing the orientation of the model as well as the material selection.

When a small arrow is following the View-X, it means that a view is stored already. To store a view in a View-X which is available, just click on the

Version 2006.0

Page 378

ProCAST User Manual

Parameters/StoreView/View-X. To retrieve (or Restore) a given view, use the "Restore" option (one can also use the Ctrl-... key where ... is the number of the view (1 to 6)). With the "Replace" option, the current view will replace the one which was previously stored. The "Reset view" (or Ctrl+0) will automatically reset the orientation of the model in the isometric view with the gravity pointing downwards. These views are stored in a file named prefix.lv (only if the option "Store the last View and Exit" is selected upon exit of ViewCAST). Please note that in View-1, it is always the last view which is active on the screen upon exit which is stored (which will replace any previously stored one). Please note that this *.lv file is not compatible with versions before 2006.0.

Material selection
The Material selection icon opens the Material selection window, which allows to define which material are display as well as the type of display.

The list of the all the available materials is displayed. On the right, the colored columns allow to specify the display attribute of each material, according to the following nomenclature : SO : Solid (displays the selected field values - e.g. Temperature) IN : Invisible (hide the selected material) WI : Wireframe (displays the selected material with wireframe only) HM : Hidden mesh (displays the FEM mesh in hidden mode) SH : Shaded material

Version 2006.0

Page 379

ProCAST User Manual

TR : Transparent material

For each material, one should click in the desired column and a "X" will be displayed. To define all the materials with the same selection, one could click on the label (e.g. click on "IN" will make all the materials invisible). Finally, when the apply button is selected, the display will be updated with the corresponding selections.

Explode materials
In the Materials selection window, there is the "Explode Materials" button, which opens the following window. For each material, it is possible to define an offset for its display. If some materials have no interface between them (EQUIV), they will be considered in the same group and they can not be exploded separately.

Version 2006.0

Page 380

ProCAST User Manual

When the Apply button is pressed, the corresponding offsets are applied, leading to a picture as follows :

Display node and element numbers

The node numbers of the surface of the mesh can be viewed with the "Ctrl-N" keys. When "Shift-Ctrl-N" keys are pressed, a small window is opening :

Version 2006.0

Page 381

ProCAST User Manual

One can enter a node or an element number. By pressing the "Node" button, all the elements which are connected to the specified node are displayed :

If the "Element" button is pressed, the nodes which belong to this element are listed :

Pick option

Version 2006.0

Page 382

ProCAST User Manual

The icon allows to pick a location on the geometry and display the node number, the node coordinates as well as the field value (e.g. Temperature, velocity magnitude, ...). One should first click on the icon (which becomes outlined in red) and then click on any point on the surface of the model. The information are displayed in a yellow frame.

In the case of Stress models (i.e. deformed models), the displayed coordinates will depend upon the "Displacement Magnitude" which is set. In order to have the real deformed coordinate, one should set a Displacement Magnitude to 1 (with a value of 0, it will correspond to the initial location of the mesh at the beginning of the calculation). The Pick option is also available in PreCAST :

Version 2006.0

Page 383

ProCAST User Manual

ViewCAST exit
Upon Exit of ViewCAST, a message is asking to the user whether he would like to keep the Last View. If Yes is answered, a file named "prefix.lv" will be stored in the working directory. This file is containing the model orientation, zoom and position, as well as the material selections.

These settings will be used for the next ViewCAST launch's.

Version 2006.0

Page 384

ProCAST User Manual

RESULTS ANALYSIS
CRITERION FUNCTIONS
ViewCAST allows to process the results which are calculated in the solver in order to built criterion functions or metallurgical results (Remark : this corresponds to the functionalities which were present in PostCAST in previous releases). The Actions menu give access to these functionalities.

R G L calculation
RGL is a generic menu which allows to calculate the solidification rate "R", the cooling rate "L" and the gradients "G", as well as any combination of those. The solidification rate corresponds to the velocity of a given isotherm (e.g. the liquidus isotherm). Thus, the user has to specify at which temperature he would like to calculate the solidification rate (see "R,G Temp" in the figure below). One can notice that it is allowed to calculate the velocity of any isotherm for other purposes than a solidification rate. The cooling rate is calculated as a linear interpolation between two temperatures (e.g. between the liquidus temperature and 10C above). Thus, the user has to specify these two temperatures (see "L Upper Temp" and "L Lower Temp" in the figure below). The gradient (magnitude) is also calculated at a given temperature, which should be specified by the user (and which corresponds to the same temperature as the one used for the solidification rate) - see "R,G Temp" in the figure below. When the RGL menu is activated, the following window opens :

Version 2006.0

Page 385

ProCAST User Manual

Firstly, the user has to choose between two methods to calculate the solidification rate. In Method 1, when each node reaches the specified temperature, a point is located along the temperature gradient some distance away and the time that it takes for the isotherm to reach that point is determined. R is then calculated as that distance divided by the difference in time. In Method 2, R is calculated as the cooling rate divided by the temperature gradient. Method 1 takes longer to compute, but it does not depend on the cooling rate. The results obtained by Method 2 are affected by the temperature levels used to calculate L. Four options are available for the calculation of the thermal gradient. One can either compute the "Total" gradient (i.e. all components of the gradient) or in either the X, Y or Z directions. Please note that the gradient is calculated at each node, when the "R,G Temp" is reached. This means that the gradient corresponds to a different time at each node.

Version 2006.0

Page 386

ProCAST User Manual

Then, the user has the possibility to combine the R, G and L variables together to obtain a criterion function (called Mapping factor). To do so, the coefficients a, b, c and d should be specified

and the Mapping factor "M" will be calculated as :

If the following parameters are used :

a b c d = 1.0 = 0.0 = 1.0 = -0.5

the Mapping factor "M" corresponds to the Nyiama criterion :

For the calculation of the Nyiama criterion, suggested values are :

L Upper Temp = Tliquidus + 2 L Lower Temp = Tsolidus R,G Temp = Tsolidus + 0.1 * (Tliquidus - Tsolidus)

Finally, the units (length and temperature) used in the RGL calculation, as well as the output format (*m.unf, Patran or I-DEAS) should be defined. When the *m.unf output format is used, the Mapping Factors (M), the cooling rates (L), the temperature gradient (G) and the Isotherm velocity (R) can be selected for visualization in the Contour Menu.

Version 2006.0

Page 387

ProCAST User Manual

Feeding length
The Feeding length provides the capability to calculate the distance between the solidus and some user defined critical temperature which represents some fraction solid beyond which feeding is impaired. This distance is then compared with a "critical feeding length," which is a simple linear function of the hydrostatic pressure. If the feeding distance exceeds the critical length, then porosity would be likely. The Feeding length was designed for the cases of directional solidification (such as DS or SX casting). Thus, it is not guaranteed that the Feeding length will work well in other cases. The critical Feeding Length is calculated as :

These constants A and B should be calibrated with experiments. The solidus temperature should be entered in the "Solidus" field (see below) and the temperature corresponding to a critical solid fraction (at which feeding is not anymore possible - typically 60-80%) should be entered in the "Critical" field.

Version 2006.0

Page 388

ProCAST User Manual

Please note that as the hydrostatic pressure is used in the Feeding length calculation, the gravity vector should be defined in PreCAST (but it is not necessary to run fluid flow). If a thermal only calculation was run and the gravity was not defined in PreCAST before the run of the calculation, it is still possible to load the case again in PreCAST, to set the gravity vector, to save the case and to exit. Then, "DataCAST -u" should be used (see the "Solver" chapter for more details on DataCAST -u, as well as the special care which should be taken in the use of this capability in order not to scratch the results). Then, the units used in the calculation, as well as the output format have to be selected. Once the Apply button is pressed, the Feeding length is calculated. It can then be viewed in the Contour Menu.

Isochrons
An Isochron is corresponding to the elapsed time from the beginning of the calculation until a specific temperature is reached. The Isochrons can be calculated in the Actions / Isochron menu.

Version 2006.0

Page 389

ProCAST User Manual

In order to select the desired temperatures, two possibilities are available : "Semi Auto" and "Specify Temps". In Semi Auto, the user can select 15 temperatures (to calculate 15 isochrons plots) by specifying the starting temperature, as well as a "delta" increment between the temperatures (then the units and the output format should also be specified).

With the "Specify Temps" mode, the user can select any temperature (up to 15 isochrons). Again, the output format and the units should be selected.

Version 2006.0

Page 390

ProCAST User Manual

Once the Apply button is pressed, the Isochrons are calculated. They can then be viewed in the Contour Menu.

Alpha case
Alpha case corresponds to the thickness of alpha case for the surface nodes adjacent to the ceramic shell in Titanium alloy investment castings. The necessary parameters, as well as the output format should be specified in the following window.

Once the Apply button is pressed, the Alpha case is calculated. It can then be viewed in the Contour Menu.

Version 2006.0

Page 391

ProCAST User Manual

SDAS (secondary dendrite arm spacing)

The secondary dendrite arm spacing (also called SDAS or 2 ) can be calculated according to the following equations, where tf is the local solidification time and M a constant which depends upon the alloy properties.

The SDAS is calculated in ViewCAST according to the following parameters.

Version 2006.0

Page 392

ProCAST User Manual

Tstart and Tend are the temperatures which are used to compute the local solidification time. Usually Tstart corresponds to the liquidus temperature, whereas Tend should be the temperature just above (e.g. 1 degree above) the first eutectic transformation. The Exponent corresponds to the power 1/3 in the above equation (it is thus recommended to set the Exponent to 0.33333) and M is the coarsening constant according to the above equation (which is alloy dependant). The units for M should be [microns3/sec]. For an Al-7%Si-0.3%Mg (A356), the coarsening constant M is equal to 680 [microns3/sec]. For an Al-2%Cu, the coarsening constant M is equal to 1400 [microns3/sec]. For an Fe-0.09%C (Low carbon steel), the coarsening constant M is equal to 29250 [microns3/sec]. For an Fe-0.6%C (Cast iron), the coarsening constant M is equal to 6050 [microns3/sec]. For an Fe-10%Ni (Stainless steel), the coarsening constant M is equal to 20600 3 [microns /sec]. Once the Apply button is pressed, the SDAS is calculated. It can then be viewed in the Contour Menu.

Version 2006.0

Page 393

ProCAST User Manual

POROSITY
To analyse the porosity in a casting, several options are available : Temperature field with the cut-off option Fraction of solid field with the cut-off option Shrinkage porosity field Niyama criteria for critical cooling rates Specific RGL criteria

Temperature field and Fraction of solid field

The porosity being primarily due to enclosed pockets of liquid, one can well observe them by looking at the Temperature or Solid fraction fields in cut-off mode (in order to visualize inside the casting during the solidification). For temperatures, it is recommended to select a cut-off value above the solidus temperature, whereas cut-off values of solid fractions below 70% shall be used.

Porosity
When POROS > 0 is used, one can visualize the Contour called "Shrinkage porosity". Values corresponding to a level below 0.01 shall be considered as microporosity (to be viewed best with slices) and values above 0.01 are considered as macroporosity (to be viewed best with cut-off mode). See the "Porosity models" section for more details.

Niyama
See the Criterion Functions section for more details.

Specific RGL criteria

One has the possibility to define a customized criterion function, with the RGL option (see the Criterion Functions section for more details).

Version 2006.0

Page 394

ProCAST User Manual

FATIGUE LIFE INDICATOR

In the case of die casting, when a stress calculation is run, ProCAST calculates automatically the fatigue life in the dies. The model is based upon a "strain-driven" approach and a power law relationship, which corresponds to low cycle fatigue. During the calculation, the accumulated plastic strain (if any) is recorded at each location of the die. Then, the life of the die is calculated as a power law of this plastic strain. If this life is shorter than the previously calculated life at this given location, this latter value is stored. If there is no plastic strain, then, the elastic strain is recorded and the life of the die is calculated with an other power law relationship. Again, if the computed life is shorter as the previously calculated (at this same location), this latter value will be used. The parameters of the model correspond to measured values for a typical die alloy (H13 steel), published by the Society of Automotive Engineers. One should note that the values are measured at room temperature and thus, the effect of temperature is only taken into account through the temperature-dependant stress properties (i.e. Youngs modulus, Yield stress, Hardening, ...). Thus, one should consider that ProCAST calculates a "Fatigue life INDICATOR" and that it does not correspond to absolute values of number of cycles. No data are required for this model and it can be used as an indicator for different steel grades generally used for molds. One should notice that it should be mainly used to compare different designs with the same mold material (e.g. to compare the effect of cooling channels or of cycling time). One should not compare the influence of different steels on the "Fatigue life indicator".

Version 2006.0

Page 395

ProCAST User Manual

HOT TEARING INDICATOR

During a stress calculation, ProCAST calculates the susceptibility to hot tearing. The hot tearing indicator allows to model cracks which are forming during the solidification (i.e. it corresponds to openings in between dendrites which are not yet totally solidified and which are opening because of a tensile stress). Hot tearing are formed only above the solidus temperature. The hot tearing indicator is a "strain-driven" model based upon the total strain which occurs during the solidification. The model is computing the elastic and the plastic strain at a given node when the fraction of solid is between CRITFS (usually 50%) and 99%. There is no parameter for this model (except CRITFS which is a general stress Run Parameter). As the amount of plastic strain will strongly depend upon the stress properties in the mushy zone, the hot tearing indicator should be used to compare different designs with the same alloy. One should not compare hot tearing indicator levels for different alloys.

Version 2006.0

Page 396

ProCAST User Manual

CRACKING INDICATOR
The "Cracking indicator" model of ProCAST corresponds to cracking occurring after completion of solidification. The model is based upon the "modified Gurson" model. It corresponds to a plastic strain driven model. The cracking indicator model couples the stress calculation with the porosity calculation. It is valid with any porosity model. The cracking indicator is corresponding to a void fraction indicator (i.e. it calculates the amount of voids created by cracks, which have nucleated and grown under the influence of stresses). Plastic strain (and plastic strain only) will allow cracks to nucleate and grow (thus to create "crack void fraction"). The presence of porosity will increase the amount of crack nucleation and growth. All the plastic strain is taken into account (including the one forming in the mushy zone). There is no inputs needed for that model. One should note that the constants which are in the modified Gurson model, are not temperature dependant (as there is no indication in the literature about how these values should change with temperature). However, the temperature dependency is taken into account in one way by the stress material properties and thus the plastic strain (e.g. there will be less plastic strain at low temperature). This model is an "Indicator" which can be used for any material. The value of the scale of the indicator will strongly depend upon the material and the stress data. Thus, one should not compare two different materials, but one should use it to compare the same material in different situation (i.e. different cooling conditions, different stress state, different porosity level). This model corresponds to a "damage" model, which couples for the first time stress calculation with defects in the casting (i.e porosity). This model is very new and should be used with care as there is very little experience in this field at this stage. However, it could be very interesting to see the effect of design decisions on the level of damage (e.g. a reduced cooling will change the amount of porosity and/or change the amount of stress and thus, it will change the amount of cracks which may appear). The cracking model is activated with the Run Parameter CRACK = 1 or 3. The default value is 0. See the "Stress Run Parameters" for more details.

Version 2006.0

Page 397

ProCAST User Manual

FRECKLES INDICATOR
The "Freckles indicator" of ProCAST is based upon a local Rayleigh Number calculation. The algorithm is described in "Development of a Freckle Predictor via Rayleigh-Number method for Single-Crystal Nickel-Based Superalloy Castings", by C. Beckermann, J.P. Gu and W. Boettinger, Metallurgical and Materials Transactions A, Vol. 31A. To activate this model, the casting material properties should have been created in PreCAST with the Thermodynamic databases (see the "Thermodynamic Databases" section for more details), using the Ni database. When the properties are calculated, a file named "prefixls.dat" is created. This file (which contains the liquid solute concentrations) should be present in the working directory in order to be able to calculate the Freckles indicator. Moreover, the Run Parameter FRECK with a value of 1 should be added manually in the p.dat file.

Version 2006.0

Page 398

ProCAST User Manual

RESULTS EXPORTS
INTRODUCTION
Different kind of calculated results can be exported in ViewCAST. The File/Export menu should be used for this purpose.

The results can be exported either in Patran or I-DEAS format, except the Stress results which can be exported in an ASCII format.

Version 2006.0

Page 399

ProCAST User Manual

GEOMETRY
When the File/Export/Geometry menu is selected, the following window opens

Firstly, the format of the output file should be selected (1). Then, as Patran or IDEAS files do not contain the unit information, it is necessary to specify in which units it is desired to export the geometry information (2). In the case of stress calculations, as the geometry is deforming, one can choose to export the deformed geometry at a given timestep (3). It is also possible to export the geometry with the "reversed displacements" (4). This latter option is useful if one would like to run a stress model with a "negative" distortion in order to see whether the final shape will correspond to one of the original drawing. In this case, one should export the model in the d.dat format.

Version 2006.0

Page 400

ProCAST User Manual

RADIATION FACES
For a radiation model, the radiation faces can be exported in the Patran or I-DEAS formats. One should also specify the units which will be used in the exported file.

Radiation faces correspond to a "surface mesh" or radiative faces, like an enclosure.

Version 2006.0

Page 401

ProCAST User Manual

TEMPERATURE
For the export of Temperatures, three options are available for the selection of the time-steps to be exported, as shown in the menu below.

Interval
With the Interval option, one can export temperatures according to the steps which are defined in the Steps menu (only the Steps selection will be used and not the Time selection) :

Then the format of the exported file (Patran or I-DEAS) and the units should be selected.

Version 2006.0

Page 402

ProCAST User Manual

Specify Steps
Any step to be exported can be selected with the "Specify Steps" option. The selected time-steps should be entered in the following list :

The above window allows also to select the format of the export (Patran or IDEAS), as well as the units. The data are exported when the Apply button is pressed.

Select Steps
The "Select Steps" option allows to choose the desired steps in the list of the available timesteps :

Version 2006.0

Page 403

ProCAST User Manual

The selected steps are highlighted in red. The above window allows also to select the format of the export (Patran or I-DEAS), as well as the units. The data are exported when the Apply button is pressed. The clear button allows to clear all the selections. Once the data are exported, a log file with the exported steps is created. This file is named prefixt.log :

Version 2006.0

Page 404

ProCAST User Manual

HEAT FLUX
The heat flux can be exported in exactly the same way as the temperature - see the "Results Exports / Temperature" section for more details.

Version 2006.0

Page 405

ProCAST User Manual

DISPLACEMENTS
The displacements can be exported in exactly the same way as the temperature see the "Results Exports / Temperature" section for more details. In order to Export the geometry with "reversed" displacements, please refer to the "Results Exports / Geometry" section for more details.

Version 2006.0

Page 406

ProCAST User Manual

STRESS
The stress results can be exported in exactly the same way as the temperature - see the "Results Exports / Temperature" section for more details. The only difference is that the stress results are exported in a "Neutral" ASCII file format in order to be imported in any stress software.

The format of this "Neutral file" is the following :

Loop on the Timesteps Timestep Number_of_nodes Loop on the Number_of_nodes node_number, x, y, z, s1, s2, s3, s4, s5, s6 End_loop End_loop

where x, y and z are the node coordinates and s1 to s6 are the six components of the stress tensor at the node. As ProCAST is calculating the stress data at the Gauss points, these values are interpolated at the nodes.

Version 2006.0

Page 407

ProCAST User Manual

PARALLEL SOLVER
INTRODUCTION
The goal of Parallel processing is to accelerate the computing time of a calculation, by distributing it on several processors. Different techniques are available and are described hereafter.
SMP : Symmetric MultiProcessing

In the SMP technique (often called Shared Memory Processing), the calculation is split in different CPU's which are sharing the same memory (see figure below). This technology is usually limited to a maximum of 32 CPU's.

DMP : Distributed Memory Processing

Distributed Memory Processing is an architecture where each CPU accesses its own memory. Data are shared between processors through message passing. The send and receive operations require actions by the processors on both ends of the communication. The Message-passing Interface, or MPI, is the software standard that has been developed by an industry/government consortium. Such a configuration (see picture below) requires a very fast network between the CPU's.

Version 2006.0

Page 408

ProCAST User Manual

SMP/DMP combined architecture

When each computer (node) has more than one CPU, sharing the same memory, one can combine the SMP and DMP principles to make a SMP/DMP combined architecture. Such a configuration (see picture below) requires a very fast network between the CPU's.

ProCAST Parallel architecture

The Parallel version of ProCAST is based upon the DMP technology, using MPI. Depending upon the machine/cluster which is used, one can use either a DMP or a combined SMP/DMP architecture. One should note that it is also possible to use a single computer with several CPU's sharing the same memory.

Version 2006.0

Page 409

ProCAST User Manual

As the MPI technology is based upon the "Message Passing Interface", it means that the CPU's have to communicate between them. This emphasizes the importance of the network linking the different nodes. One can use either a network with a 1 Gigabit/s Ethernet switch to link the nodes. It is possible to have faster communications with for instance a Myrinet network (from Myricom). This latter solution is more expensive, but the gain in performances can be quite significant. As communications are critical for the performances, it is very important to use the appropriate switch as well as the appropriate cables (e.g. for a 1 Gigabit/s Ethernet switch, one should use Category 6 cables - or above).

Version 2006.0

Page 410

ProCAST User Manual

HOW DOES PROCAST PARALLEL WORKS ?

In order to distribute the calculation between the N processors, the geometry (model) is split in N sub-domains. This partitioning is done fully automatically by the software. It is done in order to balance the load between the processors as evenly as possible (i.e. almost the same number of nodes in each sub-domain) and in such a way that the amount of communication between the processors is minimized (i.e. the least number of common nodes between the sub-domains). This principle is illustrated in the figure hereafter.

Then, each sub-domain is allocated to one processor, as shown in the figure below (each color represents a sub-domain). The communication between the subdomains is done automatically, through the fast network.

Version 2006.0

Page 411

ProCAST User Manual

The following pictures illustrate the partitioning of different cases. In the lower geometry, the partitioning for 2 and 4 CPU's is shown on the left and right, respectively.

Version 2006.0

Page 412

ProCAST User Manual

If such a partitioning is very efficient for thermal, radiation and stress calculations (as well as for flow calculations with an already filled cavity), it is not the case for filling calculations. As illustrated in the figure below (where the filling starts from the biscuit in the bottom of the picture), the first stage of the filling will occur only in the red sub-domain (CPU 1), whereas the final stage of the filling will take place in the purple sub-domain (CPU 2). This means that the CPU 2 will not do anything during a while at the beginning of the calculation, whereas CPU 1 will "sleep" near the end of the calculation. This will lead to a very poor balancing between the processors and thus a very bad performance.

Version 2006.0

Page 413

ProCAST User Manual

In order to prevent such detrimental behavior, one should rather have a partitioning like the one on the right (in the figure below), rather than the one on the left.

However, as one does not know how the cavity will be filled (this is why a calculation is performed !), a "dynamic partitioning" algorithm has been developed. At a given step, the flow front is at a given location (in blue in the figure below). The software is creating a layer ahead of the the flow front (the width of this layer corresponds to a number of elements, defined by the Run Parameter NFFWID). Then, both the liquid region and this layer are partitioned in sub-domains. Once the liquid is reaching the end of this layer (i.e. when the distance between the flow front and the end of the layer is less than NFFSAF elements), a new partitioning is performed. Thus, the frequency of the partitioning

Version 2006.0

Page 414

ProCAST User Manual

is defined by NFFWID. The smaller the value, the larger the numbers of partitioning. As the partitioning is taking some CPU time, one should specify a not too small value for NFFWID).

One should note that in the case of a filling problem, in the presence of a mold, two distinct partitioning are made. One for the thermal and radiation problem (which is done only once at the beginning of the calculation) and one for the casting domain only, which is dynamically re-partitioned. In order to perform this automatic domain decomposition, to manage the communication between the processors and to solve the linear system in a distributed fashion, several libraries are used : MIPCH or LAM/MPI A Portable Implementation of MPI (Message Passing Interface). Depending upon the platforms, MPI can be called with different names. Parmetis A Parallel graphic partitioner PetSc The Portable, Extensible Toolkit for Scientific computation

Version 2006.0

Page 415

ProCAST User Manual

USE OF THE PARALLEL SOLVER

The Parallel solver is fully compatible with the Scalar solver. This means that the input and output files are strictly identical (except two additional Run parameters for the Parallel solver called NFFWID and NFFSAF). Thus, the Pre- and Post-Processing should be made in the Scalar version, as usual. The two "Parallel" Run Parameters (NFFWID and NFFSAF) should be added manually in the p.dat file. If they are not specified, the default values of NFFWID = 10 and NFFSAF = 2 are used. Recommended values are NFFWID = 10 to 20 and NFFSAF = 2 to 5 (see the "How does ProCAST Parallel works" section for more details about the meaning of these Run Parameters). Once a case is set-up in PreCAST, the standard "DataCAST" should be called (it is not anymore necessary to use a specific DMP DataCAST executable). The Manager allows to have a direct access to the Parallel solver. To do so, one should first activate the Parallel Processing Preferences in the "Software Manager/Software configuration" window. The interactive launch of the Parallel solver is valid only for Linux and Windows at this stage. For other UNIX platforms, the run should be launched using manual commands in a Command window (see the end of this section for more details) Once the Manager is configured as decribed above, the "ProCAST" button is opening the following window for Linux :

Version 2006.0

Page 416

ProCAST User Manual

and for Windows :

Version 2006.0

Page 417

ProCAST User Manual

One can see in the above execution windows that a "Parallel" section is appearing. For the ProCAST solver, one still have in both cases the possibility to "Execute DataCAST first". On Linux, the execution in batch mode can also be used (in this case, the Batch log file can be defined - if nothing is specified, a file named DMP_runlog.txt will automatically be created (not available on Windows).

