COSMOS C11 MODEL Y ANALISIS

CHAPTER 11 Modeling and Analysis
Examples
Introduction
This chapter presents some examples on the analyses you can perform in the Basic
System. These examples cover detailed discussions of linear static analysis,
buckling, and natural frequencies and mode shape computations. Almost all
examples discussed in this chapter document commands required for model
creation, mesh generation, load and boundary condition enforcement, analysis and
postprocessing. Where possible, comparisons have been made with analytical
solutions to indicate the accuracy of COSMOS/M.
To use these examples, enter GEOSTAR, and simply execute the commands as
instructed. The computer prompts for commands are shown in 10 point Courier and
the corresponding inputs are shown in bold, as illustrated below:
Geo Panel: Propsets > Material Property
Material property set > 1
Material Property Name > EX: Elasticity modulus 1st dir
Property value > 15.5E6
Material Property Name >
(Click on cancel to terminate)
Please refer to the COSMOS/M User Guide (V. 1) for more information on using
GEOSTAR.
COSMOS/M Basic FEA System • 12/97 11-1

Chapter 11 Modeling and Analysis Examples
Input files for the examples discussed here are available in the PROBLEMS.EXE
file installed in the COSMOS/M directory (Macintosh and Unix Workstation
versions may have a separate folder or directory with a title Example Problems or
PROBS, respectively). On PCs, the file PROBLEMS.EXE can be unpacked by
typing the command PROBLEMS at the DOS prompt. For other machines, follow
the instructions provided to unpack or expand the archived file.
✍ The second part of this manual presents more than 100 verification problems on
all analysis features of the Basic System. You are recommended to use these
examples to learn more about the Basic System.
Table 5-1. List of Modeling and Analysis Examples
List of Examples
LGAP1 - Hertzian Contact Using Node-to-Node Linear Gap
Contact/Gap
Elements. (See page 11-3.)
BOND1 - Static Analysis of a U-Frame with Non-compatible
Bond
Elements. (See page 11-10.)
Linear Elastic LEFM1 -Stress Intensity Factor of a Plate with Two Edge
Fracture Cracks. (See page 11-23.)
Mechanics LFEM2 - Stress Intensity Factor of Three Point Bend
Analysis Specimen. (See page 11-29.)
Buckling BUCKL1 - Buckling of a Rectangular Plate under Uniform
Analysis Pressure. (See page 11-34.)
MODAL1 - In-Plane Effects on the Natural Frequencies of
Modal a Column. (See page 11-40.)
Analysis MODAL2 - Modal Analysis of a Square Frame with Rigid
Body Modes. (See page 11-46.)
LSUBM1 - Using Submodeling for Shell Problems.
(See page 11-51.)
LSUBM2 - Using Submodeling with Tetrahedral Elements.
Submodeling
(See page 11-54.)
LSUBM3 - Using Submodeling for Shell Problems.
(See page 11-57.)
ASME Stress ASME.GEO - ASME Code Stress Evaluation.
Requirement (See page 11-59.)
Use of Multiple
TEMP1.GEO - Multiple Thermal Load Cases
Thermal Load
(See page 11-62.)
Cases
11-2 COSMOS/M Basic FEA System • 12/97

Part 1 Linear Statics, Modal, and Buckling Analysis
Examples on Linear Static Stress Analysis

The types of linear static problems you can analyze in the Basic System include
models with various combinations of mechanical, thermal, and gravitational loads,
cracks, gaps and contact surfaces, substructuring, cyclic symmetry, and many others.
The examples discussed in this section only address some features of the linear static
analysis module STAR owing to space limitations. The verification problems
presented in the second part of this manual cover almost all aspects of linear static
modeling and analysis.
Example on Contact
Contact problems are best solved as nonlinear problems. The example in this chapter
shows the procedure to solve a simpe contact problem using linear analysis. It should
be noted that the principle of superposition, valid for regular analysis, may not hold
when using linear analysis to solve contact problems.
Example LGAP1 - Hertzian Contact Using Node-to-Node Linear (See

page
Gap Elements 11-2.)
The complete input for this example is available in the file LGAP1.GFM. The study
of stresses caused by the pressure between elastic bodies are important in the design
of train tracks, ball and roller bearings, expansion joints for steel girders, and many
other practical problems. Due to the Figure 11-1. Problem Geometry for Analysis
stresses developed at contact areas
which may be as high as the yield P
limit, the study of contact problems
is often analyzed as a nonlinear
finite element problem. However, a
good approximation for the contact
stresses can be obtained using the E = 3E5
linear node-to-node contact/gap ν = 0.3
R = 10
elements in the Basic System. The P = 100,000 R
figure below shows the half model
of a cylinder for contact stress
analysis using node-to-node gap
elements. The material properties,
dimensions and applied load are Half model of a cylinder on a flate
plate (Hertz contact problem)
also shown in the figure.

Since the complete input for this example is available to you, there is no need to
issue any command for generation of the finite element mesh. The only two
commands you need are the FILE (File > Load...) command to load the input file
and the R_STATIC (Analysis > STATIC > Run Static Analysis) command to
execute analysis. In what follows, you will be presented with information on how
the finite element model was developed for analysis. Where ever a command is
required to be executed, you are clearly instructed to do so.
To start with, copy the file LGAP1.GFM to your working directory, enter
GEOSTAR, and execute the FILE (File > Load...) command as illustrated below:
Geo Panel: File > Load...
Input file name with extension > LGAP1.GFM
Accept Defaults
The file LGAP1.GFM includes information on the problem geometry such as

points, curves, contours, and regions as well as the finite element data such as
nodes, elements, properties, loads, and analysis specifications. You will see the
finite element model constructed and displayed as the input is read in. To properly
see the model, use the Perfect Scaling icon, and the View icon to set the X-Y view.
The command ACTNUM (Control > ACTIVATE > Entity Label) can be used to
activate numbering of geometric or finite element entities to see the respective
labels generated. The geometry of the cylinder was generated by using the
command CRARC3PT (Geometry > CURVES > CIRCLES > by 3 Points) for the
curved part and CRLINE (Geometry > CURVES > Line with 2 Pts) for the straight
line. These curves were further broken using the command CRBRK (Geometry >
CURVES > MANIPULATION MENU > Break (equally)) to smaller segments to
construct regions for mesh transition from coarse to fine. The figure below shows
the curve, contour, and region labels of the cylinder. The list commands such as
PTLIST, CTLIST, RGLIST, (Edit > LIST > Points, Contours, Regions, ...) etc., can
be used to list the created geometric entities.

Figure 11-2. Curve, Contour and Region Plots of the
There are two types of elements used in this model. The PLANE2D element
models the continuous domain of the cylinder in a state of plane strain whereas the
GAP element was used to connect the cylinder to the contact surface. These
elements were defined using the command EGROUP (Propsets > Element Group).
The PLANE2D element was defined under group 1 and the GAP element was
defined under group 2. Since the PLANE2D element models plane strain behavior,
there are no section constants required for this element under real constant set 1.
The finite element mesh of the cylinder was generated using the command MA_RG
(Meshing > AUTO MESH > Regions) for all four regions. The command
NMERGE (Meshing > NODES > Merge) was used to merge the nodes at the
common boundary curves. The GAP element is a special element which needs to be
created by properly activating the required real constant sets as explained in the
following paragraph.
The figure below shows the complete finite element model of the cylinder with the
applied loading and constraints. The GAP element was defined by using the EL
(Meshing > ELEMENTS > Define Element) command which connects the
specified nodes defined by the ND (Meshing > NODES > Define) command. In the
Basic System, the gap closure tolerance (gdist) is defined as a real constant
parameter. Therefore, each gap element requires a real constant set which has to be
activated before the element is defined. For this problem, five gap elements
numbered 72 through 76 were used.

Figure 11-3. Finite Element Model of the
The figure below shows a clear view of the connection between the cylinder and the
bottom contact surface. To start with, the element group 2 was activated for gap
elements, and nodes 76 through 80 were defined at a y-coordinate of zero and x-
coordinates equal to the exact coordinates of the respective nodes of the cylinder to
which they are connected. It is important to specify the exact x-coordinates for the
nodes of the gap element as a slight eccentricity will introduce lateral components
of forces on the gap. In the next step, real constant 2 was activated and the first gap
element (label 72) was defined using the EL (Meshing > ELEMENTS > Define
Element) command. This element connects nodes 7 and 76. Similarly, the re-
maining gap elements were defined after activating the respective real constant sets.
Figure 11-4. Enlarged View of the Contact Surface

For a gap element, you need to specify the allowable relative motion between the
two nodes beyond which the gap will resist compression or tension. For
compression gaps, the most common situation is to specify the relative motion
between the two nodes as equal to the initial distance between them such that they
come into contact before resisting compression. Due to the amount of labor
involved in inputting these distances for a large number of gaps, an option is
provided in the element group definition (EGROUP (Propsets > Element Group)
command) to instruct the program to calculate and use the initial distance between
the nodes of each gap element. Refer to Chapter 4, Element Library, in the
COSMOS/M User Guide. In this case there is no need to define the real constants
associated with the gaps and any input will be ignored.
The bottom nodes of all gap elements were fixed using the DND (LoadsBC >
STRUCTURAL > DISPLACEMENT > Define Nodes) command. Node 6 which
represents the node directly in contact with the bottom surface was also fixed using
the same command. Along the vertical axis of symmetry, the symmetric boundary
conditions (UX = 0) were imposed using the DCR (LoadsBC > STRUCTURAL >
DISPLACEMENT > Define Curves) command. The force at the top was applied
using the FND (LoadsBC > STRUCTURAL > FORCE > Define Nodes) command.
There is only one material set for this problem, defined for the cylinder. The
properties E and ν were defined using the MPROP (Propsets > Material Property)
command. To check the accuracy of the solution, the computation of reaction forces
was requested using the REACTION (Analysis > Reaction) command.
After you have inspected the model, execute the command R_STATIC (Analysis >
STATIC > Run Static Analysis) to perform linear static analysis. The output file
was examined for reaction forces on the nodes in contact. As shown in the
illustration below, the vertical component of reaction forces adds up to 100,000
which is the force applied, thus providing equilibrium check.

LOAD CASE NUMBER. . . . . . . . . . . . = 1
FORCES IN THE CLOSED BOUNDARY/GAP ELEMENT(S), NCE = 3

ELEMENT FX FY FZ GAP DISTANCE NORMAL FORCE
72 0.0000E+00 0.4404E+05 0.0000E+00 0.5163E-01 0.4404E+05
73 0.0000E+00 0.3153E+05 0.0000E+00 0.2032E+00 0.3153E+05
74 0.0000E+00 0.6277E+03 0.0000E+00 0.4496E+00 0.6277E+03
TOTAL 0.0000E+00 0.7620E+05 0.0000E+00
R E A C T I O N F O R C E S
NODE CSYS RFX RFY RFZ RMX RMY RMZ

6 0 0.1373E+05 0.2380E+05 ---- ---- ---- ----
To see the stress results, the required component of stress can be plotted using
STRPLOT (Results > PLOT > Stress). The figure below shows the horizontal and
vertical components of stresses as well as the von Mises stress contour plots.
Figure 11-5. Stress Contour
Hertz (1895) developed the mathematical theory supported by experimental results

for surface stresses and deformations produced by pressure between curved bodies.
The maximum compressive stresses according to his theory occur at the center of
the contact surfaces while the maximum shear stresses occur at the interior of the
compressed parts. The maximum contact stress is given by the following equation
(Roark's Formulas for Stress and Strain, Sixth Edition, Young, 1989):

(11-1)
where p is the load per unit length and D is the diameter of the cylinder. Using the
above equation, the vertical component of stress at the contact surface is 3.2376E6
whereas the maximum vertical component of stress computed at the center of
element 5 is 4.43372E6. Because of the very small contact area in what one initially
approximates as a line or a point, contact stresses for even lighter loads can be very
high. Since in reality the stresses are highly localized and triaxial, the actual stress
intensity can be very high, and this cannot be accounted for by linear elastic contact
mechanics theories. Therefore the solutions provided by the finite element analysis
can indeed be viewed as accurate.
The command DEFPLOT (Results > PLOT > Deformed Shape) can be used to see
the deformed shape of the cylinder. The DISPLOT (Results > PLOT > Displace-
ment) command can be used to view the displacement contours. The figure below
shows the deformed shape (full and enlarged) as well as the vertical and resultant
displacement contours of the cylinder model.
Figure 11-6. Deformed Shape and Displacement Contour

(See
Example on Bond page
11-2.)
There are many modeling problems in which the compatibility of elements and
nodes at the common boundaries is often difficult to produce during mesh
generation. As recommended in Chapter 8, Modeling Guidelines, the compatibility
of elements must be always maintained, and only for exceptional cases, the bond
feature can be used to tie the non-compatible parts of the mesh.
Example BOND1 - Static Analysis of a U-Frame with Non-compatible

Elements
The input for the geometry Figure 11-7. Model for Analysis Showing
of this problem including Locations for Bond Enforcement
some geometric entities is
provided to you in the file
BOND1.GFM. In the Fixed
descriptions to follow, the Edge
procedures for defining Applied

Pressure
higher order geometric Loads
entities such as polyhedrons Surfaces for BOND
and parts, mesh generation, Enforcement
defining bonding surfaces,

applying constraints and
loads, and performing
analysis are explained in
detail. The model for
analysis consists of a u-
shaped cantilever solid
subjected to pressure loading. The figure below shows the geometry of the model
with applied boundary conditions and loads. The locations of the bonding surfaces
are also indicated since the model will be meshed in such a way that the elements
will be mismatched at these interfaces.
✍ There are about 4000 nodes, 2200 elements, and 11,700 equations in this
example. Solution of linear static analysis requires about 30 Mb of free disk
space on a IBM PC with 16 Mb of RAM. Make sure you have this disk space
available on your computer before attempting to solve this example.

As you can notice, there are four solid volumes in this model. Half of the solid
volumes will be meshed with hexahedral solid (SOLID) elements of different
element size, and the remaining half will be meshed with tetrahedron solid
(TETRA10) elements, also of different sizes. This process will result in a
mesh with non-compatible SOLID-to-SOLID, SOLID-to-TETRA10, and
TETRA10-to-TETRA10 element interfaces. In a conventional finite element
solver, this type of mesh will be unusable for analysis. However, with the bond
feature of COSMOS/M, all you need is the command BONDDEF (LoadsBC >
STRUCTURAL > BONDING > Define Bond Parameter) to connect the non-
compatible parts of the mesh together for providing compatibility, and proceed with
the analysis. The bond definition process is performed at the geometric entity level
such that the nodes and elements at the bonding interfaces are automatically
identified and interconnected by means of constraint relations. Note that the term
non-compatible here refers to the mismatching of nodes at the interfaces, and it is
different from incompatible which refers to cases where the interface has matching
nodes but the deformation fields of the elements are not compatible.
To start with, copy the file BOND1.GFM to your working directory, enter
GEOSTAR, and execute the FILE (File > Load...) command as illustrated below:
Geo Panel: File > Load...
Input file name with extension > BOND1.GFM
Accept Defaults
You will see the geometry of the model constructed and displayed on the screen as
the input commands are read. As you can notice from the geometry of the model,
there are four solid volumes in this model. The geometry created includes curves,
surfaces, regions, and two volumes. These two solid volumes created as volume
entities will be meshed with brick elements using the parametric (mapped) mesh
generation scheme (M_VL (Meshing > PARAMETRIC MESH > Volumes)
command). The remaining two solid volumes are created by using polyhedron and
part definitions and are subsequently meshed with tetrahedron elements using the
MA_PART (Meshing > AUTO MESH > Parts) command.
We will proceed with the definition of polyhedrons and parts using existing
surfaces and regions for 3-D automatic meshing. The procedures for 3-D automatic
meshing is explained in COSMOS/M User Guide. In order to define a polyhedron,
you need to first identify only those surfaces and regions that constitute the
polyhedron. This is best achieved by making use of the selection feature available
in GEOSTAR. The surfaces and regions constituting polyhedrons (also parts) 1 and
2 are shown in the figure below. Notice that surface 22 is common to both parts.

Figure 11-8. Regions and Surfaces for Part Entities 1 and 2
Before defining the first polyhedron, execute the INITSEL (Control > SELECT >
Initialize) command as illustrated below separately for surfaces and regions:
Geo Panel: Control > SELECT > Initialize

Entity Name > Surfaces
Initialization Flag > Yes
Selection Set number > 1
Entity Name > Regions

As seen from the above figure, the first polyhedron is bounded by surfaces 13, 14,
15, 16, and 22, and regions R1 and R3. You need to place these entities in the
selection list using the command SELINP (Control > SELECT > by Label) as
illustrated below:
Geo Panel: Control > SELECT > by Label
Beginning Entity > 13
Ending Entity > 16
Increment > 1


Ending Entity > 22

Increment > 1

Ending Entity > 3
Increment > 2
If you clear the screen and issue SFPLOT; and RGPLOT; (Edit > PLOT >
Surfaces, Regions) commands successively, you will see that only those surfaces
placed in the selection list will be now plotted. You can now issue the PH
(Geometry > POLYHEDRA > Define) command to define a polyhedron, and the
program will only select the surfaces and regions in the selection list to form the
entity. Specify an average element size of 5 for the first polyhedron as shown
below:
Geo Panel: Geometry > POLYHEDRA > Define
Polyhedron Label > 1
Reference entity name > Surfaces
Surface/Region Label > 1