Then, for the Parallel solver, some parameters and options should be defined, before the Run can be launched.

Firstly, the number of processors or the processors selection should be done. If the radio button "Number of Processors" is selected, the user must specify on how many processors the calculation should be distributed. Then, the machine will automatically select the processors. If the "Processors Selection" is chosen (not available on Windows), the user can define specifically on which processors the calculation should be run. The list of processors should be specified in between comas (","). A dash ("-") allows to define a list (i.e. "2-4" means "2,3,4"). Please note that the processors numbering is starting at 0. Then, the Run option should be specified : On Linux and Unix,"Options(1)" correspond to the MPI options, whereas "Options(2)" corresponds to the ProCAST Parallel options. On Windows, no options needs to be specified, except if the data are on an other disc than the C:/ drive (see the synthax hereafter). For the MPI options, it is advised to use the default options which are proposed (ssi rpi usysv for Linux). These options are improving the MPI communications when LAM-MPI is used.

Version 2006.0

Page 418

ProCAST User Manual

At this stage, the ProCAST Parallel options are only for debugging purposes and thus none should be specified by the user. If one would like to change these options, it can be done directly in the corresponding fields. The Parallel solver (on Linux only) can be launched from the "Run List" of the Manager (see the "Software Manager/Run list" section for more details). It is also possible to launch the Parallel solver "manually", from a Command Window. In such case, the syntax is the following :
For Linux :
mpirun -options(1) CPU_selection procastparallel prefix -options(2)

mpirun : MPI executable -options(1) : LAM MPI options to speed-up communications

-ssi rpi usysv

CPU_selection : selection of processors (2 possibilities) -np N selection of N processors ci,j-k list of processors procastparallel : ProCAST Parallel executable prefix : case name -options(2) : ProCAST Parallel options none should be specified at this stage To run the case in the background, one can use the following command :
mpirun -options(1) CPU_selection procastparallel prefix -options(2) > runlog.txt &

For Windows :
mpiexec -options(1) CPU_selection procastDMP.exe prefix

mpiexec : MPI executable -options(1) : Mapping options When the calculation data are on a different drive (e.g. the M: drive), one should specify it with the "-map option" under "Option(1)". The synthax is the following : "-map m:\\Disk1\Data" This means that the \\Disk1\Data directory is mounted as the "M" drive.

Version 2006.0

Page 419

ProCAST User Manual

Then, the calculation can be launched from the Manager (or from a DOS command window) from the M: drive :

Please note that the Mapping options (i.e. -map m:\\Disk1\Data) can be configured in the Installation settings of the Manager. CPU_selection : selection of processors -n N selection of N processors procastDMP.exe : ProCAST Parallel executable prefix : case name Examples :
mpiexec procastDMP.exe prefix

or
mpiexec -map m:\\Disk1\Data procastDMP.exe prefix

Version 2006.0

Page 420

ProCAST User Manual

REPEATABILITY
The goal of a parallel solver is to obtain the same results as the scalar version and regardless of the number of processors. This is called "Repeatability". One should note that Parallel processing (using an implicit solver like in ProCAST), involves the resolution at once of a linear system on distributed processors, with the appropriate communications between the processors. As a first consequence, the algorithms which are used in the parallel version (especially for the pre-conditioner) can not be exactly the same as the ones used in the scalar version. This may lead to small differences in convergence (and thus in timesteps) and in round-off. Moreover, as iterative solvers are used, the solution is never exact and it may slightly depend upon the way it is solved (which can be slightly different on 1, 2, 4 or 8 processors). Thus, one should expect to have very similar results between the scalar and the parallel version and between runs made with different numbers of processors, however, most of the time, one will not have exactly the same result. This is especially true in filling calculations where very small differences may have an effect on the filling pattern (as different meshes or different COURANT values may have too).

Version 2006.0

Page 421

ProCAST User Manual

LIMITATIONS
In ProCAST 2006.0, not all the modules and features of ProCAST have been parallelized yet. This section is describing the limitations of the current version, as well as the precautions in the use of the parallel solver. As the use of a nonparallelized feature may degrade completely the performances of the solver, these features are made non accessible and the solver will stop in such case. The tests are mainly done through the values of the corresponding Run Parameters. However please note that when a Parallel calculation is terminated, many MPI print-out may appear. This means that the ProCAST error message may be "embedded" into MPI messages. In ProCAST 2006.0, only the THERMAL, RADIATION, FLOW and STRESS modules are parallelized. The parallelization of the other modules is not planned at this stage. The following features are not parallelized (the list may not be exhaustive) and thus should not be used with the parallel solver : THERMAL module : THERMAL = 2 POROS = 4 and 8 (only POROS = 1 is available) User functions Freckles indicator FLOW module : Thixo casting with the Cut-off model. However the non-newtonian model is parallelized Interpenetrating mesh (e.g. shot piston) (PENETRATE = 1) RESERVOIR option Turbulent model

Version 2006.0

Page 422

ProCAST User Manual

MACHINE CONFIGURATION
In order to have an efficient parallel processing, it is important to set-up the machines in the right way. This section explains the basics principles of the machines installation and configuration in order to run the ProCAST Parallel solver. Please note that the following is not necessary for the Windows DMP version.
Cluster of processors

When a cluster of processors is used (e.g. Linux cluster of 4 machines, having each 2 CPU's), they should be connected through a dedicated fast network, such as a 1 Gigabit/s Ethernet or a Myrinet switch. Please note that the ProCAST Parallel installation (with the libraries) may be different if a 1 Gigabit/s Ethernet or a Myrinet switch is used. A special care should be taken upon the type of cables which are used. The cable should be designed for fast communication (for 1 Gigabit/s Ethernet, Category 6 cables - or above - should be used). Concerning the operating system, it is important to make such that the installation of each node is the same in order to optimize the performances.
SSH

During the Parallel calculation, the software is communicating with the different nodes (computers of the cluster) through "Secured shells" (ssh). Thus, the machines should be configured in order to allow "ssh" communication between the nodes. If a "paraphrase" (password) is set to activate the ssh, one should perform the two following commands when a new Command Window is used :
ssh-agent $SHELL ssh-add

then, the "paraphrase" should be given. If this operation is not done, the paraphrase will be asked for each node when the ProCAST Parallel solver will be launched. If ssh is configured without paraphrase, then, this is not necessary. If the Parallel solver is launched from the Manager, it is necessary to perform the above operation in the Command window before the Manager is launched (in this same command window).

Version 2006.0

Page 423

ProCAST User Manual

Disk mounting procedure

The disk on which the software and libraries are installed, as well as the disk(s) on which the data will be located (it can be different disks) should be mounted in such a way that they can be accessed in the same way (i.e. with the same path) from any machine (node).

Version 2006.0

Page 424

ProCAST User Manual

PARALLEL VERSION INSTALLATION

As a pre-requisite for the DMP installation, the scalar version must be installed first and the $ProCAST20060 environment variable must be set. This section describes the necessary configuration on the following platforms : Linux Windows

Linux
LAMHOME environment variable

The "lam-7.1.1" directory can be placed anywhere (but it should be accessible from all nodes in the same way). It is advised to put it in the $ProCAST20060 directory (sideways to the bin directory). The LAMHOME environment variable should be set to point to the /..../lam-7.1.1 directory. The directory $LAMHOME/bin directory should be included in the PATH (for ALL nodes) :
set path=($LAMHOME/bin $path)

LD_LIBRARY_PATH

The environment variable LD_LIBRARY_PATH should contain the path of the libraries delivered with the software (i.e. the one contained in the libDMP directory of the installation). To do so, one should specify in the .cshrc (or .tcshrc) file of ALL nodes the following :
setenv LD_LIBRARY_PATH /......../libDMP:$LD_LIBRARY_PATH

LAMBOOT

On Linux machines, the "LAM/MPI" service should be started once when the machine is booted. To do so, the command "lamboot" should be run (only on node 1). This lamboot will be valid for all further DMP runs. Sometimes, it may be needed to reset the lamboot. To do so, one should first stop the service with the "lamhalt" command. Then "lamboot" can be launched again. One should be

Version 2006.0

Page 425

ProCAST User Manual

careful that a "lamhalt" or a "lamboot" command will stop all currently running parallel jobs (using LAM).
Hostname specification

The list of machines (nodes) on which the parallel processing will be performed should be specified in a file named "lam-bhost.def". This file is located in the lam7.1.1/etc directory. One should put in the "hostnames" on which the parallel executable should run. If the machines called "hostname1" and "hostname2" contain each 2 processors, the lam-bhost.def file should contain the following lines :
hostname1 cpu=2 hostname2 cpu=2

Windows (XP and 2000 Professional only)

The Windows DMP version is designed to work for a multiple CPU's Windows machine. The supported OS are Windows XP and Windows 2000.
Files and directories installation

When the Windows_DMP.exe InstallShield will automatically install the ProCAST DMP solver in the bin directory of the installation, as well as the MPICH2 environnement. Please note that the installation of MPICH2 requires the prior installation of the patch "Microsoft .Net Framework version 1.1". If it is not installed, the following message will appear. Click Yes and load the Patch.

You can also search for this patch in the "Windowsupdate" web site of Microsoft :

Version 2006.0

Page 426

ProCAST User Manual

Version 2006.0

Page 427

ProCAST User Manual

HARDWARE AND OS
The pre-requisites and requirements for the hardware and for the operating systems (OS) are the following (at this stage only Linux and Windows are available) :

Windows
Windows XP and Windows 2000 Professional only. Please note that the installation of MPICH2 requires the prior installation of the patch "Microsoft .Net Framework version 1.1".

Linux (32 bits)

The requirement for Linux is the level of the Linux kernel and of the glibc. One should have levels higher or equal to :
kernel : glibc : >= 2.4.20 >= 2.3.2

In order to get the level of the kernel or of glibc, the following commands should be run on the machine :
rpm -q kernel rpm -q glibc

With these requirements, the ProCAST parallel version is technically compatible with any Linux operating system (e.g. RedHat, Suse, Mandrake, ...). However, the version is guaranteed (i.e. tested) only on Redhat EL3 (RedHat Entreprise Linux 3). The following versions of libraries and compilers were used :
LAM-7.1.1 Petsc-2.3 Parametis-3.1 gcc-3.1 compiler(for LAM, Petsc and Parametis) icc 8.1.026 compiler (for ProCAST)

Linux (64 bits)

The requirement for Linux is the level of the Linux kernel and of the glibc. One should have levels higher or equal to :
kernel : glibc : >= 2.4.21 >= 2.2.1

Version 2006.0

Page 428

ProCAST User Manual

In order to get the level of the kernel or of glibc, the following commands should be run on the machine :
rpm -q kernel rpm -q glibc

With these requirements, the ProCAST parallel version is technically compatible with any Linux operating system (e.g. RedHat, Suse, Mandrake, ...). However, the version is guaranteed (i.e. tested) only on Redhat EL3 (RedHat Entreprise Linux 3). The following versions of libraries and compilers were used :
LAM-7.1.1 Petsc-2.3 Parametis-3.1 gcc-3.1 compiler(for LAM, Petsc and Parametis) icc 8.1.026 compiler (for ProCAST)

Version 2006.0

Page 429

ProCAST User Manual

LAM/MPI AND MPICH COPYRIGHTS

Copyright and Software License for LAM/MPI :
Copyright (c) 2001-2004 The Trustees of Indiana University. All rights reserved. Copyright (c) 1998-2001 University of Notre Dame. All rights reserved. Copyright (c) 1994-1998 The Ohio State University. All rights reserved. Indiana University has the exclusive rights to license this product under the following license. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: 1. All redistributions of source code must retain the above copyright notice, the list of authors in the original source code, this list of conditions and the disclaimer listed in this license; 2. All redistributions in binary form must reproduce the above copyright notice, this list of conditions and the disclaimer listed in this license in the documentation and/or other materials provided with the distribution; 3. Any documentation included with all redistributions must include the following acknowledgement: "This product includes software developed at the Ohio Supercomputer Center at The Ohio State University, the University of Notre Dame and the Pervasive Technology Labs at Indiana University with original ideas contributed from Cornell University. For technical information contact Andrew Lumsdaine at the Pervasive Technology Labs at Indiana University. For administrative and license questions contact the Advanced Research and Technology Institute at 1100 Waterway Blvd. Indianapolis, Indiana 46202, phone 317-274-5905, fax 317-274-5902." Alternatively, this acknowledgement may appear in the software itself, and wherever such thirdparty acknowledgments normally appear. 4. The name "LAM" or "LAM/MPI" shall not be used to endorse or promote products derived from this software without prior written permission from Indiana University. For written permission, please contact Indiana University Advanced Research & Technology Institute. 5. Products derived from this software may not be called "LAM" or "LAM/MPI", nor may "LAM" or "LAM/MPI" appear in their name, without prior written permission of Indiana University Advanced Research & Technology Institute. Indiana University provides no reassurances that the source code provided does not infringe the patent or any other intellectual property rights of any other entity. Indiana University disclaims any liability to any recipient for claims brought by any other entity based on infringement of intellectual property rights or otherwise. LICENSEE UNDERSTANDS THAT SOFTWARE IS PROVIDED "AS IS" FOR WHICH NO WARRANTIES AS TO CAPABILITIES OR ACCURACY ARE MADE. INDIANA UNIVERSITY GIVES NO WARRANTIES AND MAKES NO REPRESENTATION THAT SOFTWARE IS FREE OF INFRINGEMENT

Version 2006.0

Page 430

ProCAST User Manual

OF THIRD PARTY PATENT, COPYRIGHT, OR OTHER PROPRIETARY RIGHTS. INDIANA UNIVERSITY MAKES NO WARRANTIES THAT SOFTWARE IS FREE FROM "BUGS", "VIRUSES", "TROJAN HORSES", "TRAP DOORS", "WORMS", OR OTHER HARMFUL CODE. LICENSEE ASSUMES THE ENTIRE RISK AS TO THE PERFORMANCE OF SOFTWARE AND/OR ASSOCIATED MATERIALS, AND TO THE PERFORMANCE AND VALIDITY OF INFORMATION GENERATED USING SOFTWARE. Indiana University has the exclusive rights to license this product under this license.

MPICH Copyright Notice:

Copyright (c) 1993 University of Chicago Copyright (c) 1993 Mississippi State University Permission is hereby granted to use, reproduce, prepare derivative works, and to redistribute to others. This software was authored by: Argonne National Laboratory Group W. Gropp: (630) 252-4318; FAX: (630) 252-5986; e-mail: gropp@mcs.anl.gov E. Lusk: (630) 252-7852; FAX: (630) 252-5986; e-mail: lusk@mcs.anl.gov Mathematics and Computer Science Division Argonne National Laboratory, Argonne IL 60439 Mississippi State Group N. Doss: (601) 325-2565; FAX: (601) 325-7692; e-mail: doss@erc.msstate.edu A. Skjellum:(601) 325-8435; FAX: (601) 325-8997; e-mail: tony@erc.msstate.edu Mississippi State University, Computer Science Department & NSF Engineering Research Center for Computational Field Simulation P.O. Box 6176, Mississippi State MS 39762 GOVERNMENT LICENSE Portions of this material resulted from work developed under a U.S. Government Contract and are subject to the following license: the Government is granted for itself and others acting on its behalf a paid-up, nonexclusive, irrevocable worldwide license in this computer software to reproduce, prepare derivative works, and perform publicly and display publicly. DISCLAIMER This computer code material was prepared, in part, as an account of work sponsored by an agency of the United States Government. Neither the United States, nor the University of Chicago, nor Mississippi State University, nor any of

Version 2006.0

Page 431

ProCAST User Manual

their employees, makes any warranty express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights.

Version 2006.0

Page 432

ProCAST User Manual

ADVANCED POROSITY CALCULATIONS

Porosity calculations are performed as a "post-processing" of the thermal results. Thus, one should first configure and run a thermal calculation as usual and then to configure and run the porosity model.

INTRODUCTION
The porosity module of ProCAST is based upon the model developed by Pequet, Gremaud and Rappaz. For further details, please refer to : Ch. Pequet, M. Gremaud, M. Rappaz, Modeling of Microporosity, Macroporosity and Pipe Shrinkage Formation during the Solidification of Alloys using a MushyZone Refinement Method, Met. Mater. Trans. 33A (2992) 2095-2106. The model can be summarized as follow (abstract of the above mentioned paper) : A microporosity model, based on the solution of Darcy's equation and microsegregation of gas, has been developed for arbitrary three-dimensional (3D) geometry and coupled with macroporosity and pipe shrinkage predictions. In order to accurately calculate the pressure drop within the mushy zone, a dynamic refinement technique has been implemented: a fine and regular Finite Volume (FV) grid is superimposed onto the Finite Element (FE) mesh used for the heat flow computations. For each time step, the cells, which fall in the mushy zone, are activated and the governing equations of microporosity formation are solved only within this domain, with appropriate boundary conditions. For that purpose, it is necessary to identify automatically the various liquid regions that may appear during solidification: open regions of liquid are connected to a free surface where a pressure is imposed; partially-closed liquid regions are connected to an open region via the mushy zone and closed regions are totally surrounded by the solid and/or mold. For partially-closed liquid pockets, it is shown that an integral boundary condition applies before macroporosity appears. Finally, pipe shrinkage (i.e., shrinkage appearing at a free surface) is obtained by integration of the calculated interdendritic fluid flow over the open region boundaries, thus ensuring that the total shrinkage (microporosity plus macroporosity and pipe shrinkage) respects the overall mass balance.

Version 2006.0

Page 433

ProCAST User Manual

The principle of the porosity calculation is the following : a) a thermal calculation is performed first b) the temperature and fraction of solid results are then processed in the Advanced Porosity module in order to calculate the macro- and micro-porosity, as well as the pipe shrinkage. c) to do so, a finer grid is automatically superimposed to calculate the pressure drop in the mushy zone (solution of the D'Arcy equation) and the nucleation and growth of pores are computed. d) the calculation of the nucleation and growth of the pores is made according to the a gas model. The Sieverts law, presented in the paper mentioned above, is generalized as a Power law in the software. In order to achieve a porosity calculation, the appropriate settings should be defined for the thermal calculation and then the porosity input file (named "prefix_poro.d") should be configured. These settings are described in the "Advanced Porosity Pre-Processing" section.

Version 2006.0

Page 434

ProCAST User Manual

ADVANCED POROSITY PRE-PROCESSING

Configuration of the thermal calculation

The configuration of the thermal model in ProCAST should be made in the standard way. The only precaution is to set the case in order to prevent piping (i.e. PIPEFS = 0). Please note that there are no restriction to activate other models during the thermal calculation, such as fluid flow or stress. However, it is important to remember that the Advanced Porosity model does not handle remelting. If this occurs, it is not guaranteed that the solution will converge.
Configuration of the porosity model

The calculation condition file (prefix_poro.d (ASCII format)) contains all the necessary information to perform a calculation. It is divided in blocks of data containing different type of information :
General information porosity.param Material properties viscosity.param coarsening.param Gas and bubble properties nucleation.growth.param gas.param Process information pressure.param feeding.param gravity.param Calculation settings grid.param calculation.param

Each block starts by the "block.name" and ends by "end". Inside a block, commands are inserted. Each command is always followed by one argument (always one argument). The order of the blocks or of the commands within a block is free.

Version 2006.0

Page 435

ProCAST User Manual

In the description hereafter, each block is presented according to the following format.

Please note that only SI units must be used for the porosity calculation definition (however the thermal calculation can be set-up as usual with any units). The frames shown hereafter in light blue correspond to examples of settings. The following sections are describing these different inputs.

General Information
porosity.param

Firstly, the type of model should be selected. As in ProCAST only the "standard" model is available, the above block should always be specified as such.

Material Properties
When the "Advanced Porosity Model" is used in conjunction with a ProCAST calculation, all the necessary thermo-physical properties (specific heat, enthalpy, latent heat, fraction of solid and density) are automatically extracted from the ProCAST input data. Thus, nothing has to be specified and no special care has to be taken (as in previous versions). Of course, in order to allow porosity calculations, one has to have a density which is defined as a function of temperature (in this case a message saying that you have used a constant "specific.mass.mixture" will be printed). This version of the "Advanced Porosity Model" does not allow to have "expanding" materials (i.e. density which increases with increasing temperatures). In addition to the ProCAST thermo-physical data, the following data must be specified :
viscosity.param

Version 2006.0

Page 436

ProCAST User Manual

For the calculation of the D'Arcy equation (i.e. flow through the mushy zone), the viscosity is an important quantity. The viscosity can be taken either as a constant, or as a function of temperature. In this latter case, an Arrhenius law is considered :

: activation.energy : reference.viscosity The following frames are showing the two possibilities available for the user : a) Temperature dependant viscosity

b) Constant viscosity

coarsening.param

An other important parameter of the calculation of the D'Arcy equation is the permeability (K) of the mushy zone, which is described by the following equation :

Version 2006.0

Page 437

ProCAST User Manual

Beside the fraction of solid (gs), the permeability is a function of the secondary dendrite arm spacing (2), which can be calculated according to the following equation :

where tL is the time at which the liquidus temperature was reached and M(t) is the coarsening constant. The coarsening constant can be calculated as follow, for a binary alloy :

The following frames are showing the two possibilities available for the user : a) The coarsening constant (M(t)) is specified

b) A constant secondary dendrite arm spacing value is specified

Gas and Bubble Properties

nucleation.growth.param

The nucleation and growth parameters of pores (i.e. bubbles) are defined in this above block. The cavitation pressure corresponds to the pressure of the liquid within a closed pocket of liquid at which macro-shrinkage will start to appear. If no cavitation pressure is defined, the Power law for the gas solubility (see gas.param) is also used for the calculation of the macro-shrinkage. The cavitation pressure is usually equal to 0.8 to 0.9 of the atmospheric pressure (i.e. 0.8-0.9 bar or 80'000 to 90'000 Pa).

Version 2006.0

Page 438

ProCAST User Manual

When a pore will nucleate, it will appear with the radius defined under "nucleation.radius". As the pore is usually nucleating on a substrate, its radius will remain constant until the bubble is not wetting anymore the substrate. Then, the pore radius will increase. Please note that this initial pore radius value should not be defined with smaller values than 10 microns. Otherwise, the pressure inside the bubble will be so high that it will never nucleate and the calculation will not run well. The nucleation of the pores is defined by a "pore.density". Finally, the gas.metal.surface.tension will determine the pressure inside a pore, due to it's curvature.
gas.param

The gas properties are defined in the "gas.param" block. Firstly, the concentration of dissolved gas in the liquid metal should be specified (gas.nominal.concentration). The units are [ccSTP/100g], which means cubic centimeters of gas (at normal conditions of pressure and temperature [i.e. STP conditions - 20C - 1 bar], per 100 grams of liquid metal. The gas.partition coefficient corresponds to the ratio of the solubility limits of the gas in the solid and in the liquid (like a partition coefficient in a phase diagram). This value should be defined between 0 and 1. In the "Standard" model, the gas concentration [gas] in the liquid at equilibrium is described by the following powerlaw : where Pp is the pressure in the pore, n is the "gas.pressure.exponent" and A(T) is a defined as :

Where Ao corresponds to the "gas.equilibrium.constant.a" and B to the "gas.equilibrium.constant.b". For di-atomic gases (such as Hydrogen [H2]), n is equal to 0.5 (i.e. 1/2).

Version 2006.0

Page 439

ProCAST User Manual

One should be careful that the initial gas concentration should not be larger than the solubility limit of the corresponding gas (in such case, bubbles would be created in the liquid metal before the pouring of the metal and the corresponding amount of dissolved gas would decrease). In order to prevent such case, a warning will be printed at the beginning of the calculation (see figure below) if the following situation arises (then it is advised to reduce the initial gas concentration or to change the gas parameters).

Process Information
pressure.param

The type of process, as well as the pressure definition are specified in this block (casting.type).

Version 2006.0

Page 440

ProCAST User Manual

Firstly, one should specify the process which is modeled, among three choices : gravity, injection or continuous.
gravity :

For all type of gravity casting (sand, permanent mold, investment, ...), the "gravity" type should be used. In this case, the software is automatically setting the pressure to the highest surface of the casting. This means that the user does not need to specify anything, except the value of the pressure (p.imposed). The highest surface is determined with the help of the gravity vector. With this mode, the metal at the top of the casting will be free to pipe, according to the value of mobility.limit.
injection :

The type "injection" has to be used for hpdc, lpdc and squeeze casting, as well as all processes where the liquid metal is injected. In this case, the injection point has to be specified with the "x,y,z.pressure.coordinate" (see below) - it is advised to select the last point to solidify as the feeding will be stopped when this point will solidify. As the metal is injected, no piping will occur, as there will be a continuous feeding of metal (however, depending upon the geometry and the thermal conditions, macroshrinkage may occur). In this case, the mobility.limit does not need to be specified and it will be automatically set to 1.
continuous :

For all continuous casting processes, the "continuous" type should be used. As for "injection", one will have a continuous feeding of metal and thus, the mobility.limit does not need to be defined. In this case, it will automatically be set to 1. Moreover, the feeding point (usually the last point to solidify) should be specified with the "x,y,z.pressure.coordinate" (see below). In addition, in the case of a vertical continuous casting, one should specify the liquid level (with p.level). In this case, the value of the applied pressure will be automatically adapted to take into account the appropriate metallostatic pressure. In the case of "injection" or "continuous", the location where the pressure is applied should be defined (whereas it is automatically set to the highest surface is the case of "gravity"). This is done with the keywords "x.pressure.coordinate", "y.pressure.coordinate" and "z.pressure.coordinate". It is strongly advised to specify a location at the injection point which will solidify last as the feeding will be stopped when the fraction of solid at this location will reach a value of 1. The applied pressure is specified with "p.imposed". This pressure is either applied on the highest surface (in the case of "gravity"), or at the "x,y,z.pressure.coordinate" location in the case of "injection" or "continuous". The value of p.imposed can be defined either by a constant or by a function of time (with a value block F(TIME)).