Average element size > 5
Tolerance > 0.0001
Redefine element size on bound > Change size
The polyhedron is a hollow entity, and it needs to be converted to a solid volume by

using the PART (Geometry > Define Part) command as illustrated below:
Geo Panel: Geometry > Define Part
Part Label > 1
Polyhedron 1 > 1
Polyhedron 2 > 1
Similarly, the surfaces and regions that constitute the second polyhedron are
identified and placed in the selection list as illustrated below:



Ending Entity > 22

Ending Entity > 4
Increment > 2
Next, proceed to define the second polyhedron with an element size of 7.5 and the
part entity as illustrated below:
Geo Panel: Geometry > POLYHEDRA > Define
Polyhedron Label > 2
Reference entity name SF or RG > Surfaces
Surface/Region Label > 22

Average element size > 7.5
Tolerance > 0.0001
Redefine element size on bound > Change size
Geo Panel: Geometry > Define Part

Part Label > 2
Polyhedron 1 > 2
Polyhedron 2 > 2
To make all surfaces and regions available for other modeling purposes, you need to
re-execute the INITSEL (Control > SELECT > Initialize) command for all entities
as illustrated below:
Entity Name > All

The procedures you executed so far focused on preparing the geometry for
meshing. To reiterate, volumes 1 and 2 to be meshed with brick elements were
predefined as volume entities and the command M_VL (Meshing > PARAMETRIC
MESH > Volumes) can be directly applied on these volumes. The remaining two
solid volumes were defined as part entities 1 and 2 for automatic mesh generation
with tetrahedron elements using MA_PART (Meshing > AUTO MESH > Parts)
command.
We will first generate the 8-node brick elements in volumes 1 and 2 using the
mapped mesh generation scheme. Select the M_VL (Meshing > PARAMETRIC
MESH > Volumes) command and specify input as illustrated below:
Geo Panel: Meshing > PARAMETRIC MESH > Volumes
Beginning Volume > 1
Ending Volume > 1
Increment > 1
Number of nodes per element > 8
Number of elements on 1st curve > 3
Number of elements on 2nd curve > 10
Number of elements on 3rd curve > 4
Accept Defaults ...
The 8-node brick elements (SOLID) generated will be assigned with element group
number 1 by default. Repeat the M_VL (Meshing > PARAMETRIC MESH >
Volumes) command for the second volume with a different number of elements
along each curve as illustrated below:
Geo Panel: Meshing > PARAMETRIC MESH > Volumes
Beginning Volume > 2
Ending Volume > 2
Increment > 1

Number of elements on 3rd curve > 6
Accept Defaults...

Before generating the tetrahedron elements in parts 1 and 2, you need to first
activate a different element group number for proper association with the generated
10-node tetrahedron (TETRA10) elements:
Geo Panel: Control > ACTIVATE > Set Entity
Set Label > Element Group
Element Set number > 2
You can now proceed to generating the mesh for parts 1 and 2 using tetrahedron
elements. Since surface 22 is common to both parts 1 and 2, and part 2 with
element size 7.5 was last defined, the element density on surface by default will be
7.5. Therefore, before meshing part 1, you need to reset the element density on all
surfaces of part 1 to 5 using the PHDENS (Meshing > MESH DENSITY >
Polyhedron Elem Size) command as shown below:
Geo Panel: Meshing > MESH DENSITY > Polyhedron Elem Size
Beginning Polyhedron > 1
Increment > 1
Average element size > 5
Use the command MA_PART (Meshing > AUTO MESH > Parts) to generate 10-
node tetrahedron elements as shown below:
Geo Panel: Meshing > AUTO MESH > Parts
Beginning Part > 1
Ending Part > 1
Increment > 1
Hierarchy check flag > Check Hierarchy
Element order > High
Number of Smoothing Iterations > 4
Repeat the PHDENS (Meshing > MESH DENSITY > Polyhedron Elem Size)
command and mesh the second part with 10-node tetrahedron elements as shown
below:
Geo Panel: Meshing > MESH DENSITY > Polyhedron Elem Size
Increment > 1
Average element size > 7.5

Geo Panel: Meshing > AUTO MESH > Parts

Beginning Part > 1
Ending Part > 1
Increment > 1
Hierarchy check flag > Check Hierarchy
Element order > High
Number of Smoothing Iterations > 4
Clear the screen and issue HIDDEN,1; (Display > DISPLAY OPTION > Hidden
Element Plot) and EPLOT (Edit > PLOT > Elements) commands to obtain an
element plot of the model without hidden lines. You can notice from the finite
element mesh that at all three common boundaries, the elements do not match. The
figure below shows a three dimensional view as well as 2-D enlarged views of the
mesh at the common boundaries.
Figure 11-9. Generated Finite Element Mesh with Non-compatible Interfaces
Enlarged view of mismatched

elements at BOND surfaces in the X-Y
plane
✍ Since the element types are incompatible, do not attempt to merge the nodes for
this problem even if they appear to be coincident at the common boundaries. At
interfaces with full compatibility, the nodes may however be merged.
You need to identify the surfaces that form the common boundary where the
elements are mismatched. Clear the screen, activate surface label display using
ACTNUM,SF,1 (Control > ACTIVATE > Entity Label) command and plot all
surfaces using SFPLOT (Edit > PLOT > Surfaces) command. You can zoom-in on
the three interfaces for bond definition so that you can clearly identify the surface

numbers for input. The figure below shows the enlarged views of the surfaces with
labels at the common boundaries.
Figure 11-10. Surfaces at the Common Boundaries
From the above figure, you will find that:

• at the SOLID-to-SOLID element interface, surfaces 7 and 9 are interconnected
and can be used for bond set 1,
• at the SOLID-to-TETRA10 element interface, surfaces 11 and 16 are
interconnected, and can be used for bond set 2,
• at the TETRA10-to-TETRA10 element interface, surface 22 connects the two
part entities and can be used for bond set 3.
The above information is helpful in bond definition using the command BONDDEF
(LoadsBC > STRUCTURAL > BONDING > Define Bond Parameter). Select this
command and specify the surface numbers for the first bond set as illustrated
below:
Geo Panel: LoadsBC > STRUCTURAL > BONDING > Define Bond
Parameter
Bonding set > 1
Primary Geometric entity type > Surface
Primary Surface > 9

Secondary Geometric entity type > Surface
Beginning Surface > 7
Ending Surface > 7
Increment > 1
Direction flag > Bi Dir
Repeat this procedure for the remaining two interfaces as shown below:

Parameter
Bonding set > 2
Primary Surface > 11

Ending Surface > 16
Increment > 1
Parameter
Bonding set > 3
Primary Surface > 22

Ending Surface > 22
Increment > 1
If you make a mistake in bond set definition, you can delete that set using the
BONDDEL (LoadsBC > STRUCTURAL > BONDING > Delete Bond Parameter)
command. Use the command BONDLIST (LoadsBC > STRUCTURAL >
BONDING > List) to verify the geometric entities bonded together. This command
will provide the bonding information between primary (or source) and secondary
(or target) entities on the screen as shown below:
Set Stype Source Ttype #Targets Targets

1 SF 9 SF 1 7
2 SF 11 SF 1 16
3 SF 22 SF 1 22
The actual constraint relations between the nodes of source and target geometric
entities are formed and computed in the analysis stage. During this process, the
bond sets defined above are further divided into subsets in which each source node
is connected to one or more target nodes in its vicinity. It is possible to obtain a
listing of the source and target nodes in the subsets, as explained a little later.

You can now proceed with applying boundary conditions and loads. Use the
command DSF (LoadsBC > STRUCTURAL > DISPLACEMENT > Define
Surface) to restrain all degrees of freedom at the left end (surface 4) as illustrated
below:
Geo Panel: LoadsBC > STRUCTURAL > DISPLACEMENT > Define
Surface
Displacement Label > ALL: All 6 DOF
Value > 0.0
Ending Surface > 4
Increment > 1
The pressure loading on the right end as well as the top surface of the right half is
applied using the PSF (LoadsBC > STRUCTURAL > PRESSURE > Define
Surface) command as illustrated below;
Geo Panel: LoadsBC > STRUCTURAL > PRESSURE > Define Surface
Pressure Magnitude > 500
Ending Surface > 18
Increment > 1
Pressure at the end of direction 1 > 500
Pressure Direction > Normal
Geo Panel: LoadsBC > STRUCTURAL > PRESSURE > Define Surface
Pressure Magnitude > 500
Ending Surface > 19
Increment > 1
Pressure Direction > Normal
The input of material properties and element group data completes the model
preparation phase:

Geo Panel: Propsets > Pick Material Lib

Material Name > A_STEEL
Unit Label > FPS
Geo Panel: Propsets > Element Group

Element group > 1
Element category > Volume
Element type (for volume) > SOLID: 8- to 20-node 3D solid element
Accept Defaults ...

Element group > 2
Element category > Volume
Element type (for volume) > TETRA10: 10-node tetrahedral solid element
Accept Defaults
Before executing linear static stress analysis, you need to activate the input print
flag using the command PRINT_OPS (Analysis > OUTPUT OPTIONS > Set Print
Options) in order to obtain a listing of primary node vs. secondary nodes bonding
information.
Geo Panel: Analysis > OUTPUT OPTIONS > Set Print Options
Displacement Print Flag > Yes
... ...
Input Print Flag > Yes
As mentioned earlier, do not merge nodes for this example before performing
analysis. Use the command R_STATIC (Analysis > STATIC > Run Static
Analysis) to execute a linear static stress analysis. The solution of this example
requires about 30 Mb of free disk space on a personal computer with 16 Mb of
memory.
After the analysis is successfully executed, you can inspect the output file
(jobname.OUT) for results. You will notice that the bond information is written in
the following form:

Joint Bond coupling information

-------------------------------
Ref.; Reference node

Types of coupling
EQ. PTC ; Point to curve
EQ. PTS ; Point to surface
Set Type Ref. NODE1 NODE2 NODE3 NODE4 NODE5 NODE6 NODE7 NODE8 NODE9 NODE10
1 PTS 3224 3075 3079 3035 3031
2 PTS 3225 3079 3083 3039 3035
... ...
The set in the above listing refers to the subsets derived from the bond sets you
earlier defined using the BONDDEF (LoadsBC > STRUCTURAL > BONDING >
Define Bond Parameter) command.
You can view the displacement and stress results after activating the required
components using the (Results > PLOT > Stress, Displacement) command. The
figures below show contours of resultant displacement and von Mises stresses.
These figures reinforce the fact that the results obtained using the bond feature are
continuous and smooth as if a fully compatible mesh were used.
Figure 11-11. Displacement and Stress

(See
Linear Elastic Fracture Mechanics Analysis Examples page
11-2.)
There are two quarter-point singular elements (8-node PLANE2D and 20-node
SOLID) available in the Basic System to model problems with cracks. When you
use these elements, the fracture parameter computed is the stress intensity factor for
the applicable mode of fracture. Other fracture parameters such as the J-integral or
the energy release (G) can be calculated based on the computed stress intensity
factors or directly compute through the J_integral option of the Basic Ststem. The
following two examples illustrate the computation of stress intensity factors for
problems with different crack geometry. It should be understood that in the
following examples, due to structural symmetry, only half or one quarter of the
structure is modeled. However, in more general cases, you may have to define a
crack element for each one of the two faces of a crack.
In the examples of this section only the use of crack element is considered in the
evaluation of the fracture mechanics characteristics. However, you may use the
J-integral option to evaluate the same for 2D problems (using TRIANG and
PLANE2D elements), as well.
Example LEFM1 -Stress Intensity Factor of a Plate with Two Edge

Cracks
The input for this problem is available to you in the file LEFM1.GEO. This
example considers the mode-I stress intensity factor solution of a plate with two
through-thickness edge cracks under far-field tensile stresses as shown in the figure
below. The elastic material properties and other dimensions are shown in the figure
below.

Figure 11-12. Problem for Stress Intensity Factor
σ∞ σ∞ = 1 0 0
E=30E6
120 ν=0.28
2W t=0.5
a=20
Crack
a a
120
Quarter model for analysis
σ∞
Plate with through-thickness

edge cracks
The symmetry of the problem enables us to use a quarter of the model for analysis.
Set the view to 2-D X-Y using the Viewing (Binocular) icon. The geometry and the
finite element mesh are created as illustrated in the following commands:
Geo Panel: Geometry > GRID > Plane
Rotation/Sweep Axis > Z
Offset on Axis > 0.0
Grid Line style > Solid
Geo Panel: Display > VIEW PARAMETER > View

Accept defaults...
Geo Panel: Geometry > SURFACES > Draw w/ 4 Coord

Surface > 1
XYZ-Coordinate of Keypoint 1 > 30,30,0
Click on Auto Scale icon ...
Geo Panel: Meshing > PARAMETRIC MESH > Surfaces

Ending Surface > 1

Increment > 1
Number of elements on the 1st curve > 12
Number of elements on the 2nd curve > 15
Spacing ratio for 1st curve > 1.0

Spacing ratio for 2nd curve > 1.0
Note that the mesh was generated with higher order elements in order to use the
quarter-point element. The PLANE2D element required to be used for this problem
is selected using the EGROUP (Propsets > Element Group) command whereas the
material and sectional property data are entered as follows:
Element Group > 1
Element Category > Area
Element Type (for area) > PLANE2D: 2D 4- to 8-node plane element
Accept Defaults ...

Material Property set > 1
Unit Label > FPS
Geo Panel: Propsets > Real Constant

Associated Element group > 1
Real Constant set > 1
Start loc. of the real const > 1

No. of real const to be entered > 2
RC1: Thickness > 0.5

RC2: Material angle (Beta) > 0.0
Symmetric boundary conditions are enforced along the vertical and horizontal axes
of symmetry. Since the crack represents a traction free edge and a discontinuity in
the geometry, there are no displacement boundary conditions enforced on the crack.
Therefore, after you apply displacement boundary conditions on the curve entity,
you need to remove the constraints placed on the crack. The following commands
illustrate:


Curves
Beginning curve > 1
Displacement Label > UY: Y translation
Value > 0.0
Ending curve >
Increment > 1

Curves
Beginning curve > 4
Displacement Label > UX: X translation
Value > 0.0
Ending curve > 4
Increment > 1
Geo Panel: LoadsBC > STRUCTURAL > DISPLACEMENT > Delete

Nodes
Beginning node > 1
Ending node > 4
Increment > 1
The far-field tensile stresses on the top edge can be applied using the PCR
(LoadsBC > STRUCTURAL > PRESSURE > Define Curves) command as shown
below:
Geo Panel: LoadsBC > STRUCTURAL > PRESSURE > Define Curves
Beginning curve > 2
Pressure magnitude > -1000
Ending curve > 2
Increment > 1
Pressure at the end of direction 1 > -1000
Pressure Direction > Normal Direction
You can now proceed to define the quarter-point element at the crack tip. As
outlined in Chapter 8, Modeling Guidelines, the command CRACK (Analysis >
STATIC > CRACK > Define Crack) is used to identify the nodes at the crack tip.
The first node is always at the crack tip, and the second node at the corner of the
same element, away from the crack tip. The third node is required for 3-D crack
problems only. The CRACK (Analysis > STATIC > CRACK > Define Crack)
command is illustrated below:

Geo Panel: Analysis > STATIC > CRACK > Define Crack
Crack number > 1
Node 1 > 5
Node 2 > 3
Node 3 >
The commands CRACKLIST and CRACKDEL (Analysis > STATIC > CRACK >
List Crack, Delete Crack) from the CRACK submenu can be used to list or delete
quarter-point elements defined. The figure below shows the finite element model
with applied loading and boundary conditions. The figure also shows the node
numbers at the crack tip and the corner of the quarter-point element.
Figure 11-13. Finite Element Model with Boundary Conditions and Load
You can now proceed to performing a linear static stress analysis of the model using
the command R_STATIC (Analysis > STATIC > Run Static Analysis). The Basic
System computes the stress intensity factors and prints them in the output file. To
see the effect of geometric singularity introduced by the crack on the stress
distribution, you can use the command STRPLOT (Results > PLOT > Stress). The
figures below show a plot of von Mises stresses as well as the deflected shape of the
model.

Figure 11-14. Stress Contours and Deformed Shape Plot of the Cracked Plate
The stress intensity factors are read from the output file (jobname.OUT). For this
example, they were found to be:
CRACK STRESS INTENSITY FACTORS FOR CRACK 1
KI = 0.885133E+03
Note that the mode-II stress intensity factor is not relevant for this problem. The
analytical solution for the mode-I stress intensity factor is as follows (Keer and
Freedman, 1973):
a a3
1.12 − 0.61 + 0.13 3
K I = σ ∞ πa W W for all a/W (11-2)
a
1−
W
The table below shows a comparison between the analytical and COSMOS/M
solutions.
Table 11-1. Comparison Between COSMOS/M and Analytical Solution for KI
KI Solution Result Error (%)
Theoretical 925.868 N/A

COSMOS/M 300 elements 885.133 4.4

Example LFEM2 - Stress Intensity Factor (See

page
of Three Point Bend Specimen 11-2.)
The input for this problem is available to you in the file LEFM2.GEO. This
example considers the mode-I stress intensity factor solution of a beam with a
through-thickness edge crack at the center subjected to a point load as shown in the
figure below. Otherwise known as a 3-point bend specimen, this type of model is
commonly used to experimentally determine the mode-I stress intensity factor. The
elastic material properties and other dimensions are shown in the figure below.
Figure 11-15. Problem for Stress Intensity Factor Computation
P
P/2=0.5
B E=30E6
ν=0.28
W=s/2 40 B=0.5
a Crack
s s 80
P/2 P/2
Thee point bend test for Model for analysis
mode-I fracture toughness
The symmetry of the problem enables us to use one half of the model for analysis.
Change the view to 2-D X-Y using the View (Binocular) icon. The geometry and
the finite element mesh are created as illustrated in the following commands:
Rotation/sweep Axis > Z
Grid Line Style > Solid

Surface > 1
Click on Auto Scale icon ...