Version 2006.0

Page 441

ProCAST User Manual

In order to define the pressure as a time table, one should replace the pressure value by a pointer (i.e. a name starting by a "*" - see below).

Then, an additional value block should be specified with the time function. The name of the value block should be the same as the one of the pointer (without the "*"). Then, the following syntax should be used :

If the level of the liquid is not corresponding to the level of the coordinate where the pressure is applied, this can be accounted for with the keyword "p.level". This has to be used in continuous casting (as the "effective" level of the liquid is changing during the process) or if a part of the casting is not modeled. When both "p.level" and "x,y,z.pressure.coordinate" are defined, the pressure applied at the p.level altitude (in absolute axis coordinated. in the direction of the gravity) corresponds to the applied pressure (defined by "p.imposed"). Below p.level, the metallostatic contribution is added to p.imposed. This can be described with the following piece of code (considering that the gravity is in the -Z direction) :
if (z > p.level) then pcell = p.imposed else pcell = p.imposed + rho * g * (p.level - z) endif where z is the current altitude of the cell for which the pressure (pcell) is calculated.

The value of p.level can be defined either by a constant or by a function of time (with a value block F(TIME)).
feeding.param

Version 2006.0

Page 442

ProCAST User Manual

The mobility.limit corresponds to the critical solid fraction from which the liquid at the top free surface can not move anymore (for piping). If the mobility.limit is very low (0.01), it means that as soon as the solid fraction is equal to 1%, the liquid can not move down anymore to feed the liquid pockets below. Thus one will have almost no piping, but macroshrinkage in the center of the casting. On the other hand, if a larger value of mobility.limit is used (0.2), the liquid can move down up to a fraction of solid of 20%. This will induce a significant piping and thus reduce the amount of macroshrinkage which may occur in the center of the casting. The value of mobility.limit is dependent upon the type of microstructure. In the case of columnar structure, the dendrite are linked to the solid shell and thus can not move too much, leading to a small value of the mobility.limit. On the other hand, in the case of equiaxed grains, they can move down more freely and thus, the mobility.limit is higher. If one would like to prevent piping (such as in high pressure die casting, squeeze casting or continuous casting), the mobility.limit is automatically set to 1.0 In this case, the liquid will continuously feed the casting. The value of mobility.limit can be defined either by a constant or by a function of time (with a value block F(TIME)). This can be used in order to model the end stage of a DC casting process. In this case, the mobility limit will be changed from 1.0 to 0.2 when no more liquid metal will be poured in the ingot. The "permeability.solid.fraction.cutoff" is used to cut-off the permeability at high solid fractions. As it can be seen in the following equation :

when the solid fraction (gs) is becoming zero, K is equal to zero. As in the D'Arcy equation, one is using 1/K, it is going to infinity. In order to prevent that, the permeability is taken as a constant for fractions of solid above the value of "permeability.solid.fraction.cutoff". Values around 0.95 to 0.98 are recommended. If values higher than 0.99 are used, automatically a value of 0.99 is set. If nothing is specified, a value of 0.98 is used. At very high solid fractions, the interdendritic flow becomes very small and the pressure drop calculation becomes very difficult (on a mathematical point of view). In order to improve the convergence of the model, the feeding calculation is stopped at a fraction of solid above "feeding.solid.fraction.cutoff". Values around 0.95 to 0.98 are recommended. If values higher than 0.98 are used,

Version 2006.0

Page 443

ProCAST User Manual

automatically a value of 0.98 is set. If nothing is specified, a value of 0.98 is used. This value has to be decreased in the case of bad convergence the model.
gravity.param

The three components of the gravity must be defined. One should note that the gravity should be parallel to one reference axis (i.e. parallel to x, y or z). Thus, two components of the gravity must be equal to zero. In addition, it is not possible to have a rotating gravity. Thus, it is not possible to use this module for Tilt or centrifugal cases.

Calculation Settings
grid.param

The "grid.spacing" corresponds to the cell size of the FVM mesh. One should select a size which is in relationship with the size of the mushy zone (in order to have several elements through the mushy zone thickness). The "grid.calculation.tolerance" is used in order to remove duplicate cells which would lie at the limit in between two elements. If a regular orthogonal FEM mesh is used and if the cell size (or grid spacing) is exactly matching the FEM mesh (i.e. FEM elements have a size of 1 cm and FVM cells have a size of 1 mm.), it happens that many cells are lying exactly on the element borders. The algorithm which is used involves long loops to guarantee no double cells. The grid.calculation.tolerance is used to check whether a cell is aligned with element borders. If this tolerance is too large, doublon checks will be made many times and the CPU time can rise very strongly. If this tolerance is too small, it is possible to create "holes" in the FVM mesh. Thus, a value of 1.e-12 is recommended. If the cellular meshing becomes very long, it is advisable to change slightly the cell size from 1.00 to 1.00001 in order to remove these alignments. If unstructured FEM meshes are used, this problem never happens. In cases, such as hpdc, sometimes, the gates between the casting and the overflows are very thin. This may lead to a lack of cells at the gate and thus, the

Version 2006.0

Page 444

ProCAST User Manual

overflow may become "isolated" form the casting (i.e. no cells in between the casting and the overflow). In such case, the porosity calculation will not be possible, as there is no way to "transmit" the pressure from the casting to the overflow. This could also happen if very thin sections, such as flashes are modeled. In this case, these "orphan" cells should be removed in order to allow the calculation to run. As this procedure of "orphan search" is CPU intensive, it can be disabled in cases where it is clear that this will not occur (for bulky parts). Otherwise, it is highly recommended to set the "grid.search.orphan" value to ON.
calculation.param

Finally, a few calculation parameters should be defined. The "region.number" (or domain number) in which the porosity calculation should be done must be specified at this time. Please note that it is possible to run a porosity calculation in more than one domain. In this case, the domain list should be specified in a value block (see below). One should however make sure that all the domains are connected. The "enthalpy.model" should always be set to zero (it is there for compatibility reasons). The "convergence.tolerance" is the value of the tolerance which is used in the pressure solver. The recommended value is 1e-10. If a calculation is not converging, it is advised to increase this value (1e-9 to 1e-8). The "maximum.iterations" corresponds to the number of iterations that are run before the calculation stops (if the convergence.tolerance is not reached before). It is advised to leave this value as such. Concerning the porosity results storage, two formats are available : a) "calcosoft" format b) "ProCAST" format

Version 2006.0

Page 445

ProCAST User Manual

In the "calcosoft" format, the porosity is stored at each FVM cell. This allows a very fine resolution of the results, but it requires the use of the specific calcosoft post-processor (see the "Advanced Porosity Post-processing" section for more details). In the "ProCAST" format, the porosity results, calculated at the FVM cells are extrapolated to the FEM nodes. This allows to view the "Advanced porosity" results in ViewCAST. The "store.type" allows to select which result files format should be stored. 0 : "calcosoft" format only 1 : "ProCAST" format only 2 : both formats The "store.frequency" parameter defines the frequency of storage of the porosity results. It is possible to store the pore fraction only (and not the pressure) with "store.pore.fraction", however, if the pressure only storage is defined (with "store.pressure"), the pore fraction will also be stored. These parameters will be valid only for the "calcosoft" format results. The ProCAST format results will be stored at the same frequency as the ProCAST temperatures, whatever values are specified for these parameters. In order to define a list of regions (to perform the porosity calculation in more than one region), the region.number should be specified by a pointer (i.e. a name starting by a "*" - see below).

Then, an additional value block should be specified with the list of regions. The name of the value block should be the same as the one of the pointer (without the "*"). Then, the following synthax should be used :

Version 2006.0

Page 446

ProCAST User Manual

Warnings

One should note that the "Advanced Porosity" module is based upon very complex physical algorithm and thus it's mathematical handling may be quite delicate in some cases. It is thus possible that the solution is not converging (or is giving error with the pre-conditioner). In such case, it is advised to either : modify the cell size, store more thermal timesteps, change (i.e. increase) the convergence.tolerance, decrease the value the feeding.solid.fraction.cutoff or of the permeability.solid.fraction.cutoff. One should also take a special care to make sure that there is no remelting. Finally, when the mushy zone is quite small (in case of large castings with low conductivity materials), the cell size should be adapted (i.e. the cell size should be smaller than the mushy zone length).

Examples prefix_poro.d input file

Example 1 : TemplateGravity_poro.d
#------------ General information -------------porosity.param: porosity.model end standard # It should always be 'standard'

#------------ Material Properties -------------viscosity.param: constant.viscosity centipoise end coarsening.param: secondary.dendrite.arm.spacing microns end $$$ # <------0.001 Pa.s = 1

$$$

# <-------

100.e-6 m

= 100

#--------- Gas and Bubble Information ---------nucleation.growth.param: nucleation.radius microns (1E-05) pore.density gas.metal.surface.tension Aluminium cavitation.pressure recommanded value end 1.E-05 1.E+09 $$$ 10000. # should NOT be smaller than 10 # recommanded value # <------0.90 value for # 10000 = 0.1 bar /

Version 2006.0

Page 447

ProCAST User Manual

gas.param: gas.nominal.concentration CCSTP/100g (for Al) gas.partition.coefficient gas.equilibrium.constant.a 3 and B = 0 ) gas.equilibrium.constant.b and B = 48822) gas.pressure.exponent end

$$$ 1.E-01 2.2e-3 0.00 0.5

# <-------

0.10 to 0.30

# typical value for Al # typical value for Al (A = 2.2e# typical value for Al (A = 1.18 # typical value for Al (n = 0.5)

#------------ Process Information -------------pressure.param: casting.type p.imposed bar for gravity end gravity 101325.0 # Gravity setting # <------- 101325.0 [Pa.s] = 1

feeding.param: mobility.limit feeding.solid.fraction.cutoff permeability.solid.fraction.cutoff end gravity.param: gx gy gz end

0.80 0.98 0.98

# Range from 0. to 1. # 0.98 : recommanded value # 0.98 : recommanded value

$$$ $$$ $$$

# <------# <------# <-------

gravity.x gravity.y gravity.z

#------------ Calculation settings ------------grid.param: grid.spacing grid.calculation.tolerance end calculation.param: region.number store.frequency store.type formats store.pore.fraction store.pressure enthalpy.model convergence.tolerance maximum.iterations end value regions data PARAMETERS $$$ $$$ $$$ end value # <------# <------# <------Porosity Domain 1 ... Porosity domain N $$$ 1.e-12 # <------0.002 = 2 mm # recommended value

*regions 1 2 on on 0 1e-10 5000

# 2 : both ProCAST and calcosoft

Example 2 : TemplateInjection_poro.d
#------------ General information -------------porosity.param: porosity.model end standard # It should always be 'standard'

#------------ Material Properties -------------viscosity.param: constant.viscosity centipoise end $$$ # <------0.001 Pa.s = 1

Version 2006.0

Page 448

ProCAST User Manual

coarsening.param: secondary.dendrite.arm.spacing microns end

$$$

# <-------

100.e-6 m

= 100

#--------- Gas and Bubble Information ---------nucleation.growth.param: nucleation.radius microns (1E-05) pore.density gas.metal.surface.tension Aluminium cavitation.pressure recommanded value end gas.param: gas.nominal.concentration CCSTP/100g (for Al) gas.partition.coefficient gas.equilibrium.constant.a 3 and B = 0 ) gas.equilibrium.constant.b and B = 48822) gas.pressure.exponent end 1.E-05 1.E+09 $$$ 10000. # should NOT be smaller than 10 # recommanded value # <------0.90 value for # 10000 = 0.1 bar /

$$$ 1.E-01 2.2e-3 0.00 0.5

# <-------

0.10 to 0.30

# typical value for Al # typical value for Al (A = 2.2e# typical value for Al (A = 1.18 # typical value for Al (n = 0.5)

#------------ Process Information -------------pressure.param: casting.type p.imposed 100 bar x.pressure.coordinate injection point y.pressure.coordinate injection point z.pressure.coordinate injection point end injection $$$ $$$ $$$ $$$ # Injection setting # <------- 10132500.0 [Pa.s] = # <------- X coord. of the # <------- Y coord. of the # <------- Z coord. of the

feeding.param: mobility.limit injection feeding.solid.fraction.cutoff permeability.solid.fraction.cutoff end gravity.param: gx gy gz end

1.00 0.98 0.98

# Is automatically set to 1 for # 0.98 : recommanded value # 0.98 : recommanded value

$$$ $$$ $$$

# <------# <------# <-------

gravity.x gravity.y gravity.z

#------------ Calculation settings ------------grid.param: grid.spacing grid.calculation.tolerance end calculation.param: region.number store.frequency store.type formats store.pore.fraction store.pressure enthalpy.model convergence.tolerance maximum.iterations end $$$ 1.e-12 # <------0.002 = 2 mm # recommended value

*regions 1 2 on on 0 1e-10 5000

# 2 : both ProCAST and calcosoft

Version 2006.0

Page 449

ProCAST User Manual

value regions data PARAMETERS $$$ $$$ $$$ end value # <------# <------# <------Porosity Domain 1 ... Porosity domain N

Example 3 : TemplateContinuous_poro.d
#------------ General information -------------porosity.param: porosity.model end standard # It should always be 'standard'

#------------ Material Properties -------------viscosity.param: constant.viscosity centipoise end coarsening.param: secondary.dendrite.arm.spacing microns end $$$ # <------0.001 Pa.s = 1

$$$

# <-------

100.e-6 m

= 100

#--------- Gas and Bubble Information ---------nucleation.growth.param: nucleation.radius microns (1E-05) pore.density gas.metal.surface.tension Aluminium cavitation.pressure recommanded value end gas.param: gas.nominal.concentration CCSTP/100g (for Al) gas.partition.coefficient gas.equilibrium.constant.a 3 and B = 0 ) gas.equilibrium.constant.b and B = 48822) gas.pressure.exponent end 1.E-05 1.E+09 $$$ 10000. # should NOT be smaller than 10 # recommanded value # <------0.90 value for # 10000 = 0.1 bar /

$$$ 1.E-01 2.2e-3 0.00 0.5

# <-------

0.10 to 0.30

# typical value for Al # typical value for Al (A = 2.2e# typical value for Al (A = 1.18 # typical value for Al (n = 0.5)

#------------ Process Information -------------pressure.param: casting.type p.imposed bar x.pressure.coordinate inlet point y.pressure.coordinate inlet point z.pressure.coordinate inlet point p.level the liquid end continuous 101325.0 $$$ $$$ $$$ *plevel # Continuous setting # <------- 101325.0 [Pa.s] = 1 # <------- X coord. of the liquid # <------- Y coord. of the liquid # <------- Z coord. of the liquid # <------- Level of the top of

feeding.param: mobility.limit feeding.solid.fraction.cutoff permeability.solid.fraction.cutoff end

1.00 0.98 0.98

# It is usually set to 1.0 # 0.98 : recommanded value # 0.98 : recommanded value

Version 2006.0

Page 450

ProCAST User Manual

gravity.param: gx gy gz end

$$$ $$$ $$$

# <------# <------# <-------

gravity.x gravity.y gravity.z

#------------ Calculation settings ------------grid.param: grid.spacing grid.calculation.tolerance end calculation.param: region.number store.frequency store.type formats store.pore.fraction store.pressure enthalpy.model convergence.tolerance maximum.iterations end value regions data PARAMETERS $$$ $$$ $$$ end value value plevel data F(TIME) $$$ $$$ $$$ $$$ $$$ $$$ end value # <------# <------# <------Porosity Domain 1 ... Porosity domain N $$$ 1.e-12 # <------0.002 = 2 mm # recommended value

*regions 1 2 on on 0 1e-10 5000

# 2 : both ProCAST and calcosoft

# Time level # Time level # Time level

Version 2006.0

Page 451

ProCAST User Manual

ADVANCED POROSITY SOLVER

Once the "Advanced Porosity" case is set-up (i.e. when the ProCAST Thermal calculation is run and when the prefix_poro.d input file is ready), one can launch the "Advanced Porosity" solver. To do so, the "ProCAST" solver button of the Manager should be used. This opening the following window :

The "Run" button should be hit (in the "Advanced Porosity Solver" section) to start the calculation. If the ProCAST Thermal calculation is not yet run, it is possible to click on the "Execute DataCAST/ProCAST first" check box before the "Run" button. It is also possible to automatically launch a serie of calculation using the "Run list" of the Manager. If the above window does not show the "Advanced Porosity Solver" section, one can activate it in the "Installation settings" window (by clicking in the "Advanced Porosity module" check box, under "Modules display") :

Version 2006.0

Page 452

ProCAST User Manual

Version 2006.0

Page 453

ProCAST User Manual

ADVANCED POROSITY POST-PROCESSING

Two different formats are available for the storage of the porosity results : a) "calcosoft" format b) "ProCAST" format In the "calcosoft" format, the porosity is stored at each FVM cell. This allows a very fine resolution of the results, but it requires the use of the specific calcosoft post-processor. In the "ProCAST" format, the porosity results, calculated at the FVM cells are extrapolated to the FEM nodes. This allows to view the "Advanced porosity" results in ViewCAST. The "store.type" allows to select which result files format should be stored (see the "Calculation Settings" section for more details). The "ProCAST" format results can be visualized in ViewCAST with the Contour "Thermal/Advanced Porosity". As for the "Shrinkage Porosity" contour, automatically the last timestep is shown in cut-off mode (above 1%) (see picture below).

Version 2006.0

Page 454

ProCAST User Manual

It is also possible to view the "Porosity Pressure" which corresponds to the Pressure drop calculated in the mushy zone in the Advanced Porosity model.

To access the "calcosoft format", one has to make a right mouse click on the "ViewCAST" button of the manager. This will open a sub-menu and one should select "Advanced Porosity results" (see below) :

This is automatically opening the "calcosoft" post-processor and the porosity field is shown. One should use the tape player to visualize the last timestep.

Version 2006.0

Page 455

ProCAST User Manual

When a cut-off value of 1% is used, the following result is obtained (which is very similar to the ViewCAST result above) :

Version 2006.0

Page 456

ProCAST User Manual

One should note that the "calcosoft" format results are more precise than the "ProCAST" format results. This is due to the fact that the Advanced porosity calculation is performed at cells (which are all shown in the "calcosoft" format, whereas these cells results are "extrapolated" to the FEM nodes for the "ProCAST" format results. If there is only a few cells which are empty (i.e. a small macroshrinkage), this may be "diluted" in the surrounding microporosity during the extrapolation to the FEM nodes. Thus, this small macroshrinkage may not be visible in the ViewCAST. Concerning the "calcosoft" format, the microporosity is shown using the yellowgreen scale, whereas the macroshrinkage is shown in orange and red. In practice the macroshrinkage are negative values (from 0 to -100), whereas microporosity are positive values (from 0 to 100). The macroporosity between 0 and 50% is shown in orange, whereas the one between 50 and 100% is shown in red. In "ProCAST" format, this distinction can not be made and both the microporosity and the macroporosity is shown (as positive values).

Version 2006.0

Page 457

ProCAST User Manual

MICROSTRUCTURES
The Microstructure models were improved in version 2006.0. In particular, the treatment of the thermal algorithm was significantly improved. This allows to have much larger timesteps than before. Please note that due to this change the maximum timestep should be kept to reasonable values (i.e. one should make sure that the solidification range is crossed in several timesteps). Concerning the prediction of the Casting strength, previously it was only a function of the grain size. It has been improved in order to be also a function of the ferrite-perlite fractions. As a consequence, the Casting strength will not be only dependent upon the thermal history near the solidification, but also upon the further cooling (i.e. it will depend upon the cooling rates between 700 and 900C).

INTRODUCTION
The microstructure module of ProCAST is now able to calculate automatically the microstructures, based upon the composition of the alloy. This can be achieved with the link of the module with thermodynamic databases. Depending upon the chemical composition, the microstructure module automatically detects the phases which will appear and the type of microstructure which should be computed (dendritic, eutectic, nodular, ...). For instance, if an Al-7%Si-0.3%Mg alloy is specified (A356), automatically, the software will detect that primary dendrites will form, followed by an interdendritic eutectic. On the other hands, if the composition of a Nodular Cast Iron (SGI) is defined, the nodule counts, the austenite radius, the pearlite and ferrite fractions will be computed, together with the corresponding mechanical properties (such as hardness, yield and tensile strength). The software will also automatically detect that if there is no Magnesium in the cast iron, the structure will be lamellar rather than spheroidal. In the same way, if the composition is hypo-eutectic, primary dendrites of austenite will form first before the eutectic precipitation. The only parameters that the user may need to specify are the nucleation parameters (see below). This is due to the fact that this is not an intrinsic property and that it may depend upon the metal treatment. Moreover, it may be necessary to define the growth kinetics of the eutectic phase (see below)

Version 2006.0

Page 458

ProCAST User Manual

As the microstructure module is linked with the thermodynamic databases, it is necessary to have the corresponding database license, for the desired base material.

Nucleation of the primary dendritic grains

The primary dendritic phase (if any) will nucleate and grow as equiaxed grains. The nucleation of these grains strongly depends upon the alloy treatment and thus the nucleation parameters should be defined. The model which is used is based upon the "Gaussian distribution" model proposed in 1987 by Rappaz et al.(Acta Metall., 35, (1987), 1487 and 2929). This model defines the relationship between the number of nuclei and the undercooling. The distribution of the nuclei with undercooling has the form of a Gaussian distribution and thus, the integral of this curve is an "S-shape" function (see graphs below).

The mathematical description of this "S-shape" curve is as follows :

with the following definitions :

Version 2006.0

Page 459

ProCAST User Manual

Thus, the nucleation behavior of the primary dendritic phase is fully defined by the three above parameters.

Nucleation of the eutectic grains

The nucleation of the eutectic grains is based upon the model proposed in 1966 by Oldfield (ASM Transaction, 598, (1966), 945). The number of nuclei is a powerlaw (Oldfield proposed a quadratic law) of the undercooling. The model is described by the following equations :

Thus, the nucleation behavior of the eutectic phase is fully defined by the two above parameters.

Growth kinetics of the eutectic grains

The eutectic is growing with a quadratic power of the undercooling :

Thus, the above constant fully defines the growth characteristics of the eutectic growth.

Default values
The following table is showing the default values which are used in the microstructure module for the models described above and for the different alloys. Of course, the nucleation data may change from one alloy to the other, due to different metal treatment. These default values corresponds to values proposed in the literature.

Version 2006.0

Page 460

ProCAST User Manual

Version 2006.0

Page 461

ProCAST User Manual

CASE SET-UP AND RESULTS

Set-up
The set-up of a microstructure calculation is very simple as the model is based upon the chemical composition. Firstly, the problem should be defined as a usual thermal case. Then, the chemical composition of the alloy should be specified in the corresponding material properties tab (see below).

For the thermal properties, only the thermal conductivity and the density should be defined in addition to the chemical composition. Then, the enthalpy needs to be calculated in PreCAST, with the Thermodynamic database (based upon the specified chemical composition), using the Lever model for Fe alloys and the Scheil model for the other systems. Thus, at the end, the thermal conductivity, the

Version 2006.0

Page 462

ProCAST User Manual

density, the enthalpy, the fraction of solid, the liquidus and solidus temperatures are defined. Finally, the Run Parameter MICRO should be set to 1 (for the users of previous versions, one should not anymore use values different from 1, as the selection of the micro model is now automatic). A value of 0 will disable the microstructure calculation. If the nucleation or growth parameters have to be changed from the default values (see Table below), the corresponding Run parameters can be modified (see the Microstructure Run Parameters section for more details).

Results
Depending upon the alloy composition, the type of microstructure will be different. As a consequence, the type of results which are computed will also be different. All the microstructure results can be visualized in the post-processing, in the "Contour/Micro" menu.
a) Dendritic primary phase and eutectic secondary phase

Most alloys are solidifying with a primary phase of dendrites, followed by interdendritic eutectic. In this case, the following quantities are calculated :

Version 2006.0

Page 463

ProCAST User Manual

The "Primary Dendrite Radius" corresponds to the primary grain size. The "Secondary Dendrite Arm Spacing" (also called SDAS) is the distance between the secondary dendrite arms of the primary phase. The "Primary Solid Fraction" corresponds to the fraction of primary phase, whereas the rest corresponds to the "Fraction of Eutectic". The "Eutectic Grain Radius" corresponds to the radius of equiaxed eutectic grains which are nucleating in between the dendrites of the primary phase. Finally, the "Eutectic Inter-lamellar Spacing" is the characteristic distance of the eutectic structure.
b) Special case of Fe-C alloys - grey iron (lamellar eutectic)

In the case of grey iron (lamellar eutectic), in addition to the quantities described above, other quantities can be calculated :

Metastable phases ("Fraction of Metastable Eutectic" and "Metastable Eutectic Grain Radius"), may appear depending upon the chemical composition and the local cooling conditions. The solid state transformations of austenite decomposition into Ferrite and Pearlite is calculated ("Fraction of Ferrite" and "Fraction of Pearlite", "Pearlite Spacing"),

Version 2006.0

Page 464

ProCAST User Manual

as well as the corresponding final mechanical properties ("Tensile Strength", "Yield Strength" and "Brinell Hardness").
c) Special case of Fe-C alloys - nodular cast iron (eutectic composition)

In the case of Nodular cast iron, nodules of graphite, surrounded by austenite are formed, instead of eutectic grains. Thus, the "Nodule count" (which corresponds to the density of graphite nodules), as well as the "Austenite Radius" and "Graphite Radius" are calculated. The solid state transformations, as well as the mechanical properties are calculated, as described above. In addition, the "Elongation" is calculated from the microstructure results.