Beginning surface > 1
Ending surface > 1
Increment > 1
Number of elements on the 1st curve > 20

Number of elements on the 2nd curve > 10
Accept Defaults...
As in the previous example, the mesh was generated with higher order elements in
order to use the quarter-point element. The PLANE2D element required to be used
for this problem and the material and sectional property data are entered as follows:
Element Group > 1
Element Category > Area
Element Type (for area) > PLANE2D: 2D 4- to 8-node plane element
Accept Defaults ...

Unit Label > FPS

Real Constant set > 1
Start loc. of the real const > 1

No. of real const to be entered > 2

RC2: Material angle (Beta) > 0.0
Symmetric boundary conditions are enforced along the vertical axis of symmetry.
Since the crack represents a free edge, there are no displacement boundary
conditions enforced on the crack. Along the vertical axis of symmetry (except at the
crack), the horizontal displacements are specified to be zero. The boundary
conditions are input as shown below:


Nodes
Beginning node > 1
Value > 0.0
Ending node > 1
Increment > 1

Nodes
Beginning node > 351
Displacement Label > UX: X translation
Value > 0.0
Ending node > 351
Increment > 1
Repeat the above command for UX at nodes 372, 413, 434, 475, 496, 537, 558,
599, 620 and 661. The point load at the top is specified as shown below:
Geo Panel: LoadsBC > STRUCTURAL > FORCE > Define Node
Force label > FY: Y force
Value > -0.5
Ending node > 661
Increment > 1
You can now proceed to define the quarter-point element at the crack tip. For this
example, the first node which is always at the crack tip is 351, and the second node
which is at the corner of the same element away from the crack tip, is 289. These
nodes are specified using the CRACK (Analysis > STATIC > CRACK > Define
Crack) command as illustrated below:
Geo Panel: Analysis > STATIC > CRACK > Define Crack
Crack number > 1
Node 1 > 351
Node 2 > 289
Node 3 > 289
The figure below shows the finite element model with loads and boundary
conditions. The figure also shows the node numbers at the crack tip and the corner
of the quarter-point element.

Figure 11-16. Finite Element Model with Boundary Conditions and Loads
You can now proceed to performing a linear static stress analysis of the model using
the command R_STATIC (Analysis > STATIC > Run Static Analysis). The effect
of geometric singularity introduced by the crack on the stresses can be viewed
using the command STRPLOT (Results > PLOT > Stress). The figures below show
a plot of von Mises stresses as well as the deflected shape of the model.
Figure 11-17. Stress Contours and Deformed Shape Plot of the Cracked Plate
The stress intensity factors are read from the output file (jobname.OUT). For this
example, they were found to be:
CRACK STRESS INTENSITY FACTORS
KI = 0.294492E+01

Once again, the mode-II stress intensity factor is not relevant for this problem. The
analytical solution for the mode-I stress intensity factor is as follows (Gross and
Srawley, 1972):
PW
K I = 3. 75 3 for 0. 4 ≤ a W ≤ 0.6 (11-3)
B(W − a) 2
The table below shows a comparison between the analytical and Basic System
solutions. A second finite element solution obtained with 800 elements, shows
convergence towards the analytical solution for mode-I stress intensity factor.
Table 11-2. Comparison Between COSMOS/M and Analytical Solution For KI
KI Solution Result Error (%)
Theoretical 3.3541 N/A


(See
Buckling Analysis Examples page
11-2.)
Buckling analysis which determines the critical buckling loads and mode shapes
can be performed on many types of practical problems in the Basic System. Various
eigenvalue extraction methods and modeling features available in the DSTAR
module will enable you to address almost all types of modeling and analysis
situations. Owing to space limitations, it is not possible to demonstrate all of these
features. The example presented here discusses the application of some of the
features outlined in Chapter 1. In addition, the second part of this manual presents
more than twelve verification problems you may find useful in understanding many
aspects of the buckling analysis capabilities available.
✍ The second part of this manual presents more than 100 verification problems on
all analysis features of the Basic System. You are recommended to use these
examples to learn more about the Basic System.
Example BUCKL1 - Buckling of a Rectangular Plate

Under Uniform Pressure
The input for this example is available in the file BUCKL1.GEO. This example
considers the eigenvalue buckling of a rectangular plate under uniform pressure.
The material properties and dimensions of the plate are shown in the figure below.
Two opposite edges of the plate are simply supported whereas the remaining two
edges are clamped. The plate is subjected to a uniform pressure loading of unit
intensity so that the computed eigenvalues will directly yield the critical buckling
loads. Otherwise, the eigenvalues are treated as multipliers to the actual buckling
loads.

Figure 11-18. Problem for Critical Buckling Load and Applied Boundary Conditions
E = 30E6 a = 60
ν = 0.3 b = 20
t = 0.3
a
Simply supported
Clamped
b
Simply supported
Rectangular plate under uniform compression
One of the clamped edges of the plate is subjected to uniform pressure. However, in
order to induce buckling in the finite element model, the x-component of the
displacements was released on this edge whereas the other clamped edge was
restrained with respect to all components of displacements. The figure above also
shows the applied boundary conditions on the plate in 3-D space.
The geometry and the finite element mesh of the plate are created as illustrated
below:

Surface > 1

Beginning surface > 1
Ending surface > 1
Increment > 1

Accept Defaults

Click on Auto Scaling icon to view the model clearly. The simply supported
boundary condition on two edges of the plate was modeled by restraining the z-
component of displacements. The pressure loading and the displacement boundary
conditions are applied as shown in the commands below:
Curves
Beginning curve > 1
Displacement label > UZ: Z translation
Value > 0.0
Ending curve > 4
Increment > 1

Curves
Beginning curve > 3
Displacement label > ALL: All 6 DOF
Value > 0.0
Ending curve > 3
Increment > 1

Curves
Beginning curve > 4
Displacement label > UY: Y translation
Value > 0.0
Ending curve > 4
Increment > 1
Displacement label > RX: X rotation

Displacement label > RY: Y rotation
Displacement label > RZ: Z rotation
Geo Panel: LoadsBC > STRUCTURAL > PRESSURE > Define Curves
Beginning curve > 4
Pressure magnitude > 1
Ending curve > 4
Increment > 1
Pressure direction > Normal Direction

The plate will be modeled with SHELL4 elements selected using the EGROUP
(Propsets > Element Group) command. The material and sectional property data
are entered using MPROP (Propsets > Material Property) and RCONST (Propsets
> Real Constant) commands as follows:
Element group > 1
Element category > Area
Element type (for area) > SHELL4: 4-node thin shell element
OP1: Type > QUAD4

... ...
Accept Defaults
Geo Panel: Propsets > Material Property

Material property name > EX: Elasticity modulus 1st dir
Property value > 30E6
Material property name > NUXY: Poisson ratio 2nd to 1st

Property value > 0.3

Real constant set > 1
Start loc. of the real const. > 1

No. of real const. to be entered > 6
... ...
Accept Defaults
The number of buckling modes as well as the method of eigenvalue extraction is

specified using the A_BUCKLING (Analysis > FREQUENCY/BUCKLING >
Buckling Options) command. We will input for the first five modes of buckling
using Lanczos method for this problem as shown below:
Geo Panel: Analysis > FREQUENCY/BUCKLING > Buckling Options
Number of Eigenvalues > 5
Method > Lanczos method
... ...
Accept Defaults

The command R_BUCKLING (Analysis > FREQUENCY/BUCKLING > Run

Buckling) can now be issued to compute the critical buckling loads. The results of
execution are written to the output file, jobname.OUT, and you can examine this file
for buckling load factors and the corresponding mode shape deflections. The
critical buckling loads can also be listed on the screen using the FREQLIST (Results
> LIST > Natural Frequency) command. The listing below shows the buckling
load factors for the first five modes:
B U C K L I N G E I G E N V A L U E (S)
by
L A N C Z O S A L G O R I T H M
EIGENVALUE EIGENVALUE
NUMBER
1 0.2712670E+05
2 0.2775493E+05
3 0.3398611E+05
4 0.3840657E+05
5 0.5011349E+05
The analytical solution for this problem can be found in Roark's Formulas for
Stress and Strain (Sixth Edition, Young, 1989) and is given below for a/b = 3:
(11-4)
where b is the length of the shorter side, and t is the thickness. The following table
shows the comparison between the analytical and the Basic System solutions.
Table 11-3. Comparison Between COSMOS/M and Analytical Solution
First Buckling Mode Solution Result Error (%)
Analytical 26925.82 N/A

As seen in the above table, the Basic System solutions are very accurate. The
command DEFPLOT (Results > PLOT > Deformed Shape) can be used to plot the
buckling mode shapes. The first three buckling modes are shown below in the
isometric viewing position.

Figure 11-19. Buckling Mode Shapes of the Plate In 3-D Space
You can change the viewing position to 2-D space to see the lateral deformations
clearly. The figure below shows the three buckling modes corresponding to the
above figure in the side view position.
Figure 11-20. Corresponding Buckling Mode Shapes In 2-D Space

(See
Modal Analysis Examples page
11-2.)
Modal analysis which determines the natural frequencies and mode shapes can be
performed on many types of practical problems in the Basic System. Various
eigenvalue extraction methods and modeling features available in the DSTAR
module will enable you to address almost all types of modeling and analysis
situations. Owing to space limitations, it is not possible to demonstrate all of these
features. Two examples presented here (MODAL1 and MODAL2) discuss the
application of some of the features outlined in Chapter 1. In addition, the second
part of this manual presents many verification problems you may find useful in
understanding many aspects of the modeling and analysis capabilities.
Example MODAL1 - In-plane Effects on the Natural Frequencies of a

Column
The input for this Figure 11-21. Problem for Modal Analysis
problem is available to with In-Plane Effects
you in file
MODAL1.GEO. For P=1000
slender structures with
in-plane loading, the
natural frequencies are
significantly altered
depending on the type E=30E6
of preload applied. As A=1.0
55
I=0.01
explained in Chapter 2, ρ= 7 . 2 4 6 E - 4
Mathematical
Background, compressive
loads tend to decrease the
natural frequencies
whereas tensile preloads
increase them. The effect Problem Geometry Finite Element Model
of a preload on the natural
frequencies is significant
for the first few modes and its influence gradually diminishes for higher modes.
This example quantitatively demonstrates the effects of compressive and tensile
preloads on the natural frequencies with respect to those obtained with no preloads.

The geometry and the finite element mesh for this problem are created as illustrated
in the following commands:
... ...
Accept Defaults
Geo Panel: Geometry > CURVES > Draw Polyline

Curve > 1
Geo Panel: Meshing > PARAMETRIC MESH > Curves

Beginning curve > 1
Ending curve > 1
Increment > 1
Number of elements on each curve > 8

Spacing ratio > 1.0
The column will be modeled using 2-D beam elements (BEAM2D). This element is
selected using the EGROUP (Propsets > Element Group) command whereas the
material and sectional property data are entered as follows:
Element group > 1
Element category > Line
Element type (for line) > BEAM2D: 2D elastic beam element
Accept Defaults


RC1: Cross-sectional area > 1.0

RC2: Moment of inertia IZ > 0.01
RC3: Depth > 1.0


Material name > A_STEEL
Unit label > FPS
The displacement boundary conditions at the top and bottom are enforced using the
DND (LoadsBC > STRUCTURAL > DISPLACEMENT > Define Nodes)
command where as the vertical load on the column is specified using the FND
(LoadsBC > STRUCTURAL > FORCE > Define Nodes) command. In the first
run, we will obtain natural frequencies in the presence of a compressive preload.
The following lines illustrate the use of these commands:
Nodes
Beginning node > 9
Displacement label > ALL: All 6 DOF
Value > 0.0
Ending node > 1
Increment > 1

Nodes
Beginning node > 9
Displacement label > UX: X translation
Value > 0.0
Ending node > 9
Increment > 1
Geo Panel: LoadsBC > STRUCTURAL > FORCE > Define Nodes
Beginning node > 9
Value > -1000
Ending node > 9
Increment > 1
The effect of a preload on the natural frequencies is specified by activating the in-
plane effects flag in the A_FREQUENCY (Analysis > FREQUENCY/BUCKLING
> Frequency Options) command as illustrated below. The number of eigenvalues
are also input in the same command. We will investigate the first five modes of the
column.

Geo Panel: Analysis > FREQUENCY/BUCKLING > Frequency Options

Number of frequencies > 5
Method > Subspace iteration method
Maximum number of iterations > 16
Sturm sequence flag > No
Shift flag > No Eigenvalue Shift
Shift value > 0.0
In-plane effect flag > Yes
... ...
Accept Defaults
You can use the command R_FREQUENCY (Analysis > FREQUENCY/

BUCKLING > Run Frequency) to compute the first five natural frequencies of the
column in the presence of a compressive preload. After the analysis is successfully
executed, you can use the command FREQLIST (Results > LIST > Natural
Frequency) to list the computed natural frequencies on the screen.
Delete the load at the top of the column and define a tensile preload as illustrated
below:
Geo Panel: LoadsBC > STRUCTURAL > FORCE > Delete Nodes
Beginning node > 9
Ending node > 9
Increment > 1
Beginning node > 9
Value > 10000
Ending node > 9
Increment > 1
Before executing another normal modes analysis, you need to activate a flag so that
the results of the analysis you last performed are not over-written with the new
results. Use the command PRINT_OPS (Analysis > OUTPUT OPTIONS > Set
Print Options) as illustrated below:

Geo Panel: Analysis > OUTPUT OPTIONS > Set Print Options
Displacement Print Flag > Yes
... ...
Output Flag > Append
Re-execute the problem using the command R_FREQUENCY (Analysis >

FREQUENCY/BUCKLING > Run Frequency) to compute the first five natural
frequencies of the column in the presence of a tensile preload.
Delete the load at the top of the column using the command FNDEL (LoadsBC >
STRUCTURAL > FORCES > Delete Nodes) and deactivate the in-plane effects
flag in the A_FREQUENCY (Analysis > FREQUENCY/BUCKLING > Frequency
Options) command. Execute another analysis to obtain results without the effects
of preloads. Again, you can use the command FREQLIST (Results > LIST >
Natural Frequency) to list the computed natural frequencies. The results are also
written to the output file (jobname.OUT).
The table below shows a summary of results for the first five modes. The results
clearly show the effects of tensile and compressive preloads on the natural
frequencies. As stated earlier, the influence of preloads on the natural frequencies
decrease for higher mode shapes.
Table 11-4. Comparison of Natural Frequencies With and Without Preloads
Mode In-plane Load - Frequency (Hz)

Number Compressive No Tensile
1 11.71 16.44 20.03
2 48.43 53.26 57.68
3 105.92 110.83 115.54
4 183.02 187.90 192.66
5 275.20 279.88 284.47
The effect of tensile and compressive preloads on the natural frequencies are better
represented in the following figure.

Figure 11-22. Effect of Tensile and Compressive Preloads on the Natural

Frequencies
300
Compressive Load
250
Frequency (Hz)
No Load
200
150 Tensile Load
100
50
0
1 2 3 4 5
Mode Number
The DEFPLOT (Results > PLOT > Deformed Shape) command can be used for
plotting the free vibration modes of the column. The figure below shows the first
three modes for the column with no preloads.
Figure 11-23. Free Vibration Modes of the Column

Example MODAL2 - Modal Analysis of a Square Frame (See

page
with Rigid Body Modes 11-2.)
The input for this problem is available to you in file MODAL2.GEO. The analysis
problem consists of a box-shaped frame with material and cross sectional properties
and dimensions as shown in the figure below. The frame which lies in 2-D space is
unsupported, and it is required to determine the first ten natural frequencies and
mode shapes.
Figure 11-24. Problem for Modal Analysis with Rigid Body Modes
60
E=30E6
A=1.0
60 I=0.01
ρ =7.246E-4
Problem Geometry Finite Element Model
The geometry and the finite element mesh for this problem are created as illustrated
in the following commands:
Geo Panel: Geometry > CURVES > Draw Polyline

Curve > 1

Geo Panel: Meshing > PARAMETRIC MESH > Curves

Beginning curve > 1
Ending curve > 4
Increment > 1
Number of elements on each curve > 5

Spacing ratio > 1.0
Click on the View (Binocular) icon and change the view to X-Y. You can scale the
image to fit the viewport by clicking on the Auto Scale icon. The frame will be
modeled with 2-D beam elements (BEAM2D). This element is selected using the
EGROUP (Propsets > Element Group) command whereas the material and
sectional property data are entered as follows:
Element group > 1
Egroup category > Line
Element type (for line) > BEAM2D: 2D elastic beam element
Accept Defaults


RC1: Cross-sectional area > 1

RC2: Moment of inertia (IZ) > 0.01
RC3: Depth > 1
... ...
Accept Defaults

Material name > A_STEEL
Unit label > FPS

Since the finite element mesh for each curve is generated independently, there will
be two nodes at the corners. For compatibility requirements, these nodes have to be
merged into one. The NMERGE (Meshing > NODES > Merge) and NCOMPRESS
(Edit > COMPRESS > Nodes) commands are used to merge the nodes and
consecutively number them, respectively.
Since the behavior of the finite model is confined to 2-D space, you can expect to
see three rigid body modes: two translations and a rotation. As described in Chapter
2, Mathematical Background, problems with rigid body modes can be dealt with in
two ways: using an eigenvalue shift, or a soft spring addition, specified using the
A_FREQUENCY (Analysis > FREQUENCY/BUCKLING > Frequency Options)
command. We will use both options and compare the results.
We will compute the first ten free vibration modes and mode shapes for this
problem. The number of eigenvalues as well as the eigenvalue shift are input in the
A_FREQUENCY (Analysis > FREQUENCY/BUCKLING > Frequency Options)
command as illustrated below:
Method > S: Subspace iteration
Shift flag > Shift by program
Shift value > 0.0
... ...
Accept Defaults
The command R_FREQUENCY (Analysis > FREQUENCY/BUCKLING > Run

Frequency) is next executed to perform modal analysis of the square frame. After
the analysis is successfully executed, you can use the command FREQLIST
(Results > LIST > Natural Frequency) to list the computed natural frequencies on
the screen. The table below shows the mode numbers and the corresponding natural
frequencies using the shift option. The first three modes shown represent rigid body
modes with negligible frequencies. This table also shows the frequencies obtained
using the soft spring option, explained a little later.