The Nodular cast iron (SGI) model is activated as soon as there is a non zero amount of Mg in the chemical composition. Please note that Mg is not an element which is present in the Computherm thermodynamic database as it will have a negilible effect on the computed material properties. However, it should be specified in order to trigger the SGI model.
d) Special case of Fe-C alloys - nodular cast iron (hypo-eutectic composition)

In the case of a hypo-eutectic nodular cast-iron, the same properties as described above are obtained, in addition to the primary phase calculation of the austenite dendrites.

Version 2006.0

Page 465

ProCAST User Manual

The Nodular cast iron (SGI) model is activated as soon as there is a non zero amount of Mg in the chemical composition. Please note that Mg is not an element which is present in the Computherm thermodynamic database as it will have a negilible effect on the computed material properties. However, it should be specified in order to trigger the SGI model.
e) Special case of Fe-C alloys - steel

In the case of a steel with the composition of carbon equivalent less than 0.53 , the following properties are obtained.

The fraction of peritectic is the solid fraction formed from the reaction of liquid and the existing primary solid phase. The fraction of proeutectoid refers to the fraction of proeutectoid ferrite or cementite formed from the austenite phase as a function of time during the solid phase transformation. The carbon equivalent value controls which type of the proeutectoid phase will form (the "carbon equivalent" corresponds to %C + (%Si + %P)/3 ).
References

For cast iron, the mechanical properties calculations are based upon the microstructure, according to two following papers : Stefanescu et al, Proceedings of the 4th Decennial International Conference on Solidification Processing, Sheffield, (July 1997), 609. Goettsch et al, Metallurgical and Materials Transactions, 25A:5, (1994), 1063.

Version 2006.0

Page 466

ProCAST User Manual

EXAMPLES
In order to illustrate the application of the microstructure module on the solidification of nodular cast iron, two calculations were performed with two different chemical compositions. To do so, a very simple geometry was used, as shown below. The casting is cooled from the right with a chill, whereas the rest is in a sand mold.

This set-up produces a full range of cooling rates, as shown in the cooling curves hereafter.

Version 2006.0

Page 467

ProCAST User Manual

The following figure is showing the two alloys which were used, with the corresponding chemical compositions. The only difference between the two calculations is the amount of Carbon (from 3.2% to 3.5%).

Version 2006.0

Page 468

ProCAST User Manual

The above figure is showing the different kind of results which are automatically computed for the two alloys. One can see that for the Alloy B, there is no "Primary phase", as the alloy is lying on the eutectic composition. The following figures are showing the comparison of the different results. On the left, the Alloy 1 is shown (3.2% C - Hypo-eutectic) and the Alloy 2 is shown on the right (3.5% C - Eutectic).
Eutectic phase fraction

Nodule counts

Version 2006.0

Page 469

ProCAST User Manual

Austenite Radius

Ferrite fraction

Version 2006.0

Page 470

ProCAST User Manual

Pearlite fraction

Brinell Hardness

Elongation

Version 2006.0

Page 471

ProCAST User Manual

Tensile strength

Yield strength

Secondary Dendrite Arm Spacing (SDAS)

Version 2006.0

Page 472

ProCAST User Manual

One can see that the SDAS is only available for Alloy 1, as there is no dendrites (i.e. no primary phase) in Alloy 2.

Version 2006.0

Page 473

ProCAST User Manual

IRON AND STEEL

In order to run iron micro modeling, the following set-ups are needed in PreCAST and in the p.dat file :

a.

b.

Run Parameters/Micro: i. MICRO=1 ii. Select Default Values (Gray Iron or Ductile Iron) iii. Based on the inoculation level change the values accordingly 1. EUNUCL (For ductile iron, default value is 2000 ) 2. EUPOWER (For ductile iron, default value is 2.5 ) 3. Run Parameters/Thermal i. MACROFS ii. PIPEFS iii. FEEDLEN p.dat file (for SGI only) i. MOLDRIG (0~1, change it according to the hardness of mold) ii. GRAPHITE (0~1, default=1.0)

c.

Introduction to Iron and Steel

The study of the micro structure of steels and irons must start with the iron-carbon equilibrium diagram. Many of the basic features of this system (Fig. 1) influence the behavior of even the most complex iron alloys. For example, the phases found in the simple binary Fe-C system persist in complex steels. The iron-carbon diagram provides a valuable foundation on which to build knowledge of both plain carbon and alloy steels in their immense variety.

Version 2006.0

Page 474

ProCAST User Manual

Figure 1. The iron-carbon diagram. It should first be pointed out that the normal equilibrium diagram really represents the metastable equilibrium between iron and iron carbide (cementite). Cementite is metastable, and the true equilibrium should be between iron and graphite. Graphite occurs extensively in cast irons (2-4 wt % C), but it is usually difficult to obtain this equilibrium phase in steels (0.03-1.5 wt %C). Therefore, the metastable equilibrium between iron and iron carbide should be considered for steel, because it is relevant to the behavior of most steels in practice. It is convenient to combine the effect of the silicon with that of the carbon into a single factor which is called the carbon equivalent (CE): CE=%C+%Si/3 It is called iron if CE is greater than 2, otherwise it is called steel.

Steel
When the weight percent sum of all elements other than Fe is more than 5% (such as some alloy steel or stainless steel), only equiaxed dendrite model will be activated.

Version 2006.0

Page 475

ProCAST User Manual

1) Carbon content less than 0.53% (such as Alloy A in Fig. 1)

When the melt alloy cools from an initial temperature higher than the liquidus temperature down to a temperature slightly below it, the primary delta dendrite phase begins to nucleate in the liquid until a recalescence occurs due to the heat released from the growing nuclei. During the recalescence, the nucleation ceases and the nuclei grow rapidly into dendritic grains and soon impinge on each other at the end of recalescence. Then the growth of delta dendritic grains is replaced by the coarsening of delta dendritic arms. When the temperature reaches to the peritectic temperature, the peritectic transformation starts if there is still liquid available. Otherwise, the austenite phase precipitates from the delta phase until all of the delta phase is transformed into the austenite phase. Lastly, when the temperature of the casting is cooled down to the alpha phase transformation temperature, the alpha phase precipitates from the austenite phase to the eutectoid temperature. Below the eutectoid temperature, the graphite of Fe3C phase nucleates initially on the boundary of the austenite grains and then the coupled growth of alpha and Fe3C phases leads to the formation of pearlite phase.
2) Carbon content is greater than 0.53% (such as Alloy B in Fig. 1)

When the melt alloy cools from an initial temperature higher than the liquidus temperature down to a temperature slightly below it, the primary austenite dendrite phase begins to nucleate in the liquid. The austenite phase will grow until the end of solidification. There is no peritectic reaction here. When the temperature of the casting is cooled down to the alpha phase(C<0.8%) or cementite phase(C>0.8%) transformation temperature, the alpha phase or cementite phase precipitates from the austenite phase to the eutectoid temperature. Below the eutectoid temperature, the coupled growth of alpha and Fe3C phases leads to the formation of pearlite phase. The eutectoid temperature is around 723C while the eutectoid composition is 0.80% C. Slowly cooling alloys containing less than 0.80% C, hypo-eutectoid ferrite is formed from austenite in the range 910-723C with enrichment of the residual austenite in carbon, until at 723C the remaining austenite, now containing 0.8% carbon transforms to pearlite, a lamellar mixture of ferrite and iron carbide (cementite). In austenite with 0.80 to 2.06% carbon, cooling in the temperature interval 1147C to 723C, cementite first forms progressively depleting the austenite in carbon, until at 723C, the austenite contains 0.8% carbon and transforms to pearlite. Steels with less than about 0.8% carbon are thus hypo-eutectoid alloys with ferrite and pearlite as the prime constituents, the relative volume fractions being determined by the lever rule which states that as the carbon content is increased, the volume percentage of pearlite increases, until it is 100% at the eutectoid composition. Above 0.8% C, cementite becomes the hyper-eutectoid phase, and a similar variation in volume fraction of cementite and pearlite occurs on this side of the eutectoid composition.

Version 2006.0

Page 476

ProCAST User Manual

Cast Iron
The term cast iron, like the term steel, identifies a large family of ferrous alloys. Cast irons are multicomponent ferrous alloys. They contain major (iron, carbon, silicon), minor (<0.01%), and often alloying (>0.01%) elements. Cast iron has higher carbon and silicon contents than steel. Because of the higher carbon content, the structure of cast iron, as opposed to that of steel, exhibits a rich carbon phase. Depending primarily on composition, cooling rate and melt treatment, cast iron can solidify according to the thermodynamically metastable Fe-Fe3C system or the stable Fe-graphite system. When the metastable path is followed, the rich carbon phase in the eutectic is the iron carbide; when the stable solidification path is followed, the rich carbon phase is graphite. Referring only to the binary Fe-Fe3C or Fe-C system, cast iron can be defined as an iron-carbon alloy with more than 2% C. Important notice is that silicon and other alloying elements may considerably change the maximum solubility of carbon in austenite. The formation of stable or metastable eutectic is a function of many factors including the nucleation potential of the liquid, chemical composition, and cooling rate. The first two factors determine the graphitization potential of the iron. A high graphitization potential will result in irons with graphite as the rich carbon phase, while a low graphitization potential will result in irons with iron carbide. The metastable phase amount has both direct and indirect effects on the properties of ductile iron castings. Increasing the volume percent of hard, brittle carbide increases the yield strength, but reduces the tensile strength and elongation, of ductile iron castings. Because there is no graphite expansion for the metastable phase, the formation of carbide increases the likelihood of internal casting porosity. The two basic types of eutectics - the stable austenite-graphite or the metastable austenite-iron carbide (Fe3C) - have wide differences in their mechanical properties, such as strength, hardness, toughness, and ductility. Therefore, the basic scope of the metallurgical processing of cast iron is to manipulate the type, amount, and morphology of the eutectic in order to achieve the desired mechanical properties. The structure of the matrix is essentially determined by the cooling rate through the eutectoid temperature range. Slow cooling rates prevalent in heavy sections promote the transformation of ferrite.

Classification
Historically, the first classification of cast iron was based on its fracture. Two types of iron were initially recognized:

Version 2006.0

Page 477

ProCAST User Manual

White iron: Exhibits a white, crystalline fracture surface because fracture occurs along the iron carbide plates; it is the result of metastable solidification (Fe3C eutectic) Gray iron: Exhibits a gray fracture surface because fracture occurs along the graphite plates (flakes); it is the result of stable solidification (Graphite eutectic).

With the advent of metallography, and as the body of knowledge pertinent to cast iron increased, other classifications based on microstructural features became possible: Graphite shape: Lamellar (flake) graphite (FG) as shown in Fig. 4.1 and 4.2, spheroidal (nodular) graphite (SG) as shown in Fig. 5.1 and 5.2, compacted (vermicular) graphite (CG), and temper graphite (TG); temper graphite results from a solid-state reaction (malleabilization.) Matrix: Ferritic, pearlitic, austenitic, martensitic, bainitic (austempered).

The correspondence between commercial and microstructural classification, as well as the final processing stage in obtaining common cast irons, is given in Fig. 2.

Fig.2. Basic microstructures and processing for obtaining common commercial cast irons

Version 2006.0

Page 478

ProCAST User Manual

Graphite is a hexagonal-close pack form of carbon that can grow in both the liquid and solid forms of iron. In theory, in irons above the eutectic composition of carbon, the graphite first nucleates in the liquid, and then continues to grow in the solid. In irons below the eutectic composition, the carbon does not start to grow until the iron reaches eutectic temperature. As seen in a micro, the larger nodules are from growth initiated in the liquid, and the smaller nodules are from growth that does not start until solidification temperatures are reached. The graphite nodules continue to grow as the iron cools, so the amount of growth that occurs in the liquid is smaller than what would be assumed by the micro. The expansion from the graphite that grows in the liquid pushes liquid back into the riser, and does not offset shrinkage. So in order to minimize shrinkage, it is necessary to maximize the late formation of graphite without having to reduce the actual amount of graphite. Understanding what happens in a non-steady state solidification of Ductile Iron suggests a way that this can be done. In a hypoeutectic mode of solidification, austenite forms as a solid with a lower than average carbon content. This increases the carbon content of the remaining liquid until it reaches the eutectic composition. Likewise, in a hypereutectic mode of solidification, graphite nodules form in the liquid, removing carbon from the liquid until it is reduced to the eutectic composition

Fig. 3 Phase diagram showing movement of carbon concentration in liquid metal as iron solidifies. It can be seen from the diagram on the previous page, which the maximum amount of carbon that can be formed in late graphite is determined by the eutectic composition, and as long as the iron is at eutectic or above, the amount of late graphite will be the same.

Version 2006.0

Page 479

ProCAST User Manual

Iron and Steel models

The following micro models are available in ProCAST:

1. Equiaxed Dendrite Model

The model is based upon the model proposed in 1987 by Rappaz et al. This model defines the relationship between the number of nuclei and the undercooling. The distribution of the nuclei with undercooling has the form of a Gaussian distribution and thus, the integral of this curve is an "S-shape" function. Following nucleation, the dendrite tip growth is controlled by the supersaturation at the dendrite tip. This means that the tip growth is based on the total undercooling at the tip. As the tip grows, the solid fraction in each grain is not known from the tip position. In fact, the fraction solid is less than the fraction of the grain obtained from the tip position. At each time, the new tip concentration and fraction solid are known from a thermal and solute balance at the scale of the grain. The tip growth velocity is obtained from the Lipton-Glicksman-Kurz model, which simulates the growth of an isolated dendrite tip. The tip continues to grow until it reaches the end of the grain. At this point, solid fraction is still less than unity. However, mixing of the solute is complete at this stage. Therefore, a back diffusion type equation can be used to calculate solid fraction. If the phase diagram has a terminal reaction, e.g., eutectic, the remaining liquid gets rapidly transformed into a solid structure. It is assumed that the temperature of the grain is uniform and curvature undercooling is neglected. Also it is considered that the

Version 2006.0

Page 480

ProCAST User Manual

thermal undercooling is negligible. The growth of the dendrite tip is controlled mainly by solute diffusion. Therefore, only solutal undercooling is considered.

2. Eutectic Model
This model can be applicable to both regular and irregular eutectics. In the case of regular eutectics, growth of both phases of the eutectic structure is non-faceted in nature. For irregular eutectic, the growth process of one of the phases is faceted. Growth of the faceting phase requires considerably higher entropy of fusion. Examples of faceted growth are graphite growth in stable austenite/graphite eutectic and Silicon in Al-Si eutectic. The metastable austenite/cementite eutectic is an example of non-faceted/non-faceted type eutectic growth. Growth of both the stable and metastable eutectic are addressed here. Growth of the stable eutectic usually proceeds at a higher temperature. A higher cooling rate results in the formation of a metastable eutectic. This model assumes bulk heterogeneous nucleation at foreign sites which are already present within melt or intentionally added to the melt by inoculation. The nucleation of the eutectic grains is based upon the model proposed in 1966 by Oldfield (ASM Transaction, 598, (1966), 945). The number of nuclei is a power law (Oldfield proposed a quadratic law) of the undercooling. The growth of the grains is controlled by thermal undercooling at the solid/liquid interface. Solutal undercooling is neglected here since solute diffusion during eutectic solidification is negligible. The thermal undercooling is given by the difference between the eutectic temperature and the actual solid/liquid interface temperature.

3. Ductile Iron Eutectic Model

The eutectic growth process in ductile iron is a divorced growth of austenite and graphite, which do not grow concomitantly. At the beginning of the liquid/solid transformation, graphite nodules nucleate in the liquid and grow in the liquid to a small extent. The formation of graphite nodules and their limited growth in liquid depletes the melt locally of carbon in the vicinity of the nodules. This facilitates the nucleation of austenite around the nodules, forming a shell. Further growth of these nodules is possible by diffusion of carbon from the melt through the austenite shell. Once the austenite shell is formed around each nodule, the diffusion equation for carbon through the austenitic shell is solved in 1--D spherical coordinates. The boundary conditions are known from the phase diagram because thermodynamic equilibrium is maintained locally. Conservation of mass and solute is maintained in each grain. Because of the density variation resulting from the growth of austenite and graphite, the expansion/contraction of the grain is taken into account by allowing the final grain size to vary. Toward the end of solidification, the grains impinge on each other. This is taken into consideration by using the Johnson-Mehl approximation.

4. Gray/White Iron Eutectic Model

Version 2006.0

Page 481

ProCAST User Manual

This model is a special case of eutectic growth model and is applicable to cast gray/white iron only. In cast iron, one may obtain both gray and white iron depending on the melt composition and cooling conditions. Given a controlled melt composition, the most important factor that will determine whether a given region will solidify as white or gray is the cooling rate. It has been observed that for a specific melt composition and solidification condition, there exists a parameter called a critical cooling rate. If a region of a casting solidifies with a cooling rate higher than the critical cooling rate, then it will be white. The reverse is the case for gray iron.

5. Ductile Iron Eutectoid Model

This model can be used during the eutectoid transformation while describing the phase transformation of ductile iron to room temperature after solidification. The eutectoid reaction leads to the decomposition of austenite into ferrite and graphite for the case of the stable eutectoid and to pearlite for the metastable eutectoid transformation. Usually, the metastable eutectoid temperature is lower than the stable eutectoid temperature. Slower cooling rates result in more stable eutectoid structure. If the complete transformation of austenite is not achieved when the metastable temperature is reached, pearlite forms and grows in competition with ferrite. Growth of Ferrite: Even though ferrite can form either from the breakdown of pearlite or from the direct decomposition of austenite, it is assumed here that ferrite results only from the latter source. The following assumptions are made for modeling the growth of ferrite: 1) The austenite to ferrite transformation occurs at steady state and is controlled by carbon diffusion. 2) The ferrite grains grow as spherical shells within austenite grains and the number of ferrite grains is equal to the number of graphite nodules. 3) Thermodynamic equilibrium exists at graphite/ferrite and ferrite/austenite interfaces. These are defined by equilibrium solvus lines extended below the equilibrium eutectoid temperatures. 4) Diffusion from the ferrite/austenite interface towards austenite is neglected as diffusion coefficients and concentration gradients in austenite are small compared to those in ferrite. Nucleation and Growth of Pearlite: The nucleation of pearlite usually occurs at austenite grain boundaries. It has been demonstrated that pearlite colonies grow either as spheres or hemispheres following nucleation. By the movement of high mobility (i.e., low interface energy) incoherent interfaces, these colonies can grow edgewise or sidewise into the austenite. This means that pearlite grows in competition with ferrite until austenite is completely transformed.

Version 2006.0

Page 482

ProCAST User Manual

Transformation of austenite into pearlite is usually modeled with an Avrami equation because the study of nucleation of pearlite is difficult, especially under continuous cooling conditions. Also, pearlite grains impinge on each other at an early stage, especially at a relatively high cooling rate.

6. Gray Iron Eutectoid Model

The gray iron eutectoid transformation model is based on the approach used for gray iron eutectic. Nucleation and growth of ferrite takes place once the temperature drops below the stable eutectoid temperature. If the transformation of austenite is not complete when the metastable eutectoid temperature is reached, then nucleation and growth of pearlite takes place. The nucleation and growth rate expressions for pearlite are the same as those for the ductile iron eutectoid model.

7. Peritectic Transformation Model

In a peritectic transformation, liquid reacts with an existing solid phase to form a new solid phase. In conventional models, the new solid is assumed to form at the interface between the parent liquid and solid phases. Once the new solid phase is formed, further reaction between the parent phases is limited by the layer of solid formed. Hence the rate of reaction is controlled by the diffusion of solute through the shell of the transformed product. It has been suggested by some researchers that the peritectic transformation may be achieved through a liquid layer in between the parent and the product solid phases. This mechanism has been adopted in the present model. For example, in the case of steel, the austenite phase forms initially at the root of the dendrite arms of the delta phase and grows along the delta/liquid interface. The speed of this growth is the same as that with which liquid moves toward the delta phase. The diffusion problem can be simplified as the liquid layer is very thin and the diffusion of carbon in the liquid is very rapid so that the carbon concentration gradient in the liquid is negligible.

8. Solid Transformation Models

This model is only applicable to the Fe-C system and is used for tracking the fraction transformed for the cases of gamma to ferrite and gamma to cementite. Prior to reaching the eutectoid temperature, some of the austenite phase may transform into alpha phase as part of a pro-eutectoid transformation. If you started with a hypereutectoid composition, the pro-eutectoid transformation will be from austenite (gamma) to cementite. Both of these pro-eutectoid transformations are addressed by the current model. The wt% carbon equivalent determines whether the gamma to ferrite or gamma to cementite transformation will be used. Both of these models require that the equiaxed dendrite model be chosen for the initial liquid/solid phase transformation.

Version 2006.0

Page 483

ProCAST User Manual

Density calculation during iron alloys solidification:

During gray or ductile solidification, the densities of the different phases are computed according to the composition of the phase at that particular temperature. In the calculation, the graphite expansion is included. The densities are updated at each time based on the temperature. The porosity calculation is based upon the calculated density (see the "SGI Porosity model" section for more details). The input density curve in PreCAST will not affect the calculated porosity formation if micro model is activated.

Iron micro model inoculation setup:

In ProCAST, you can modify the phase nucleation depending on the inoculation practice in the foundry for iron alloys solidification simulation. The inoculation will be good if it is in-mold or in-stream with additional melt pretreatment. It is not as good if the inoculation is only in-ladle. You can change the run parameters to represent the practice of different inoculation. The higher value of EUNUCL (a micro run parameter) represents better inoculation. The default value is set to 2000 which should represent pretty good inoculation. Metastable phase can form in the faster cooling area if the inoculation is not good enough.

Graphite precipitation:
One more micro run parameter for iron alloys solidification simulation is added called graphite precipitation (GRAPHITE in p.dat) which tells the degree of graphite precipitation during solidification. It varies from 0 to 1. 1 means that the graphite expansion potential is completely considered in the simulation so the casting will have a relatively low tendency of shrinkage. 0 means that the graphite expansion does not occur hence there is no compensation for the shrinkage of the liquid during solidification by graphite expansion. During the micro calculation, the computed expansion part of the density (as a function of the phases present) is multiplied by GRAPHITE. If GRAPHITE = 0, there will be no expansion, whereas with a value of 1, the full expansion contribution will be taken into account in the density. This density is only used if a porosity calculation (POROS = 1) is made during the microstructure calculation (see the "SGI Porosity model" section for more details). The value of GRAPHITE will not affect at all the computed microstructure. The default value is set to 1.0. This parameter is used to adjust the porosity formation to the real foundry condition.

Case studies

Version 2006.0

Page 484

ProCAST User Manual

A simple geometry casting is used to illustrate the application of the microstructure module on the solidification of different alloys. The casting is cooled from the right with a chill (higher heat transfer coefficient). The two bigger surfaces (back and front) are symmetry. All the other faces are adiabatic.

1. Al 4.9wt%Si
For this alloy, as the temperature cools down, the primary phase forms first. The possibility to have and the amount of eutectic phase depends on the cooling rate. With faster cooling rate, there is less amount of primary phase but more eutectic phase. Fig. 6 shows the comparison of current calculation with some experiment and other modeling results for the solidification of this alloy.

Fig. 6 Fraction of Eutectic 1) Primary dendrite radius The PRIMARY DENDRITE RADIUS provides the current position of the dendrite tip and volume fraction of the dendritic grain as it varies with time. At the end of the primary solidification, this parameter will equal the final grain radius.

Version 2006.0

Page 485

ProCAST User Manual

Primary dendrite radius 2) Primary solid fraction This reflects the volume fraction of the primary phase formed during solidification. As stated above, there is more primary phase for lower cooling.

3) Eutectic grain radius (grain radius of the eutectic phase)

Version 2006.0

Page 486

ProCAST User Manual

4) Fraction of eutectic

5) Eutectic interlamellar spacing The INTER-LAMELLAR SPACING parameter determines the fineness of the eutectic. Smaller values of this parameter provide better mechanical properties.

Version 2006.0

Page 487

ProCAST User Manual

2. Ductile Iron (GGG 60)

This is a eutectic ductile iron. There is no primary austenite phase but only eutectic phase during solidification. If the Carbon Equivalent is less then the eutectic composition, the first phase comes out would be primary austenite phase then eutectic. The AUSTENITE RADIUS and GRAPHITE RADIUS provide the instantaneous values of the solidified grain size and nodule size respectively. At the end of solidification, they provide a correct description of the final grain size and nodule size. AUSTENITE RADIUS AND GRAPHITE RADIUS can be related to the mechanical properties of castings. Density is calculated from the local fraction of graphite, austenite, and liquid, thus capturing the contraction and expansion behavior of ductile iron. As explained earlier, the stable eutectoid growth refers to the decomposition of austenite into ferrite and graphite and the metastable eutectoid growth refers to the decomposition of austenite into pearlite, which is a coupled growth of ferrite and cementite. The properties of the iron depend strongly on the relative amounts of ferrite and pearlite in the matrix. As the pearlite content increases, tensile and yield strengths also increase, but at the cost of ductility. Ferrite content controls fracture toughness and dynamic properties of iron. The FRACTION OF FERRITE and FRACTION OF PEARLITE give the relative amount of stable and metastable eutectoid structures. The pearlite/ferrite ratio can be related to tensile strength through its effect on matrix microhardness. Usually, a finer pearlite grain size is associated with a finer interlamellar spacing with better mechanical properties. The following run parameters are used. EUNUCL=1000 EUPOWER=2.5

Version 2006.0

Page 488

ProCAST User Manual

1)

Nodule counts

2)

Austenite radius

3)

Graphite radius

Version 2006.0

Page 489

ProCAST User Manual

4)

Pearlite Spacing

5)

Fraction of Ferrite

Version 2006.0

Page 490

ProCAST User Manual

6)

Fraction of Pearlite

7)

Fraction of Eutectic

Version 2006.0

Page 491

ProCAST User Manual

8)

Fraction of Metastable Phase

9)

Tensile Strength

Version 2006.0

Page 492

ProCAST User Manual

10)

Yield Strength

11)

Elongation

Version 2006.0

Page 493

ProCAST User Manual

12)

Hardness

We can see that it is not 100% stable phase (Eutectic) every where for this casting after solidification. On the higher cooling area (right side), the metastable phase (ledeburite) formed. The fraction of stable phase plus the fraction of metastable phase is 1. The hard brittle metastable phase can form at higher cooling area when the inoculation is not good enough. The metastable phase can increase the yield strength but reduce the tensile strength of the ductile iron castings.

3. Gray Iron (GG 20)

Version 2006.0

Page 494

ProCAST User Manual

Mechanical properties of the cast grey iron part are a function of the stable and metastable eutectic volume fractions and grain sizes. FRACTION OF EUTECTIC gives the amount of the gray eutectic, whereas FRACTION OF METASTABLE EUTECTIC gives the amount of the white eutectic. In most cases, the gray structure is more desirable as it gives improved tensile strength and ductility. The EUTECTIC GRAIN RADIUS parameter gives the gray eutectic grain radius. The INTER LAMELLAR SPACING parameter calculates the spacing of the gray eutectic. 1) Fraction of Primary phase

2) Fraction of Eutectic

Version 2006.0

Page 495

ProCAST User Manual

3) Fraction of Metastable Eutectic

4) Eutectic grain Radius

Version 2006.0

Page 496

ProCAST User Manual

5) Eutectic Interlamelar Spacing

6) Fraction of Ferrite

Version 2006.0

Page 497

ProCAST User Manual

7) Fraction of Pearlite

8) Pearlite Spacing

Version 2006.0

Page 498

ProCAST User Manual

9) Tensile Strength

10) Yield Strength

Version 2006.0

Page 499

ProCAST User Manual

11) Brinell Hardness

Metastable phase can form if the cooling is high enough.

4) Carbon Steel (Plain Carbon AISI 1008 steel)

From the phase diagram, we know that the first phase formed during solidification is primary austenite. The peritectic reaction happens when the temperature drops to the peritectic temperature. Usually, peritectic growth is limited by the formation of the solid transformed product at the reacting liquid/solid phase boundary. The PERITECTIC FR. OF SOLID parameter gives the volume fraction of solid

Version 2006.0

Page 500

ProCAST User Manual

resulting from this reaction. It is important to know the amount of the phase formed through this reaction, as it usually forms as a surface layer on the primary dendritic solid phase. As temperature cools down, pro-eutectoid ferrite or cementite will form depends on the composition until when the temperature reaches to the eutectoid temperature. Below the eutectoid temperature, the pearlite forms. FRACTION OF PROEUTECTOID PHASE refers to the fraction of proeutectoid ferrite or cementite formed from the austenite phase as a function of time. The carbon equivalent value controls which type of the proeutectoid phase (ferrite or cementite) will form. If the carbon content is less than the eutectoid composition 0.8%, the proeutectoid phase would be ferrite phase, otherwise would be cementite phase. 1) Primary dendrite arms

2)

Primary solid fraction

Version 2006.0

Page 501

ProCAST User Manual

3) Secondary dendrite arm spacing

4) Fraction of peritectic

Version 2006.0

Page 502

ProCAST User Manual

5) Fraction of eutectoid

6) Fraction of proeutectoid

Version 2006.0

Page 503

ProCAST User Manual

Version 2006.0

Page 504

ProCAST User Manual

"2-D"
INTRODUCTION
The current version of ProCAST does not give access anymore to 2-D modeling. However, as it may be useful to run quick tests or for inverse modeling, "pseudo" 2-D capabilities were introduced. The principle is that 2-D calculations are replaced by 3-D calculations on one layer of elements. To do so, the 2-D mesh generator was modified in order to automatically create this layer of element, in order to reproduce either a 2-D cartesian situation or a 2-D axi-symmetric case. On the left (see figure below), the original 2-D mesh is transformed in a "one layer" 3-D mesh (right) for cartesian calculation (only a small thickness is modeled).

On the left (see figure below), the original 2-D mesh is transformed in a "one layer" 3-D mesh (right) for axisymmetric calculation (only a small sector is modeled).

Version 2006.0

Page 505

ProCAST User Manual

Version 2006.0

Page 506

ProCAST User Manual

MESHING
To call the 2-D mesh generator, one should do a RIGHT mouse click on the "MeshCAST" menu in the manager and the following sub-menu will appear

One can then select between 2-D or 3-D meshing. If MeshCAST-2D is selected, a new windows opens with the 2-D meshing interface. This interface was adapted from the 2-D mesh generator of CALCOSOFT. It allows to mix triangles and quadrangles in the same mesh and to control quite well the size and the shape of the mesh. A brief description of the 2D mesh generator is made hereafter. Firstly, the MeshCAST-2D window opens and the user can select between a new project (i.e. a new mesh) or to load a previous project.

Version 2006.0

Page 507

ProCAST User Manual

When a 2D mesh is generated, a file named "prefix.restart" is created. This allows to restore either the Geometry, the Domains, or everything up to the mesh itself. It is also possible to retrieve ProCAST 2-D meshes generated in the previous versions of ProCAST.

Then, the drawing area should be defined. The units should be selected (1), and the size of the drawing area must be defined (2). A grid is superimposed to the drawing area (to help for the drawing). Its characteristics can be defined (3). Finally, the type of problem (cartesian or axisymmetric) has to be selected (4). The "Continue" button should be used to go to the next window (5).

Version 2006.0

Page 508

ProCAST User Manual

The drawing can start at this stage. The type of entity to draw should be selected with the corresponding icon (1). Then, this entity can be drawn in the graphics area (2). One can change between the select and deselect mode with the corresponding button (3). The check box in-between is to activate the select or deselect all mode. One should notice that the intersections should be defined by the user, with the left icon in (4). One should select the two lines which are intersecting and then press this icon. Once the drawing is finished, the domains definition can be done in the next window, after pressing the "Continue" button (5).

Version 2006.0

Page 509

ProCAST User Manual

Before meshing, the different domains should be defined. One should first "Add" a new domain (1) - the corresponding domain number will appear in the list (2). Then, the contour of the domain should be selected in the graphics area (3). This operation should be repeated for each domain. Then, one can go in the "Meshing" window with the "Continue" button (4).

Version 2006.0

Page 510

ProCAST User Manual

The different domains are listed in (1). This mesh generator allows to define both structured meshes (quadrangles) or free triangular meshes (2). Firstly, one should select the domain to be meshed (1). Please note that structured meshes should be meshed first. Then the type of mesh should be selected (2). If a structured mesh is selected, one should define the four faces defining this domain (3). Then, the number of elements along each of the four faces should be defined (3). Please note that it is possible to have more than one line to define a face. Finally, the mesh button can be pressed (4) to generate the mesh.

In the case of "unstructured" mesh (i.e. triangles), the operation is more simple as one should just select the domain to be meshed (1), select the "triangle" mode (2), define the average mesh size (3) and mesh (4). Once all the meshes are generated, one can "Continue" to the next window.

Version 2006.0

Page 511

ProCAST User Manual

The last operation which can be done is the grouping of different domains ("combine"). This operation is optional. One should highlight all the domains which should be grouped into one (1) and then the "Combine" button should be pressed (2). Once this is done, it is possible to save the mesh (3). At this stage, automatically, the 2D mesh is extruded into a 3D one.

Version 2006.0

Page 512

ProCAST User Manual

In the case of axisymmetric cases, it is important to make sure that the axis of symmetry is well at y = 0. Due to round-off errors, one may have slightly negative coordinated of the nodes which are on the axis. In this case, the "Tools/Check Axisymmetric" menu should be activated.

The software than is asking for a confirmation for this operation.

We are now ready to set-up the case in PreCAST.

Version 2006.0

Page 513

ProCAST User Manual

CASE SETTING
After meshing, the "2-D" extruded mesh can be loaded in PreCAST. If the "cartesian" option was selected, the mesh will look like that :

If the "axisymmetric" option was selected, the mesh will look like that :

Version 2006.0

Page 514

ProCAST User Manual

Such cases should be set-up exactly in the same way as 3-D cases. One should just take care of the following points : A symmetry boundary condition should be set on the two large faces of the model, as shown hereafter.

Version 2006.0

Page 515

ProCAST User Manual

The boundary conditions should be defined on the small faces :

Version 2006.0

Page 516

ProCAST User Manual

CAFE-3D
Pre-processing
The set-up and the use of the CAFE-3D module is fully described in the CALCOSOFT-3D manual (this manual is available in PDF format in the ProCAST installation, in the dat/manuals/pdf directory). This section is summarizing the specific aspects of the CAFE model related with ProCAST. The CAFE model can be used in 3 different modes : "post-processing" mode weak coupling mode full coupling mode

In all cases, the ProCAST thermal case should be set as usual in PreCAST. In addition to the standard set-up, one should specify in PreCAST the "nucleation" boundary conditions. To do so, in the "Boundary Conditions/Assign surface" menu, one should "Add" nucleation BC's in the list (one entry per type of nucleation BC).

Version 2006.0

Page 517

ProCAST User Manual

Then, the different surfaces of the casting should be selected and stored in the corresponding nucleation entry (as for any other BC). Please note that no data should be allocated to these nucleation entries, as this will be done in the separate CAFE pre-processor.

If the "full" or "weak" coupling mode is used (for the CAFE calculation), the two following run parameters should be set as : THERMAL = 2 MICRO = 1024

Version 2006.0

Page 518

ProCAST User Manual

Version 2006.0

Page 519

ProCAST User Manual

Before starting the CAFE set-up (in the CALCOSOFT-3D CAFE pre-processor), one should run first DataCAST. This will create the unformatted mesh file (prefixg.unf), which is necessary for the loading into the CAFE pre-processor. Then, the CAFE-3D pre-processor can be started. To launch it, one should make a right click in the Manager and select the "CAFE pre-processor".

Please refer to the CAFE-3D manual for details about the CAFE-3D set-up.

Version 2006.0

Page 520

ProCAST User Manual

The CAFE-3D pre-processor will create all the necessary inputs for the microstructure calculation. All these inputs are stored in an ASCII file called "prefix_mica.d". This file can also be edited by hand.

Solver
To run the CAFE calculation, one should press the "ProCAST" button of the Manager which will open the following window :

Version 2006.0

Page 521

ProCAST User Manual

Run in "Post-processing mode"

To run a "Post-processing" CAFE calculation, click on the Run button under "CAFE Solver (Post-processing mode)". You have the choice to run DataCAST and ProCAST first or not. In all cases, the ProCAST result files should be converted in the CAFE format (i.e. epf format of CALCOSOFT) before starting the CAFE calculation.
Run in "Full and weak coupling modes"

To use the Full or weak coupling mode, the standard ProCAST executable must be launched (one can in this case either use the Standard Run button or the one under "CAFE Solver (full-weak coupling mode)". The CAFE will be activated as the Run parameter MICRO=1024 is defined. The distinction between weak and full coupling is defined by the parameter "nheat" (card 70) in the prefix_mica.d input file (weak coupling : nheat = 0 / full coupling : nheat = 1).

Post-processing
To view the CAFE results, the CALCOSOFT-3D post-processor must be used. It can be accessed directly from the ProCAST Manager with a right click on the "ViewCAST" button.

Version 2006.0

Page 522

ProCAST User Manual

The use of the CALCOSOFT-3D post-processor is explained in details in the CALCOSOFT-3D manual (which is available in the software installation in the dat/manuals directory).

Version 2006.0

Page 523

ProCAST User Manual

INVERSE MODELING
INTRODUCTION
The Inverse Modeling module of Procast allows to determine, from simple and well controlled temperature measurements, thermophysical data or boundary conditions. The principle of direct and inverse modeling is presented in the figure hereafter.

Direct calculation
In a standard Procast calculation (i.e. a "direct" calculation), one has the following inputs and outputs :
Inputs :

geometry thermophysical properties initial conditions boundary conditions thermal history

Outputs :

Version 2006.0

Page 524

ProCAST User Manual

Inverse calculation
The principle of inverse modeling is to provide the thermal history as an input (i.e. measured cooling curves at several locations) and to have one of the four inputs described above as an output (i.e. unknown). Thus, two possibilities are proposed in the Procast inverse module :
a) Thermophysical calculation

The "thermophysical calculation" consists in the determination of unknown thermophysical data of a material, such as volumic specific heat, thermal conductivity and/or latent heat. Thus, the inputs and outputs are :
Inputs :

geometry initial conditions boundary conditions thermal history thermophysical properties

Outputs :

b) Boundary conditions calculation

The "boundary conditions" calculation consist in the determination of unknown boundary conditions, such as interface heat transfer coefficient, film coefficient, heat flux and/or emissivity. Thus, the inputs and outputs are :
Inputs :

geometry thermophysical properties initial conditions thermal history boundary conditions

Outputs :

A thermophysical calculation and a boundary conditions calculation cannot be mixed in the same analysis (i.e. it is not possible to determine a specific heat and a heat transfer coefficient within the same calculation). In this manual, the unknowns will be called by the generic name of "properties" or "beta value". A "property" (or a "beta value") can thus be a specific heat, a thermal

Version 2006.0

Page 525

ProCAST User Manual

conductivity, a latent heat, an interface heat transfer coefficient, a film coefficient, a heat flux or an emissivity.

Version 2006.0

Page 526

ProCAST User Manual

MODEL SET-UP
Before running any inverse calculation, first, the "direct" problem has to be set-up. For the setting of a direct ProCAST calculation, please refer to the Pre-processing chapter. Once the direct problem is set-up, the file prefixd.dat contains all the information about : geometry (mesh) thermophysical properties of materials interface heat transfers boundary conditions initial conditions run parameters

The inverse modeling set-up intend to specify on which data the inverse calculation will be done and to configure the run parameters for the calculation. PreCAST allows to set-up the inverse problem in an interactive manner (with the "inverse" menu), in order to create the prefixid.dat file.

One have first the choice to select among a "Material Properties" problem or a "Boundary Conditions" problem.

Material Properties
If one wants to run a "Material Properties" inverse calculation, the following panel should be activated (with the "Activate" check box) and configured. For each material present in the model, the user has to decide whether he would like to run an inverse calculation on the Specific heat (Cp), the thermal conductivity (K) and/or the latent heat (L).

Version 2006.0

Page 527

ProCAST User Manual

By default, all the properties are deactivated ("N" stands for No). By clicking on the "green N", it will toggle to "Y" (for Yes). In the example below, it is intended to run an inverse calculation on the Aluminium alloy specific heat and latent heat (no inverse calculation on the thermal conductivity).

Please note that one should run an inverse calculation on one material domain only. If the property (like the specific heat) is defined in the material property file as a constant, the inverse calculation will be done on this constant value. If the value is defined as a temperature function, the inverse calculation will be performed for each temperature of this function.
Important

For an inverse calculation involving "Specific heat" determination, the RUN PARAMETER "THERMAL" should be set to 2 ( THERMAL 2 ). This is mandatory, otherwise, the solution will not converge. In any other configuration not involving specific heat, any thermal model can be used (THERMAL = 1 or 2).

Boundary conditions

Version 2006.0

Page 528

ProCAST User Manual

If one wants to run a "Boundary Conditions" inverse calculation, the following panel should be activated (with the "Activate" check box) and configured. On the left, all the available interfaces are listed, whereas on the right, all the "Heat" boundary conditions are present. For interfaces, one should decide on which interface(s) it is planed to perform an inverse calculation (by toggling between N and Y). In the example hereafter, the inverse calculation is run on the interface 3. For Boundary conditions, one has the choice to perform an inverse calculation on either the heat transfer coefficient (H), the heat flux (Q) and/or the emissivity (E).

If the property (like the interface heat transfer coefficient) is defined in the database as a constant, the inverse calculation will be done on this constant value. If the value is defined as a temperature (or time) function, the inverse calculation will be performed for each temperature (time) of this function.

Inverse settings
Finally, some setting should be defined. The node numbers, corresponding to the measurement points must be specified. It is important to note that the node numbers should be determined in ViewCAST, on results (as the creation of interfaces in PreCAST, after MeshCAST is adding new nodes and the optimization of PreCAST is renumbering the nodes).

Version 2006.0

Page 529

ProCAST User Manual

Moreover, the order with which the nodes are entered into the list should be the same as the order of storage of the nodes in the prefixim.dat file. If the measurements are in different units than C, one should specify it with the "Measurement Units" setting.

Then, the Inverse "Run Parameters" should be specified at this stage.

tau : Time constant for the filtering of the measurements

In order to remove small perturbations which might occur during the measurements, the curves are filtered using the time constant, tau . The units of TAU are [s]. Default value : 1
sigma : Weighting coefficient for temperature

The weighting coefficient for temperature should be kept small in order to have a good convergence. A value of O.1 C has always given good results. Default value : 0.1
conv : Convergence tolerance

The convergence will be reached when the relative variation, between two iterations, of each property will be smaller than the value, conv. Default value : 0.05

Version 2006.0

Page 530

ProCAST User Manual

varb

: Variation of each beta value during an iteration

During an iteration, the beta values will be perturbed one after the others in order to determine the sensitivity coefficients of each property. To do so, each beta value will be changed by a given amount corresponding to the value of varb multiplied by the beta value. Values of varb between 0.05 - 0.2 are convenient and correspond to a variation of 5 to 20% of the beta values Default value : 0.1
itermax : maximum number of iterations before the stop of the calculation

It is useful to define the maximum number of iteration at which the calculation is stopped. In some cases, if the tolerance is too small, the variation of the beta values will not be within the tolerance, although the calculation would have converged. A value between 15 and 30 iterations is reasonable. Default value : 15

Beta values settings

The structure of the prefixid.dat file is described in the File formats section. For each beta value (i.e. for each unknown property), the following parameters are defined (default values are automatically proposed by precast, see below):
ibindex :

material number, interface number, boundary number (this corresponds to the number which appears before the material, interface or boundary condition item in the corresponding window)
itunit :

unit code for the temperature or time "tbeta" used for temperature-dependant or time-dependant properties
ibunit :

unit code for the beta value

tbeta :

temperature or time corresponding to the beta value (zero when a CONSTANT value is used)
beta :

beta value (unknown property), initial guess the initial guess corresponds to the value initially defined in the HEAT boundary condition of the direct calculation.
ibeta :

flag to indicate what is the type of the beta value (constant, time- or temperaturedependant) : 1 : The type of the beta value is CONSTANT 2 : The type of the beta value is TEMPERATURE dependant 3 : The type of the beta value is TIME dependant
weight :

weighting coefficient for the beta value

betamin :

Version 2006.0

Page 531

ProCAST User Manual

minimum accepted value for beta

betamax :

maximum accepted value for beta

relax :

maximum variation of the beta value between two iterations (i.e. if relax=3, the beta value can increase at maximum by 3 times the previous value or can be no smaller than 1/3 of its previous value within the same iteration). PreCAST proposes default values that may be changed by the user in the prefixid.dat file (by editing the prefixid.dat file). The default values are :
ibeta :

This value is automatically set to 1, 2 or 3. This value should not be changed by the user.
tbeta :

This value prompts the time or temperature value of the corresponding beta value. This value can not be changed by the user. If ibeta = 1 (CONSTANT type). a zero value is written.
beta :

This value corresponds to the initial guess. By default, it is the value which was in the corresponding dataset.
weight :

By default this value is set equal to initial guess of the beta value. This value should be of the same order of magnitude than the final beta value or larger (for a good convergence).
betamin :

This value is the minimum allowed value for beta. It should be set to the physical minimum that can be expected. By default, betamin is set to 1/100 of beta
betamax :

This value is the maximum allowed value for beta. It should be set to the physical maximum that can be expected. By default, betamax is set to 100 times beta
relax :

By default relax is set to 3. If a property is defined as both temperature and time dependant, it is strongly advised not to perform an inverse calculation in the same time on both the temperature-dependant beta values and the time-dependant beta values.

Additional information concerning the Inverse Settings :

weight : weighting coefficient for the beta value. This weighting coefficient determines the importance of the allowed variation of the beta value during an iteration. If the value is large, the beta value could vary in a large extent. If this value is small, the beta value could not vary much. In fact, for practical use, this value should be in the order of magnitude of the beta value in order to give satisfactory results.

Version 2006.0

Page 532

ProCAST User Manual

betamin : minimum accepted value for beta For physical reasons, the beta values should not be smaller than a specified value, betamin (i.e. a heat transfer coefficient should not be negative !). This limit is set in order to speed up the convergence. betamax : maximum accepted value for beta For physical reasons, the beta values should not be larger than a specified value, betamax (i.e.. a thermal conductivity of steel can not exceed a value of pure copper or pure silver !). This limit is set in order to speed up the convergence. relax : maximum variation of the beta value between two iterations In order to reduce divergence problems of the inverse calculation, it is possible to limit the variation of the beta values from one iteration to the next one. A value of relax=3 would mean that the new beta value will be limited to 3 times the previous one or 1/3 of the previous one. The value of relax should be strictly larger than 1. If the relax value is rather small, the convergence will be obtained in more iterations, however, the convergence should be obtained more smoothly.

Version 2006.0

Page 533

ProCAST User Manual

INVERSE RUN
In order to run an inverse calculation, the following files should be present : prefixd.dat (ProCAST input file) prefixp.dat (ProCAST Run parameters prefixid.dat (Inverse input file) prefixim.dat (Inverse measurements)

Then, the calculation can be started. To do so, a right click on the "ProCAST" button on the manager is opening a sub-menu, which allows to start the inverse solver.

It could also be launched manually from a Command Window as :

inverse prefix

Automatically, DataCAST is launched and the inverse calculation is starting. The status of the inverse calculation can be viewed with the "invstat prefix" command (from a second Command Window).
Output Files :

During the execution of the calculation, several files are created, beside the usual ProCAST files : prefixir.dat This file contains the evolutions of the beta values along the iterations, as well as the evolution of the residual. prefix.stat This file contains information of the status of the calculation (see below)prefixic.dat

Version 2006.0

Page 534

ProCAST User Manual

This file contains the calculated temperature evolutions (of the last calculated iteration) corresponding to the locations of the measured curves. prefix.tmp This file is a temporary file created during the calculation. This file is needed only during the calculation and is useless for the user. This file is automatically deleted at the end of the calculation However, if the calculation is interrupted by the user before completion, it should be deleted manually by the user. prefix.list When the calculation is run in the background mode, the printout which appears normally on the screen is listed in this file. This file is not written in a continuous mode, but only when the buffer of the machine is unloaded (thus, this may take a few minutes after the beginning of the calculation before having something in this file).
Monitoring the calculation :

As inverse calculation could be quite long and are often run in "background mode", there are three ways to monitor the status of the calculation. The monitoring of the calculation is a good way to check whether the calculation is stable and whether it is converging in good conditions. The three monitoring options are the following : a) Status of the calculation : the following command :
invstat prefix

prints the status of the calculation (it is equivalent to "prostat" in Procast). The number of iterations already calculated is printed as well as the beta values of the last iteration. This can also be accessed from the manager with a right click on the "Status" button.

Version 2006.0

Page 535

ProCAST User Manual

The residual of the current iteration is also printed. The residual is the average temperature difference between the measured and the calculated curves (average on all the steps and all the curves, at the given iteration). The residual will decrease towards zero along the iterations. The monitoring of the residual is a good way to see how the convergence is reached. In addition, it is mentioned if the calculation is still under progress or if it is completed (if the calculation was interrupted by the user, it will still display that the calculation is under progress). This command should be called in the current directory where the prefix* files are present (this is not like "prostat" which can be called from any directory). b) Evolution of the beta values The beta values obtained at each iteration are stored in the file prefixir.dat Thus, during the calculation, it is possible to visualize the evolution of each beta value (from the first to the current iteration) and see whether the calculation is converging or not. The format of the file prefixir.dat is the same than the one of the measurement. Curve "i" of the file prefixir.dat correspond the ith beta value. In addition, the residual is stored as the last curve of this file. Thus, it is possible to graph the evolution of the residual as function of the iterations. c) Comparison between the measurements and the calculated curves The goal of inverse modeling is to find the best set of beta values with which the calculated temperature curves will match the measured ones. At the end of each iteration, the calculated curves corresponding to the location of the measurement points are stored in a file labeled prefixic.dat This file is updated at each iteration. Thus, it is possible to visualize during the calculation (with ViewCAST) the superimposition of the measured curves and the calculated ones. Of course, during the first iteration, the match will be bad, however, at the end of the calculation, the match should be quite good. Final result At the end of the calculation, the final beta values correspond to the values of the last iteration in the file prefixir.dat. At the end of the calculation, it is advisable to check the quality of the convergence. To do so, the file prefixir.dat can be graphed (see section "Monitoring the calculation" above). The convergence is good if the beta values are not changing much during the last iterations and if the residual is small (the residual has the unit of degrees and a residual of "1" means that the average difference between the measured and calculated curves is 1 degree).

Version 2006.0

Page 536

ProCAST User Manual

Depending on the convergence tolerance which is chosen, it is possible that the beta values stay almost constant (but oscillate around an average value), however, the calculation is continued until the maximum number of iterations which was specified is reached. In such case, it is possible to reduce the maximum number of iterations (for a further run) in order to reduce the computing time, without changing the quality of the result. In order to check the quality of the determined beta values, one can superimpose the measured and the calculated curves (see section "Monitoring the calculation" above). The quality of the match will be an indication of the quality of the final beta values, as given by the residual.
Remarks

Inverse modeling involves many direct ProCAST calculations. At each direct ProCAST calculation, the results are written in the result file at each time-step. This is resulting in many disk access. If the calculation in performed in a network, it is advisable to run the calculation in a directory which is on the closest disk from the CPU, otherwise, these disk access through a network may slow down significantly the calculation. It is always possible to run the calculation in the directory /tmp and then move the final results to the home directory of the user.
Limits

The following limits have been set in the inverse solver :

Maximum Maximum Maximum Maximum number number number number of beta values : 100 time-steps for the measurements : 500 of thermocouples : 20 of iterations : 100

Version 2006.0

Page 537

ProCAST User Manual

FILE FORMATS
Inverse settings file
All the inverse settings are stored in the file prefixid.dat. Like the file prefixp.dat the user may change this file "by hand" with an editor and then run again inverse. The format of this inverse settings file is described hereafter.
Definitions : ntcunit :

unit code for the temperature measurements

ntccode :

code which specifies how the thermocouple locations are defined : ntccode = 0 : coordinates of the thermocouple ("0" is not implemented in the current version, use only ntcode=1) ntccode = 1 : node number corresponding to the thermocouple location
node :

node number in the mesh used for the calculation corresponding to the measurement points (ntccode=1) (the order of the node numbers should correspond to the order of the cooling curves in the file prefixim.dat)
imat :

material number where the thermocouple is located (ntccode=0)

ixyzunit :

unit code for the coordinates when ntccode=0

x,y,z :

x, y and z coordinates of the thermocouple

tau :

time constant for the filtering of the measurements [s]

sigma :

weighting coefficient for temperature [K]

conv :

convergence tolerance [-]

varb :

variation of each beta value during an iteration [-]

itermax :

maximum number of iterations before the stop of the calculation

ibindex(i) :

material number, interface number, boundary number

itunit(i) :

unit code for the temperature tbeta(i) ibunit(i) :unit code for the beta value, beta(i)
tbeta(i) :

temperature corresponding to the beta value, beta(i)

beta(i) :

beta value (unknown property), initial guess

ibeta(i) :

Version 2006.0

Page 538

ProCAST User Manual

flag to indicate if the beta value will be taken into account in the inverse calculation 1: the beta value is an unknown of the problem and will be taken into account in the calculation (it is a constant value) 2: the beta value is an unknown of the problem and will be taken into account in the calculation (it is a temperature-dependant value) 3: the beta value is an unknown of the problem and will be taken into account in the calculation (it is a time-dependant value)
weight(i) :

weighting coefficient for the beta value

betamin(i) :

minimum accepted value for beta

betamax(i) :

maximum accepted value for beta

relax(i) :

maximum variation of the beta value between two

iterations

All the parameters and variables defined above are defined in groups (or families), which are labeled by keywords. The keywords are : MEASUREMENTS PARAMETERS THERMAL_CONDUCTIVITY SPECIFIC_HEAT LATENT_HEAT INTERFACE_HEAT_TRANSFER EXTERNAL_FILM_COEFFICIENT EXTERNAL_HEAT_FLUX EXTERNAL_EMISSIVITY
Format : MEASUREMENTS ntcunit Loop on the thermocouples (i) ntccode(i) if ntccode(i)=0 imat(i) ixyzunit(i) x(i) y(i) z(i) else if ntccode(i)=1 node endif End Loop on the thermocouples PARAMETERS tau sigma [real] [real] [integer] [integer] [integer] [integer] [real] [real] [real] [integer]

Version 2006.0

Page 539

ProCAST User Manual

conv varb itermax

[real] [real] [integer]

For all the following keywords, the structure of data is the same. Only the necessary keywords should be specified, which correspond to the desired beta values. THERMAL_CONDUCTIVITY SPECIFIC_HEAT LATENT_HEAT INTERFACE_HEAT_TRANSFER EXTERNAL_FILM_COEFFICIENT EXTERNAL_HEAT_FLUX EXTERNAL_EMISSIVITY
Loop on the beta values (i) ibindex(i) [integer] itunit(i) [integer] ibunit(i) [integer] tbeta(i) [real] beta(i) [real] ibeta(i) [integer] weight(i) [real] betamin(i) [real] betamax(i) [real] relax(i) [real] End Loop on the beta values

As the data are read by a "Free Format Reader", data can be placed in any fashion (lines, columns,...), as long as the sequence of the data corresponds to the specified one. However, each line should not exceed 256 characters. Moreover, comments can be included in the data file after a "#" sign (all the characters between the "#" sign and the end of the line will be ignored and can be used as comments).

Inverse measurement files : X-Y format

The format of the measurement file (prefixim.dat), as well as the files prefixir.dat and prefixic.dat is the following (it corresponds to the prefix.tt file format of ViewCAST) : Sequential formatted file (ASCII)
nx, ny, node#1,node#2,node#3, .... , node#ny x(1),y(1,1),y(2,1),...,(ny-1,1),y(ny,1) x(2),y(1,2),y(2,2),...,(ny-1,2),y(ny,2) x(1),y(1,3),y(2,3),...,(ny-1,3),y(ny,3) ... ...

Version 2006.0

Page 540

ProCAST User Manual

x(1),y(1,nx),y(2,nx),...,(ny-1,nx),y(ny,nx) Description : nx :

number of values on the X-axis (this corresponds to the number of time-steps if the X-axis is time)
ny :

number of curves
node#i :

node number corresponding to the curves. These number can be any integer numbers (as they are not used by the inverse calculation). The only requirement is that ny integer values are present. The reason of these values is that it corresponds to the format which is compatible with PostCAST.
x(j) :

value "j" on the X-axis

y(i,j) :

value "j" on the Y-axis for curve "i"

Version 2006.0

Page 541

ProCAST User Manual

INVERSE APPENDIX
Inverse model

Version 2006.0

Page 542

ProCAST User Manual

Inverse calculation involves a large number of direct calculations

Inverse modeling involves many direct calculation. For each iteration, it is necessary to perform one direct calculation for each unknown of the problem, plus the "original" calculation. Thus, if the problem to solve has N unknowns and M iterations are needed to reach convergence, it will involve (N+1) x M calculations. For instance if one has 10 unknowns and it is necessary to calculate 15 iterations, it make a total of (10+1) x 15 = 165 direct calculations ! This shows that in order to keep reasonable computing times, it is necessary that each direct calculation is small and quick. Indeed, if one direct calculation is lasting 1 minute, the inverse problem example shown above will take 165 minutes to solve (about 3 hours). If instead, one direct calculation is lasting 2 days of CPU time, the inverse calculation will take 1 year !!!

"Pseudo 2D modeling" - simple calculations

Considering the above conclusion about short CPU times for the direct calculation, it shows that it is necessary to run the inverse calculation on simple

Version 2006.0

Page 543

ProCAST User Manual

geometries. Thus, the use of 2D geometries is strongly recommended (or is even in most cases "mandatory"). One should realize that it is impossible to consider to perform an inverse calculation on a real casting geometry, as the CPU time for one direct calculation will be just too large. Thus, inverse modeling involves the design of specific experiments, which are used for this purpose only. Thus requirement is reasonable as the unknowns (i.e. the boundary conditions or the material properties) are usually not depending upon the small geometrical details or of the complex shapes. One should however be careful to design an experiment which is representative of the behavior of the real part.

How to design the appropriate experiment

As mentioned above, the experimental set-up should be designed so as to be as representative as possible from the real process where the result of the inverse calculation are to be used. There are no generic rules about the design of inverse experiments as it depends upon the goals that are set. The main and most important rule is "Make it simple". This means that one should stick to 2D models and that if possible only one type of unknown should be determine at a time. In principle, there are no limitation in the inverse modeling towards the maximum number of unknowns. However, it is strongly advised to determine one thing at a time. For instance, design one experiment for the determination of the heat transfer coefficient between the water and the mould in the case of a cooling channel and design an other experiment for the determination of the casting-mould interface heat transfer coefficient. Although it is possible in the principle to determine both together, it is advisable to simplify the problem by decomposing it. The quality of the result of an inverse calculation is proportional to the quality of the inputs of the inverse model. As a reminder, the 4 inputs which are necessary for the inverse modeling are : a) geometry b) initial conditions c) temperature measurements d1) material properties (in case of boundary condition determination) d2) boundary conditions (in case of material properties determination) It is thus of primary importance to make sure that the quality of these 4 inputs are as good as possible.
a) geometry

Version 2006.0

Page 544

ProCAST User Manual

The best way to have an accurate geometry description in 2D is to consider 2D axi-symmetric problems. Thus, the modeling of cylinders or tubes is recommended. One should be careful that the running system and the feeding system must also be axi-symmetrical. In some cases, the use of 2D cartesian can also be applied with success for the modeling of plates or the determination of boundary conditions on flat surfaces.
b) initial conditions

Usually, within the direct solver, it is only possible to set easily constant initial temperatures. Thus, it is important to perform an experiment where the initial temperatures of the different domains are as constant and well known as possible. If the filling stage is very important (as one can loose a significant amount of superheat during the filling), one should consider whether the design is appropriate or if a lower average constant initial temperature can be considered.
c) temperature measurements

The quality of the temperature measurements are of primary importance in order to have a good result. Thus, one should be careful to use appropriate thermocouples (not too thin or not too big). The time response of the thermocouples is also important if one wants to capture fast transient phenomena. When measuring in a solid material, one should ensure that there is an intimate contact between the tip of the thermocouple and the material. Concerning the placement (location) of the thermocouples, please refer to the next section.
d1) material properties (in case of boundary condition determination)

As it is not always easy to know well the material properties of the materials which are used in the real process, it may be advisable to use different materials for the inverse experiments. For instance if one wants to determine the heat transfer coefficient at the casting-mould interface in a high pressure die casting mould (with an aluminium alloy casting and a steel mould), it is possible to cast a well known Al-Si7%-Mg0.3% in a block made out of a well known steel alloy. The idea is to use similar materials where the materials properties are better known, as the heat transfer coefficient will not depend upon slight differences in material properties. For instance, in the example shown above, it is not important to take a steel which is resistant at high temperatures (like it is needed for an hpdc die), as the inverse experiments will be run only a few times (unlike the hpdc die which will undergo many cycles, under high loads).
d2) boundary conditions (in case of material properties determination)

This point is probably the most critical in inverse modeling when one wants to determine the thermophysical properties of a material. Unlike case (d1) above where it is possible to use well known materials, in this case, one has to find a method to determine the boundary conditions in the experimental set-up. Usually, one have to run a first inverse model to determine the boundary conditions and then to run a second inverse calculation for the material properties determination.

Version 2006.0

Page 545

ProCAST User Manual

An other solution is to use some of the measured temperatures as boundary conditions (fixed temperature boundary conditions). However, in such a case, one can only determine the diffusivity of the material (the diffusivity is the ratio of the thermal conductivity over the volumic specific heat). It is thus impossible to separate the thermal conductivity from the specific heat (without knowing the flux).

Model first your experiment

As it was mentioned in the previous section, it is not possible to give generic rules about design of inverse experiments. However, the user is able to check the quality of its design by himself. It is strongly recommended to model first the experimental set-up before even building it. Of course, as the properties or boundary conditions, such a modeling has to be done with guessed values (one should notice that it is important to have at least an idea about the order of magnitude of the unknown value before starting to explore inverse modeling techniques). Based upon this first modeling, it is possible to run a few additional calculations by modifying the guessed values. This corresponds to some kind of "sensitivity analysis". Based upon the results of these calculations, it will be possible to see whether the calculated temperature curves at different locations are sensitive or not to the variation of the unknown (or guessed) values. This provides a useful mean of determining where are the best locations to place the thermocouple. Indeed, one would like to have the largest temperature variations when the unknown values are changing slightly (for instance, it is useless to place a thermocouple at the first floor of a building to measure the influence of the opening of a window at the 15th floor !). Of course, it is advised to place thermocouples as close as possible from the location where the property has to be determine. For instance, if one wants to determine an interface heat transfer coefficient, it is advisable to place thermocouples quite close from that interface, on either sides. However, one should be careful not to place thermocouples too close from the surface as it is very difficult to measure accurately very close from surfaces. Thus, it is better to have a better quality measurement, a little bit further away from the surface, than a low quality measurement, very close from the surface. Concerning the number of thermocouples, one shall first say that the computing time does not depend at all upon the number of thermocouples (it only depends upon the number of unknowns). However, experience shows that it is advisable to use a small number of thermocouples (less then 10) to obtain good results. Thus, the user has to concentrate its attention about where are the optimum locations of the thermocouples rather than "flooding" the experiments with many thermocouples. Moreover, the inverse solver is based upon a "least square fit" on all the thermocouples, in a global fashion. Thus, if some thermocouples are

Version 2006.0

Page 546

ProCAST User Manual

"useless" and provide not very accurate measurements, they may degrade the final answer. Finally, the advantage of modeling first the experiments, before running them, is also to show how much CPU time will be needed to run one direct calculation. Thus, it will show whether the set-up is appropriate or not for fast inverse calculations.

Run a direct calculation with your initial guesses and compare it with the measurements
At this stage, the experimental set-up is built and measurements were performed. Before running the inverse analysis, it is advisable to re-run once the direct calculation with the initial guesses of the unknown. Then, the calculated temperature curves corresponding to the location of the thermocouples can be compared to the measured curves in the experimental set-up. If the calculated curves are quite close from the measured ones, one can run the inverse calculation. If on the opposite, the calculated curves are very far away from the measured ones, it is good to re-run once more the direct calculation with new guesses which seems more appropriate. This "manual" inverse modeling may speed-up considerably the inverse calculation as convergence will be reached much faster.

Start by simple calculations and then increase the complexity (i.e. the number of unknowns)
Inverse modeling can be used for the determination of quite complex boundary conditions (e.g. time- or temperature-dependent heat transfer coefficients, temperature-dependant equivalent specific heat,) which may involve a rather large number of unknowns. In order to reach a good and quick convergence of the inverse calculation, it is advisable to decompose the problem in several simpler stages. In order to illustrate this, let's consider the example of the determination of specific heat, the thermal conductivity and the latent heat (all in the same time) of an Al-alloy. One expects that the thermal conductivity and the specific heat will be both function of the temperature. However, in a first stage, it is good to run an inverse calculation with constant properties. This means that the inverse modeling is looking for one value of specific heat, one value of thermal conductivity and one value of latent heat. Figure A1.1 shows the result of such a calculation with the superimposition of the measured and calculated curves. The corresponding constant material properties are also shown. One can see that the correspondence between the measurements and the calculation is only fair, which is not a surprise as constant properties were considered. This first calculation provides a way to

Version 2006.0

Page 547

ProCAST User Manual

"calibrate" the unknowns. Then, it is possible to refine those unknowns in a second inverse calculation. Figure A1.2 shows the superimposition of the calculated and measured curves, when thermal conductivity is described by a temperature function using 3 unknowns and the specific heat by a temperature function using 4 unknowns (the latent heat is still described by one value). The initial guesses used in this second inverse calculation corresponded to the result of the first inverse calculation. One can see on figure A1.2 that the correspondence between the calculation and the measurements are much better in this second case. Then a third inverse calculation was performed, using a 6 points specific heat curve and a 5 points thermal conductivity curve (an one latent heat). Again the initial guesses used in this latter calculation were taken as interpolation of the results of the previous calculation. Figure A1.3 shows a very good correlation between measurements and calculation. One shall notice that in the above example, it would certainly been possible to find directly the final result (of Figure A1.3). However, in complex problems, it is not always possible to start with a reasonable initial guess and thus, convergence can be hard to find (or even impossible in some cases).

Figure A1.1 Comparison of the measured and calculated curves obtained with 1 values of thermal conductivity, 1 specific heat values and 1 latent heat.

Version 2006.0

Page 548

ProCAST User Manual

Figure A1.2 Comparison of the measured and calculated curves obtained with 3 values of thermal conductivity, 4 specific heat values and 1 latent heat.

Figure A1.3 Comparison of the measured and calculated curves obtained with 6 values of thermal conductivity, 5 specific heat values and 1 latent heat.

Temperature- or Time-dependant boundary conditions or properties

Usually, the thermophysical properties are temperature-dependent and the boundary conditions may be time- or temperature-dependent too.

Version 2006.0

Page 549

ProCAST User Manual

In such a case, the user has to decide the type of dependency that should be used in the inverse calculation. To illustrate that, one can use the example of the determination of the equivalent specific heat of an alloy (as shown in Figure A1.4). One can see that the curve is almost constant above and below liquidus, whereas it exhibits very strong variations between the liquidus and solidus In such a case, the user has to define himself the temperature points at which the unknowns will be calculated. As in the above mentioned example, these points should be selected in between the solidus and liquidus temperatures (with a reasonable number of points in order to describe the curve). Above the liquidus or below the solidus, one can consider constant values or variable values as desired. Concerning the number of points to be selected, one should select enough points, but not too many. Indeed, if too many points are selected, one can obtain oscillations of the final beta values.

Figure A1.4 Equivalent specific heat obtained by inverse modeling, showing the tabulation points (tabulated temperatures) that were chosen.

What to do if it does not converge ?

In some cases, it is possible that the solution does not converge. This can be seen if the residual is not decreasing or is even increasing, if the correspondence between the measured and calculated curves is completely wrong or if the beta values are oscillating. Usually, convergence should be obtained within 15-20 iterations. If more iterations are needed, it means that something goes wrong. In such case, one should check the following points :

Version 2006.0

Page 550

ProCAST User Manual

Are the thermocouples well defined ? Is the order of the list of the thermocouple corresponding to the order of the measured curves in the measurement file ? In case of time- or temperature-dependent beta values, are the tabulation points appropriate (for instance, if one is looking for a variation between 500 and 700 and the tabulation points are chosen between 600 and 800, it might not work). Are they any (or enough) measurements in the range of the tabulation points? (for instance if one looks for a temperature dependant heat transfer coefficient between 100 and 800C and that the measurements are only between 400 and 700C, the values below 400 and above 700C will never converge). Are the initial temperatures of the model corresponding to the ones of the measurements (or are the initial measured temperatures constant ?). Is the geometry of the model and the known boundary conditions well defined ? Are the material properties of the different materials used in the measurements well known ?

Remarks and Tips

In a thermophysical calculation, it is possible to determine in the same time specific heat, thermal conductivity and latent heat properties However, one has to be careful that these properties are not totally independent as the diffusivity is the ratio of the thermal conductivity over the specific heat and that the specific heat and the latent heat are both contained in the enthalpy. The calculation will converge much faster if the initial beta values (i.e. initial guess) are the closest for the final values. It is thus advised to run first a direct calculation with the initial guess in order to check that the calculated curves are not too far from the measured ones. The time-steps of the measurements can be completely free and is totally independent from the one used in the calculation. For time- or temperature dependant properties, one has to perform the inverse calculation on all the values of the table. For specific heat determination, the RUN PARAMETER "THERMAL" has to be set to 2 ( THERMAL 2 ). In all other cases, not involving the specific heat determination, any thermal model can be used (THERMAL = 1 or 2).

Version 2006.0

Page 551

ProCAST User Manual

INPUT-OUTPUT FILES
This section is listing the different files which are created and used by the different modules of ProCAST.

MeshCAST files
PREFIX.gmrst PREFIX.sm PREFIX.stlsm PREFIX.igs PREFIX.iges PREFIX.stp PREFIX.step PREFIX.xmt_txt PREFIX.x_t PREFIXd.dat PREFIX_pre_sh.sm PREFIX.patran PREFIX.out PREFIX.ideas PREFIX.unv PREFIX.nastran PREFIX.ansys PREFIX.stl PREFIX.bstl PREFIX.mesh PREFIX_sub.sm PREFIX_sub_act.sm PREFIX.elem PREFIX.node PREFIX.wrk merge PREFIX.ceg merge usr_cmnds *.gif mesh.print PREFIX.psm IO IO O I I I I I I I I I IO I I I I I I IO O O O O O O IO Binary ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII Restart file for meshing Surface mesh Surface mesh from STL file IGES file IGES file STEP file STEP file Parasolid file Parasolid file ProCAST restart file Shell meshing recovery file Patran file (volume mesh) Patran file (surface mesh) I-DEAS file (volume mesh) I-DEAS file (surface mesh) Nastran file (volume mesh) ANSYS file (volume mesh) STL ASCII file STL Binary file Volume mesh Surface mesh extracted from volume mesh An active subset of original surface mesh Ansys element file Ansys nodal file Temporary working file for surface mesh Temporary working file for surface mesh User defined icons GIF capture

PreCAST files
PREFIXd.dat IO ASCII General model setup (including the geometry), also used for Restarts PREFIXp.dat IO ASCII Run Paramerters file, also used for Restarts PREFIXg0.dat IO ASCII Cut-off values per domain for Thixo casting bc.db I ASCII Boundary conditions database matl.db I ASCII Material properties database stress.db I ASCII Stress properties database intf.db I ASCII Interface heat transfer database proc.db I ASCII Process database CT_Al.db I ASCII Al Thermodynamic database CT_Fe.db I ASCII Fe Thermodynamic database CT_Ni.db I ASCII Ni Thermodynamic database CT_Mg.db I ASCII Mg Thermodynamic database

Version 2006.0

Page 552

ProCAST User Manual

CT_Ti.db I PREFIXt.unf I condition PREFIX.rnm I PREFIXd.unf I extraction PREFIXid.dat IO PREFIX.out I PREFIX.unv I PREFIX.mesh I predefined_%d_p.dat default_p.dat

ASCII Binary Binary Binary ASCII ASCII ASCII ASCII ASCII ASCII

Ti Thermodynamic database Temperature results for extracting initial Node renumbering Time step information, used for IC Inverse data PATRAN mesh neutral file IDEAS mesh universal file Meshcast file Pre-defined p.dat files Default values of the p.dat file

INPUT files (for the ProCAST solver - generated by DataCAST)

These files are also Output files for ViewCAST
PREFIXd.dat IO ASCII General model set-up (including the geometry), also used for Restarts PREFIXg.unf O Binary Model description (Mesh, Material Settings , BC settings) PREFIXt.unf O Binary Temperature results file, storing initial values PREFIXs.unf O Binary Solid node coordinates for moving elements PREFIXo.unf O Binary Enclosure node coordinates PREFIX.rnm O Binary Node renumbering information, used by PreCAST for temperature extraction PREFIXff.unf O Binary Free faces information (for faster postprocessing) PREFIXd.out O ASCII DataCAST synopsis information file

OUTPUT files (for ViewCAST)

PREFIXa.unf PREFIXb.unf PREFIXc.unf PREFIXci.unf PREFIXcv.unf PREFIXcp.unf PREFIXd.unf PREFIXe.unf PREFIXemg.unf PREFIXes.unf PREFIXf.unf PREFIXfl.unf PREFIXfr.unf PREFIXfs.unf PREFIXft.unf PREFIXgn.unf PREFIXgs.unf PREFIXht.unf PREFIXk.unf PREFIXmg.unf PREFIXn.unf PREFIXo.unf PREFIXp.unf PREFIXpa.unf PREFIXpp.unf PREFIXpv.unf PREFIXprt.unf PREFIXq.unf PREFIXr.unf PREFIXs.unf O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Start and end time of solidification Gas porosity Shrinkage porosity file Cracking indicator Pore volume Contact pressure Time step - step Turbulent dissipation rate Electric field VON MISES stresses Fluid fraction (FVOL) Fatigue life Freckle indicator Solid fraction Filling (wetting) time Strain values at the gauss points Stress values at the gauss points Hot tearing indicator Turbulent kinetic energic file Magnetic field Non newtonian shear rate and viscosity Nodal coordinates offset for enclosure Pressure Peak acceleration for core blowing Peak pressure for core blowing Peak velocity for core blowing Particle traces Heat flux Compressible density Solid node coordinates for moving elements

Version 2006.0

Page 553

ProCAST User Manual

PREFIXtv.unf PREFIXx.unf PREFIXy.unf PREFIXz.unf PREFIXu.unf PREFIXv.unf PREFIXw.unf PREFIX.vf PREFIX.view PREFIX.serr PREFIXp.out

O O O O O O O O O O O

Binary Binary Binary Binary Binary Binary Binary ASCII ASCII ASCII ASCII

Turbulent viscosity X displacements Y displacements Z displacements X velocity Y velocity Z Velocity View factors after symmetrization View factor groups View factor row sum errors ProCAST run information file

Files necessary for a RESTART

(some of these files are needed for ViewCAST)
PREFIXAmtx.unf integral method PREFIXbs0.unf PREFIXbs1.unf PREFIXcr.unf PREFIX.ctoc PREFIX.fic PREFIX.fom PREFIXfv.unf indicator PREFIX.glue PREFIXgp.unf PREFIXmdc.unf PREFIX.mcf PREFIX.mrv PREFIXm1.unf PREFIXm2.unf PREFIXm4.unf Eutectic Growth PREFIXm16.unf PREFIXm32.unf PREFIXm64.unf PREFIXm128.unf PREFIXm256.unf Gamma/Cementite PREFIXm512.unf PREFIXpn.unf PREFIX.pr PREFIX.rmp PREFIX.sel PREFIX.shk PREFIXsm.unf PREFIXsr.unf PREFIXst.unf PREFIXsth.unf PREFIX.t2q PREFIXtb.unf PREFIXthx.unf PREFIXl.prf PREFIXlp.unf O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Magnetic potential used with boundary Back stress 0, used in kinematic hardening Back stress 1, used in kinematic hardening Curing reaction for RTM Radiation view factor information Virtual mold information Lost foam information Fraction of void, used in cracking Status of glue nodes in lost foam Contact gap Stored mass and diffusion matrices Stored coefficients for momentum equation Lost foam information Micromodel, Ductile Iron Eutectic Micromodel, Primary Dendrite Micromodel, Coupled Stable/Metastable Micromodel, Micromodel, Micromodel, Micromodel, Micromodel, Gray/White Iron Eutectic Ductile Iron Eutectoid Gray Iron Eutectoid Peritectic Transformation Delta/Gamma, Gamma/Alpha,

Micromodel Average penetration, penalty number Pressure relaxation info Radiation face remapping Highest filled element (FREESURF=2) Shrinkage data, obsolete Slave-master data Load residual Thermal load Last thermal load Tri-to-quad (Radiation) Turbulence data, obsolete Thixocasting viscosity Temperature gradient at the solidus Filling BC shut-off times

Files generated by ViewCAST

PREFIXes.unf PREFIXs1.unf PREFIXs2.unf O O O Binary Binary Binary Effective stress results Principal stress 1 results Principal stress 2 results

Version 2006.0

Page 554

ProCAST User Manual

PREFIXs3.unf PREFIXss.unf PREFIXsp.unf PREFIXsx.unf PREFIXsy.unf PREFIXsz.unf PREFIXsxy.unf PREFIXsyz.unf PREFIXszx.unf PREFIXepn.unf PREFIXac.unf PREFIXac.ntl PREFIXdas.unf PREFIXdas.log PREFIXdas.ntl PREFIXm.unf PREFIXm.log PREFIXm.ntl PREFIXi.unf PREFIXi.log PREFIXi.ntl PREFIXl.unf PREFIXl.log PREFIXl.ntl PREFIXr.ntl PREFIXg.ntl PREFIXf.ntl PREFIXe.ntl PREFIXs.ntl PREFIXt.ntl PREFIXt.log PREFIXq.ntl PREFIXq.log PREFIXp.ntl PREFIXp.log PREFIXv.ntl PREFIXv.log PREFIXd.ntl PREFIXd.log PREFIX.tt PREFIX.lv PREFIX.clip *.gif

O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O

Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary Binary ASCII ASCII Binary ASCII ASCII Binary ASCII ASCII Binary ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII

Principal stress 3 results Maximum shear stress Average normal stress Normal stress in X direction Normal stress in Y direction Normal stress in Z direction Shear stress on X-Y plane Shear stress on Y-Z plane Shear stress on Z-X plane Effective plastic strain Alpha case Alpha case neutral file Secondary dendrite arm spacing SDAS log SDAS neutral file Mapping factor (RGL) Mapping factor log Mapping factor neutral file Isochrons Isochrons log Isochrons neutral file Feeding length Feeding length log Feeding length neutral file Radiation face geometry Mesh neutral file Radiation face to group neutral file Radiation row sum error neutral file Stress results neutral file Temperature neutral file Temperature log file Heat flux neutral file Heat flux log file Pressure neutral file Pressure log file Velocity neutral file Velocity log file Displacement neutral file Displacement log file Temperature-time Last view Slice data GIF file

Miscellaneous
(Some files are created for internal purposes)
PREFIX.cm PREFIX.mtx bem.out PREFIXc.out debug.out dbg.out mem_usage poros.out PREFIX.rhs PREFIX.lhs PREFIX.sol PREFIXrf.out PREFIX.rhs_v sconv PREFIXf.out PREFIX.phs

Version 2006.0

Page 555

ProCAST User Manual

PREFIX_rho.dat PREFIX_k.dat PREFIX_mu.dat PREFIX_tdd.dat PREFIX.err PREFIXt.out

Version 2006.0

Page 556

ProCAST User Manual

TIPS & TRAPS

RESTART
What should we do when we would like to restart a calculation after some data modifications ? If the content of the d.dat file was not modified, one should not run DataCAST and it is possible to run directly ProCAST. Typically, this happens if the content of the p.dat file only was modified. If PreCAST is loaded, the d.dat file is not modified only if the Run Parameters are changed. If anything else is changed in PreCAST, it is necessary to run first DataCAST and then ProCAST. When the d.dat file is modified, DataCAST has first to be run and then ProCAST should be launched. When one would like to CONTINUE a calculation from a given timestep, DataCAST with the -u option should be used. This is mandatory in order not to erase the existing result files (for more details please refer to the "Solver" section of the "Run of the calculation" chapter.

Version 2006.0

Page 557

ProCAST User Manual

CONVERGENCE PROBLEMS
If a calculation does not converge or has problem to converge (this can be seen in the p.out file - please refer to the Troubleshooting section for more details), the following operations can be tried : The Thermal model does not converge check the thermal material properties (in particular if an enthalpy curve is defined, check that the curve is strictly in an ascending order) modify the solver parameters from TDMA to CGSQ (with the CGSQ Run parameter) Increase the convergence thermal Run parameter CONVT from 1 to 2 (maximum 5) Modify the mesh in order to have a better quality The fluid flow model does not converge (velocities of pressures) modify the solver parameters from TDMA to CGSQ (with the CGSQ Run parameter) Increase the convergence flow Run parameter CONVV from 0.05 to 0.1 (maximum 0.5) Modify the mesh in order to have a better quality

Version 2006.0

Page 558

ProCAST User Manual

STRESS CALCULATIONS
Stress calculations might be rather long (in terms of CPU time), as the six components of the stress tensor have to be computed, in addition to Temperature, Pressure and the three components of the velocity. In order to reduce CPU time, it is recommended to use as much as possible the "Vacant" and "Rigid" options, as well as the symmetries. The contact algorithm does not take into account friction. Stress calculations are starting as soon as the fraction of solid is larger than a critical fraction defined by the Run Parameter CRITFS. By default, CRITFS = 0.5. It is strongly advised to run separately filling calculations and stress calculations in order to save CPU time and memory.

Version 2006.0

Page 559

ProCAST User Manual

GAPS IN STRESS MODELS

During a thermo-mechanical calculation, gaps may form between the different domains (e.g. between the casting and the mold). ProCAST automatically accounts for the modification of the interface heat transfer coefficient when gaps are forming. The model which is used assumes that radiation and heat conduction through an air gap is progressively replacing the transfer due to the contact (see the model in the figure below). If VACUUM is set to 1, the heat conduction part is deactivated. When there is a contact, the heat transfer coefficient is increased as a function of the pressure.

The figure below illustrates the change of the resulting interface heat transfer coefficient, as a function of the air gap width.

Version 2006.0

Page 560

ProCAST User Manual

Version 2006.0

Page 561

ProCAST User Manual

STRESS VISUALIZATION
Stresses are rather complex to view, as it corresponds to a tensor (with 6 components). ProCAST offers different possibilities to visualize these results, in addition to each of the components of the stress tensor.

Effective Stress
There are several ways to evaluate the local stress state with a single value. The formula below shows how the Effective Stress in ViewCAST is calculated (e.g. in one dimension, the effective stress corresponds to the stress itself). To simplify, one could consider that it corresponds to some kind of "average" value of the stress. As shown in the equation below, the Effective Stress is always positive and thus, it does not allow to determine whether the state of stress is in compression or in tension. To determine that, the other viewing possibilities should be used.

The stress components which are in the square root correspond to the principal stresses.

Average Normal Stress

The following figure is showing the stress state (or the stress tensor) of a unit volume oriented according to the X-, Y- and Z-coordinates. The Average Normal Stress corresponds to the average of the three normal components. This value can be either negative (mainly compressive state) of positive (mainly tensile state).

Version 2006.0

Page 562

ProCAST User Manual

Maximum Shear Stress

The shear stresses components are highlighted in red in the figure above. The "maximum shear stress" is defined by the equation below (where the stress components corresponds to the maximum and the minimum normal stresses).

Principal Stresses

Version 2006.0

Page 563

ProCAST User Manual

Considering the stress state of a given unit volume, it is possible to find a unit cube with a different orientation, where all the shear stresses are zero. Thus, only the normal stresses are remaining and they are called principal stresses. The principal stress 1 is the larger value and the principal stress 3 is the smaller value. Positive values show a tensile state, whereas negative values show a compressive state.

Version 2006.0

Page 564

ProCAST User Manual

TUTORIALS
GENERAL INTRODUCTION
ProCAST is a modular system and allows the coupling of various modules.

Flow Chart
Several steps are necessary in ProCAST in order to generate a FEM model, to setup the calculation, to run the analysis and finally, to interpret the results.

The following picture illustrates the flow chart of ProCAST.

Version 2006.0

Page 565

ProCAST User Manual

Meshcast is the fully automatic 2D and 3D tetrahedral mesh generator of ProCAST. It provides facilities for CAD reading, geometry quality checks and repairs. An automatic shell generator is included in MeshCAST and is especially dedicated to investment casting. PreCAST is the pre-processor of ProCAST and allows the user to read a finite element model, to assign material properties to the different model components, to set the interface and boundary conditions and to define the initial conditions. DataCAST compiles the settings and provides the necessary files for the calculation. ProCAST runs the analysis. ViewCAST is the post-processor and allows the user to visualize the results as ProSTAT reports the status of the calculation.

The user is guided through these different operation with the help of the file manager (see below):

Version 2006.0

Page 566

ProCAST User Manual

MeshCAST
Meshcast is the fully automatic 2D and 3D tetrahedral mesh generator of ProCAST. It provides facilities for CAD reading, geometry quality checks and repairs. An automatic shell generator is included in MeshCAST and is especially dedicated to investment casting. Meshcast is interfaced with various CAD file formats. In addition to ProCAST files, MeshCAST can read the following geometries: iges, step, parasolids and stl surface meshes: patran and ideas solid meshes: patran and ideas

Version 2006.0

Page 567

ProCAST User Manual

Usually the user will read a geometry and check its quality (the geometry will be eventually repaired, simplified or improved). The surface mesh will be generated and the solid mesh will be generated based on the surface mesh. The quality of the solid mesh is checked and the solid mesh can be optimized. The solid mesh is then ready to be read in PreCAST, the pre-processor of ProCAST.

PreCAST
PreCAST is the pre-processor of ProCAST and allows the user to read a finite element model, to assign material properties to the different model components, to set the interface and boundary conditions and to define the initial conditions.

Version 2006.0

Page 568

ProCAST User Manual

ProCAST
DataCAST compiles the settings and provides the necessary files for the calculation. ProCAST runs the analysis.

ViewCAST
ViewCAST is the post-processor and allows the user to visualize and analyse the results.

Version 2006.0

Page 569

ProCAST User Manual

Version 2006.0

Page 570

ProCAST User Manual

RESTART
The goal of this tutorial is to explain how to continue or restart a ProCAST calculation. The reason to restart a calculation is multiple:

It is important to remind that the *p.dat file contains run parameters information as the *d.dat file contains calculation data as the material properties, the mesh information, the boundary, interface and initial conditions.

A. NATURAL STOP:

Version 2006.0

Page 571

ProCAST User Manual

B. HARDWARE PROBLEM:

C. SOFTWARE & CONVERGENCE:

Version 2006.0

Page 572

ProCAST User Manual

D. CALCULATION INTERRUPTION (Ctrl C)

AS A SUMMARY:

Version 2006.0

Page 573

ProCAST User Manual

The -U option is available in the File Manager under Datacast:

Version 2006.0

Page 574

ProCAST User Manual

BOUNDARY CONDITIONS
In order to assign the boundary conditions, i.e. all conditions applied to the external side of the model, the user will proceeded as follow:

Go in the Boundary Conditions menu. Using the Add-> button, choose the required Boundary conditions for your model. Using the selection/deselection tools, select the area of the model where the corresponding Boundary conditions will be applied. Do not forget to save the selection of the area of the model. The surface of the selected area is displayed. This is useful in order to know the filling area and to calculate the right filling velocity. Assign the right Boundary condition entry from the database The Boundary condition database can be used, checked and enriched. Let's detail the above items. Go (1) in the Boundary Conditions menu. Press (2) Add -> in order to look for the boundary conditions you need.

Version 2006.0

Page 575

ProCAST User Manual

Select in the Boundary Condition list the required BCs.

The selected BCs will be displayed in the top right table.

Use (3) the selection/deselection tools in order to select the area of the model where the corresponding Boundary conditions will be applied. Do not forget to Store (4) the selection.

Version 2006.0

Page 576

ProCAST User Manual

Please note that once you pressed the Store button (4), the surface of the selected area is displayed (5). This is useful in order to know the filling area and to calculate the right filling velocity. In the velocity Boundary condition type, you have a velocity calculator tool based on the filling type and the inlet area:

Assign (6) then the right Boundary condition entry from the database (7).

Version 2006.0

Page 577

ProCAST User Manual

The database entry is then indicated in the top right table (DB Entry) of the preprocessor interface.

The Boundary condition database (7) can be checked and enriched. The read button allows the user to check the content of a Boundary condition. The Add-> button allows the addition of a Boundary condition among the available list. It is also possible to copy and modify an existing Boundary condition. The use of the Del button will delete the selected entry. It is possible to sort the Boundary conditions per Boundary type.

Version 2006.0

Page 578

ProCAST User Manual

Please note that some boundary conditions can be constant or time and/or temperature dependent. In this case one has to be careful as the constant multiplies the content of the table (if defined) and both tables multiplies each other.

The following picture is describing the selection/deselection icons.

Version 2006.0

Page 579

ProCAST User Manual

The heat flux through an interface is defined in ProCAST as follow.

Version 2006.0

Page 580

ProCAST User Manual

SYMMETRY
In order to optimize the calculation cpu time, one should take advantage of the rotational or mirror symmetries of the geometry to be modeled.

In order to define the symmetry (axis of rotation or plane of symmetry), go in the Geometry menu and press the Symmetry option.

Version 2006.0

Page 581

ProCAST User Manual

The axis of symmetry is defined with two points as the coordinates of three points are necessary to define the plane of symmetry. In order to get the coordinates of the symmetry, activate the 'Get Co-ord' button and click on the axis of symmetry nodes or on the nodes of the symmetry plane. The cursor must be in the corresponding X1, X2 or X3 coordinate input box. Do not forget to press Apply once the axis or th plane of symmetry is defined.

Version 2006.0

Page 582

ProCAST User Manual

TIME STEPS

Version 2006.0

Page 583

ProCAST User Manual

Version 2006.0

Page 584

ProCAST User Manual

MATERIAL PROPERTIES
The goal of this tutorial is to help the user with the definition of material properties which are required for the different material types: mould sand mould in case of lost foam process filter casting for Newtonian flow casting for non-Newtonian flow (semi-solid infiltration) exothermic

As it is not always easy to have the material properties of multi-component alloys, the user should take advantage of the Thermodynamic databases available with ProCAST. Thermodynamic databases will provide the fraction of solid and enthalpy curves and solidus and liquidus temperatures of the casting alloy based on the alloy composition. Most of the material properties are temperature dependent. Data can be entered as a constant or can be tabulated.

In case the data are tabulated, a linear interpolation is applied between the defined point and a horizontal (constant) extrapolation is applied (except for the enthalpy).

Version 2006.0

Page 585

ProCAST User Manual

Casting material
In order to perform a heat (with liquid to solid phase change) and fluid analysis in a casting material, the following material properties are required (two possibilities): A. Thermal: Thermal conductivity [W/mK] Density [Kg/m3] Specific heat [J/KgK] Latent heat [J/Kg] Fraction of solid [-] Solidus and Liquidus temperatures [C] Fluid: Viscosity [centipoise]

B. Thermal:

Version 2006.0

Page 586

ProCAST User Manual

Thermal conductivity [W/mK] Density [Kg/m3] Enthalpy curve [J/Kg] Fraction of solid [-] Solidus and Liquidus temperatures [C] Fluid: Viscosity [centipoise]

Please note that in order to predict porosity, the density of the material must be temperature dependent as shown below. The density change between solidus and liquidus temperature will represent the shrinkage fraction during solidification.

Version 2006.0

Page 587

ProCAST User Manual

In case the fluid flow behavior is shear rate dependent (non-Newtonian), the user can describe the viscosity as a function shear rate and temperature with the help of the Carreau Yasuda relationship:

or with the Power Cut-off law:

Version 2006.0

Page 588

ProCAST User Manual

For the use of the Power Cut-off viscosity relationship, please see the Thixo casting section.

Exothermic material
An exothermic material is defined with the following thermal data: Thermal: Thermal conductivity [W/mK] Density [Kg/m3] Specific heat [J/KgK]

Version 2006.0

Page 589

ProCAST User Manual

In order to take into account the energy released by the exothermic reaction, an exothermic model is available in the material interface. Select the Exothermic menu and enter: The burnt fraction as a function of time The ignition temperature The exothermic energy

Version 2006.0

Page 590

ProCAST User Manual

Filter material
In order to perform a heat and fluid flow analysis in a filter material, the following material properties are required: Thermal: Thermal conductivity [W/mK] Density [Kg/m3] Specific heat [J/KgK] Fluid -> Filter: Void fraction [-] Specific Area [1/cm] Pressure drop (optional)

Mold material
In order to perform a heat analysis in a mold material, the following material properties are required: Thermal: Thermal conductivity [W/mK] Density [Kg/m3]

Version 2006.0

Page 591

ProCAST User Manual

Specific heat [J/KgK]

In case of a sand mold for the Lost Foam process, the permeability of sand must be defined under the Fluid section. The following material properties are thus required: Thermal: Thermal conductivity [W/mK] Density [Kg/m3] Specific heat [J/KgK] Fluid -> Fluid Permeability: Permeability [cm2]

Thermodynamic databases

Version 2006.0

Page 592

ProCAST User Manual

In case of a binary alloy, it is not too difficult to extract the solidification path from the corresponding phase diagram using a Lever or a Scheil relationship. But most of the time, industrial casting alloys are composed of multi-element.

Let's compare the solidification behavior of two slightly different alloys in order to illustrate the importance of having the right data.

Version 2006.0

Page 593

ProCAST User Manual

Version 2006.0

Page 594

ProCAST User Manual

Thermodynamic databases are calculating the fraction of solid and enthalpy curves and solidus and liquidus temperatures on the basis of the alloy composition.

The minimum of the Gibb's free energy curve (or the linear combination of two curves) at temperature T will provide the stable phase diagram.

Version 2006.0

Page 595

ProCAST User Manual

In order to use the Thermodynamic databases, (1) click on the COMPOSITION button, (2) enter the base element of the alloy and then (3) fill the composition table in weight %.

Version 2006.0

Page 596

ProCAST User Manual

(4) click then on apply and (5) select the Scheil or Lever rule.

The available base systems are Al, Fe, Ni, Ti and Mg. The available alloying elements are described in the "Databases/Thermodynamic Databases" section.

Version 2006.0

Page 597

ProCAST User Manual

CYCLING MODELLING

Version 2006.0

Page 598

ProCAST User Manual

In PreCAST there are two techniques to model cycling: A. the Standard method and B. the Die Combo technique. This appears in the Interfaces menu when one want to Add-> a new interface condition. The Standard or Die Combo technique choice is here proposed.

The first technique consists of the modeling of the different cycling sequences with the help of time dependent interface heat transfer coefficients and heat boundary conditions.

Version 2006.0

Page 599

ProCAST User Manual

This classical technique has the advantage of offering all flexibility and is able to model any situation. However in case of process optimization, it can be quite tedious to change all tabulated interface heat transfer coefficients and heat boundary conditions if the cycling timing has to be changed. The Die Combo technique facilitate the cycling modeling through the use of dedicated interfaces.

Version 2006.0

Page 600

ProCAST User Manual

The different cycle times are directly entered in the Run Parameter Cycles menu. And the interface heat transfer coefficient and heat boundary conditions are directly entered in the Die Combo interface.

Version 2006.0

Page 601

ProCAST User Manual

Once such a cycling calculation (thermal only) was done, it is interesting to extract the distribution of the mold temperature and to use it as initial temperature for a filling or stress analysis. Just copy and rename the *d.dat and *p.dat files. Open PreCAST and modify the settings in order to model a filling (Fluid Flow) or stress analysis. In order to extract the temperature distribution, go in the Extract option of the Initial Conditions menu of PreCAST.

Using the Browse button, look for the Cycling (thermal only) calculation and set the Step number at which the temperature distribution will be extracted.

Version 2006.0

Page 602

ProCAST User Manual

Version 2006.0

Page 603

ProCAST User Manual

VIRTUAL MOLD
The virtual mold should be used in order to save cpu time. The thermal aspect of the mold is taken into account (heat diffusion, local heating, heat saturation) but as the mold is not meshed, the user cannot visualize the temperature distribution in the mold. Please note that a zero velocity Boundary condition should be set to the surface of the casting in case the virtual mold is used. The virtual mold is defined in the Geometry menu.

Enter the dimensions of the virtual mold. Please make sure that the virtual mold will include the complete casting geometry. As the external boundary of the virtual mold is considered as adiabatic (no heat exchange), the size of the virtual mold should be set large enough. Using the Check Geom option you can have the external dimension of the geometry. This helps to define the size of the virtual mold.

Press Apply in order to generate the virtual mold.

Version 2006.0

Page 604

ProCAST User Manual

The calculated thermal depth can be displayed using the Show Depth button and the scale of the thermal depth can be adjusted using the Set Scale button. It is also possible to remove the virtual mold.

Version 2006.0

Page 605

ProCAST User Manual

The mold material is now displayed in the Material menu and material properties can be assigned to the virtual mold.

Version 2006.0

Page 606

ProCAST User Manual

An interface heat transfer coefficient has to be defined between the casting and the mold.

And an initial temperature has to be defined for casting and mold.

Version 2006.0

Page 607

ProCAST User Manual

Version 2006.0

Page 608

ProCAST User Manual

PROCESS TEMPLATES
The goal of these tutorials is to help the user with the typical settings and run parameters. The coupling of Thermal and Fluid flow modules allows the user to model any process, i.e. gravity casting, sand casting, high pressure die casting, low pressure die casting, investment casting ... But each process requires some specific settings. Basically the user will set in Precast: GEOMETRY: a solid mesh MATERIALS: a casting and mold material INTERFACE: an interface heat transfer coefficient between casting and mold BOUNDARY CONDITIONS: a heat and pouring conditions PROCESS: a gravity vector and moving solids INITIAL CONDITIONS: initial temperatures and the casting cavity will be set as empty RUN PARAMETERS: general, heat, cycling, radiation, flow and turbulence Please note that the user is able to use predefined templates for the run parameters, i.e. the parameters for the process of interest can be defined only once at the beginning.

Version 2006.0

Page 609

ProCAST User Manual

HPDC-Cycling
The goal of this tutorial is to show you step by step how to setup a high pressure die casting model in ProCAST. The geometry used for this tutorial is composed of a: casting top die bottom die insert meshed cooling channels unmeshed cooling channels

Version 2006.0

Page 610

ProCAST User Manual

The goal here is to run first a thermal analysis with several cycles in order to predict the number of necessary cycles to reach steady state. In a second step, the temperature distribution of the mold components will be used as initial temperature for a fluid flow or stress analysis. In this tutorials both options for the cooling channels are considered: meshed and not meshed channels. The goal is to illustrate the set-up of both options. The interfaces will be treated with the Diecombo technique.

Load the hpdc.mesh model in PreCAST

Version 2006.0

Page 611

ProCAST User Manual

Enter first the case name under the 'Case' field of the file manager and press the PreCAST menu. PreCAST will load first a ProCAST file (d.dat) if it is present in the current directory. If PreCAST does not find any ProCAST file, it will look for a MeshCAST file (.mesh) with the corresponding prefix. And if there is no MeshCAST file, you will have to use the Open file menu of PreCAST in order to look for the right file.

Version 2006.0

Page 612

ProCAST User Manual

Version 2006.0

Page 613

ProCAST User Manual

Check the geometry PreCAST indicates automatically the number of materials, the total number of nodes and of elements. The units and the global size of the model are shown. These information and the volume of the different components are available in the File menu -> Check geometry options.

Assign material properties to the casting, mold and inserts. In this case, Aluminium is assigned to the casting material, Steel H13 to the top and bottom dies and to the insert and water is assigned to the meshed cooling channels.

Version 2006.0

Page 614

ProCAST User Manual

Please note that the temperature of the casting material will be re-initialized before each cycle as the temperature of the mold components at cycle N-1 will be taken

Version 2006.0

Page 615

ProCAST User Manual

into account for cycle N. It is therefore important to set the right Material type for each component.

Create and Assign the Interfaces between the components In this case all interfaces have to be converted from EQUIV to COINC. In order to validate this choice, press the Apply button.

Version 2006.0

Page 616

ProCAST User Manual

Bottom Die - Cooling channels (Standard interface): Between the bottom die and the meshed cooling channels, a constant heat transfer coefficient of 2000 W/m2K will be set.

Version 2006.0

Page 617

ProCAST User Manual

Die - Die interface (Diecombo): Between the top die and a bottom die a diecombo interface will be set.

Version 2006.0

Page 618

ProCAST User Manual

Version 2006.0

Page 619

ProCAST User Manual

Bottom Die - Insert interface (Diecombo): Between the bottom die and the insert, a diecombo interface is assigned with a heat transfer coefficient of 1000W/m2K.

Version 2006.0

Page 620

ProCAST User Manual

Bottom Die - Casting interface (Diecombo):

Top Die - Casting interface (Diecombo):

Version 2006.0

Page 621

ProCAST User Manual

Please note that the casting will be 'attached until ejection' to the top die.

Top Die - Insert interface (Diecombo):

Version 2006.0

Page 622

ProCAST User Manual

Casting - Insert interface (Diecombo):

Version 2006.0

Page 623

ProCAST User Manual

Assign the Boundary conditions

Air cooling condition will be assigned to the external surface of the mould

Version 2006.0

Page 624

ProCAST User Manual

Heating channels (200C) will be assigned to the surfaces of the non-meshed channels using a heat transfer coefficient off 2000W/m2K and an external

Version 2006.0

Page 625

ProCAST User Manual

temperature of 200C. The inner surface of the non-meshed channels are selected using the propagate selection tool.

An imposed temperature of 20C is applied all around the meshed cooling channels.

Version 2006.0

Page 626

ProCAST User Manual

Version 2006.0

Page 627

ProCAST User Manual

Set the gravity In order to set the gravity vector, go in the Process menu.

Set the constant Initial conditions

Version 2006.0

Page 628

ProCAST User Manual

Set the run parameters In the Preferences menu, select the hpdc Cycling option. The right and dedicated run parameters will be automatically set.

Version 2006.0

Page 629

ProCAST User Manual

Version 2006.0

Page 630

ProCAST User Manual

Version 2006.0

Page 631

ProCAST User Manual

Version 2006.0

Page 632

ProCAST User Manual

Save and Exit PreCAST

Version 2006.0

Page 633

ProCAST User Manual

10

Run DataCAST and ProCAST

11

Visualize the results in ViewCAST

Version 2006.0

Page 634

ProCAST User Manual

Version 2006.0

Page 635

ProCAST User Manual

HPDC-Flow
The goal of this tutorial is to show you step by step how to setup a high pressure die casting model in ProCAST. The geometry used for this tutorial is composed of a: casting top die bottom die insert meshed cooling channels unmeshed cooling channels

The goal here was first to run a thermal analysis with several cycles in order to predict the number of necessary cycles to reach steady state. In a second step, the temperature distribution of the mold components will be used as initial temperature for a fluid flow or stress analysis. In this tutorials we are going to extract the die temperature distribution of HPDCCycling in order to use it as initial temperature.

Load the hpdcd.dat model in PreCAST

Version 2006.0

Page 636

ProCAST User Manual

Create a new calculation directory and copy from your hpdc-cycling directory the hpdcd.dat and hpdcp.dat files.

Load in PreCAST the hpdcd.dat file:

Version 2006.0

Page 637

ProCAST User Manual

Modification of the data MATERIALS: the casting material is set as empty, but the same properties will be used INTERFACES: no modifications compared to the HPDC-cycling case, the same Diecombo interfaces will be used BOUNDARY CONDITIONS: an imposed temperature and velocity is applied on the casting inlet.

Version 2006.0

Page 638

ProCAST User Manual

PROCESS: no modifications, the same gravity vector is used INITIAL CONDITIONS: the extract procedure is applied on the top die, bottom die and insert domains

Version 2006.0

Page 639

ProCAST User Manual

The temperature distribution at the beginning of cycle number 10 will be considered. In ViewCAST or in the hpdcp.out file it can be determined that this time corresponds to step 765. Using the Browse tool, you can look for the HPDCCycling directory where the temperature results of the cycling calculation are.

Version 2006.0

Page 640

ProCAST User Manual

The temperature distribution can be displayed in order to check that the extract procedure worked.

Version 2006.0

Page 641

ProCAST User Manual

RUN PARAMETERS: General and Thermal are not modified, The number of Cycles is set to 1 and Flow is set to 3. The frequency for saving the velocity results (FLOW) is set equal to TFREQ (frequency for saving the temperature results).

Version 2006.0

Page 642

ProCAST User Manual

Save and Exit

Version 2006.0

Page 643

ProCAST User Manual

Run the calculation

Visualize the results Please note that at step zero, the temperature of the die is not uniform as we used the calculated temperature distribution obtained before cycle ten in the HPDCcycling calculation.

Version 2006.0

Page 644

ProCAST User Manual

Version 2006.0

Page 645

ProCAST User Manual

Version 2006.0

Page 646

ProCAST User Manual

LPDC
The goal of this tutorial is to show step by step how to setup a low pressure die casting model in ProCAST.

Version 2006.0

Page 647

ProCAST User Manual

The geometry used for this tutorial is composed of a: casting left and right side dies top and bottom cores

Load the lpdcd.dat model in PreCAST

Version 2006.0

Page 648

ProCAST User Manual

Enter first the case name under the 'Case' field of the file manager and press the PreCAST menu. PreCAST will load first a ProCAST file (d.dat) if it is present in the current directory. If PreCAST does not find any ProCAST file, it will look for a MeshCAST file (.mesh) with the corresponding prefix. And if there is no MeshCAST file, you will have to use the Open file menu of PreCAST in order to look for the right file.

Version 2006.0

Page 649

ProCAST User Manual

Check the geometry

Version 2006.0

Page 650

ProCAST User Manual

PreCAST indicates automatically the number of materials, the total number of nodes and of elements. The units and the global size of the model are shown. These information and the volume of the different components are available in the File menu -> Check geometry options.

Define the symmetry plane

Version 2006.0

Page 651

ProCAST User Manual

Version 2006.0

Page 652

ProCAST User Manual

Version 2006.0

Page 653

ProCAST User Manual

Assign material properties to the casting, mold and cores. In this case, CuAluminium Bronze is assigned to the casting material, Steel_H13 to both parts of the dies and Sand_Silica to the core domains.

Create and Assign the Interfaces between the components In this case all interfaces are COINC. In case the model would have EQUIV interfaces, they should be converted from EQUIV to COINC. In order to validate this choice, do not forget to press the Apply button.

Version 2006.0

Page 654

ProCAST User Manual

Once the interfaces are created, assign a heat transfer value to each interface. Typical values are 500 W/m2K between casting and cores and 1000 W/m2K between casting and steel dies.

Version 2006.0

Page 655

ProCAST User Manual

Assign the Boundary conditions

Version 2006.0

Page 656

ProCAST User Manual

A natural air cooling heat Boundary condition will be applied all around the dies and a pressure Boundary condition together with a temperature Boundary condition will be imposed at inlet of the gating system. Please note that the simulation model might have a different height compared to the current casting if the full filling column was not taken into account in the simulation model. In such a case the pressure ramp should be shifted accordingly. Please also note that the pressure which is applied to the inlet for the filling sequence will be the difference between the reference pressure which will be defined in the run parameters and this pressure BC. In this case where a total pressure of 0.2 bar is applied, a zero reference pressure will be defined in the FLOW run parameters.

Set the gravity In order to set the gravity vector, go in the Process menu.

Version 2006.0

Page 657

ProCAST User Manual

Set the constant Initial conditions

Version 2006.0

Page 658

ProCAST User Manual

Set the run parameters

In the Preferences menu, select the LPDC filling option. The right and dedicated run parameters will be automatically set.

Version 2006.0

Page 659

ProCAST User Manual

Version 2006.0

Page 660

ProCAST User Manual

Version 2006.0

Page 661

ProCAST User Manual

10

Save and Exit PreCAST

Version 2006.0

Page 662

ProCAST User Manual

11

Run DataCAST and ProCAST

12

Visualize the results in ViewCAST

Version 2006.0

Page 663

ProCAST User Manual

Version 2006.0

Page 664

ProCAST User Manual

Version 2006.0

Page 665

ProCAST User Manual

Gravity-Sand
This tutorial will guide you step by step in the ProCAST set-up of a sand casting simulation. The geometry used for this tutorial is composed of a: casting sand box cores

Version 2006.0

Page 666

ProCAST User Manual

Load the sand.mesh model in PreCAST

Please enter first the case name under the 'Case' field of the file manager and press the PreCAST menu.

Version 2006.0

Page 667

ProCAST User Manual

PreCAST will load first a ProCAST file (d.dat) if it is present in the current directory. If PreCAST does not find any ProCAST file, it will look for a MeshCAST file (.mesh) with the corresponding prefix. And if there is no MeshCAST file, you will have to use the Open file menu of PreCAST in order to look for the right file.

Version 2006.0

Page 668

ProCAST User Manual

Check the geometry PreCAST indicates automatically the number of materials, the total number of nodes and of elements. The units and the global size of the model are shown. These information and the volume of the different components are available in the File menu -> Check geometry options.

Version 2006.0

Page 669

ProCAST User Manual

Assign material properties to the casting, mold and cores. In this case, Ductile Iron is assigned to the casting material, Sand silica to the mold and Steel_H13 to the core domains.

Version 2006.0

Page 670

ProCAST User Manual

Create and Assign the Interfaces between the components In this case all interfaces have to be converted from EQUIV to COINC. In order to validate this choice, press the Apply button.

Version 2006.0

Page 671

ProCAST User Manual

Version 2006.0

Page 672

ProCAST User Manual

Once the interfaces are created, assign a heat transfer value to each interface. Typical values are 500 W/m2K between casting and sand and 1000 W/m2K between casting and steel cores.

Assign the Boundary conditions

Version 2006.0

Page 673

ProCAST User Manual

A natural air cooling heat Boundary condition will be applied all around the sand box, a velocity Boundary condition will be applied and a temperature Boundary condition will be imposed to some nodes of the top surface of the down sprue.

Version 2006.0

Page 674

ProCAST User Manual

Set the gravity

Version 2006.0

Page 675

ProCAST User Manual

In order to set the gravity vector, go in the Process menu.

Set the constant Initial conditions

Version 2006.0

Page 676

ProCAST User Manual

Set the run parameters In the Preferences menu, select the Gravity filling option. The right and dedicated run parameters will be automatically set.

Version 2006.0

Page 677

ProCAST User Manual

In the General run parameters panel, DTMAX is set to 5 seconds in this case.

Version 2006.0

Page 678

ProCAST User Manual

In the Thermal run parameters panel, TFREQ is set to 5 here.

Version 2006.0

Page 679

ProCAST User Manual

In the Flow run parameters panel, VFREQ is set to 5 (equal to TFREQ).

Version 2006.0

Page 680

ProCAST User Manual

Do not forget to press Apply in order to validate the Run Parameters selection.

Save and Exit PreCAST

Version 2006.0

Page 681

ProCAST User Manual

10

Run DataCAST and ProCAST

11

Visualize the results in ViewCAST

Version 2006.0

Page 682

ProCAST User Manual

Position the model using the picture manipulation tools (zoom, drag, rotate, ...).

Version 2006.0

Page 683

ProCAST User Manual

Select the casting material.

Visualization of the flow behavior using the 'tape player' buttons.

Version 2006.0

Page 684

ProCAST User Manual

Analysis of the velocity vectors. Possible mold erosion can be studied.

Version 2006.0

Page 685

ProCAST User Manual

Temperature map in a cross section through the total model:

Version 2006.0

Page 686

ProCAST User Manual

Isolated pockets of liquid are displayed using the Fraction of Solid cut-off option.

Investment
The goal of this tutorial is to show you step by step how to setup an investment casting model in ProCAST. The geometry used for this tutorial is composed of a: casting a shell mold an enclosure

Version 2006.0

Page 687

ProCAST User Manual

Remark: It is possible to model the preheating of the shell mold in the furnace using the same model as the one for fluid flow modeling. The temperature distribution of the shell mold is then used as initial condition for the fluid flow calculation using the extract procedure. To do so:

Load the investment.mesh model in PreCAST

Version 2006.0

Page 688

ProCAST User Manual

Please enter first the case name under the 'Case' field of the file manager and press the PreCAST menu. PreCAST will load first a ProCAST file (d.dat) if it is present in the current directory. If PreCAST does not find any ProCAST file, it will look for a MeshCAST file (.mesh) with the corresponding prefix. And if there is no MeshCAST file, you will have to use the Open file menu of PreCAST in order to look for the right file. If this Case name is not entered, you will have to go in the File menu of PreCAST and open the right file.

Version 2006.0

Page 689

ProCAST User Manual

Check the geometry PreCAST indicates automatically the number of materials, the total number of nodes and of elements. The units and the global size of the model are shown. These information and the volume of the different components are available in the File menu -> Check geometry options.

Version 2006.0

Page 690

ProCAST User Manual

Please note that there is an enclosure for the radiation calculation.

Version 2006.0

Page 691

ProCAST User Manual

Definition of the symmetry planes

Two mirrors of symmetry will be considered here. The Get Co-ord button is useful to pick the node coordinates on the planes of symmetry. You can check that the symmetry was well applied if the symmetry is set in the Boundary conditions.

Version 2006.0

Page 692

ProCAST User Manual

Assign material properties to the casting and shell mold.

Version 2006.0

Page 693

ProCAST User Manual

Inconel 718 and a shell material are applied to the casting and shell mold.

Create and Assign the Interfaces between the components In this case all interfaces have to be converted from EQUIV to COINC. In order to validate this choice, press the Apply button.

Version 2006.0

Page 694

ProCAST User Manual

Assign the interface heat transfer coefficient

Version 2006.0

Page 695

ProCAST User Manual

Assign the Boundary conditions

In addition to the symmetry Boundary condition, a heat transfer coefficient with View Factors ON will be applied on the surface of the shell. Velocity and temperature will be applied to the top surface of the downsprue.

Version 2006.0

Page 696

ProCAST User Manual

Remarks:

Version 2006.0

Page 697

ProCAST User Manual

A temperature and emissivity has to be applied to the enclosure (Boundary conditions ->Assign Enclosure).

Version 2006.0

Page 698

ProCAST User Manual

Check that the normals of the enclosure are all pointing inside the enclosure. In case it is not the case, use the Align or Reverse options.

Version 2006.0

Page 699

ProCAST User Manual

Set the gravity In order to set the gravity vector, go in the Process menu.

Set the constant Initial conditions

Version 2006.0

Page 700

ProCAST User Manual

Set the run parameters

In the Preferences menu, select the Gravity filling option. The right and dedicated run parameters will be automatically set.

Version 2006.0

Page 701

ProCAST User Manual

Version 2006.0

Page 702

ProCAST User Manual

Version 2006.0

Page 703

ProCAST User Manual

Version 2006.0

Page 704

ProCAST User Manual

10

Save and Exit PreCAST

Version 2006.0

Page 705

ProCAST User Manual

11

Run DataCAST and ProCAST

Useful tip:

Version 2006.0

Page 706

ProCAST User Manual

12

Visualize the results in ViewCAST

Filling results (temperature map)

Version 2006.0

Page 707

ProCAST User Manual

Temperature distribution during solidification

Symmetry re-construction in the Post-processor

Version 2006.0

Page 708

ProCAST User Manual

Cross section (temperature map)

Version 2006.0

Page 709

ProCAST User Manual

Prediction of pockets of liquid using the fraction of solid field and the cut-off option

Filter
This tutorial will guide you step by step in the ProCAST set-up of a sand casting simulation with a filter material. The geometry used for this tutorial is composed of a: casting filter chill A virtual mold will be generated for the sand mold.

Version 2006.0

Page 710

ProCAST User Manual

This geometry will also allow the illustration of symmetry settings.

Load the filter.mesh model in PreCAST

Please enter first the case name under the 'Case' field of the file manager and press the PreCAST menu. PreCAST will load first a ProCAST file (d.dat) if it is present in the current directory. If PreCAST does not find any ProCAST file, it will look for a

Version 2006.0

Page 711

ProCAST User Manual

MeshCAST file (.mesh) with the corresponding prefix. And if there is no MeshCAST file, you will have to use the Open file menu of PreCAST in order to look for the right file.

Check the geometry PreCAST indicates automatically the number of materials, the total number of nodes and of elements. The units and the global size of the model are shown. These information and the volume of the different components are available in the File menu -> Check geometry options.

Version 2006.0

Page 712

ProCAST User Manual

Symmetry definition

Version 2006.0

Page 713

ProCAST User Manual

In order to define the coordinates of the two points which are defining the axis of symmetry, press the 'Get Co-ord' button and then click with the mouse on a mesh points located on the axis of symmetry. You can check that the symmetry boundary conditions are well set. Go in the Boundary conditions menu and check that you have the symmetry well defined.

Virtual Mold generation Check first the global size of the model.

Version 2006.0

Page 714

ProCAST User Manual

Generate a virtual mold larger than the model. The virtual mold has to be large enough as the external boundary of the virtual mold is considered as adiabatic.

Version 2006.0

Page 715

ProCAST User Manual

You can then visualize the thermal depth.

Assign material properties Please note that the virtual mold will be taken into account and you have thus to assign material properties to the mold.

Version 2006.0

Page 716

ProCAST User Manual

A filter material is assigned to the filter.

Version 2006.0

Page 717

ProCAST User Manual

In addition to the thermal conductivity, the density and the specific heat, a filter is defined by its Void Fraction and Surface Area (it is also possible to define the Pressure Drop properties - optional).

Create and Assign the Interfaces between the components In this case only the interface between the chill and casting has to be converted from EQUIV to COINC. It is important to note that no interface between the casting and the filter is created. One should thus leave the EQUIV interface type between casting and filter. In order to validate this choice, press the Apply button.

Version 2006.0

Page 718

ProCAST User Manual

Assign then heat transfer coefficients between the chill and the casting and between all components and the virtual mold. Please note that no heat transfer

Version 2006.0

Page 719

ProCAST User Manual

coefficient is defined between the casting and the filter. This will be defined in the Assign Volume Boundary Conditions menu.

Assign the Boundary Conditions The symmetry Boundary condition was automatically set as symmetry was defined previously.

Version 2006.0

Page 720

ProCAST User Manual

A zero velocity Boundary condition has to be applied all around the casting as no coincident meshed mold is considered. A velocity Boundary condition and an imposed temperature are applied to the top of the gating system.

Version 2006.0

Page 721

ProCAST User Manual

The interface heat transfer coefficient is defined in the Assign Volume -> Filter Heat condition.

Version 2006.0

Page 722

ProCAST User Manual

Set the gravity

Define the Initial Conditions

Version 2006.0

Page 723

ProCAST User Manual

10

Set the Run Parameters

Version 2006.0

Page 724

ProCAST User Manual

In the Preferences menu, select the Gravity filling option. The right and dedicated run parameters will be automatically set.

In the General run parameters panel, DTMAX is set to 5 seconds and TSTOP is set to1000C in this case. This means that the calculation will be stopped as soon as the casting domain temperature is below 1000C.

Version 2006.0

Page 725

ProCAST User Manual

Version 2006.0

Page 726

ProCAST User Manual

Version 2006.0

Page 727

ProCAST User Manual

Version 2006.0

Page 728

ProCAST User Manual

11

Save and Exit PreCAST

Version 2006.0

Page 729

ProCAST User Manual

12

Run DataCAST and ProCAST

13

Visualize the results in ViewCAST

Version 2006.0

Page 730

ProCAST User Manual

Version 2006.0

Page 731

ProCAST User Manual

Version 2006.0

Page 732

ProCAST User Manual

Version 2006.0

Page 733

ProCAST User Manual

Stress
This tutorial will guide you step by step in the ProCAST set-up of a stress model. The geometry used for this tutorial is composed of a: casting die

Version 2006.0

Page 734

ProCAST User Manual

Load the stress.mesh model in PreCAST

Please enter first the case name under the 'Case' field of the file manager and press the PreCAST menu. PreCAST will load first a ProCAST file (d.dat) if it is present in the current directory. If PreCAST does not find any ProCAST file, it will look for a MeshCAST file (.mesh) with the corresponding prefix. And if there is no MeshCAST file, you will have to use the Open file menu of PreCAST in order to look for the right file.

Version 2006.0

Page 735

ProCAST User Manual

Version 2006.0

Page 736

ProCAST User Manual

PreCAST indicates automatically the number of materials, the total number of nodes and of elements. The units and the global size of the model are shown. These information and the volume of the different components are available in the File menu -> Check geometry options.

Assign material properties to the casting and mold materials.

Version 2006.0

Page 737

ProCAST User Manual

And assign the corresponding stress properties.

Version 2006.0

Page 738

ProCAST User Manual

Assign the Interfaces between the components. Please note that the mesh was already optimized in MeshCAST and thus the interfaces are already created (they are already set as COINC).

An interface heat transfer coefficient of 1000W/m2K is selected in the database and assigned between the casting and the dies and between top and bottom die.

Version 2006.0

Page 739

ProCAST User Manual

Assign the Boundary conditions

Version 2006.0

Page 740

ProCAST User Manual

A natural air cooling heat Boundary condition is assigned all around the die and a zero displacement Boundary condition is applied to the top and bottom surface of the die.

Version 2006.0

Page 741

ProCAST User Manual

Set the gravity In order to set the gravity vector, go in the Process menu.

Version 2006.0

Page 742

ProCAST User Manual

Set the constant Initial conditions

Version 2006.0

Page 743

ProCAST User Manual

Set the run parameters

Version 2006.0

Page 744

ProCAST User Manual

Version 2006.0

Page 745

ProCAST User Manual

Please be careful to set PIPEFS to zero. Do not forget to press Apply in order to validate the Run Parameters selection.

Save and Exit PreCAST

Version 2006.0

Page 746

ProCAST User Manual

10

Run DataCAST and ProCAST

11

Visualize the results in ViewCAST

Version 2006.0

Page 747

ProCAST User Manual

Temperature results:

Version 2006.0

Page 748

ProCAST User Manual

Effective stress results:

Version 2006.0

Page 749

ProCAST User Manual

Y-Displacements:

Version 2006.0

Page 750

ProCAST User Manual

Hot tearing prediction:

Non-Coincident model
This tutorial will guide you step by step in the ProCAST set-up of a noncoincident model. We can use non-coincident meshes in ProCAST. This facilitates the assembly of the different model components but you need to be careful as in such a case the interfaces between the components are defined according to a tolerance. If all components are well positioned and the mesh densities are not too different, the thermal or flow predicted results will be good. But one should thus run a thermal calculation first and check if the temperature distribution is correct. If the interfaces are not good, i.e. not within the tolerance (1/10 of the smallest element length), then one won't have heat exchange at some interfaces. It is then possible to change this tolerance value (see the Interfaces assignment section for more details). The above test procedure has then to be run again. If you run a fluid flow analysis with a non-coincident model, do not forget to assign a zero velocity boundary condition all around the casting. The geometry used for this tutorial is composed of a: casting

Version 2006.0

Page 751

ProCAST User Manual

sand mold

Load the different mesh components, c1.mesh, m1.mesh and m2.mesh. No prefix is entered in this case in the file manager (Case field) as you are going to load the models using the Multiple Mesh option in the File menu

Version 2006.0

Page 752

ProCAST User Manual

Version 2006.0

Page 753

ProCAST User Manual

Check the geometry PreCAST indicates automatically the number of materials, the total number of nodes and of elements. The units and the global size of the model are shown. These information and the volume of the different components are available in the File menu -> Check geometry options.

Version 2006.0

Page 754

ProCAST User Manual

Assign material properties to the casting and to both mold sides

Version 2006.0

Page 755

ProCAST User Manual

Create and Assign the Interfaces between the components One coincident interface was recognized in this case (even if both models are not coincident) between both mold components because one element is most probably exactly coincident on each mold side.

Version 2006.0

Page 756

ProCAST User Manual

In case of a non-coincident model one has to turn the coincident interface type in non-coincident and add the other non-coincident interfaces. Please note that the casting is the Master and the Mold is the Slave.

Version 2006.0

Page 757

ProCAST User Manual

Do not forget to Apply the non-coincident interfaces and to assign an interface heat transfer coefficient between the different interfaces.

Assign the Boundary conditions

Version 2006.0

Page 758

ProCAST User Manual

Version 2006.0

Page 759

ProCAST User Manual

Version 2006.0

Page 760

ProCAST User Manual

Version 2006.0

Page 761

ProCAST User Manual

As a summary the boundary conditions are: Heat: air cooling all around the mould

Version 2006.0

Page 762

ProCAST User Manual

Velocity: zero velocity all around the casting (this necessary in case of a nocoincident mold, in case of virtual mold or in case there is no mold) Velocity: -y velocity at the top of the downsprue. Temperature: Pouring temperature of the metal assigned at the top of the downsprue.

Set the gravity In order to set the gravity vector, go in the Process menu.

Set the constant Initial conditions

Version 2006.0

Page 763

ProCAST User Manual

Set the run parameters In the Preferences menu, select the Gravity filling option. The right and dedicated run parameters will be automatically set.

Version 2006.0

Page 764

ProCAST User Manual

Version 2006.0

Page 765

ProCAST User Manual

Version 2006.0

Page 766

ProCAST User Manual

Version 2006.0

Page 767

ProCAST User Manual

Save and Exit PreCAST

Version 2006.0

Page 768

ProCAST User Manual

10

Run DataCAST and ProCAST

Version 2006.0

Page 769

ProCAST User Manual

11

Visualize the results in ViewCAST

Version 2006.0

Page 770