Table 11-5. Comparison of Natural Frequencies Using

Eigenvalue Shift and Soft Spring Option
Option Frequency (Hz)

Mode Number
Shift Soft Spring
1 2.40086e-06 1.70049e-03
2 9.05758e-06 1.70051e-03
3 1.25760e-05 1.70122e-03
4 5.08984e+00 5.08985e+00
5 8.84410e+00 8.84410e+00
6 1.71626e+01 1.71626e+01
7 1.71626e+01 1.71626e+01
8 2.00285e+01 2.00285e+01
9 3.74398e+01 3.74398e+01
10 4.47586e+01 4.47586+01
The command DEFPLOT (Results > PLOT > Deformed Shape) can be used to plot
the mode shapes. The rigid body modes can also be computed using the soft spring
option. As explained in Chapter 2, Mathematical Background, the addition of soft
spring stiffness results in each diagonal term of the structural stiffness incremented
by a small value to provide numerical stability. In the Basic System, you can vary
the value of soft spring stiffness. The soft spring addition is specified as shown
below:
Method > S: Subspace iteration
Shift flag > No Eigenvalue Shift
Shift value > 0.0
In-plane effect > No
Tolerance > 0.000010
Soft Spring flag > Yes
... ...
Accept Defaults
Use the command PRINT_OPS (Analysis > OUTPUT OPTIONS > Set Print
Options) to append the new analysis results to the previous one in the output file.
Next, use the command R_FREQUENCY (Analysis > FREQUENCY/BUCKLING

> Run Frequency) to compute the natural frequencies and mode shapes of the
square frame. After the analysis is successfully executed, you can use the command
FREQLIST (Results > LIST > Natural Frequency) to list the computed natural
frequencies on the screen. As shown in the previous table, the first three modes
representing rigid body displacements have negligible natural frequencies. You can
notice from the table that the natural frequencies for the deformable modes (mode 4
onwards) are identical with those obtained using the eigenvalue shift option. The
figure below shows the three rigid body modes obtained using the soft spring
option.
Figure 11-25. Rigid Body Modes Computed Using Soft Spring Option
Rigid Body Mode 1 Rigid Body Mode 2 Rigid Body Mode 3
The DEFPLOT (Results > PLOT > Deformed Shape) command can be used in
either approach for plotting the deformable free vibration modes of the square
frame. The figure below shows the fourth, fifth and six free vibration modes.
Figure 11-26. Deformable Free Vibration Modes of the Square Frame

Using Soft Spring Option

(See
Examples on Submodeling page
11-2.)
Submodeling is a feature of STAR which can be used to improve the solution at

locally critical areas without having to rerun for the whole model. This can reduce
the CPU time substantially for large problems.
Example LSUBM1 - Using Submodeling for Shell Problems
The input for the modeling Figure 11-27. Model Geometry for the Analysis
part of this example is
available in the file
LSUBM1.GEO. The figure
below shows a 3-D view of
the model constrained at
one end and subjected to
forces at the other end.
After the initial run of the

model the von Mises stress
distribution viewed from a
more convenient
orientation, is displayed in
the following figure.
Change the view to X-Z
using the Viewing icon.
Geo Panel: Results >

PLOT > Stress
Load case number > 1
Component > VON
Layer number > 1
Coordinate system > 0
Stress Flag > Nodal Stress
Face of element > Top Face
Select Contour and Accept Defaults

Figure 11-28. von Mises Stress Distribution of the Original Run
To improve the results at the areas of high stress concentrations, four portions
surrounding the constrained holes are selected by repeated use of SELWIN (Control
> SELECT > by Windowing) command (see the LSUBM1.GEO file for selected
elements):
After the element selection is made to define the submodel, use the SUBMODEL
(Analysis > STATIC > Define SubModel) command to activate submodeling and
refine the local mesh. Run STAR to solve for the submodel.
Geo Panel: Analysis > STATIC > Define SubModel
Submodeling option > On
Push boundary node flag > No
Geo Panel: Analysis > STATIC > Run Static Analysis
The following figure displays von Mises stress results. Note that you need to bring
back the rest of the model (not analyzed with the submodel) to visualize results on
the entire model.
Entity name > Elements
... ...
Accept Defaults

Geo Panel: Results > PLOT > Stress

Component > VON
Layer number > 1
Figure 11-29. von Mises Stress Distribution After Running the Submodel
The maximum von Mises stresses calculated by the submodel shows a modification
of about 10% over the original run (diagram).
If you deactivate submodeling (use the SUBMODEL (Analysis > STATIC > Define
SubModel) command to deactivate submodeling) and run the program for the
entire model it takes more than 20 minutes to complete the job on a 66 Hz, 486 PC,
whereas it takes only 3 minutes to run for the submodel. Time savings increase
substantially as the problem size becomes larger. More interestingly, the maximum
von Mises stress obtained in the entire model run (including the refined mesh) is
almost the same as the one calculated by the submodel. The accuracy obtained by
the submodel depends on how far are the boundaries of the submodel from the
stress concentration areas.

Example LSUBM2 - Using Submodeling with Tetrahedral Elements (See

page
11-2.)
The input for the model of this example is available in the file LSUBM2.GEO. The
figure below shows von Mises stress distribution of the joint after the initial run.
Figure 11-30. von Mises Stresses for the Initial Run
470.1000
3.493800
A view along the Y-direction (X-Z plane) is chosen for a better selection of the
submodel region, using the Viewing icon.
Geo Panel: Control > SELECT > by Windowing
... Select Desired Area ...
(Note: alternatively, you may use the SELRANGE (Control > SELECT > by
Range) command to select partially through the depth of the model).


Figure 11-31. Region Selected for Submodeling
The following figure displays the von Mises stresses after running for the
submodel. You can change the view to isometric by using the Viewing icon.
... ...
Accept Defaults

Component > VON
Layer number > 1

Figure 11-32. von Mises Stresses for the Submodel
517.2600
2.908900
Maximum von Mises stresses are increased form the initial value of 470 psi to 517
psi after solving for the submodel.

Example LSUBM3 - Using Submodeling for Shell Problems (See

page
11-2.)
The input for this problem is provided to you in the file LSUBM3.GEO. The model
represents a car dashboard subjected to surface pressure and concentrated loads.
The figure below shows von Mises stresses after the initial run. To improve the
results at the lower part of the model, the SELWIN (Control > SELECT > by
Windowing) command with the circular window option is used to select a circular
area to define the submodel.
Figure 11-33. von Mises Stress for the Initial Run

(For initial run ...)
Geo Panel: Control > SELECT > by Windowing
(Select critical area for submodeling)

(Run analysis on the submodel)


... ...
Accept Defaults

Component > VON
Layer number > 1
Figure 11-34. Local von Mises Stresses for the Submodel

(refined area is the actual submodel)
The maximum von Mises stress increased from the initial value of 2.61E5 psi to
2.72E5 psi.

(See
Example on ASME Stress Requirement page
11-2.)
The stresses are reevaluated in the following example on the bases of the redundant
forces and moments along a section associated with shell-type structures in order to
be able to compare them with allowable stresses, Sm, 1.5 Sm and 3 Sm, specified
by the ASME code section III and in accordance with Gordon, 1976 (refer to
Chapter 2 Mathematical Backgrounds, as well as to Appendix C).
Example ASME.GEO - ASME Code Stress Evaluation
The input for this problem is Figure 11-35. A Section Defined Through
provided to you in file the Nozzle Thickness
ASME.GEO. The stresses are
evaluated at the two ends of a CL
section defined on a cross- 1"
sectional slice of an
axisymmetric nozzle subjected 2"
to internal pressures according
to the ASME code stress
2"
evaluations. This examples is P
2"
intended to show the general
use of this capability. However,
4"
in a typical problem you may 1 2 1"
need to consider several

sections in order to find the 3" 2"
maximum stresses. A section is
defined by the two nodes at its
ends. These nodes could be 5"
defined any where within the
model and are not confined to
the FEA mesh (i.e., they do not have to be associated with any elements). To find a
proper orientation for sections at irregular areas such as in the nozzle to shell
junctions please refer to Appendix C.
In the following only the commands related to ASME code stress requirement are
discussed.

In order to use this option, activate the ASME flag in the A_STRESS (Analysis >
STATIC > Stress Analysis Options) command:
Geo Panel: Analysis > STATIC > Stress Analysis Options
AISC code check > No
... ...
ASME code check flag > Yes
Geo Panel: Analysis > STATIC > ASME CODE > Define ASME Section
ASME section > 1
First node > 55
Second node > 75
Number of points > 25
Radius of curvature: 0=plane > 0 axisym -1 straight > -1
Thickness direction bending stress flag > Yes
In the above command, the section is labeled one. You may define up to 300
sections at different locations of the structure. By the entries for second and third
prompts, the section is defined at two ends by nodes 55 and 75. By choosing the
default as 25 points for the fourth prompt, the program is instructed to consider
only 25 integration points along the section. However, for a section with a very
large stress gradient you may consider up to 65 points (i.e., you may check stress
variation by plotting von Mises stresses along the section using the LSECPLOT
(Results > PLOT > Path Graph) command).
The fifth prompt, “Radius of curvature” indicates whether the computation is

considered for the plane type of problems (such as plane stress, plane strain or for
axisymmetric problems approximated by Cartesian formulation) or true
axisymmetric formulation such as the one considered in the present example. If you
consider the structure at section 1 to be locally a cylinder with straight walls (inner
and outer walls), then it is proper to consider the radius of curative to be infinite
(see Appendix C, Subsection “Axisymmetric Formulation” for definition of the
curvature-radius). This can be represented as -1 or equivalently a large number for
that prompt. However, if you want to consider the curvature at the lower part of the
section, you may do so, by choosing the radius of curvature of the average mid-wall
centerline which in this model can be approximated to be equal to 2 inches.
The last prompt is used to consider or ignore the thickness direction bending
stresses (see Appendix C for more details). You may use “A_LIST, STRESS”
(Analysis > List Analysis Option) to check the status of this option as well as
“ASMESECLIS and ASMESECDEL” (Analysis > STATIC > ASME CODE > List

ASME Section, Delete ASME Section) to list or delete the sections. After running
the analysis, the results are printed at the end of output file which you may compare
with the allowable stresses required by the ASME code.
Since all the computations for the ASME code stresses are performed during the
stress calculation, you do not need to perform the displacement calculation part of
the solution, every time you need to rerun for modified or added sections. In that
case simply reissue the R_STRESS (Analysis > STATIC > Run Stress Analysis)
command. Furthermore, if you wish to define the section along a path which do not
coincide with the natural boundary of the element mesh, you may do so by first
defining two nodes anywhere within the model and then associate them to the two
ends of the desired section. In addition to the static analysis, the ASME code stress
evaluation option is also available with the Advanced Dynamic module for the time
history analysis.

(See
Example on the Use of Multiple Thermal Load Cases page
11-2.)
Thermal effects can be considered in all of the primary load cases and may be input
directly by the user as nodal temperatures or may be read from a prior heat transfer
analysis. In the following example the application of both types of thermal loadings
is illustrated.
Example TEMP1.GEO - Multiple Thermal Load Cases
The input for this problem is provided to you in file TEMP1.GEO. The model of an
axisymmetric nozzle is subjected to five different loading environments, each
represented by a load case as follows:
a. Load case 5: Figure 11-36. Axisymmetric Nozzle
Concentrated force at the tip
of the nozzle plus nodal
temperatures from step 6 of CL
a prior transient heat transfer 1"
analysis.
b. Load case 10: 2"
Nodal temperatures defined
at certain nodes. 2"
c. Load case 20: 2"
Nodal temperatures defined

4"
at certain nodes. 2 3 1"
d. Load case 30:

Nodal temperatures from 3" 2"
step 20 of the prior transient

heat transfer analysis.
5"
e. Load case 40:
Nodal temperatures from
step 10 of the prior transient
heat transfer analysis.

Figure 11-37. Nozzle Heated by Fluid
Temp
300
150
60
Time (HR)
0.02 0.12
0.1 0.2
Assuming the nozzle is heated by a fluid flow with temperature variation as shown
in the figure above, the first step is to solve the transient heat transfer problem using
HSTAR or FFE Thermal. To review the commands pertinent to thermal part of the
problem, refer to the TEMP1.GEO file. After you solve the thermal problem,
activate the thermal loading flag:
Geo Panel: Analysis > STATIC > Static Analysis Options
Loading flag > T
... ...
Accept Defaults
To define load case number 5:

Set label > LC
Load case set number > 5

(Activate load case 5)
Force label > FX: X force
Value > 100
Ending node > 101
Increment > 1
(Apply a nodal force of 100 at node 101)

Geo Panel: LoadsBC > LOAD OPTIONS > Read Temp as Load
Time step label > 6
(Assign temperature profile from time step 6 of heat transfer analysis
to load case 5)

Set label > LC

Geo Panel: LoadsBC > THERMAL > TEMPERATURE > Define Nodes
Value > 200
Ending node > 104
Increment > 1
(Define a temperature of 200 at nodes 101 through 104)

Set label > LC

Geo Panel: LoadsBC > THERMAL > TEMPERATURE > Define Nodes
Value > 400
Ending node > 104
Increment > 1
(Define a temperature of 400 at nodes 101 through 104)


Set label > LC

Time step label > 20
(Add the thermal effect from the heat transfer time step 20)

Time step label > 10
(To consider the thermal effect from the heat transfer time step 10)
Once all the load cases are defined, you may run the static analysis by issuing the
R_STATIC (Analysis > STATIC > Run Static Analysis) command. Postprocessing
will be available for all load cases as usual.


COSMOS C11 MODEL Y ANALISIS

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

COSMOS C11 MODEL Y ANALISIS

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

COSMOS C11 MODEL Y ANALISIS

Uploaded by

Copyright:

Available Formats

CHAPTER 11 Modeling and Analysis

COSMOS/M Basic FEA System • 12/97 11-1

Table 5-1. List of Modeling and Analysis Examples

11-2 COSMOS/M Basic FEA System • 12/97

Examples on Linear Static Stress Analysis

Example LGAP1 - Hertzian Contact Using Node-to-Node Linear (See

COSMOS/M Basic FEA System • 12/97 11-3

The file LGAP1.GFM includes information on the problem geometry such as

11-4 COSMOS/M Basic FEA System • 12/97

Figure 11-2. Curve, Contour and Region Plots of the

COSMOS/M Basic FEA System • 12/97 11-5

Figure 11-3. Finite Element Model of the

Figure 11-4. Enlarged View of the Contact Surface

11-6 COSMOS/M Basic FEA System • 12/97

COSMOS/M Basic FEA System • 12/97 11-7

LOAD CASE NUMBER. . . . . . . . . . . . = 1

FORCES IN THE CLOSED BOUNDARY/GAP ELEMENT(S), NCE = 3

NODE CSYS RFX RFY RFZ RMX RMY RMZ

Figure 11-5. Stress Contour

Hertz (1895) developed the mathematical theory supported by experimental results

11-8 COSMOS/M Basic FEA System • 12/97

Figure 11-6. Deformed Shape and Displacement Contour

COSMOS/M Basic FEA System • 12/97 11-9

Example BOND1 - Static Analysis of a U-Frame with Non-compatible

procedures for defining Applied

defining bonding surfaces,

11-10 COSMOS/M Basic FEA System • 12/97

COSMOS/M Basic FEA System • 12/97 11-11

Figure 11-8. Regions and Surfaces for Part Entities 1 and 2

Geo Panel: Control > SELECT > Initialize

Entity Name > Regions

Geo Panel: Control > SELECT > by Label

11-12 COSMOS/M Basic FEA System • 12/97

Ending Entity > 22

Geo Panel: Control > SELECT > by Label

Surface/Region Label > 1

The polyhedron is a hollow entity, and it needs to be converted to a solid volume by

COSMOS/M Basic FEA System • 12/97 11-13

Geo Panel: Control > SELECT > Initialize

Geo Panel: Control > SELECT > by Label

Geo Panel: Control > SELECT > by Label

Surface/Region Label > 22

Geo Panel: Geometry > Define Part

11-14 COSMOS/M Basic FEA System • 12/97

Number of elements on 1st curve > 3

Number of elements on 2nd curve > 10

Number of elements on 3rd curve > 4

Accept Defaults ...

Number of elements on 1st curve > 4

COSMOS/M Basic FEA System • 12/97 11-15

Element Set number > 2

11-16 COSMOS/M Basic FEA System • 12/97

Geo Panel: Meshing > AUTO MESH > Parts

Figure 11-9. Generated Finite Element Mesh with Non-compatible Interfaces

Enlarged view of mismatched

COSMOS/M Basic FEA System • 12/97 11-17

Figure 11-10. Surfaces at the Common Boundaries

From the above figure, you will find that: