Modeling Guide

Introduction
Modeling Guide
Overview of Model Generation
2
Overview of Model Generation
This document gives a general overview of the steps used to create a finite element analysis model. For
detailed information on commands, refer to the SimXpert Quick Reference. You may also refer to the
Example Problems or Workspace Guides for sample problems with step by step instructions.
Model Generation
Modeling Guide
Geometry Overview
4
Geometry Overview
Geometry for FEA is created for the purpose of describing the physical shape of the model to be analyzed.
It forms the base upon which the finite element mesh will be created.
Geometry can be created in SimXpert, imported from CAD, or directly accessed from CAD files with a
live link to the CAD code.
Units
SimXpert interprets all dimensions and input data with respect to a system of units. It is important to set
the appropriate units prior to importing any unitless analysis files (such as a Nastran Bulk Data file) or
creating materials, properties, or loads. You can control the system of units by selecting Units Manager
from the Options Editor in the Tools menu. If you import a file that contains units, SimXpert will
convert them into those specified in the Units Manager.
Create
Point
In addition to Geometric points, vertices on curves, surfaces, and solids are also considered points. You
can also create Mesh Control Points which mark locations where nodes will be placed when meshing is
performed.
Points can be suppressed so that a node will not be forced to be placed at that location when meshing.
3
5
Model Generation
Elements and Meshing
Elements and Meshing
This section describes the different element types supported in SimXpert, as well as the procedures used
to create these elements.
Check your Workspace Guide to determine which elements are supported for your analysis type.
Element Types
There are four basic element groups in SimXpert:
Scalar (0-D)
Line (1-D)
Surface (2-D)
Solid (3-D) elements.
There are also a number of specialty elements. A list of the elements currently supported by SimXpert,
including a brief description of each, follows.
Scalar (0D) Elements
Concentrated Mass - Three dimensional mass and/or inertia element located at a node. Used to
represent parts of a structure which contain mass but do not add stiffness.
Spring - Used to provide a specified stiffness between a nodal degree of freedom and ground.
Damper - Used to provide viscous damping between a nodal degree of freedom and ground.
Line (1D) Elements
Bar - Uniaxial element with tension, compression, torsion and bending. It is used to model
general beam/frame structures for linear analysis.
Beam - Uniaxial element with tension, compression, torsion and bending. It can be tapered and
have different properties along its axis. The shear center can be offset from the neutral axis.
Used to model both linear and nonlinear beam/frame structures and beams with open sections.
Curved Beam or Pipe - Curved beam, pipe, or elbow element with an arc for the neutral axis.
Useful for modeling piping systems.
Bushing - Vibration control device that has impedance values (stiffness and damping) that are
frequency dependent and they recommended as a replacement for spring elements.
Damper - Used to provide viscous damping between two selected degrees of freedom or
between one degree of freedom and ground.
Spring - Used to provide a specified stiffness between two selected degrees of freedom. Other
properties can include a damping coefficient and a stress coefficient to be used in stress
recovery. It is recommended to create this element between two coincident nodes.
Gap - Nonlinear element with different tension, compression and shear stiffness. It is used to
represent surfaces or points which can separate, close or slide relative to each other.
Modeling Guide
Elements and Meshing
6
Scalar Mass - Useful for the selective representation of inertia properties, such as occurs when a
concentrated mass is effectively isolated for motion in one direction only.
Rod - Uniaxial element with tension, compression and torsional stiffness. No bending or shear.
Generally, used to model trusses.
Tube - Rod element with tubular cross section which is often used to model pipes.
Plot Only - Non-structural element used to represent structural features that are not being
analyzed but aid in the visualization of the model.
Surface (2D) Elements
Plate/Shell - Two dimensional plate or shell element used to represent any three dimensional
structure. Resists membrane, shear and bending forces. Used to model structures comprised of
thin plate shells.
Tria - Three or six noded element (linear or quadratic element, respectively).
Quad - Four or eight noded element (linear or quadratic element, respectively).
Cohesive - Four or eight noded element used to simulate the process of delamination (linear
or quadratic element, respectively).
Axisymmetric - (Three or six) or (four or eight) noded element. The element can be linear or
fully nonlinear. Also, the element can be cohesive, for which it is either four or eight noded
(linear or quadratic element, respectively).
Shear Panel - Resists only shear forces. Used to model structures which contain very thin elastic
sheets, typically supported by stiffeners. This is a four noded element.
Acoustic Absorber - Defines a frequency dependent acoustic absorber element. Used in coupled
fluid structural analysis.
Solid (3D) Elements
Solid - Three dimensional solid element used to represent any three dimensional structure.
Specialty Elements
Rigid - Include both rigid and interpolation elements. General use rigid elements consist of
RBAR, RROD, RBE2, RBE1, and RSSCON. Of these elements, the RBAR and RBE2 are the
most commonly used. Interpolation elements consist of the RBE3 and RSPLINE. The RBE3 is a
linear interpolation element often used to distribute either loading or mass. It is can also be used
to interpolate the motion of a set of independent nodes and assign the motion from the
interpolation to a reference node (dependent node). The RSPLINE is often used to model mesh
transitions.
Weld - General purpose connector element which can connect non congruent meshes. Used in
linear analysis.
Contact - Used to identify bodies which are potential candidates for contact during analysis.
Bodies can be rigid or deformable. Self contact is also permitted.
7
Model Generation
Elements and Meshing
FAQs
1. Why should I use coincident nodes for spring elements?
In 1-D elements such as a bar, a positive axial force or stress indicates that the member is in tension. This
is not the case for scalar elements. Since the scalar elements do not have any geometry, the sign of the
output force or stress does not necessarily indicate compression or tension in the spring. The force in the
spring is recovered as follows:
Simply by reversing the order of the endpoints of the spring the sign of the force and stress output for
these elements reverse. For example, consider the spring shown in Figure 0-1 where D1 and D2 are scalar
displacements.
Figure 0-1 Spring Before and After Deformation
For D2 >D1,
If the spring is defined from node 1 to node 2 the axial force would be negative. If the spring is defined
from node 2 to node 1 the axial force would be positive.
Whenever a spring element is connected across two nodes, it is good modeling practice to make sure that
the nodes are coincident and use the same output coordinate system. If the nodes are not coincident, a
difference in displacement may be caused by a rigid body rotation. This situation causes forces in the
spring and a net moment on the structure. This condition is an example of an internal constraint, which
is allowed because it is valid for some types of modeling. Unless it is desired to have these internal
constraints, use coincident nodes. Since the length of the spring is zero when coincident nodes are used,
no internal moments are created. Also, the springs degrees of freedom are in the output coordinate
system.
2. Why is a bushing recommended as a replacement for spring elements?
As discussed in the previous question, if spring elements are used and the geometry is not aligned
properly, internal constraints may be induced. Bushing elements have all the features of the spring
F K U
1
U
2
( ) =
Before
Deformation
1 2
1
After Deformation
D2 D1
Modeling Guide
Elements and Meshing
8
elements without the drawback of the potential internal constraint. Furthermore, since a spring element
needs to be defined for each connected degree of freedom, several spring elements can be reduced to a
single bushing element.
Mesh Size Control
Mesh sizes can be set interactively using Seed in the Automesh group under the Meshing tab. It is
possible to specify a default mesh size for surfaces, or the size of an element at a vertex. For curves there
are several options for creating mesh seeds. The options are as follows:
Uniform: Causes elements to be distributed evenly along a curve.
One Way Bias: Specifies the mesh sizing for a given curve with an increasing or decreasing
element edge length.
Two Way Bias: Specifies the mesh sizing for a given curve with symmetric non-uniform
element edge lengths.
Curvature Based: Specifies the mesh sizing for a given curve with the element edge length
controlled by curvature of the curve.
Tabular: For specifying either parametric locations (place seeds at parametric locations of the
curve), arc length ratio locations (place seeds at arc length ratio locations of the curve arc), or
node/point locations (place seeds at the location of selected nodes or points) for mesh seeds.
A mesh seed consists of two adjacent control points. To create linear finite elements while meshing , a
mesh seed defines a single element edge. The two control points cause the creation of two finite element
nodes at opposite ends of the element edge.
When meshing to create parabolic finite elements (elements with mid-side nodes), two mesh seeds are
used to define one element edge. An example of this is a geometric solid that is seeded with Number of
Elements =10 and then is meshed with parabolic tetrahedral elements, five elements are created along
the seeded edge of the solid.
9
Model Generation
Elements and Meshing
In addition, it is possible to define mesh control points on curves or surfaces to ensure that a node is
placed at that location. This is done using the FEM Tool Control > Points/Curves in the Misc group
under the Meshing tab.
Mesh should have high density in areas of large stress gradients.
Nodes along common edges of adjoining geometry entities may need to match. If these nodes are not
coincident, equivalencing nodes will not cause all of the adjacent elements to become connected. If the
adjacent elements are not connected, the model will have free edges or free faces at these locations. For
a linear analysis, always merge coincident nodes before analyzing the model using Equivalence in the
Modify group under theNodes/Elements tab.
Element Creation
In SimXpert you can use a number of methods to create elements:
Automesh
Mesh Geometry: Used to create a set of elements on curves, surfaces, or solids.
Curves: Ageometric curve is defined parametrically (one parameter). The meshing is done in
parametric space using 1D elements.
Mesh seed on
curve, Number
of elements =10
Modeling Guide
Elements and Meshing
10
Surfaces: A geometric surface is defined either parametrically or non-parametrically. The
mesher either follows the parametric directions (surface defined parametrically), or initially
follows the perimeter of the surface, subsequently meshing inward following the inner most
perimeter defined by the created element edges (surface defined non-parametrically).
Solids: Controls automatic meshing of solid geometry or volumes (in space). The meshing can
be tetrahedral, or by sweeping, it can be hexahedral. Tetrahedral meshing is for a geometric
solid or a set of contiguous tria elements that completely enclose a volume. The meshing is
done, for a solid, by surface meshing its faces with tria elements, then creating one layer of
tetrahedral elements that are congruent to the tria elements. Then, another layer of tetrahedral
elements is created that is congruent to the first layer of tetrahedral elements. This is continued
until the solid is filled with tetrahedral elements. The meshing for a set of contiguous tria
elements is similar to that for a geometric solid, but the mesher begins with the set of tria
elements.
Hexahedral meshing is for a geometric solid, where the meshing is done by sweeping a set of
2D quadrilateral elements from one solid face (starting face) to the opposing face (ending
face). This means that the solid must be sweepable, or 2 1/2/ D. The starting and ending face
(opposing faces), are connected with a set of four edged faces (linking faces). Each of the four
edged faces must have two of its opposing edges congruent with an edge of the starting or
ending face. The mesher creates on one of the opposing faces a 2D quadrilateral element
mesh, then sweeps it to the other opposing face, creating a hexahedral element mesh between
the opposing faces.
Non-geometry meshing commands.
Skin - Creates shell elements on the free faces of 3D solid elements.
2-3-4 line mesh - Produces mesh between curves.
3-4 Point mesh - Produces mesh between virtual curves created by selecting end points.
On Mesh - Used to modify an existing mesh. Useful for fixing or cleaning-up a distorted
area of a mesh .
Create elements one at a time. Useful for simple models, line elements, and to fix areas of distorted
elements.
Copy existing elements using the Transform option of the Tools menu.
Project elements onto plane, mesh, curve, surface.
Reflect (mirror) elements through a plane.
Reorient - Move elements with respect to coordinate systems.
Scale - Create a scaled copy of the elements around a given location.
Rotate - Move elements a specified angle.
Translate - Move elements a specified distance.
Sweep Shells into Solids or Node pairs into Quads using the FEM based group under the Meshing tab.
Revolve - Revolve around a vector.
11
Model Generation
Elements and Meshing
Extrude - Drag along a vector.
Glide - Drag along a curve.
Loft - Create linear mesh between destination and source elements.
Normal - Drag elements along normals.
Flange on Mesh - Drag node pairs along a specified angle.
Surface Meshing Guidelines
Examine your model before you begin the meshing process to determine which method of mesh
generation is most applicable. Following are the guidelines for meshing surfaces.
For Surface from Automesh under the Meshing tab, there are several methods of meshing.
Auto Decide: Meshes the surface(s) and uses the meshing algorithm which gives the best
quality elements.
Paver: The mesh will be a paved mesh. The paving algorithm creates elements around any
edges and boundaries of the surface first, then fills with elements in the interior of the
surface, and finally smooths the mesh.
Mapped: The mesh will be a mapped (iso) mesh. Elements are created in rows and
columns using intersecting iso-lines.
Minimal: Using this will cause the creation of the smallest surface mesh (fewest number of
2D elements) possible, considering adjacent meshes and mesh seeds. Only one surface can
be meshed at a time for this option. Currently (2010), minimal meshing only works with
CATIA geometry (surfaces and curves).
Most meshes involve creating geometry first. If you have small features in your geometry that
are not critical to the analysis, you can suppress them in SimXpert.
You can use the FEM Based meshing Revolve / Extrude / Glide commands to generate Surface
elements from node pairs.
Utilize symmetry whenever possible to reduce model size. Model size (and therefore run time)
is significantly reduced if the loading/constraints are also symmetrical. If loads/constraints are
not symmetrical, you can use the Tools >Transform >Reflect command to mirror the mesh
through a plane.
Remember, you may also want to use Translate from the Element group of the Nodes/Elements
tab or one of the commands under Tools >Transform to replicate elements instead of
performing more surface or boundary meshing.
Use the Equivalence command in the Modify group of the Nodes/Elements tab to merge
coincident nodes and connect the meshes.
Use the View, Highlight FE Boundary command to verify that you do not have any unwanted
free edges in the model.
Modeling Guide
Elements and Meshing
12
Solid Meshing Guidelines
Often you can avoid using volumes or solids by extruding or revolving planar elements to create
solid elements. If your part has a consistent third dimension, use one of the Fem based sweep
methods under the Meshing tab to create solid elements.
Use Skin from the Automesh group if planar elements are required on faces of solid elements.
Once the solid elements are created, planar elements can be created by skinning.
If you have solid models with holes or other complicated features, use the solid mesher. This
mesher creates a surface mesh first, so all items applicable to surface meshing apply.
You can associate a property with the mesh as it is being created. Properties can also be assigned to parts
or to parent geometry.
Minimal Meshing
As previously mentioned, if minimal meshing is performed the smallest surface mesh (fewest number of
2D elements) will be created for the surface (only one surface can be meshed at a time for this option.).
The mesh created will be congruent to adjacent meshes and mesh seeds.
Also, it is possible to simultaneously mesh curves on the perimeter, and interior to the perimeter, of the
surface that is being minimally meshed. For this to occur, it is necessary to associate the curves to the
surface. The meshes created for associated curves will be congruent to the mesh created for the surface.
These two things are requested using the surface meshing form, Surface from Automesh under the
Meshing tab.
Curve minimal meshing -- specify what type of curves (curves at the perimeter of the surface and
hard curves internal to the perimeter) are to be meshed.
None: No curves will be meshed. Only the surface will be meshed, and it will be meshed
minimally.
Boundary curves: Curves at the perimeter of the surface are meshed to be congruent with the
surface mesh that is created.
Boundary curves and hard curves: Curves at the perimeter of the surface and the hard curves
internal to the perimeter are meshed to be congruent with the surface mesh that is created.
Hard curve association -- specify how the curves will be selected for association to the surface.
None: Do not associate any curves to the surface. Any already associated curves will be
recognized (used) by the surface mesher.
Auto: Automatically associate curves on the surface (if the distance between a curve and the
surface is less than the tolerance, the curve is considered as being on the surface) to the
surface.
Selected: Set the form for selecting what curves are to be associated to the surface.
Curves to mesh: Select the curves the are to be associated to the surface.
Normally, a surface to be minimal meshed is four-sided without any holes.
13
Model Generation
Elements and Meshing
For hard curves, they need to divide the surface being meshed into a number of patch shapes. Each patch
shape should be a three or four sided region. If seeds or existing nodes are defined for the hard curves,
the numbers of seeds or nodes on the opposite sides of each patch should be equal. A node will be created
at the intersection of two hard curves. There can be problems if the two ends of a hard curve do not touch
the perimeter of the surface -- the ends of a hard curve are outside or inside of the surface perimeter.
Currently (2010), minimal meshing only works with CATIA geometry -- all the surfaces and curves
being minimally meshed must be of CATIA type.
A typical scenario is for the aerospace industry, that involves meshing a fuselage or wing model. Here,
a single (large) surface represents the entire top skin of the wing. Similarly, a single surface represents
the entire bottom skin of the wing. Between the two skins are individual ribs and spars (represented by
curves) that act as stiffeners.
When meshing the top or bottom surface (skin) of the wing, usually it is desired to have only a single
element span the surface (skin) between the stiffener locations (at each skin and rib, or skin and spar
sections). In such cases, it is important that only one element be generated on the surface (skin) between
the stiffeners. The minimal mesh option satisfies this requirement.
Users want to create a minimal mesh on a model with a single tool and the least possible number of
picking and selecting.
Element Shape Quality
Once you have created a surface mesh or solid mesh, check the quality of the elements using the tools
under the Quality tab. This can be done using the Plots, Checks, or Edit/Fix Element groups (e.g. select
Quality Index from the Plots group under the Quality tab).
Using the methods from the Plots group causes the elements in the model canvas to be colored depending
on the method used and the quality of the elements. The methods that can be used are described as
follows:
Quality Index: Color-codes the model elements based on a weighted composite (weighted sum)
of all the selected quality checks results. The quality checks tests that are to be used are specified
in the form which is accessed by selecting Check Quality from the Checks group under the
Quality tab. The name of the form is ShowMeshQuality.
Modeling Guide
Elements and Meshing
14
To use this form specify 1) Application Region (Model, Scene, Select Elements), 2)
entity/element types (i.e. Geometry, Rigid Elem, 2D Elem, 3D Elem), and 3) the Values (For
example, Geomerty Verification Values, Quality Values) that are to be used (that is select All,
Min. Edge Length, Min. Surface Area, Aspect Ratio, J acobian Ratio). The Values that are
selected will cause the corresponding test to be performed when performing the Quality Index
test (select Quality Index from the Plots group under the Quality tab). The Quality Index is
calculated as the normalized sum of the product of the Quality Results (quality test results for
Quality Values) and the Weight factor; Quality Index =normalized E ((activated Quality
Results) * (Weight factor)).
Aspect Ratio: The aspect ratio is generally the ratio of the length to the width of an element.
The ratio is calculated differently depending on the type of element.
Tria:
This is the same as the Patran evaluator. Let V1, V2 and V3 be three vertices of a triangle, and
M1, M2 and M3 be the bisectors of three edges on the triangle. Let Edge 1 be the edge from
V1 to V2, and Segm 1 be the line segment from V3 to M1. First, we calculate two intermediate
aspect ratios on Edge1.
The first intermediate aspect ratio is the ratio of length of Edge 1, l1, to the height, h1. The
ratio is then multiplied by (Sin 60
0
) such that a perfect element in the shape of an
equilateral triangle will equal one. The ratio is inverted if it is less than one.
The second intermediate aspect ratio is the ratio of the distance from M3 to the Segm 1, h2, to
the length of Segm 1, l2. The ratio is then multiplied by 4 * . It is inverted if it is less
than one.
The aspect ratio on Edge 1 is the maximal value of these two intermediate ratios.
3 2
3 2
15
Model Generation
Elements and Meshing
This procedure is repeated for the remaining two edges of the triangle, and the largest value
is retained as the aspect ratio for the triangle.
If Normalize is selected on the verification form, then the aspect ratio is inverted such that it
becomes less than or equal to one. This inverted aspect ratio is subtracted from one to yield
the aspect factor. An equilateral triangle will have an aspect factor of 0.
Quad:
This is the same as the Patran evaluator. The aspect ratio for a quad is derived from one test
proposed by Robinson and Haggenmacher. This test is based on projection plane created by
first bisecting the four element edges, creating a point on the plane at the vector average of
the corners. The x-axis extends from the point to the bisector on Edge 2. The ratio is
determined as the ratio of the length from the origin to the bisector of Edge 2 and the length
from the origin to the bisector of Edge 3. If the ratio is less than 1.0, it is inverted.
If Normalize is selected on the verification form, then the aspect ratio is inverted such that it
becomes less than or equal to one. This inverted aspect ratio is subtracted from one to yield
the normalized aspect ratio. A square element will have a normalized aspect ratio of 0.
Modeling Guide
Elements and Meshing
16
Tet:
This is the same as the Nastran evaluator. The aspect ratio for a tet element is computed by
taking the ratio of the height of a vertex to the square root of the area of the opposing face.
This value is then manipulated in one of two ways, depending on whether the Normalize
parameter is selected on the verification form.
If Normalize is NOT selected, the maximum height to area value is multiplied by a factor
, which is the ratio of height to edge length for an equilateral tetrahedron. This
result is reported as the Aspect Ratio. An equilateral tet will report a value of 1.
Aspect Ratio = .
If Normalize IS selected, the maximum height to area value is inverted and subtracted from 1.
Aspect Factor =
Hex:
This is the same as the Nastran evaluator. The aspect ratio is calculated as the ratio of the
distance between opposing faces. This distance is determined by treating each HEX face as if
it were a warped quadrilateral. Each face is processed to produce a projected plane. The
distances between the centerpoints of all three pairs of opposing faces are compared.The
aspect ratio is determined by taking the maximum distance between any two faces and
dividing it by the minimum distance between any two faces.
If Normalize is selected on the verification form, then the aspect ratio is inverted such that it
becomes less than or equal to one. This inverted aspect ratio is subtracted from one to yield
the aspect factor. A cubic element will have an aspect factor of 0.
C 0.805927 =
Max C
f
h
i
A
i
( ), i =1, 2, 3, 4
1 1 MaxC
f
h
i
A
i
( ) ), i =1, 2, 3, 4
17
Model Generation
Elements and Meshing
Wedge:
This is the same as the Nastran evaluator. Natran averages the two triangular faces of the
wedge element to obtain a mid-surface. The aspect ratio of this triangular mid-surface is
computed . Next the height (h
1
) of the wedge is compared to the maximum edge
length of the mid-surface (h
4
).
If the height of the wedge is greater than the maximum edge length then the aspect ratio for
the wedge element equals the mid-surface aspect ratio multiplied by the maximum edge
length divided by the distance between the triangular faces (h
3
).
If the height of the wedge is less than the maximum edge length then the aspect ratio for the
wedge element equals either the mid-surface aspect ratio or the maximum edge length divided
by the distance between the triangular faces, whichever is greater.
If Normalize is selected on the verification form, then the aspect ratio is inverted such that it
becomes less than or equal to one. This inverted aspect ratio is subtracted from one to yield
the aspect factor. An equilateral wedge element will have an aspect factor of 0.
3h
2
2h
1
( )
Modeling Guide
Elements and Meshing
18
J acobian Ratio: This is approximately the ratio of the maximum value of | J | to the minimum
value of | J | , where | J | is the determinant of the J acobian matrix, [ J (r,s,t) ] , of an element.
Tria, Quad, Tet, Hex, Wedge -
The J acobian Ratio for all element topologies is computed as follows:
J
max
and J
min
are respectively the maximum and minimum values of the determinant of the
J acobian matrix, [ J ] , computed at the integration points. The J acobian ratio is more
complicated than indicated by the above equation for the J acobian Ratio. Hence, the
approximately =symbol is used. The J acobian matrix, [ J ] , for a 2D element is the 2x2 matrix
in the following equation relating the partial derivatives of the shape function, N, with respect
to the element parametric coordinates c and q, to the derivatives of the shape function with
respect to the element cartesian coordinates x and y.
A similar relationship exists for a 3D element.
For J acobian Minimum:
Tria:
See note for Tet element.
Jacobian Ratio J
max
J
min
~
cN cc
cN cq
cx cc cy cc
cx cq cy cq
cN cx
cN cy
=
19
Model Generation
Elements and Meshing
Quad:
See note for Tet element.
Tet:
Relies upon same general J acobian evaluator used in Patran and defined in the following
book:
O.C. Zienkiewicz, "The Finite Element Method" McGraw-Hill, 1985. 3rd edition. Chapter
8 : Curved, Isoparametric Elements and Numerical Integration. Section 8.5 : "Evaluation of
Element Matrices (Transformation in Xi, Eta, Zeta Coordinates). Pages 189-192.
Minimum value of the J acobian is for both the Gauss-Legendre quadrature integration points
and nodal positions.
Wedge:
See note for Tet element.
Hex:
See note for Tet element.
Skew - this is a measure of how much an element or element face is skewed. A skewed element
or element face looks like it has been sheared.
Tria -
This is the same as the Sofy evaluator. It is calculated as the minimum of the angles formed
by each of the vertices with the mid-points of its opposite edge with that edge.
Quad -
The skew for a quadrilateral element is computed as follows:
Skew =90 - o
If o <30
o
a skew message is issued.
o is an acute angle formed by two intersecting lines, with each line connecting the mid-point
of the opposite edges. o is the same as obtained by the Nastran evaluator.
Tet -
There is no corresponding evaluator in SimXpert.
Hex -
There is no corresponding evaluator in SimXpert.
o
Modeling Guide
Elements and Meshing
20
Wedge:
There is no corresponding evaluator in SimXpert.
Taper: This is a measure of how much an element or element face is tapered.
Tria:
There is no corresponding evaluator in SimXpert.
Quad: The taper ratio for a quadralateral element is computed as follows:
This is the same as the Nastran evaluator. The taper ratio test compares the area of the triangles
formed at the four corner nodes (a triangle is formed from a corner node and the two nodes
attached to it via the two element edges connected to the corner node) to the total area of the
element. The maximum of the four values is used as the taper ratio.
The first two triangles, and corresponding areas, are formed from Node 1 and 3. For example,
Triangle 1: Node 1, 2, 4; with area A
1.
The second two triangles, and corresponding areas, are formed from Node 2 and 4.
Taper (i) =| A
i
/ (A
quad element
/ 2) - 1

| , i =1,2,3,4.
A
i
is the i-th area (A
1
, A
2
, A
3
, or A
4)
; and A
quad element
is the total area of the quadralateral
element.
The maximum of the four values is used as the taper ratio.
1 2
3 4
A
1
A
3
1 2
3 4
A
2
A
4
21
Model Generation
Elements and Meshing
Tet -
There is no corresponding evaluator in SimXpert.
Hex -
There is no corresponding evaluator in SimXpert.
Wedge -
There is no corresponding evaluator in SimXpert.
Warpage - this is a measure of how much an element or element face is out of plane.
Tria -
There is no corresponding evaluator in SimXpert.
Quad -
There is a distinction between: Warp Angle and Warp Factor.
Warp Factor:
This is the same as the Nastran evaluator. This test, for a quadralateral element, evaluates how
far out of plane the four corner node points are by determining the distance of each node point
from a "mean" plane passing near the locations of the four node points.
The corner node points are assumed to be, alternately, H units above, then H units below the
mean plane.
Modeling Guide
Elements and Meshing
22
If the lengths of the diagonals of the element are denoted by D
1
and D
2
, the warping
coefficient is obtained from the equation WF =H / ((D
1
+D
2
) / 2). If this value exceeds the
tolerance, an informational message is produced.
Warp Angle:
The highest angle between the normals of triangular areas of the Quad as divided by a diagonal
of opposing corner nodes. This is the same as the sofy evaluator.
Tet:
There is no corresponding evaluator in SimXpert.
Hex:
There is no corresponding evaluator in SimXpert.
Wedge:
There is no corresponding evaluator in SimXpert.
Length (W) - this color-codes the element edges based on edge length.
Min Length =minimum edge length of any side (neglect mid side node for higher-order
elements).
Max Length =maximum edge length of any side (neglect mid side node for higher-order
elements).
Length (Max) - this color-codes the element edges based on maximum edge length.
Max Length =maximum edge length of any side (neglect mid side node for higher-order
elements).
Length (Min) - this color-codes the element edges based on minimum edge length.
Min Length =minimum edge length of any side (neglect mid side node for higher-order
elements).
Quad Min Angle - color-codes based on the minimum internal angle for quadralateral elements.
H
H
H
H
23
Model Generation
Elements and Meshing
Angles are computed for each corner (neglecting mid side nodes of higher order elements)
from the edge vectors meeting at each corner (vertex of face).
Quad Max Angle - color-codes based on the maximum internal angle for quadralateral elements.
Angles are computed for each corner (neglecting mid side nodes of higher order elements)
from the edge vectors meeting at each corner (vertex of face).
Tri MinAngle - color-codes based on the minimum internal angle for triangular elements.
Angles are computed for each corner (neglecting mid side nodes of higher order elements)
from the edge vectors meeting at each corner (vertex of face).
Tri Max Angle - color-codes based on the maximum internal angle for triangular elements.
Angles are computed for each corner (neglecting mid side nodes of higher order elements)
from the edge vectors meeting at each corner (vertex of face).
Weld Norm - color codes based on the weld connection normality. Weld refers to a 2 noded
element (1D element) that is connected at each of its ends (nodes) to a shell element. The weld
connection normality is defined as the angle between the axis of the 1D element and the surface
of the shell elements. If the weld connection normality is less than its ideal value of 90, the
value of the Plate RZ Stiffness Factor may have to be adjusted. This is generally considered to
be a bad modeling practice.
Chordal Deviation - color-codes based on the maximum distance between the center of the
element edge and a true arc.
The deviation is measured at the center of the element edge
Dependency Loops - displays the number of rigid loops in the model in the message region.
The message region displays the quality check results.
Using the methods from the Checks group causes information about the quality of the model elements
to be displayed in forms and/or the message window. The methods that can be used are described as
follows:
Check Quality - this form (ShowMeshQuality) is used to specify the parameter values that
define elements of various quality (i.e. Good, Failing), and display the quality of the elements.
The form is displayed using Check Quality from the Check group under the Quality tab. An
application region (domain containing elements whos quailty is to be checked/verified) must be
specified. There are three choices for this; they are Model (a model consisting of all the
elements), Scene (all the elements in a particular scene), and Select Elements.
After this is done, select the tabs that correspond to the quality tests that needs to be performed
(i.e. Geometry, 2D Elem, Elem Length). Shown below is the form for 2D Elem.
Modeling Guide
Elements and Meshing
24
In the first column are the names of the types of tests that can be selected (i.e. Warp Angle Quad,
Warp Factor Quad, Aspect Ratio, J acobian Ratio). The tests that are to be performed must be
selected using the checkbox. The second column (Ideal) has the assigned ideal value for each test
(i.e. Warp Angle Quad =0.0, Warp Factor Quad=0.0, Aspect Ratio=1.0, J acobian Ratio=1.0).
25
Model Generation
Elements and Meshing
The remaining columns; Good, Failing, Weight, and Good Rating; are for specifying what the
quality of an element must be for it to be a good element, a failed element, how much a quality
test is weighted (for the weighted sum of the Quality Values), and what a good rating is for a
given quality test.
After the Quality Values are entered into the form, click the Results >> button to cause the form
to change to Quality Results so the results can be calculated. Now, click the Apply button, then
click the << Values button. The form cells will be populated with the number of elements that
correspond to each quality test.
Modeling Guide
Elements and Meshing
26
For example, for the Taper Quad test out of 95 elements 80 were good, and 15 were poor. Also,
quality results are displayed in the history window.
27
Model Generation
Elements and Meshing
More information can be obtained by right-clicking in a cell for a quality test (e.g. Taper Quad)
for Good elements, then selecting, among several options, Highlight.
Modeling Guide
Elements and Meshing
28
The elements in the canvas that are of good taper quad quality are highlighted. It is easy to
identify the 15 elements that are of poor taper quad quality.
Aspect Ratio - calculate the aspect ratio for 2D or 3D elements. This is done using Aspect Ratio
from the Checks group under the Quality tab. The results are displayed in the history window.
29
Model Generation
Elements and Meshing
Tri int angle - calculate the minimum internal angle for triangular elements. This is done using
Tri int angle from the Checks group under the Quality tab. The results are displayed in the
history window.
Quad int angle - calculate the minimum internal angle for quadralateral elements. This is done
using Quad int angle from the Checks group under the Quality tab. The results are displayed in
the history window.
J acobian - calculate the J acobian for any type of element. This is done using J acobian from the
Checks group under the Quality tab. The results are displayed in the history window.
Length - calculate the minimum and maximum length for the set of all the elements. This is done
using Length from the Checks group under the Quality tab. The results are displayed in the
history window.
It may be necessary to set the values for ideal, good, and failing in the ShowMeshQuality / Elem
Length form.
Skew - calculate the skew for the elements. This is done using Skew from the Checks group
under the Quality tab. The results are displayed in the history window.
Taper - calculate the taper for the elements. This is done using Taper from the Checks group
under the Quality tab. The results are displayed in the history window.
Warpage - calculate the warp for the elements. This is done using Warpage from the Checks
group under the Quality tab. The results are displayed in the history window.
Modeling Guide
Elements and Meshing
30
Angle Degen - calculates all the angular measures (e.g. Warp Angle Quad) for the elements, then
compare the results with the Angle Degeneracy value set in the ShowMeshQuality form. This is
done using Angle Degen from the Checks group under the Quality tab. The results are displayed
in the history window.
Length Degen - calculates all the length measures (e.g. Min. Length) for the elements, then
compare the results with the Length Degeneracy value set in the ShowMeshQuality form. This is
done using Length Degen from the Checks group under the Quality tab. The results are displayed
in the history window.
Connectivity - determines if there are badly connected elements. The results are displayed in the
history window.
Dependency - determines if there are any elements with bad dependency. The results are
displayed in the history window.
Duplicates - determines if there are duplicate elements. An element is a duplicate element if it
uses all the nodes that another element is using. A message is shown in the history window. Also,
the duplicate elements are hightlighted in the model canvas.
31
Model Generation
Elements and Meshing
Free - determines if there are free elements. An element is free if it is not connected to other
elements, even if is coincident with another element. A message is shown in the history window.
Also, the free elements are hightlighted in the model canvas.
Nastran GeomCheck - using this causes the quality of the elements to be determined using the
Nastran geometry checker code. The results are displayed in the history window.
Tet10 Linearization - linearization of tet10 elements causes the elements midside nodes to be
moved so the corresponding edges become straight. The elements are still tet10s, but their
edges are straight as for tet4 elements. The tet10 linearization occurs only for the tet10 elements
that fail the J acobian tests (J acobian Minimum and J acobian Ratio).
Miscellaneous - using this causes the results of various checks to be displayed in a window, and
the SimXpert history window.
Modeling Guide
Elements and Meshing
32
33
Model Generation
Elements and Meshing
% Tria - using this causes the percent of triangular elements of the total number of elements to
be calculated, and the quailty of the triangular elements to be determined. For this example there
are 16 triangular elements of a total of 103 shell (quad and tria) elements.
Show - display, or hide, elements that have quality within a specified range.
Specify the minimum and maximum quality for the range. Specify if the elements are to be
displayed or hidden. Click Apply. Show the element labels.
Modeling Guide
Elements and Meshing
34
As can be seen, for this example, the elements with quality from 51 to 100 are displayed, and the
elements with quality less than 51 are not displayed.
Geom Assoc - using this causes a check to determine if some elements are not associated to
geometry, or some elements nodes are not asscociated to geometry descendants.
Specify how the elements are to be selected under Element Selection Method; 1) Direct - select
by elements, or 2) By Geometry Association - select by geometry, i.e. curve, surface. Select
elements. Click Apply. Observe any message in the history window.
Find Faces - highlights the faces of solid elements that are on the faces of geometric solids.
35
Model Generation
Elements and Meshing
Select the faces of geometric solids. Click Apply. Observe the highlighted solid element faces.
Using the methods from the Edit/Fix Elements group causes the quality of the elements to be improved.
The methods that can be used are described as follows:
Fix Elements
Normals - this form allows the user to perform various operations involving element normals.
The options are as follows:
Action Type: Display
Show the element normals
Modeling Guide
Elements and Meshing
36
Hide the element normals
Action Type: Edit
Fix normals using reference element
Reference element normal is up, and the other normals are down.
Click Apply.
37
Model Generation
Elements and Meshing
Reverse normals of set of elements
Enhance Quality by Smoothing - this form allows the user to increase the quality of the elements
by performing a smoothing operation.
Modeling Guide
Elements and Meshing
38
Fix Angle Degeneracy - this form allows the user to reduce the element angle degeneracy.
Fix Length Degeneracy - this form allows the user to reduce the element length degeneracy.
Manual - this form allows the user to move the position of a node manually.
39
Model Generation
Elements and Meshing
1. Under Move Node Along select how you would like to move the node:
Element: move the node within a plane of connected elements.
Node Normal: move the node along a vector that is the average of connected element normals.
If desired, enter the maximum allowable distance to move the node in the Max. normal length
text box.
Element Surface: move the node within a virtual surface that is fit thought pertinent elements.
2. Click in the Node text box and select the node you would like to drag.
3. Drag to reposition the node.
4. Click the middle mouse button (MMB) to finalize the node position.
Notes and Tips
1. You may only move a node within the outside boundary surrounding all of its adjacent elements.
2. By default, nodes cannot be dragged across model features. Feature lines can be displayed by
selecting View >Highlight Boundaries. If you would like to drag nodes across model features
you can toggle from Fix Geometry to Change Geometry by using the pull-down menu for the
Element Fringes icon on the Element Render toolbar. Feature Angle can be set using User
Options (Tools >Options >Workspaces >Structures >Entity Options >Element.Structures).
3. The selected node will follow the cursor location until you click the middle mouse button.
4. The Location entry will only change as the position of the node is changed by moving the mouse
cuursor. It is not possible to manually change the Location entry.
5. This command works well when elements are color-coded based on a selected quality parameter.
As you drag the node, the element color will update in real time to reflect the current element
quality. You can select a specific quality parameter or a weighted composite of quality checks to
color-code your model using a selection (i.e. Quality Index (Plots-Quality Index) , Aspect Ratio,
J acobian) from the Plots group under the Quality tab.
Before moving node.
Modeling Guide
Elements and Meshing
40
Moving the node.
41
Model Generation
Elements and Meshing
Detach Element from Existing Nodes - this form allows the user to cause new nodes to be used
to define a set of elements (the new nodes are not used by elements outside the set). Thus, the
elements are detached from a mesh. It can be verified that the elements are not connected to a
mesh by using Free from the Checks group under the Quality tab.
Misc.
Internal Angle - this form allows the user to create a fringe plot of element internal angles.
Shell Thickness - this form allows the user to create a fringe plot of element thickness.
Modeling Guide
Elements and Meshing
42
Laminate Verification - verifies laminate composites.
Weld Length - shows messages about weld length.
Weld Normal - shows messages about weld normal.
All By Element - shows the quality of a single element.
43
Model Generation
Elements and Meshing
QCHECK (shell only) - quick checks for only shell elements.
Note: you can set the maximum distortion criteria using Check Quality from the Checks group under the
Quality tab. Fix all distorted elements if possible before adding any loads or constraints. This is
especially important if the distorted elements are in a critical region of the model.
Modeling Guide
Elements and Meshing
44
Improve Quality by Parts or Sets of Elements
Improves the quality of elements. The element quality is improved by remeshing the domain of the
elements which have poor quality. Use Repair in the Edit group under the Meshing tab. The parameters
that are used are
Type: Part (elements associated to the Part) or Elements (set of elements)
Parts: Select parts that have associated meshes that need to be replaced by remeshing the domain
they are in. The meshes may have some elements of poor quality that may need to be replaced.
Elements: Select the elements that are to be replaced by remeshing the domain they are in.
Remesh Triangles: If checked, the current triangular elements are to be replaced either by quad
elements if possible, or by better quality tria elements.
Remesh Hole Rings: If checked, the elements around holes are replaced by better quality
elements.
Split
Splits selected elements into multiple elements, thus possibly creating a better quality mesh. Use Split in
the Edit group under the Meshing tab. The parameters that are used are:
Split: Elements (splits the selected elements along either direction or both, depending on the
Pattern selected) or Along Line (splits elements along a line you define).
Option: On Element (splits the elements on the element surface) or On Surface(forces split
elements to adhere to the original geometric surface).
Elements: Select the elements that are to be split.
Mesh Transition - If checked, the adjoining elements are split to keep the mesh congruent.
Pattern Type: Select the pattern the elements that are created by splitting are to have. This is only
used for Split: Elements.
Pattern (button) - Click this button to see what the elements will look like after the splitting.
Find and Fix Cracks
Determines if a shell mesh is not continuous for all elements in the current scene and highlights problem
elements. Optionally, merges nodes or splits elements to fix cracks. Use Cracks in the Edit group under
the Meshing tab. The parameters that are used are:
Maximum edge distance: Maximum boundary/edge-to-boundary/edge distance. If the (actual)
distance is less than this parameter value, the boundary/edge-to-boundary/edge is not considered
to be a crack, otherwise it is a crack.
Maximum edge angle: Two boundary edges must be reasonably parallel to be considered to be a
crack. If the angle between two adjacent edges is greater than the Maximum edge angle value,
the space between the edges is not considered to be a crack.
45
Model Generation
Elements and Meshing
Fix Cracks: If checked, merge nodes or split elements to fix cracks; or if unchecked, cracks in
the model are highlighted, but not fixed.
Auto Geometry Cleanup
The Auto cleanup command in the Misc group under the Meshing tab is used to suppress curves at the
interface between two adjacent surfaces. The curves are suppressed so the surface mesher will not
automatically create nodes along them. The surfaces must be three or four sided. The parameters that are
used are:
Surfaces: Select surfaces, some of whose curves are to be suppressed.
Element Size: Specify the size of the finite elements that are to be created during meshing.
Surface cross section length tolerance: For two congruent, adjacent surfaces, each with four
edges, SimXpert will determine the smallest edge length for the four edges emanating from the
interface, L. If L is less than the specified tolerance, the curve along the interface of the two
surfaces will be suppressed.
Small surface area tolerance: For two congruent, adjacent surfaces, each with four edges,
SimXpert will determine the area of both surfaces. The smallest area, S, will be determined. If S
is less than the specified tolerance, the curve along the interface of the two surfaces will be
suppressed.
Short curve length tolerance: If the length of curves associated to the selected surfaces is less
than the specified tolerance, they will be suppressed.
Small angle tolerance: For each surface in the list, SimXpert will determine the smallest interior
angle, A. If A is less than the specified tolerance, the shortest edge at the vertex of each angle
will be suppressed.
Break Angle: For two congruent, adjacent surfaces, each with four edges, SimXpert will
determine the intersection angle, A. If A is less than the specified Break Angle, the curve along
the interface of the two surfaces will be suppressed.
Lock / Unlock
The Control >Lock / Unlock command in the Misc group under the Meshing tab is used to prevent
geometric edges/curves from being modified. For example, if a curve is locked then selected for
suppression, it cannot be suppressed. Subsequently, if the same curve is unlocked, it can be suppressed.
Collapse / Expand Edges
The Edit Edges command in the Edit group under the Meshing tab is used to
Merge two nodes that define an edge or
Expand (split) along the edge defined by the two nodes. The method to use is Method:
Collapse or Expand. The parameters that are used are:
Modeling Guide
Elements and Meshing
46
Nodes/Edge: Select two nodes to define an edge or select an edge. The nodes can be coincident.
The nodes do not have to belong to the same finite element edge.
Method:
Collapse: Collapses the edge formed by the two nodes.
Expand: Ssplits the edge formed by the two nodes.
Propagate: If checked, extends the collapse/expansion along all connected elements until the end
of the mesh path or a feature boundary is reached.
Force Editing: If checked, if a node that is associated with a hard point (vertex) can be moved,
otherwise the node cannot be moved.
Creating a Surface Mesh From an Existing Surface Mesh
The On Mesh command in the Automesh group under the Meshing tab is used to create a surface mesh
from an existing surface mesh. The parameters that are used are:
Elements: select the set of surface elements that are to be re-meshed.
Element Size: edge length of the new elements.
Mesh Type: Mixed (Quad and Tria), Tria only.
Mesh Method: Paving, Mapped.
Seed Type:
Uniform: The mesher will create new boundary nodes based on input global edge length.
Existing Boundary: All boundary edges on input mesh will be preserved.
Defined Boundary: The mesher will use all the nodes selected in the Feature Selection sub-
form to define the boundary of the output mesh. No other boundary nodes will be created.
Delete Input Mesh: Option to delete original elements.
Element property: Option to specify the property to be associated to the new elements.
Add to part: Option to associate the new mesh to a desired part.
Curvature check: Option to adjust the mesh density to control the deviation between the input
mesh and the straight element edges on the output mesh.
Feature recognition: Option to define features on the input mesh automatically based on feature
edge angles and vertex angles, and preserve them on the output mesh.
Feature selection: Allows for manual selection of entities to be preserved in the new mesh.
Washers around holes: Used to create layers of elements that are well shaped around holes.
Remesh Elements Retaining Features
The Remesh command in the Edit group under the Meshing tab is used to create a new mesh on an
existing shell mesh without removing features, hard points, or connected rigid elements. The parameters
that are used are:
47
Model Generation
Elements and Meshing
Elements: Select those elements to be re-meshed.
Feature angle: Specify the angle between adjacent element normals that if exceeded will define a
feature line.
Free edge size: Edge length for free edges for those elements with free edges.
Keep better quality only: Option to keep the new elements only if they are of better quality
otherwise the older mesh will be retained.
Automatically Remesh CAD Geometry and Update LBCs
When geometry is imported into SimXpert, the user will want to mesh it and possibly create LBCs that
are associated to it. After this is done, the geometry may be changed. If the geometry is modified in a
CAD program, then imported into a SimXpert database, it will be necessary to remesh the geometry and
create new LBCs. Doing this manually could involve a considerable amount of work. There are several
ways to deal with this task. One way is to use SimXperts capability of detecting changes in the geometry,
then automatically re-meshing as shown in the examples below.
Follow these steps to update SimXpert automatically on your CAD model:
1. Execute SimXpert.
2. Import CATIA geometry.
3. Mesh the geometry and create any LBCs, associating them to the geometry.
4. Exit SimXpert.
5. Execute CATIA.
6. Open the geometry in CATIA.
7. Make changes to the geometry.
8. Save the changes to the CATIA database, then exit CATIA.
9. Execute SimXpert, then open the database corresponding to the CATIA geometry.
10. SimXpert will detect that the geometry is different and will automatically remesh changed
regions. Also, the corresponding LBCs will be updated.
Possible variations to the above steps:
Variation 1
1. Mesh geometry in SimXpert, then exit SimXpert.
2. Open the geometry in CATIA.
3. Create new geometry in CATIA, then exit.
4. Open the geometry (original and new) in SimXpert.
5. In SimXpert the original geometry will retain its mesh and the new geometry will be part of the
database.
Variation 2
Modeling Guide
Elements and Meshing
48
1. Mesh some of the geometry in SimXpert, then exit SimXpert.
2. Open the geometry in CATIA.
3. Delete the geometry that was meshed in SimXpert, then exit CATIA.
4. Open the geometry (original minus deleted) in SimXpert.
5. The geometry that was deleted in CATIA is gone, and the mesh that was associated to the deleted
geometry is gone.
The second way to do this is to use Smart Update. File >Smart Update is used to access geometry that
is currently being used in a CAD program. Any changes to the geometry in the CAD program will be
indicated to SimXpert. If the geometry was previously meshed and had LBCs assigned to it, it will be re-
meshed and the LBCs modified. Using this approach requires both SimXpert and the CAD program to
be executing simultaneously.
Element Modification
To modify an element you can select one or more elements from the window and click Edit >Properties
from the main menu.
Click Edit >Properties.
Selected
elements
49
Model Generation
Elements and Meshing
The form for modifying the element properties of the selected elements will appear.
Selecting a multiple number of elements to modify their properties will display only common properties,
and the modified properties will be applied to all the selected elements.
Instead of selecting Edit >Properties from the top menu, you can right-click the selected elements, then
select Properties form the drop-down menu.
Modeling Guide
Connections
50
Connections
Connections can be created in the Connection group under the LBCs tab.
A connection in SimXpert defines the location of the connection between parts. Before creating a
connection, a Connection Type must first be selected. Creating a Connection alone does not create any
elements. Once a Connection has been created, it must be Rendered to produce a specific connection
type such as a RBE2 or CWELD.
Connection Groups
Connections are stored in Connection Groups. A connection group references up to four parts that will
be connected when the Connections in it are rendered. A Maximum Length tolerance defines the
maximum distance between parts for which to generate connections.
The diagram above is an analogy between Parts and Connection Groups. Parts are similar to Connection
Groups. Parts can contain geometry entities. Connection Groups contain Connections. Geometry is
similar to Connections in that neither are exported to a solver file. Geometry and Connections can be re-
used across different disciplines (e.g. Structures and Crash). Also, as compared in the boxed regions of
the diagram, meshing is similar to rendering. Meshing has several algorithms - some for surfaces and
some for solids, while rendering has several methods for each connection type. Meshing and Rendering
are both procedures that create actual elements that can be exported to a solver file.
Along with the various Connection Types, there are also two Connection Group Options to choose from
when creating connections.
Elements
Elements
Render
Meshing
Parts Connection
Group
Geometry
Connections
51
Model Generation
Connections
1. Auto Create Connection Group - Determines Parts that lie within the tolerance of the connection.
If no Connection Group exists for those parts, a new Connection Group will be created.
2. Use Current Connection Group - Connections are placed into the current Connection Group
(Current group can be set by Right-Clicking the group name in the Model Browser and selecting
Set Current from the context menu).
Creating Connections
In order to create a connection, it is first necessary to select one of the following seven Connection Types.
1. Spot Weld
2. Bolt
3. Hinge Pin
4. Seam Weld
5. Adhesive Curve
6. Adhesive Surface
7. Trim Mass
Spot Weld
Spot Weld creates a general connection between shell parts.
To create a spot weld connection, select the points defining the spot weld locations. The locations do not
need to be at existing nodes.
There are also various rendering options for Spot Weld:
ACM2 - Solid HEXA element connected with RBE3 elements
RBE2 - Node-to-Node Rigid element
CFAST - Mesh independent CFAST element
CWELD - Mesh independent CWELD element
RBAR - Node-to-Node Rigid Bar element
Generic - Node-to-Node multi-element connector defined by a SimXpert Generic Template
For the Node-to-Node connections, if the connection location does not coincide with a node on the
surface mesh, near nodes will be moved or the shell elements split.
Bolt
Bolt creates a rigid connection between aligning holes in shell element parts. It is rendered with RBE2
elements which create a spider web at each hole by linking the elements around the perimeter.
To create a bolt connection, select any node on the perimeter of one of the holes. The aligning holes on
the other parts will be automatically detected.
Modeling Guide
Connections
52
Bolt has two rendering options:
Clevis Search - Searches for a co-linear hole in the same part and connects if found
Release Rotate Dof - CBAR elements will be created instead of the linking RBE2s with the pin
flag releasing the rotational DOF about the bolt axis at one end
Hinge Pin
Hinge Pin is a connection that can have rotational degrees of freedom released. It is effectively the same
as a Bolt Connection, but instead has the Release Rotate DOF option selected by default.
Seam Weld
Seam Weld is a rigid connection representing a seam weld between parts. It is typically used for lap joints
and T-connections. It is rendered with multiple RBE2 elements. Seam Weld also automatically handles
mismatched meshes by element splitting and node adjustments.
To create a seam weld connection, select a series of locations to define a polyline along which the seam
weld will be created.
Adhesive Curve
Adhesive Curve creates a patch between two parts defined by a curve. Also, it generates solid HEXA
elements connected to the parts through RBE3 elements.
To create an adhesive curve connection, select a series of locations to define a polyline representing one
side of the adhesive area.
For Adhesive Curve, there are various rendering options such as:
Mesh Thickness - The thickness of the adhesive region (solid elements)
Mesh Width - The width of the adhesive region
Mesh Width Offset - The offset of the adhesive region from the adhesive curve
Mesh Length - The length of the individual solid elements
Youngs Modulus - Material property for the adhesive region
Poissons Ratio - Material property for the adhesive region
Adhesive Surface
Adhesive Surface is very similar to Adhesive Curve Connection, but in Adhesive Surface, region is
defined by selecting elements.
To create an adhesive surface connection, select shell elements on one part that defines the adhesive area
For Adhesive Surface, there are various rendering options such as:
Thickness - The thickness of the adhesive region (solid elements)
Youngs Modulus - Material property for the adhesive region
53
Model Generation
Connections
Poissons Ratio - Material property for the adhesive region
The Mesh Width and Length are assumed from the element selection. The Mesh Width Offset is zero.
Trim Mass
Trim Mass represents a lumped mass connected to various nodes.
To create a trim mass connection select a location for the mass element followed by the locations to
connect the mass to the structure.
For Trim Mass, there are various rendering options such as:
Mass - The lumped mass
Search Tolerance - Distance to search for nodes from the selected locations
Stiffness Type - Select from RBE2 (rigid) or RBE3 (interpolation) element to connect the mass
to the structure
Modeling Guide
Coordinate Systems
54
Coordinate Systems
Before describing the geometry of a structure, a coordinate system must exist. SimXpert has a built-in
rectangular Cartesian system called the basic coordinate system. It is the default coordinate system and
has an ID of zero. The orientation of the basic coordinate system is indicated by the triad in the lower left
corner of the graphics window. The triad shows orientation only, it is not located at the origin of the basic
coordinate system.
Sometimes it is convenient to use local coordinate systems for specifying loads, and or boundary
conditions. For example, a certain node may have a roller support placed in an inclined plane. A local
coordinate system with one of its axes normal to the inclined plane needs to be created and used to specify
the fixity (SPC) of the displacement component along the direction normal to the inclined plane.
Basic
coordinate
system
orientation
CONSTRAINT CONSTRAINT
55
Model Generation
Coordinate Systems
Local coordinate systems can be in cartesian, cylindrical or spherical systems.
You can create local coordinate systems by selecting Cartesian, Cylindrical, or Spherical from the
Coordinate Systemgroup under the Geometry tab. There are numerous methods to create local
coordinate systems in SimXpert:
1. 3 Points: Three points are used to define the coordinate system. The first point corresponds to the
location of origin. The second point defines the point on a specified axis and the third point
defines a point in a specified plane.
2. Euler: Creates a coordinate system through three specified rotations about the axes of an existing
coordinate system.
3. Normal: Creates a coordinate system with its origin at a point location on a surface. A specified
axis is normal to the surface.
4. Two Vectors: Creates a coordinate system with its origin at a designated location and two of the
coordinate frame axes are defined using vectors
5. Advanced: Location and orientation can be independently defined. There are four different ways
to define the location of the origin of the coordinate system: Geometry, Point/Node, Coordinate
System, and Center of Part. Further, the orientation can also be defined three ways: Global, Two
Axes, and Coordinate System.
A Cartesian coordinate system consists of three mutually perpendicular axes (X, Y, & Z axis) which
intersect at the origin. The three directions of a Cartesian coordinate system may be referred to as X,Y,Z;
u, v, w; or 1,2,3 respectively.
Coordinate
System
Direction
1
Direction
2
Direction
3 1-3 plane
Cartesian x y z x-z (y=0)
Cylindrical r z r-z ( =0)
Spherical r r- ( =0)
Cartesian
Cylindrical
Spherical
u u
u | | u
Modeling Guide
Coordinate Systems
56
In a cartesian coordinate system, the principal axes 1, 2 and 3 correspond to the X, Y and Z axes,
respectively. Points in space are entered in the order: x-coordinate, y-coordinate and z-coordinate. The
principal axes of a cartesian coordinate system and a point, P(x, y,z) are shown in Figure 0-2.
Figure 0-2 Cartesian Coordinate System
The graphical representation for creating a coordinate system using 3 Points method is shown in
Figure 0-3.
Figure 0-3 Coordinate system using 3 Points method
In Figure 0-4, the normal method to create a cartesian coordinate system is graphically represented. The
origin of the coordinate system is at a node on a surface. The z-axis direction is normal to the surface by
using right-hand rule and crossing the surfaces U direction with the V direction.The x-axis is parallel to
the U direction.
x
y
z
Axis 3
Axis 2
Axis 1
Y
Z
X
P =(x, y, z)
Second, a point location on the specified
axis
First, a point location at the Origin
Third, a point location in the speci-
fied plane
57
Model Generation
Coordinate Systems

Figure 0-4 Cartesian Coordinate system using Normal method
The graphical representation for creating a coordinate system using 2 vector method is shown in
Figure 0-5
Figure 0-5 Coordinate System using 2 Vector Method
u
v
X
Y
Z
Y
Z
X
First, a point location
for origin
Second, two points to define
the Vector for Axis-1
Third, two points to
define the Vector for
Axis-2
(0, 0, 0) (0, 0.9, 0)
(0.5, 0, 0)
(1, 0, 0)
(0, 1, 0)
Modeling Guide
Loads and Boundary Conditions
58
Loads and Boundary Conditions
Loads and boundary conditions describe the physical environment of the model to be analyzed. LBCs
can be created under the LBCs tab.
59
Model Generation
Material
Material
SimXpert supports the following material properties:
Isotropic.
Has the same properties in every direction.
Infinite number of planes of material property symmetry.
Two independent elastic constants.
Orthotropic 2D, 2D axisymmetric, and 3D.
Has the same properties in some directions, and different properties in other directions.
For 3D, there are three orthogonal planes of material property symmetry; nine independent
elastic constants; no interaction between normal and shear stresses.
Anisotropic 2D and 3D.
Has different properties for different directions.
For 3D, there are no planes of material property symmetry; 21 independent elastic constants.
For nonlinear analysis, supported properties are
Elasto plastic
Elastic behavior when total stress is less than yield stress of the material.
Plastic behavior when total stress is less than yield stress of the material.
Visco elastic (nonlinear elastic)
Elastic in the classical sense -- upon unloading, the stress-strain curve is retraced with no
permanent deformation/offset. The material is initially isotropic.
Viscous effects are modeled.
If the appropriate parameters are specified, it is possible to define a hypoelastic material.
Visco plastic
This is time-dependent cyclic plasticity.
Creep
This is time-dependent, inelastic behavior, and can occur at any stress level (below or above
the yield stress of the material).
Temperature dependent material properties are also allowed, for both linear and nonlinear analysis.
Material Overview
A wide variety of materials are encountered in stress analysis problems, and for any one of these
materials a range of constitutive models is available to describe the materials behavior. We can broadly
classify the materials of interest as those which exhibit almost purely elastic response, possibly with
some energy dissipation during rapid loading by viscoelastic response (the elastomers, such as rubber or
Modeling Guide
Material
60
solid propellant); materials that yield, and exhibit considerable ductility beyond yield (such as mild steel
and other commonly used metals, ice at low strain rates, and clay); materials that flow by rearrangement
of particles which interact generally through some dominantly frictional mechanism (such as sand); and
brittle materials (rock, concrete, ceramics).
Table 0-1 Common Material Characteristics
Material Characteristics Examples Models
Composites
(MATi,
MATORT,
PCOMP)
Anisotropic:
1) Layered,
21 Constants
2)Fiber Reinforced,
One dimensional strain in fibers
Aircraft panels
Tires, glass/epoxy
Composite
continuum
elements
Creep
(MATVP)
Strains increasing with time under
constant load. Stresses decreasing
with time under constant
deformations. Creep strains are
non-instantaneous.
Metals at high
temperatures, polymide
films, semiconductor
materials
ORNL
Norton
Maxwell
Elastic
(MATi,
MATORT)
Stress functions of instantaneous
strain only. Linear
load-displacement relation.
Small deformation
(below yield) for most
materials: metals, glass,
wood
Hookes Law
Elastoplasticity
(MATEP)
Yield condition flow rule and
hardening rule necessary to
calculate stress, plastic strain.
Permanent deformation
uponunloading.
Metals
Soils
von Mises
Isotropic
Cam -Clay
Hills Anisotropic
ds
ij
C
ijkl
dc
kl
=
S
E
2
--- T
t
CT 1 ( ) =
61
Model Generation
Material
Constitutive Models
A single material may contain multiple constitutive models. Each constitutive model characterizes
distinct ranges of the materials response. The constitutive models in MSC Nastran Implicit
Nonlinear contain a range of linear and nonlinear material models that can address or approximate the
material response of most commonly encountered materials. The constitutive models in MSC Nastran
Implicit Nonlinear can be accessed by any of the solid or structural elements. The models are assessed
independently at each constitutive calculation point (the numerical integration points in the elements).
Thus, the constitutive models are concerned only with a single calculation point. The element then
provides an estimate of the kinematic solution to the problem at the point under consideration.
Hyperelastic
(MATHE)
Stress function of instantaneous
strain. Nonlinear load-
displacement relation. Unloading
path same as loading.
Rubber Mooney
Ogden
Arruda-Boyce
Gent
Hypoelastic Rate form of stress-strain law Concrete Buyukozturk
Viscoelastic
(MATVE)
Time dependence of stresses in
elastic material under loads. Full
recovery after unloading.
Rubber,
Glass, industrial
plastics
Simo Model
Narayanaswamy
Viscoplastic
(MATVP)
Combined plasticity and creep
phenomenon
Metals
Powder
Power law
Shima Model
Table 0-1 Common Material Characteristics (continued)
Material Characteristics Examples Models
Modeling Guide
Material
62
MSC Nastran Implicit Nonlinear Material Entries
The following material bulk data entries are available in SOL 600. Each of these options are summarized
in the sections of this chapter and detailed in the MSC NASTRAN QRG. All standard MSC Nastran
materials are also available in SOL 600.
The following sections describe how to model material behavior in MSC Nastran Implicit Nonlinear.
Modeling material behavior consists of both specifying the constitutive models used to describe the
material behavior and defining the actual material data necessary to represent the material. Directional
dependency can be included for materials other than isotropic materials. Data for the materials can be
entered into MSC Nastran Implicit Nonlinear either directly through the input file or by user subroutines,
or material models may be defined in the SimXpert Materials Application. Each section of this chapter
discusses various options for organizing material data for input. Each section also discusses the
constitutive (stress-strain) relation and graphic representation of the models and includes
recommendations and cautions concerning the use of the models.
Bulk Data Entry Description
MATEP Specifies elasto-plastic material properties.
--MATTEP Specifies temperature-dependent elasto-plastic material properties.
MATF Specifies failure model properties for linear elastic materials.
MATG Specifies gasket material properties to be used in MSC Nastran Implicit
Nonlinear (SOL 600) only.
--MATTG Specifies gasket material property temperature variation to be used in MSC
Nastran Implicit Nonlinear (SOL 600) only.
MATHE Specifies hyperelastic (rubber-like) material properties for nonlinear (large strain
and large rotation) analysis in MSC Nastran Implicit Nonlinear (SOL 600) only.
--MATTHE Specifies temperature-dependent properties of hyperelastic (rubber-like)
materials (elastomers) in MSC Nastran Implicit Nonlinear (SOL 600) only.
MATED Specifies damage model properties for hyperelastic materials in MSC Nastran
Implicit Nonlinear (SOL 600) only.
MATORT Specifies elastic orthotropic material properties for 3-dimensional and plane
strain behavior for linear and nonlinear analyses in MSC Nastran Implicit
Nonlinear (SOL 600) only.
--MATTORT Specifies temperature-dependent properties of elastic orthotropic materials for
linear and nonlinear analyses used in MSC Nastran Implicit Nonlinear
(SOL 600) only.
MATVE Specifies isotropic visco-elastic material properties in MSC Nastran Implicit
Nonlinear (SOL 600) only.
--MATTVE Specifies temperature-dependent visco-elastic material properties in terms of
Thermo-Rheologically Simple behavior in MSC Nastran Implicit Nonlinear
(SOL 600) only.
MATVP Specifies viscoplastic or creep material properties to be used for quasi-static
analysis in MSC Nastran Implicit Nonlinear (SOL 600) only.
63
Model Generation
Material
Linear Elastic
MSC Nastran Implicit Nonlinear is capable of handling problems with any combination of isotropic,
orthotropic, or anisotropic linear elastic material behavior.
The linear elastic model is the model most commonly used to represent engineering materials. This
model, which has a linear relationship between stresses and strains, is represented by Hookes Law.
Figure 0-6 shows that stress is proportional to strain in a uniaxial tension test. The ratio of stress to strain
is the familiar definition of modulus of elasticity (Youngs modulus) of the material.
E (modulus of elasticity) =(axial stress)/(axial strain) (0-1)
Figure 0-6 Uniaxial Stress-Strain Relation of Linear Elastic Material
Experiments show that axial elongation is always accompanied by lateral contraction of the bar. The ratio
for a linear elastic material is:
=(lateral contraction)/(axial elongation) (0-2)
This is known as Poissons ratio. Similarly, the shear modulus (modulus of rigidity) is definedas:
(shear modulus) =(shear stress)/(shear strain) (0-3)
A Poissons ratio of 0.5, which would be appropriate for an incompressible material, can be used for the
following elements: Herrmann, plane stress, shell, truss, or beam. A Poissons ratio which is close (but
not equal) to 0.5 can be used for constant dilation elements and reduced integration elements in situations
which do not include other severe kinematic constraints. Using a Poissons ratio close to 0.5 for all other
elements usually leads to behavior that is too stiff. A Poissons ratio of 0.5 can also be used with the
updated Lagrangian formulation in the multiplicative decomposition framework using the standard
displacement elements. In these elements, the treatment for incompressibility is transparent.
S
t
r
e
s
s
Strain
E
1
v
G
Modeling Guide
Material
64
Isotropic Materials
Most linear elastic materials are assumed to be isotropic (their elastic properties are the same in all
directions). For an isotropic material, every plane is a plane of symmetry and every direction is an axis
of symmetry. It can be shown that for an isotropic material:
(0-4)
The shear modulus can be easily calculated if the modulus of elasticity and Poissons ratio
areknown.
Specifying Isotropic Material Entries
Isotropic material models are designated with the MAT1 Bulk Data entry in the MSC Nastran Input File.
References
MAT1 in the MSC NASTRAN QRG.
SimXpert Materials Application Input Data
To define an isotropic material in SimXpert:
1. Select Materials and Properties > Isotropic.
Isotropic linear elastic material models support the following material data:
The material density, used to define the mass of the structure, and the damping value are used in dynamic
loadings, while the expansion coefficient is used to identify the thermal strains.
Entry Description
MAT1 Defines the material properties for linear isotropic materials.
Isotropic-Linear Elastic Description
Youngs Modulus Defines the elastic modulus. This property is generally required. May vary
with temperature via a material table.
Shear Modulus Defines the shear modulus. This property is generally not required. May
vary with temperature via a material table.
Poissons Ratio Defines the Poissons ratio. This property is generally required. May vary
with temperature via a material table.
Density Defines the mass density. This property is optional.
G E 2 1 v + ( ) ( ) =
G E v
65
Model Generation
Material
Orthotropic Materials
An orthotropic material has three mutually orthogonal planes of symmetry. With respect to a coordinate
system parallel to these planes, the constitutive law for this material is given by the following more
general form of Hookes Law:
3-D Orthotropic
Due to symmetry of the compliance matrix, E11 =E22 , E22 =E33 , and E33 =
E11 . Using these relations, a general orthotropic material has nine independent constants:
E11, E22, E33, , , , G12, G23, G31
These nine constants must be specified in constructing the material model.
2-D Orthotropic
Orthotropic material models can be used with 2-D elements, such as plane stress, plane strain, and
axisymmetric elements. For example, the orthotropic stress-strain relationship for a plane stress
element is:
(0-5)
Specifying Orthotropic Material Entries
c
11
c
22
c
33

12

23

13
1 E
1
( ) u
12
( ) E
1
( ) u
13
( ) E
1
( ) 0 0 0
u
12
( ) E
1
( ) 1 E
2
( ) u
23
( ) E
2
( ) 0 0 0
u
13
( ) E
1
( ) u
23
( ) E
2
( ) 1 E
3
( ) 0 0 0
0 0 0 1 G
12
( ) 0 0
0 0 0 0 1 G
23
( ) 0
0 0 0 0 0 1 G
13
( )
o
11
o
22
o
33
t
12
t
23
t
13
=
v
21
v
12
v
32
v
23
v
13
v
31
v
12
v
23
v
31
Note:
The inequalities E22 > E33, E11 > E22, and E33 > E11 must be satisfied in
order for the orthotropic material to be stable. This is checked by MSC Nastran
Implicit Nonlinear.
v
23
v
12
v
31
C
1
1 v
12
v
21
( )
-----------------------------
E
1
v
21
E
1
0
v
12
E
2
E
2
0
0 0 1 v
12
v
21
( )G
=
Modeling Guide
Material
66
2-D and 3-D orthotropic materials are characterized in MSC Nastran using the following bulk
dataentries.
References
MAT3 in the MSC NASTRAN QRG.
MAT8 in the MSC NASTRAN QRG.
MATORT in the MSC NASTRAN QRG.
SimXpert Materials Application Input Data
To define an orthotropic material in SimXpert:
1. Select Materials and Properties > 2D or 3D Orthotropic
The required properties for orthotropic linear elastic material models vary based on dimension, element
type, and thermal dependencies. 3-D orthotropic material models require the following material data
(2-D requires a reduced set).
Entry Description
MAT3 Defines the material properties for linear orthotropic materials used by the
CTRIAX6 element entry.
MAT2
MAT8
Defines the material property for an orthotropic material for solids and
isoparametric shell elements.
MATORT Specifies elastic orthotropic material properties for three-dimensional and
plane strain behavior for linear and nonlinear analyses in MSC Nastran Implicit
Nonlinear (SOL 600) only in a more general way than MAT2 or MAT8.
Orthotropic-Linear Elastic Description
Youngs Modulus 1/2/3 Defines the elastic moduli in the elements coordinate system. This is
required data. May vary with temperature via a material table.
Poissons Ratio 12 Defines the Poissons ratios relative to the elements coordinate
system. This is required data. May vary with temperature via a defined
material field.
Density Defines the mass density which is an optional property.
Shear Modulus 12/23/31 Defines the shear moduli relative to the elements coordinate system.
This is required data. May vary with temperature via a defined
material field.
67
Model Generation
Material
Anisotropic Materials
Anisotropic material exhibits different elastic properties indifferent directions. The significant
directions of the material are labeled as preferred directions, and it is easiest to express the material
behavior with respect to these directions.
The stress-strain relationship for an anisotropic linear elastic material can be expressed as
(0-6)
The values of

(the stress-strain relation) and the preferred directions (if necessary) must be defined
for an anisotropic material.
Specifying Anisotropic Material Entries
Anisotropic materials are characterized in MSC Nastran using the following bulk data entries.
References
MAT2 in the MSC NASTRAN QRG.
SimXpert Materials Application Input Data
To define anisotropic material in SimXpert:
1. Select Materials and Properties>2D or 3D Anisotropic>.
Anisotropic linear elastic material models support the following material data.
Nonlinear Elastic
Hypoelastic - Isotropic
The hypoelastic model is able to represent a nonlinear elastic (reversible) material behavior. For this
constitutive theory, MSC Nastran Implicit Nonlinear assumes that
Entry Description
MAT2 Defines the material properties for linear anisotropic materials for two-
dimensional elements.
Anisotropic-Linear Elastic Description
Stress-Strain Matrix
components, Cij
Defines the upper right portion of the symmetric stress-strain matrix
relative to the elements coordinate system.
Density Defines the mass density which is an optional property.
o
ij
C
ijkl
c
kl
=
C
ijkl
Modeling Guide
Material
68
(0-7)
where is a function of the mechanical strain and is a function of the temperature.
The stress and strains are true stresses and logarithmic strains, respectively, when used in conjunction
with the updated Lagrange and large displacement options.
When used in conjunction with the large displacement option only, Equation (0-7) is expressed as
(0-8)
where are the Green-Lagrangian strain and second Piola-Kirchhoff stress, respectively.
Thismodel can be used with any stress element, including Herrmann formulation elements.
The tensors and may be defined by user subroutine HYPELA. In order to provide an accurate
solution, should be a tangent stiffness evaluated at the beginning of the iteration. In addition, the total
stress should be defined as its exact value at the end of the increment. This allows the residual load
correction to work effectively.
In user subroutine HYPELA2, besides the functionality of HYPELA, additional information is available
regarding the kinematics of deformation. In particular, the deformation gradient ( ), rotation tensor
( ), and the eigenvalues ( ) and eigenvectors ( ) to form the stretch tensor ( ) are also provided.
This information is available only for the continuum elements namely: plane strain, generalized plane
strain, plane stress, axisymmetric, axisymmetric with twist, and three-dimensional cases.
Hyperelastic - Isotropic
Hyperelastic models are specified using either the MATHP or MATHE bulk data entries and are used to
describe the behavior of materials that exhibit elastic response up to large strains, such as rubber, solid
propellant, and other elastomeric materials. These materials are described in terms of a strain energy
potential, U, which defines the strain energy stored in the material per unit of volume in the initial
configuration as a function of the strain at that point in the material.
Elastomeric materials are elastic in the classical sense. Upon unloading, the stress-strain curve is retraced
and there is no permanent deformation. Elastomeric materials are initially isotropic. Figure 0-7 shows a
typical stress-strain curve for an elastomeric material.
o

ij
L
ijkl
c

kl
g
ij
+ =
L g
S

ij
L
ijkl
E

kl
g
ij
+ =
E S ,
L g
L
F
R N U
69
Model Generation
Material
Figure 0-7 A Typical Stress-Strain Curve for an Elastomeric Material
Calculations of stresses in an elastomeric material requires an existence of a strain energy function which
is usually defined in terms of invariants or stretch ratios. Significance and calculation of these kinematic
quantities is discussed next.
Characteristics of Elastomeric Materials
Most solid rubberlike materials are nearly incompressible. Their bulk modulus is several orders of
magnitude larger than their shear modulus. For applications where the material is not highly confined,
the assumption that the material is fully incompressible is usually a good approximation. In cases where
the material is highly confined (such as in an O-ring), modeling the compressibility can be important for
obtaining accurate results. In either case, the use of hybrid (mixed formulation) elements is
recommended for this type of material in all but plane stress cases.
Elastomeric foams on the other hand are elastic but very compressible.
Elastomeric materials are considered to be isotropic in nature with random orientation of the long
chainmolecules.
Strain Energy Potential and Representative Models
Calculations of stresses in an elastomeric material requires an existence of a strain energy function which
is usually defined in terms of invariants or stretch ratios.
In the rectangular block in Figure 0-8, , , and are the principal stretch ratios along the edges
of the block defined by
(0-9)
o
,

S
t
r
e
s
s
c, Strain
100%

1

2

3

i
L
i
u
i
+ ( ) L
i
=
Modeling Guide
Material
70
Figure 0-8 Rectangular Rubber Block
In practice, the material behavior is (approximately) incompressible, leading to the constraint equation
the strain invariants are defined as
(0-10)
Depending on the choice of configurations, for example, reference (at ) or current ( ),
you obtain total or updated Lagrange formulations for elasticity. The kinematic measures for the two
formulations are discussed next.
Total Lagrangian Formulation
The strain measure is the Green-Lagrange strain defined as:
(0-11)
where is the right Cauchy-Green deformation tensor defined as:
(0-12)
in which is the deformation gradient (a two-point tensor) written as:
L3
2L2
L2
L1

1
L
1
Undeformed
Deformed
3L3

3
1 =
I
1

1
2

2
2

3
2
+ + =
I
2

1
2

2
2

2
2

3
2

3
2

1
2
+ + =
I
3

1
2

2
2

3
2
=
t 0 = t n 1 + =
E
ij
1
2
--- C
ij
o
ij
( ) =
C
ij
C
ij
F
ki
F
kj
=
F
kj
71
Model Generation
Material
(0-13)
The J acobian is defined as:
(0-14)
Thus, the invariants can be written as:
(0-15)
in which is the permutation tensor. Also, using spectral decomposition theorem,
(0-16)
in which the stretches are the eigenvalues of the right Cauchy-Green deformation tensor, and
the eigenvectors are .
Updated Lagrange Formulation
The strain measure is the true or logarithmic measure defined as:
(0-17)
where the left Cauchy-Green or finger tensor is defined as:
(0-18)
Thus, using the spectral decomposition theorem, the true strains are writtenas:
(0-19)
F
kj
cx
k
cX
j
-------- =
J
J
1

3
det C
ij
( )
1
2
---
= =
I
1
C
ii
=
I
2
C
ij
C
ij
C
ii
( )
2
( )
2
--------------------------------------- =
I
3
1
6
--- e
ijk
e
pqr
C
ip
C
jq
C
kr
det C
ij
( ) = =
(implied sum on i)
e
ijk
C
ij

A
2
N
i
A
N
j
A
=

A
2
C
ij
N
i
A
c
ij
1
2
--- ln b
ij
=
b
ij
b
ij
F
ik
F
jk
=
c
ij
1
2
---
A
ln ( )n
i
A
n
j
A
=
Modeling Guide
Material
72
where is the eigenvectors in the current configuration. It is noted that the true strains can also be
approximated using first Pad approximation, which is a rational expansion of the tensor, as:
(0-20)
where a polar decomposition of the deformation gradient is done into the left stretch tensor and
rotation tensor as:
The J acobian is defined as:
(0-21)
and the invariants are now defined as:
(0-22)
It is noted that either Equation (0-15) or Equation (0-22) gives the same strain energy since it is scalar
and invariant. Also, to account for the incompressibility condition, in both formulations, the strain energy
is split into deviatoric and volumetric parts as:
(0-23)
Mooney-Rivlin Model
The generalized Mooney-Rivlin model for nearly-incompressible elastomeric materials is written as:
(0-24)
where and are the first and second deviatoric invariants.
n
i
A
c
ij
2 V
ij
o
ij
( ) V
ij
o
ij
+ ( )
1
=
F
ij
V
ij
R
ij
F
ij
V
ik
R
kj
=
J
J
1

3
det b
ij
( )
1
2
---
= =
I
1
b
ii
=
I
2
1
2
--- b
ij
b
ij
b
ii
( )
2
( ) =
I
3
1
6
--- e
ijk
e
pqr
b
ip
b
jq
b
kr
det b
ij
( ) = = and
W W
deviatoric
W
volumetric
+ =
W
deviatoric
gmr
C
mn
I
1
3 ( )
m
I
2
3 ( )
n
n 1 =
N

m 1 =
N

=
I
1
I
2
73
Model Generation
Material
Jamus-Green-Simpson Model
A particular form of the generalized Mooney-Rivlin model, namely the third order deformation (tod)
model, is implemented in MSC Nastran Implicit Nonlinear (SOL 600). This is one of the few places
where the formulation for SOLs 106 and 129 may be more appropriate because they can use up to fifth
order terms. However, the Ogden formulation (below) is usually better for large strain behavior than even
the fifth order Mooney-Rivlin.
(0-25)
where is the deviatoric third order deformation form strain energyfunction,
are material constants obtained from experimental data.
Simpler and popular forms of the above strain energy function are obtained as:
(0-26)
Ogden Model
The form of strain energy for the Ogden model in MSC Nastran Implicit Nonlinear is,
(0-27)
where are the deviatoric stretch ratios while , , and are the
material constants obtained from the curve fitting of experimental data.
The Ogden model is usually applied to slightly compressible materials. If no bulk modulus is given, it is
taken to be virtually incompressible. This model is different from the Mooney model in several respects.
The Mooney material model is with respect to the invariants of the right or left Cauchy-Green strain
tensor and implicitly assumes that the material is incompressible. The Ogden formulation is with respect
to the eigenvalues of the right or left Cauchy-Green strain, and the presence of the bulk modulus implies
some compressibility. Using a two-term series results in identical behavior as the Mooney mode if:
, , , and
W
devratoric
tod
C
10
I
1
3 ( ) C
01
I
2
3 ( ) C
11
I
1
3 ( ) I
2
3 ( ) C
20
I
1
3 ( )
2
C
30
I
1
3 ( )
3
+ + +
+
=
W
deviatoric
tod
C
10
C
01
C
11
C
20
C
30
, , , ,
W
deviatoric
mr
C
10
I
1
3 ( ) C
01
I
2
3 ( ) + = Mooney-Rivlin
W
deviatoric
nh
C
10
I
1
3 ( ) =
Neo-Hookean
W
deviatoric
ogden

k
o
k
------
1
o
k

2
o
k

3
o
k
3 + + ( )
k 1 =
N

i
o
k
J

o
k
3
-----

i
o
k
= C
mn

k
o
k

1
2C
10
= o
1
2 =
2
2C
01
= o
2
2 =
Modeling Guide
Material
74
Arruda-Boyce Model
In the Arruda-Boyce strain energy model, the underlying molecular structure of elastomer is represented
by an eight chain model to simulate the non-Gaussian behavior of individual chains in the network. The
two parameters, and ( is the chain density, is the Boltzmann constant, is the
temperature, and is the number of statistical links of length l in the chain between chemical crosslinks)
representing initial modules and limiting chain extensibility and are related to the molecular chain
orientation thus representing the physics of network deformation.
As evident in most models describing rubber deformation, the strain energy function constructed by
fitting experiment data obtained from one state of deformation to another fails to accurately describe that
deformation mode. The Arruda-Boyce model ameliorates this defect and is unique since the standard
tensile test data provides sufficient accuracy for multiple modes of deformation.
Figure 0-9 Eight Chain Network in Stretched Configuration
The model is constructed using the eight chain network as follows:
Consider a cube of dimension with an unstretched network including eight chains of length
, where the fully extended chain has an approximate length of Nl. A chain vector from the
center of the cube to a corner can be expressed as:
(0-28)
Using geometrical considerations, the chain vector length can be written as:
(0-29)
and
nkO N n k O
N

2
o
0

3
o
0

1
o
0
C
1
k
j
i
o
0
r
0
Nl =
C
1
o
0
2
------
1
i
o
0
2
------
2
j
o
0
2
------
3
k + + =
r
chain
1
3
------- Nl
1
2

2
2

3
2
+ + ( )
1 2
=
75
Model Generation
Material
(0-30)
Using statistical mechanics considerations, the work of deformation is proportional to the entropy
change on stretching the chains from the unstretched state and may be written in terms of the chain length
as:
(0-31)
where is the chain density and is a constant. is an inverse Langevin function correctly accounts
for the limiting chain extensibility and is defined as:
(0-32)
where Langevin is defined as:
(0-33)
With Equation (0-30) through Equation (0-33), the Arruda-Boyce model can be written
(0-34)
Gent Model
Also, using the notion of limiting chain extensibility, Gent proposed the following constitutive relation:
(0-35)
where
(0-36)

chain
r
chain
r
0
------------
1
3
------- I
1
( )
1 2
= =
W nkON
r
chain
Nl
------------
\
|
| ln
|
| sinh
--------------
.
|
OC

+ =
n C

|
| L
1
r
chain
Nl
------------
\ .
| |
=
| ( ) |
1
|
--- coth =
W
dev
Arruda-Boyce
nkO
1
2
--- I
1
3 ( )
1
20N
---------- I
1
2
9 ( )
11
1050N
2
------------------ I
1
3
27 ( ) + + =
19
7000N
3
------------------ I
1
4
81 ( )
519
673750N
4
------------------------ I
1
5
243 ( ) | + +
W
dev
Gent
EI
m

6
------------
I
m
I
m
I
1
*

---------------- log =
I
1
*
I
1
3 =
Modeling Guide
Material
76
The constant is independent of molecular length and, hence, of degree of crosslinking. The model
is attractive due to its simplicity, but yet captures the main behavior of a network of extensible molecules
over the entire range of possible strains.
The volumetric part of the strain energy is for all the rubber models in MSC Nastran Implicit
Nonlinear is:
(0-37)
when is the bulk modulus. It can be noted that the particular form of volumetric strain energy is chosen
such that:
1. The constraint condition is satisfied for incompressible deformations only; for example:
(0-38)
2. The constraint condition does not contribute to the dilatational stiffness.
This yields the constraint function as:
(0-39)
upon substitution of Equation (0-39) in Equation (0-35) and taking the first variation of the variational
principle, you obtain the pressure variable as:
(0-40)
The equation has a physical significance in that for small deformations, the pressure is linearly related to
the volumetric strains by the bulk modulus .
The discontinuous or continuous damage models discussed in the models section on damage can be
included with the generalized Mooney-Rivlin, Ogden, Arruda-Boyce, and Gent models to simulate
Mullins effect or fatigue of elastomers when using the updated Lagrangian approach. In the total
Lagrangian framework however, this is available for the Ogden model only.
EI
m
W
volumetric
9K
2
------- J
1
3
---
1
\ .
|
| |
2
=
K
f I
3
( )
>0 if I
3
0 >
0 = if I
3
1 =
<0 if I
3
0 <

f I
3
( ) 3 I
3
1
6
-- -
1
\ .
|
| |
=
p 3K J
1
3
---
1
\ .
|
| |
=
K
77
Model Generation
Material
Foam Model
Sometimes elastomeric materials show large volumetric deformations. For this type of behavior, the
models discussed above are not appropriate. Instead, the foam model expressed by:
(0-41)
should be used. In contrast to the Ogden model, the first part of the foam strain energy function is not
purely deviatoric. The material constants provide additional flexibility to describe the material
behavior also for a large amount of compressibility.
Updated Lagrange Formulation for Nonlinear Elasticity
The Mooney-Rivlin, Ogden, Arruda-Boyce, Gent and Foam models may be used either in the total
Lagrange or updated Lagrange framework. This is selected using the PARAM,MARUPDAT. For plane
stress analysis the total Lagrange procedure will always be used.
The updated Lagrangian rubber elasticity capability can be used in conjunction with both continuous as
well as discontinuous damage models. Thermal, as well as viscoelastic, effects can be modeled with the
current formulation. While the Mooney model can account for the temperature dependent material
properties, the Ogden model does not support the temperature dependence at this time. The singularity
ratio of the system is inversely proportional to the order of bulk modulus of the material due to the
condensation procedure.
A consistent linearization has been carried out to obtain the tangent modulus. The singularity for the case
of two- or three-equal stretch ratios is analytically removed by application of LHospitals rule. The
current framework with an exact implementation of the finite strain kinematics along with the split of
strain energy to handle compressible and nearly incompressible response is eminently suitable for
implementation of any nonlinear elastic as well as inelastic material models. In fact, the finite
deformation plasticity model based on the multiplicative decomposition, is
implemented in the same framework.
To simulate elastomeric materials, incompressible element(s) are used for plane strain, axisymmetric,
and three-dimensional problems for elasticity in total Lagrangian framework. These elements can be
used with each other or in combination with other elements. For plane stress, beam, plate or shell
analysis, conventional elements can be used. For updated Lagrangian elasticity, both conventional
elements (as well as Herrmann elements) can be used for plane strain, axisymmetric, and three-
dimensional problems.
Experimental Determination of Hyperelastic Material Parameters
In order to determine the material parameters to be used, like Mooney coefficients, Ogden moduli,
relaxation times, etc., experiments must be carried out. In this section, the laboratory tests of which data
can be used to fit the material parameters will be described. Once the test data is available the
W

n
o
n
------
1
o
n

2
o
n

3
o
n
3 + + ( )

n
|
n
----- 1 J
|
n
( )
n 1 =
N

+
n 1 =
N

=
|
n
F F
e
F
u
F
p
=
Modeling Guide
Material
78
Experimental Data Fitting module in MSC SimXpert can be used to calculate appropriate coefficient
values.
For a homogeneous material, homogeneous deformation modes suffice to characterize the material
constants. MSC Nastran Implicit Nonlinear accepts test data from the following deformation modes:
Uniaxial tension and compression.
Biaxial tension and compression.
Planar tension and compression (also known as pure shear).
Simple Shear
Volumetric tension and compression
79
Model Generation
Material
Figure 0-10 Test Data
Uniaxial Test
1
2
3
1
2
3
1
2
3
1
2
3
Uniaxial Test Data
Biaxial Test Data
Planar Test Data
Volumetric Test Data
Modeling Guide
Material
80
Probably the most popular test is the uniaxial test (see Figure 0-11). This test can be used in tension as
well as in compression, both for incompressible and (slightly) compressible elastomeric materials. The
shape of the specimen used in compression will usually be less slender than the shape used in tension.
Within the region indicated by the dashed line, the state of deformation will be homogeneous, where the
deformation can be described by:
, (0-42)
while the corresponding engineering stresses are given by:
, (0-43)
in which is the applied force and is the cross sectional area of the undeformed specimen in the
- -plane, within the region indicated by the dashed line.
Figure 0-11 Uniaxial (Tensile) Test
Necessary input for the curve fitting program in MSC SimXpert consists of at least engineering strain
( ) versus engineering stress ( ) data points. In case of (slightly) compressible materials,
information about the volume changes is also needed. This data can be given either in terms of the area
ratio or the volume ratio. The area ratio is defined by the current cross sectional area over the original
cross sectional area . Similarly, the volume ratio is defined by the current volume over the
undeformed volume . Notice that the volume ratio and the area ratio are related by:

1
1 e
11
+ = =
2

3
J = =
o
11
o
F
A
0
------ = = o
22
o
33
0 = =
F A
0
E
2
E
3
E
1
E
3
E
2
F F
e
11
o
11
A
A
0
V
V
0
81
Model Generation
Material
If, for a particular elastomeric material, both a tensile and a compression test have been performed, all
the data points should be collected into one data file. The layout of a data file containing uniaxial test
data is given in the figure below. The columns may be separated by either spaces or commas. For (nearly)
incompressible material behavior, the third column can be omitted.
Figure 0-12 Layout of Data File for a Uniaxial Test
Equi-Biaxial Test
The equi-biaxial tensile test outlined in Figure 0-13 can be used to obtain, within the region indicated by
the dashed line, a homogeneous state of deformation defined by:
Figure 0-13 Equi-biaxial (Tensile) Test
, (0-44)
with corresponding engineering stresses:
, (0-45)
with being the original cross sectional area of the elastomeric sheet in the direction perpendicular
to the applied forces, which is assumed to be the same in the - -plane and the - -plane.
For compressible elastomers, volumetric information is needed. For the equi-biaxial test, this can be
given in terms of a thickness ratio or, similar to the uniaxial test, a volume ratio. The thickness ratio is
defined as the current sheet thickness over the original sheet thickness . The relation between the
thickness ratio and the volume ratio is:
V
V
0
------ J
A
A
0
------ 1 e
11
+ ( ) = =
e
11
o
11
A A
0
e
11
o
11
V V
0

or

1

2
1 e
11
+ 1 e
22
+ = = = =
3
J
2
=
o
11
o
22
o
F
A
0
------ = = = o
33
0 =
A
0
E
1
E
3
E
2
E
3
t t
0
Modeling Guide
Material
82
(0-46)
The layout of a data file for an equi-biaxial tensile test is given in Figure 0-13.
Planar Shear Test
A state of planar shear, also sometimes called pure shear, can be obtained by clamping and stretching an
elastomeric rectangular sheet of material, as indicated in Figure 0-14.
Figure 0-14 Planar Shear Test
Except for the vicinity of the free edges and the clamps, the state of strain can be found to be substantially
uniform, according to:
, , (0-47)
where the known stress components are given by:
F
F
F F
E
1
E
3
E
2
V
V
0
------ J
t
t
0
---- 1 e
11
+ ( )
2
= =

1
1 e
11
+ = =
2
1 =
3
J

--- =
83
Model Generation
Material
, (0-48)
in which is the cross sectional area of the undeformed specimen in the - -plane. Notice that
the engineering strain is zero, but that the corresponding engineering stress depends on the
material behavior.
(0-49)
(0-50)
Simple Shear Test
A test which, compared to the above mentioned tests, leads to a more complex kinematic description, is
the simple shear test (see Figure 0-15).Upon introducing the shear strain , the coordinates in the
deformed configuration are given by:
, , (0-51)
which yields for the deformation gradient:
F F
E
1
E
3
E
2
o
11
o
F
A
0
------ = = o
33
0 =
A
0
E
1
E
3
e
22
o
22
oU T
S
o
S
=
T
S

S
c
c
U 2
S

S
3

\ .
| |
I
1
c
cU
I
2
c
cU
+
\ .
|
| |
= =

x
1
X
1
X
2
+ = x
2
X
2
= x
3
X
3
=
Modeling Guide
Material
84
(0-52)
Figure 0-15 Simple Shear Test
Notice that , irrespective of the value of , from which it can be concluded that a simple
shear test is a constant volume test.
Based on Equation (0-51), Equation (0-52) and Figure 0-15, the engineering strain tensor and the right
Cauchy-Green strain tensor can be evaluated as:
(0-53)
(0-54)
According to Equation (0-54), the principal stretch ratios follow from the principal values of andread:
F
1 0
0 1 0
0 0 1
=
2F
E
1
E
3
E
2 atan
det F ( ) 1 =
e
0 2 0
2 0 0
0 0 0
=
C
1 0
1
2
+ 0
0 0 1
=
C
85
Model Generation
Material
, (0-55)
It can easily be verified that , which again shows that the simple shear test is a constant
volume test. The relevant engineering stress is given by:
(0-56)
with being the cross sectional area of the undeformed specimen in the - -plane.
The layout of a data file containing measurements of a simple shear test is given in Figure 0-16.
Figure 0-16 Layout of Data File for a Simple Shear Test
Volumetric Test
Although a uniaxial, equi-biaxial and planar shear test can be used to obtain information about the
volumetric behavior, for compressible materials an additional volumetric test may be preferable. This is
especially true for slightly compressible materials, since volumetric data from other tests other than a
volumetric one may easily be inaccurate (because most of the deformation is deviatoric). Two commonly
used volumetric tests are outlined in Figure 0-17. In Figure 0-17a, a cylindrical specimen is compressed
in a cylindrical hole. This test can be successfully applied for slightly compressible materials. In
Figure 0-17b, a specimen is deformed by compressing the surrounding fluid. This volumetric test can
also be used for highly compressible materials.
Figure 0-17 Volumetric Tests
For a volumetric test, the direct true stress components are assumed to be equal to the hydrostatic
pressure and given by:
(0-57)

1 2 ,
1

2
2
---- 1

2
4
---- + + =
3
1 =

3
1 =
o
12
F
A
0
------ =
A
0
E
1
E
3
2e
12
= o
12
p
T
11
T
22
T
33
F
A
p
------ = = =
Modeling Guide
Material
86
in which denotes the area of the piston in the - -plane. The deformation can be expressed in
terms of an engineering strain and corresponding stretch ratio , which can be determined from the
measured volume change according to:
(0-58)
Based on according to Figure 0-17b, the engineering stress follows from:
(0-59)
Notice that only in the case of Figure 0-17b the engineering strain and the engineering stress are
equal to the direct components of the engineering strain and the engineering stress tensor.
The layout of the data file corresponding to a volumetric test is given in Figure 0-18. Notice that because
of Figure 0-17b, the entries of the first and the third column are not independent.
Figure 0-18 Layout of Data File for a Volumetric Test
Relaxation Test
F
F
F
(a) (b)
E
2
E
3
E
1
A
p
E
2
E
3
e
e 1
V
V
0
------
3
1 J
3
1 = = =
o
o T
11

2
=
e o
87
Model Generation
Material
The basic feature of a relaxation test is that the force or stress response to a prescribed fixed displacement
or deformation is measured as a function of time. A relaxation test for a large strain elastomeric material
is indicated in Figure 0-19. By measuring the force needed for a displacement at different time
intervals, the decay of the strain energy as a function of time can be determined. For linear elastic
isotropic material, similar tests can be performed to get information about the shear modulus and/or the
bulk modulus as a function of time. In order to properly measure the instantaneous values, application
of the prescribed displacement should occur sufficiently fast. It should be noted, due to the assumption
introduced in equation Equation (0-94), that for large strain visco-elastic materials the magnitude of (the
instantaneous value of) the strain energy is not important, since every energy term in the Prony series
expansion is related to the instantaneous strain energy using a scalar multiplier. The data does not need
to be equispaced in time. Usually, at the beginning of the relaxation experiment the measurements are
done at smaller time intervals than at the end of the experiment.
Figure 0-19 Relaxation Test
If, for linear visco-elastic materials, instead of a relaxation test only a creep test can be performed, the
creep data must be transformed into relaxation data. Converting creep data into relaxation data can be
done using a numerical integration scheme, but is not part of MSC Nastran Implicit Nonlinear.
Hyperelastic Foam Properties
Elastomeric foams are cellular solids that have the following primary mechanical characteristics:
They can deform elastically up to large strain: up to 90 percent strain in compression. In most
applications, this is the dominant mode of deformation.
e o
V V
0

Au
Au
Modeling Guide
Material
88
Their porosity permits very large volumetric changes. This is in contrast to solid rubbers, which
are approximately incompressible.
Cellular solids are made up of interconnected networks of solid struts or plates which form the
edges and faces of cells. Foams are made up of polyhedral cells that pack in three dimensions.
The foam cells can either be open (e.g., sponge) or closed (e.g., flotation foam). Common
examples of elastomeric foam materials are cellular polymers such as cushions, padding, and
packaging materials which utilize the excellent energy absorption properties of foams - for a
certain stress level, the energy absorbed by foams is substantially greater than by ordinary stiff
elastic materials.
The figure below shows a typical compressive stress-strain curve for elastomeric foam.
Figure 0-20 Typical Compressive Stress-Strain Curve
Three stages can be distinguished during compression:
At small strains (less than 5 percent) the foam deforms in a linear elastic manner, due to cell wall bending.
This is followed by a plateau of deformation at almost constant stress, caused by the elastic buckling of
the columns or plates which make up the cell edges or walls. In closed cells, the enclosed gas pressure
and membrane stretching increase the level and slope of the plateau.
Finally, a region of densification occurs, where the cell walls crush together, resulting in a rapid increase
of compressive stress. Ultimate compressive nominal strains of 0.7 to 0.9 are typical.
The tensile deformation mechanisms for small strains are similar to the compression mechanisms but
differ for large strains. The figure shows a typical tensile stress-strain curve.
STRAIN
S
T
R
E
S
S

Densification
Plateau: Elastic buckling
of cell walls
Cell wall bending
89
Model Generation
Material
Figure 0-21 Typical Tensile Stress-Strain Curve
There are two stages during tension:
At small strains the foam deforms in a linear, elastic manner, due to cell wall bending, similar to that
incompression.
The cell walls rotate and align, resulting in rising stiffness. The walls are substantially aligned at a tensile
strain of about one-third. Further stretching results in increased axial strains in the walls.
At small strains for both compression and tension, the average experimentally observed Poisson's ratio,
v, of foams is one-third. At larger strains it is commonly observed that Poisson's ratio is effectively zero
during compression - the buckling of the cell walls does not result in any significant lateral deformation.
However, during tension, v is nonzero, which is a result of the alignment and stretching of the cell walls.
The manufacture of foams often results in cells with different principal dimensions. This shape
anisotropy results in different loading responses in different directions. However, the foam model does
not take this kind of initial anisotropy into account.
Determination of Foam Material Parameters
The response of the material is defined by the parameters in the strain energy function, U, so that it is
necessary to determine these parameters to use the foam model. MSC SimXpert contains a capability
for obtaining the
i
, o
i
and |
i
for the foam model with up to six terms (N=6) directly from test data. It
isusually best to obtain data from several experiments involving different kinds of deformation, over
therange of strains of interest in the actual application, and to use all of these data to determine
theparameters. Since the properties of foam materials can vary significantly from one batch to another,
all of the experiments should be performed on specimens taken from the same batch of material.
STRAIN
S
T
R
E
S
S

Cell wall bending
Cell wall alignment
Modeling Guide
Material
90
Uniaxial, Equibiaxial and Planar Deformations
The deformation modes are characterized in terms of the principal stretches,
i
, and the volume ratio, J .
The elastomeric foams are not incompressible, so that J =
1

3
!=1. The transverse stretches,
2
and/or

3
, are independently specified in the test data either as individual values from the measured lateral
deformations or through the definition of an effective Poissons ratio.
Uniaxial mode:
1
=
U
,
2
=
3
, J =
U

2
2
Equibiaxial mode:
1
=
2
=
B
, J =
B
2

3
Planar mode:
1
=
P
,
2
=1, J =
P

3

The three deformation modes above use a single form of the nominal stress-stretch relation,
(0-60)
where T
L
is the nominal stress and L
L
is the stretch in the direction of loading. Because of the
compressible behavior, the planar mode does not result in a state of pure shear. In fact, if the effective
Poissons ratio is zero, planar deformation is identical to uniaxial deformation.
Simple Shear Deformation
Simple shear is described by the deformation gradient
(0-61)
where is the shear strain. For this deformation, J =det F =1. A schematic illustration of simple shear
deformation is shown in Figure 0-22.
The nominal shear stress T
S
is:
(0-62)
where
j
=are the principal stretches in the plane of shearing, related to the shear strain, , by:
(0-63)
T
L

L
c
cU 2

L
-------

i
o
i
-----
L
o
i
J
o
i
|
i

\ .
| |
i 1 =
N

= =
F
1 0
0 1 0
0 0 1
=
T
S
c
cU 2
2
j
2
1
\ .
| |

2

------------------------------------

i
o
i
-----
j
o
i
1
\ .
| |
i 1 =
N

)

`


j 1 =
2

= =

1 2 ,
1

2
2
----- 1

2
4
----- + + =
91
Model Generation
Material
.
Figure 0-22 Simple Shear Test
The stretch in the direction perpendicular to the shear plane is L
3
=1. The transverse (tensile) stress, T
T
,
developed during simple shear deformation due to the Poynting effect, is
(0-64)
Volumetric Deformation
The volumetric deformation mode consists of all principal stretches being equal,

1
=
2
=
3
=
V
, J =
V
3
.
The pressure-volumetric ratio relation is
(0-65)
A volumetric compression test is illustrated Figure 0-23.
The pressure exerted on the foam specimen is the hydrostatic pressure of the fluid and the decrease in
the specimen volume is equal to the additional fluid entering the pressure chamber. The specimen is
sealed against fluid penetration.
Figure 0-23 Volumetric Compression Test Setup
Difference in Compression and Tension Deformation
2F
E
1
E
3
E
2 atan
T
T
c c
cU
2
j
2
1
\ .
| |
2
j
4

j
2

2
2 + ( )
------------------------------------------

i
o
i
-----
j
o
i
1
\ .
| |
i 1 =
N

)

`


j 1 =
2

= =
p
J c
cU 2
J
---

i
o
i
----- J
o
i
3
-----
J
o
i
|
i

\ .
|
|
|
| |
i 1 =
N

= =
Modeling Guide
Material
92
For small strains (less than 5 percent), foams behave similarly for both compression and tension.
However, we have seen that at large strains, the deformation mechanisms differ for compression
(buckling and crushing) and tension (alignment and stretching). Accurate modeling with the FOAM
option therefore requires that the experimental data used to define the material parameters correspond to
the dominant deformation modes of the actual problem being analyzed.
If compression dominates in the problem, the pertinent tests are:
Uniaxial compression.
Simple shear.
Planar compression (if Poissons ratio ).
Volumetric compression (if Poissons ratio ).
If tension dominates, the pertinent tests are:
Uniaxial tension.
Simple shear.
Biaxial tension (if Poissons ratio ).
Planar tension (if Poissons ratio ).
Lateral strain data can also be used to define the compressibility of the foam. Measurement of the lateral
strains may make other tests redundant, e.g., providing lateral strains for a uniaxial test eliminates the
need for a volumetric test. The foam model may not accurately fit Poisson's ratio if it varies significantly
between compression and tension.
F
F
F
(a) (b)
E
2
E
3
E
1
v 0 =
v 0 =
v 0 =
v 0 =
93
Model Generation
Material
Experimental Data Fitting
Least Squares Fit
The equations derived above for T
U
, T
B
, and T
S
, with the assumption of material incompressibility,
allow the material parameters C
ij
and
i
, o
i
to be determined from the experimentally measured stress-
strain relationships in the uniaxial, equibiaxial, and planar loading tests. A least squares fit, which
minimizes the relative error in stress, is used for this purpose. The equation for T
S
alone will not
determine the constants uniquely. The planar test data input must be augmented by either or both of the
other two types of test data to determine the material parameters.
The Ogden potential is linear in the coefficients
i
but strongly nonlinear in terms of the exponents o
i
,
thus necessitating use of a nonlinear least squares procedure. For the nominal stress-nominal strain data
pairs, the error measure, E, is minimized by E =sum(i=1to n)(1-T
i
th
/T
i
test2
), where T
i
test
is a stress value
from the test data and T
i
th
comes from one of the nominal stress expressions derived above.
The foam parameters
i
, o
i
, |
i
are determined from the experimentally measured stress-strain
relationships in the various loading tests described above. A least squares fit, which minimizes the
relative error in stress, is used for this purpose.
The foam potential is linear in the coefficients
i
but strongly nonlinear in terms of the exponents o
i
and
|
i
thus necessitating use of a nonlinear least squares procedure. For the n nominal stress-nominal strain
data pairs, the error measure E is minimized by E =sum(i=1to n)(1-T
i
th
/T
i
test2
, where T
i
test
is a stress
value from the test data and T
i
th
comes from one of the nominal stress expressions derived above.
Minimizing the relative error in stress implies that the error in slope (modulus) is minimized;
minimization of the absolute error would decrease the error at larger strains, at the expense of the
accuracy at small strains.
Alternative Method for Determination of Constants for Moderate Strains
Since the polynomial form with N=1 is very commonly used for cases where the nominal strain is not
too large, an alternative method of finding the material constants, assuming incompressibility, is to use
the uniaxial test data as follows. The nominal strain in the direction of loading in the uniaxial test is
c
U
=
U
-1. Expanding the equation for T
B
in terms of c
U
, using the Mooney-Rivlin form, and neglecting
terms of higher than second-order in c
U
, gives
T
U
=6c
U
(C
10
+C
01
-(C
10
+2C
01
)c
U
).
This is a parabola: the slope of this curve at the origin (the effective Youngs modulus at zero strain) is
6(C
10
+C
01
); this slope, together with the second-order term -6(C
10
+2C
01
)c
U
2
, defines the constants C
10

and C
01
.
Modeling Guide
Material
94
If compressibility should be modeled, then, under pure pressure loading, the compressible model with
N=1 gives, to first-order in the volumetric strain c
V
=3c
11
,
p=-(2 / D
1
)c
V
,
so that, at small nominal strains, the bulk modulus is defined as:
K=(2 / D
1
)
Hyperelastic Models in MSC Nastran
Various options are provided for defining the material properties. The first (available in both MSC
SimXpert and MSC Nastran) is to give the parameters of the polynomial form and , or the
parameters of the Ogden form and as functions of the temperature. The second is to give
the value of N, and give experimental stress-data for up to four simple tests: uniaxial, equilibrium, planar
and, if the material is compressible for volumetric compression test. MSC Nastran Implicit Nonlinear
will then compute the or and the . This method is available for N = 1 and N=2 for the
polynomial form and up to N = 6 for the Ogden form, and does not allow the properties to be temperature
dependent.
In either case, you should be careful about defining the or : especially when N > 1, the
behavior at higher strains is strongly sensitive to the values of the or , and unstable material
behavior may result if these values are not correctly defined. When some of the coefficients are strongly
negative, instability at higher strain levels is likely to occur.
Because the properties of rubber-like materials can vary significantly from one sample to another, it is
important that test data are taken from experiments on the same sample (or samples cut from the same
sheet), regardless whether the or are computed by the user or by the built-in method.
This material option can be used by itself, or can be combined with viscoelasticity to define time
dependent hyperelastic behavior. It cannot be combined with other material options such as plasticity or
creep. It may be used with the pure displacement formulation elements or with the hybrid (mixed
formulation) elements. Because elastomeric materials are usually almost completely incompressible,
fully integrated pure displacement method elements are not recommended for use with this material,
except for plane stress cases. If fully or selectively reduced integration displacement method elements
are used with the almost incompressible form of this material model in anything except plane stress
analysis, a penalty method is used to impose the incompressibility constraint. This can sometimes lead
to numerical difficulties, and the fully or selectively reduced integrated hybrid formulation elements
are therefore recommended.
Specifying Hyperelastic Material Entries
Nonlinear hyperelastic materials are characterized in MSC Nastran with the following Bulk Dataentries:
N A
ij
, D
i
N
i
o
i
, , D
i
A
ij

i
o
i
, | | D
i
A
ij

i
o
i
, | |
A
ij

i
o
i
, | |
A
ij

i
o
i
, | |
95
Model Generation
Material
.
References
MATHP in the MSC NASTRAN QRG
MATHE in the MSC NASTRAN QRG.
SimXpert Materials Application Input Data
To define a hyperelastic material in SimXpert:
1. Select Materials and Properties > Mooney-Rivlin, Odgen/Hyperfoam, or Arruda-
Boyce/Gent.
2. Select Coefficient or Test Data as the Data Type.
3. Enter parameter values.
4. Click OK.
Hyperelastic material models support the following material data.
General Hyperelastic
Entry Description
MATHP Specifies material properties for use in fully nonlinear (large strain and large
rotation) hyperelastic analysis of rubber-like materials (elastomers).
MATHE Specifies hyperelastic (rubber-like) material properties for nonlinear (large
strain and large rotation) analysis in (SOL 600) only.
Hyperelastic -General
Hyperelastic
Description
Data Type Coefficient
Density Defines the mass density which is an optional property.
Destort. Def. Coeff. A10, A01 Strain energy densities as a function of the strain invariants in the
material. May vary with temperature via a material table. This
option consolidates several of the Marc hyperelastic material
models.
Data Type Test Data
Density Defines the mass density which is an optional property.
Simple Tension/Compress. Data Table identification number of a TABLES1 entry that contains
simple tension/compression data to be used in the estimation of
the material constants Aij. xi values in the TABLES1 entry must
be stretch ratios and yi values must be values of the
engineering stress . Stresses are negative for compression
and positive for tension. (Integer >0 or blank)
l l
0

F A
0

Modeling Guide
Material
96
Mooney-Rivlin
Equivalent Tension Data Table identification number of a TABLES1 entry that contains
equibiaxial tension data to be used in the estimation of the
material constants Aij. xi values in the TABLES1 entry must be
stretch ratios . yi values must be values of the engineering
stress . is the current length, F is the current force, is
the initial length and is the cross-sectional area. In the case of
pressure of a spherical membrane, the engineering stress is given
by where P = current value of the pressure and
=initial radius and thickness. (Integer >0 or blank)
Simple Shear Data Table identification number of a TABLES1 entry that contains
simple shear data to be used in the estimation of the material
constants Aij. xi values in the TABLES1 entry must be values of
the shear tangent and yi values must be values of the
engineering shear stress . (Integer >0 or blank)
Pure Shear Data Table identification number of a TABLES1 entry that contains
pure shear data to be used in the estimation of the material
constants Aij. xi and yi values in the TABLES1 entry must be
stretch ratios and values of the nominal stress .
is the current length, F is the current force, and are the
initial length and cross-sectional area, respectively in the
1-direction. (Integer >0 or blank)
Pure Volum. Compress. Data Table identification number of a TABLES1 entry that contains
pure volumetric compression data to be used in the estimation of
the material constants Di. xi values in the TABLES1 entry must
be values of the volume ratio where is the
stretch ratio in all three directions; yi values must be values of the
pressure, assumed positive in compression. (Integer >0 or blank)
Hyperelastic -Mooney-Rivlin Description
Data Type Coefficient
Bulk Modulus (K) Defines the Bulk Modulus.
Density Defines the mass density which is an optional property.
Destort. Def. Coeff. A10, A01 Strain energy densities as a function of the strain invariants in the
material. May vary with temperature via a material table. This
option consolidates several of the Marc hyperelastic material
models.
Data Type Test Data
Hyperelastic -General
Hyperelastic
Description
l l
0

F A
0
l l
0
A
0
Pr
0

2
2t
0
r
0
t
0
,

F A
0

1
l l
0
= F A
0
l
l
0
A
0
J
3
= l l
0
=
97
Model Generation
Material
Ogden
Bulk Modulus (K) Defines the Bulk Modulus.
Density Defines the mass density which is an optional property.
Simple Tension/Compress. Data Table identification number of a TABLES1 entry that contains
simple tension-compression data to be used in the estimation of
the material constants Cij, , , and . The x-values in the
TABLES1 entry must be stretch ratios and y-values must be
values of the engineering stress . is the initial length and
is the initial cross-sectional area.
Equivalent Tension Data Table identification number of a TABLES1 entry that contains
equibiaxial tension data to be used in the estimation of the
material constants Cij, , , and . The x-values in the
TABLES1 entry must be stretch ratios and
y-values must be values of the engineering stress . is the
initial length and & is the initial cross-sectional area. (Integer >0
or blank)
Simple Shear Data Table identification number of a TABLES1 entry that contains
simple shear data to be used in the estimation of the material
constants Cij or , , and . The x-values in the TABLES1
entry must be values of the shear strain and y-values must be
values of the engineering shear stress. (Integer >0 or blank)
Pure Shear Data Table identification number of a TABLES1 entry that contains
pure shear data to be used in the estimation of the material
constants Cij, , , and . The x and y values in the
TABLES1 entry must be stretch ratios and the values
of the nominal stress . and are the initial length and
cross-sectional area, respectively, in the l-direction.
Pure Volum. Compress. Data Table identification number of a TABLES1 entry that contains
pure volumetric compression data to be used in the estimation of
the material constant K. The x-values in the TABLES1 entry must
be values of the volume ration where is the
stretch ratio in all three directions; y-values must be values of the
pressure, assumed positive in compression.
Hyperelastic-Ogden Description
Data Type Coefficient
Bulk Modulus (K) Defines the Bulk Modulus.
Density Defines the material mass density.
Hyperelastic -Mooney-Rivlin Description

k
o
k
|
k
l l
0

F A
0
l
0
A
0

k
o
k
|
k
l l
0

F A
0
l
0

k
o
k
|
k

k
o
k
|
k

1
l l
0
=
F A
0
l
0
A
0
J
3
= l l
0
=
Modeling Guide
Material
98
Number of terms The number of terms in the Ogden expression. There can be from
oen to five terms.
Ogden Terms, Modulus (k)
in the Ogden equation.
Ogden Terms, Deviatoric
Exponent (k)
in the Ogden equation.
Data Type Test Data
Bulk Modulus (K) Defines the Bulk Modulus.
Density Defines the mass density which is an optional property.
Number of terms The number of terms in the Ogden expression. There can be from
one to five terms.
Simple Tension/Compress. Data Table identification number of a TABLES1 entry that contains
simple tension-compression data to be used in the estimation of
the material constants Cij, , , and . The x-values in the
TABLES1 entry must be stretch ratios and y-values must be
values of the engineering stress . is the initial length and
is the initial cross-sectional area.
Equivalent Tension Data Table identification number of a TABLES1 entry that contains
equibiaxial tension data to be used in the estimation of the
material constants Cij, , , and . The x-values in the
TABLES1 entry must be stretch ratios and
y-values must be values of the engineering stress . is the
initial length and & is the initial cross-sectional area.
Simple Shear Data Table identification number of a TABLES1 entry that contains
simple shear data to be used in the estimation of the material
constants Cij or , , and . The x-values in the TABLES1
entry must be values of the shear strain and y-values must be
values of the engineering shear stress. (Integer >0 or blank)
Pure Shear Data Table identification number of a TABLES1 entry that contains
pure shear data to be used in the estimation of the material
constants Cij, , , and . The x and y values in the
TABLES1 entry must be stretch ratios and the values
of the nominal stress . and are the initial length and
cross-sectional area, respectively, in the l-direction.
Pure Volum. Compress. Data Table identification number of a TABLES1 entry that contains
pure volumetric compression data to be used in the estimation of
the material constant K. The x-values in the TABLES1 entry must
be values of the volume ration where is the
stretch ratio in all three directions; y-values must be values of the
pressure, assumed positive in compression.
Hyperelastic-Ogden Description

k
o
k

k
o
k
|
k
l l
0

F A
0
l
0
A
0

k
o
k
|
k
l l
0

F A
0
l
0

k
o
k
|
k

k
o
k
|
k

1
l l
0
=
F A
0
l
0
A
0
J
3
= l l
0
=
99
Model Generation
Material
Hyperfoam
Hyperelastic-Hyperfoam Description
Data Type Coefficient
Bulk Modulus (K) Defines the Bulk Modulus.
Density Defines the material mass density.
Number of terms The number of terms in the Ogden expression. There can be from
one to five terms.
Hyperfoam Terms, Modulus (k)
in the Foam equation.
Hyperfoam Terms, Deviatoric
Exponent (k)
in the Foam equation.
Hyperfoam Terms, Volumetric
Exponent (k)
in the Foam equation.
Data Type Test Data
Bulk Modulus (K) Defines the Bulk Modulus.
Density Defines the mass density which is an optional property.
Number of terms The number of terms in the Ogden expression. There can be from
one to five terms.
Simple Tension/Compress. Data Table identification number of a TABLES1 entry that contains
simple tension-compression data to be used in the estimation of the
material constants Cij, , , and . The x-values in the
TABLES1 entry must be stretch ratios and y-values must be
values of the engineering stress . is the initial length and
is the initial cross-sectional area.
Equivalent Tension Data Table identification number of a TABLES1 entry that contains
equibiaxial tension data to be used in the estimation of the material
constants Cij, , , and . The x-values in the TABLES1 entry
must be stretch ratios and y-values must be values of the
engineering stress . is the initial length and & is the initial
cross-sectional area.
u
k
o
k
|
k

k
o
k
|
k
l l
0

F A
0
l
0
A
0

k
o
k
|
k
l l
0

F A
0
l
0
Modeling Guide
Material
100
Arruda-Boyce
Gent
Simple Shear Data Table identification number of a TABLES1 entry that contains
simple shear data to be used in the estimation of the material
constants Cij or , , and . The x-values in the TABLES1
entry must be values of the shear strain and y-values must be values
of the engineering shear stress. (Integer >0 or blank)
Pure Shear Data Table identification number of a TABLES1 entry that contains pure
shear data to be used in the estimation of the material constants Cij,
, , and . The x and y values in the TABLES1 entry must be
stretch ratios and the values of the nominal stress .
and are the initial length and cross-sectional area,
respectively, in the l-direction.
Pure Volum. Compress. Data Table identification number of a TABLES1 entry that contains pure
volumetric compression data to be used in the estimation of the
material constant K. The x-values in the TABLES1 entry must be
values of the volume ration where is the stretch
ratio in all three directions; y-values must be values of the pressure,
assumed positive in compression.
Hyperelastic-Arruda-Boyce Description
Data Type Coefficient
Bulk Modulus (K) Defines the Bulk Modulus.
Density This defines the material mass density.
Arruda-Boyce Material Const.
Number of Statistic Links
NKT Chain density times Boltzmann constant times temperature.
Chain Length Average chemical chain cross length.
Coefficient of Thermal
Expansion
Defines the instantaneous coefficient of thermal expansion. This
property is optional. May vary with temperature via a material table.
Hyperelastic-Gent Description
Data Type Coefficient
Bulk Modulus Defines the Bulk Modulus.
Density This defines the material mass density.
Hyperelastic-Hyperfoam Description

k
o
k
|
k

k
o
k
|
k

1
l l
0
= F A
0

l
0
A
0
J
3
= l l
0
=
101
Model Generation
Material
Tensile Modulus Defines standard tension modulus (E).
Maximum 1st Invariant
Defines .
Data Type Test Data
Bulk Modulus Defines the Bulk Modulus.
Density This defines the material mass density.
Tensile Modulus Defines standard tension modulus (E).
Maximum 1st Invariant
Defines .
Simple Tension/Compress. Data Table identification number of a TABLES1 entry that contains
simple tension-compression data to be used in the estimation of the
material constants Cij, , , and . The x-values in the
TABLES1 entry must be stretch ratios and y-values must be
values of the engineering stress . is the initial length and
is the initial cross-sectional area.
Equivalent Tension Data Table identification number of a TABLES1 entry that contains
equibiaxial tension data to be used in the estimation of the material
constants Cij, , , and . The x-values in the TABLES1 entry
must be stretch ratios and y-values must be values of the
engineering stress . is the initial length and & is the initial
cross-sectional area.
Simple Shear Data Table identification number of a TABLES1 entry that contains
simple shear data to be used in the estimation of the material
constants Cij or , , and . The x-values in the TABLES1
entry must be values of the shear strain and y-values must be values
of the engineering shear stress. (Integer >0 or blank)
Pure Shear Data Table identification number of a TABLES1 entry that contains pure
shear data to be used in the estimation of the material constants Cij,
, , and . The x and y values in the TABLES1 entry must be
stretch ratios and the values of the nominal stress .
and are the initial length and cross-sectional area,
respectively, in the l-direction.
Pure Volum. Compress. Data Table identification number of a TABLES1 entry that contains pure
volumetric compression data to be used in the estimation of the
material constant K. The x-values in the TABLES1 entry must be
values of the volume ration where is the stretch
ratio in all three directions; y-values must be values of the pressure,
assumed positive in compression.
Hyperelastic-Gent Description
I
1
*
I
1
*
I
1
3 =
I
1
*
I
1
*
I
1
3 =

k
o
k
|
k
l l
0

F A
0
l
0
A
0

k
o
k
|
k
l l
0

F A
0
l
0

k
o
k
|
k

k
o
k
|
k

1
l l
0
= F A
0

l
0
A
0
J
3
= l l
0
=
Modeling Guide
Material
102
Viscoelastic
The material models discussed in previous sections are considered to be time independent. However,
rubber materials often show a rate-dependent behavior and can be modeled as viscoelastic materials.
Viscoelasticity can be applied:
To determine the current state of deformation based on the entire time history of loading.
To characterize small strain and large strain problems.
With other material options for linear elastic response (small strain) and hyperelastic response
(large strain).
To include temperature dependencies.
For isotropic, anisotropic, and incompressible materials.
Small Strain Viscoelasticity
In the stress relaxation form, the constitutive relation can be written as a hereditary integral formulation
(0-66)
The functions are called stress relaxation functions. They represent the response to a unit applied
strain and have characteristic relaxation times associated with them. The relaxation functions for
materials with a fading memory can be expressed in terms of Prony or exponential series.
(0-67)
in which is a tensor of amplitudes and is a positive time constant (relaxation time). In the
current implementation, it is assumed that the time constant is isotropic. In Equation (0-67),
represents the long term modulus of the material.
The short term moduli (describing the instantaneous elastic effect) are then given by
(0-68)
The stress can now be considered as the summation of the stresses in a generalized Maxwell model
(Figure 0-24)
o
ij
t ( ) G
ijkl
t t ( )
dc
kl
t ( )
dt
------------------dt
0
t
}
G
ijkl
t ( )c
kl
0 ( ) + =
G
ijkl
G
ijkl
t ( ) G
ijkl

G
ijkl
n
exp t
n
( )
n 1 =
N

+ =
G
ijkl
n

n
G
ijkl

G
ijkl
0
G
ijkl
0 ( ) G
ijkl

G
ijkl
n
n 1 =
N

+ = =
103
Model Generation
Material
(0-69)
where
(0-70)
(0-71)
Figure 0-24 The Generalized Maxwell or Stress Relaxation Form
For integration of the constitutive equation, the total time interval is subdivided into a number of
subintervals ( ) with time-step . A recursive relation can now be derived
expressing the stress increment in terms of the values of the internal stresses at the start of the
interval. With the assumption that the strain varies linearly during the time interval h, we obtain the
increment stress-strain relation as
(0-72)
where
(0-73)
and
(0-74)
o
ij
t ( ) o
ij

t ( ) o
ij
n
t ( )
n 1 =
N

+ =
o
ij

G
ijkl

c
kl
t ( ) =
o
ij
n
G
ijkl
n
exp t t ( )
n
| |
dc
kl
t ( )
dt
------------------dt
0
t
}
=
E
1
q
1
q
1
c
E
0
t
i
=q
i
/E
i
E
2
q
2
q
2
E
i
q
i
q
i
q
E
t
m 1
t
m
, h t
m
t
m 1
=
o
ij
n
Ao
ij
t
m
( ) G
ijkl

|
n
h ( )G
ijkl
n
n 1 =
N

+ Ac
kl
o
n
h ( )o
ij
n
t
m
h ( )
n 1 =
N

=
o
n
h ( ) 1 exp h
n
( ) =
|
n
h ( ) o
n
h ( )
n
h =
Modeling Guide
Material
104
In MSC Nastran Implicit Nonlinear, the incremental equation for the total stress is expressed in terms of
the short term moduli (See Equation (0-68)).
(0-75)
Note that the set of equations given by Equation (0-75) can directly be used for both anisotropic and
isotropic materials.
Isotropic Viscoelastic Material
For an isotropic viscoelastic material, MSC Nastran Implicit Nonlinear assumes that the deviatoric and
volumetric behavior are fully uncoupled and that the behavior can be described by a time dependent shear
and bulk modules. The bulk moduli is generally assumed to be time independent; however, this is an
unnecessary restriction of the general theory.
Both the shear and bulk moduli can be expressed in a series
(0-76)
(0-77)
with short term values given by
(0-78)
(0-79)
Let the deviatoric and volumetric component matrices and be given by
Ao
ij
t
m
( ) G
ijkl
0
1 |
n
h ( ) { }G
ijkl
n
n 1 =
N

Ac
kl
t
m
( ) o
n
h ( )o
ij
n
n 1 =
N

= t
m
h ( )
G t ( ) G

G
n
exp t
d
n

\ .
| |
n 1 =
N

+ =
K t ( ) K

K
n
exp t
v
n

\ .
| |
n 1 =
N

+ =
G
0
G

G
n
n 1 =
N

+ =
K
0
K

K
n
n 1 =
N

+ =
t
d
t
v
105
Model Generation
Material
The increment set of equations is then given by
(0-80)
and
(0-81)
Note that the deviatoric and volumetric response are fully decoupled.
Note that the algorithm is exact for linear variations of the strain during the increment. The algorithm is
implicit; hence, for each change in time-step, a new assembly of the stiffness matrix is required.
Anisotropic Viscoelastic Material
Equation (0-75) can be used for the analysis of anisotropic viscoelastic materials.
t
d
4 3 2 3 2 3 0 0 0
2 3 4 3 2 3 0 0 0
2 3 2 3 4 3 0 0 0
0 0 0 1 0 0
0 0 0 0 1 0
0 0 0 0 0 1
=
t
v
1 1 1 0 0 0
1 1 1 0 0 0
1 1 1 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
=
Ao t
m
( ) G
0
1 |
d
n
h ( ) | |G
n
n 1 =
N
d

)

`


t
d
Ac t
m
( ) =
o
d
n
h ( )o
d
n
t
m
h ( )
n 1 =
N
d

o
v
n
h ( )o
v
n
t
m
h ( )
n 1 =
N
v

K
0
1 |
v
n
h ( ) | |K
n
n 1 =
N
v

)

`


t
v
Ac t
m
( )
Ao
d
n
t
m
( ) |
d
n
h ( )G
n
t
d
Ac t
m
( ) o
d
n
h ( )o
d
n
t
m
h ( ) =
Ao
v
n
t
m
( ) |
v
n
h ( )K
n
t
v
Ac t
m
( ) o
v
n
h ( )o
v
n
t
m
h ( ) =
Modeling Guide
Material
106
Also, a complete set of moduli (21 components) can be specified in the HOOKVI user subroutine.
Referencing a local coordinate system or use of the ORIENT user subroutine can be used to define a
preferred orientation both for the short time moduli and the amplitude functions .
Incompressible Isotropic Viscoelastic Materials
Incompressible elements in MSC Nastran Implicit Nonlinear allow the analysis of incompressible and
nearly incompressible materials in plane strain, axisymmetric and three-dimensional problems. The
incompressibility of the element is simulated through the use of an perturbed Lagrangian variational
principle based on the Herrmann formulation.
The constitutive equation for a material with no time dependence in the volumetric behavior can be
expressed as
(0-82)
(0-83)
The hydrostatic pressure term is used as an independent variable in the variational principle. The
Herrmann pressure variable is now defined in the same way as in the formulation for time independent
elastic materials.
(0-84)
The constitutive Equation (0-82) and Equation (0-83) can then be rewritten
(0-85)
where
(0-86)
G
ijkl
0
G
ijkl
n
Ao
ij
t
m
( ) 2 G
ijkl
0
1 |
n
h ( ) | |G
ijkl
n
n 1 =
N

)

`


Ac
kl
t
m
( )
1
3
---Ac
pp
t
m
( )o
kl
=
o
n
h ( ) o'
ij
( )
n
t
m
( )
n 1 =
N

1
3
---o
kk
o
ij
+
Ao
pp
t
m
( ) 3K
0
Ac
pp
t
m
( ) =
H
o
pp
2G
0
1 v
0
+ ( )
------------------------------- =
Ao
ij
t
m
( ) 2G
e
Ac
ij
v
-
Ho
ij
+ ( ) o
n
h ( ) o'
ij
( )
n
t
m
h ( )
n 1 =
N

=
G
e
G
0
1 |
n
h ( )G
n
| |
n 1 =
N

=
107
Model Generation
Material
(0-87)
Large Strain Viscoelasticity
For an elastomeric time independent material, the constitutive equation is expressed in terms of an
energy function . For a large strain viscoelastic material, Simo generalized the small strain
viscoelasticity material behavior to a large strain viscoelastic material. The energy functional then
becomes
(0-88)
where are the components of the Green-Lagrange strain tensor, internal variables and the
elastic strain energy density for instantaneous deformations. In MSC Nastran Implicit Nonlinear, it is
assumed that , meaning that the energy density for instantaneous deformations is given by the
third order J ames Green and Simpson form or the Ogden form.
The components of the second Piola-Kirchhoff stress then follow from
(0-89)
The energy function can also be written in terms of the long term moduli resulting in a different set of
internal variables
(0-90)
where is the elastic strain energy for long term deformations. Using this energy definition, the
stresses are obtained from
(0-91)
Observing the similarity with the equations for small strain viscoelasticity the internal variables can be
obtained from a convolution expression
v
-
G
0
1 v
0
+ ( ) G
e
1 2v
0
( )
3G
e
---------------------------------------------------------------- =
W
E
ij
Q
ij
n
( )
0
E
ij
( ) Q
ij
n
E
ij
n 1 =
N

I
n
Q
ij
n
( )
n 1 =
N

+ =
E
ij
Q
ij
n

0
W =
S
ij
c
cE
ij
---------
c
0
cE
ij
--------- Q
ij
n
n 1 =
N

= =
T
ij
n
E
ij
T
ij
n
, ( )

E
ij
( ) T
ij
n
E
ij
n 1 =
N

+ =

S
ij
c

E ( )
cE
ij
------------------- T
ij
n
n 1 =
N

+ =
Modeling Guide
Material
108
(0-92)
where are internal stresses obtained from energy functions.
(0-93)
Let the total strain energy be expressed as a Prony series expansion
(0-94)
If, in the energy function, each term in the series expansion has a similar form, Equation (0-94) can
berewritten
(0-95)
where is a scalar multiplier for the energy function based on the short termvalues.
The stress-strain relation is now given by
(0-96)
(0-97)
(0-98)
Analogue to the derivation for small strain viscoelasticity, a recursive relation can be derived expressing
the stress increment in terms of values of the internal stresses at the start of the increment.
T
ij
n
S

ij
n
0
t
}
t ( )exp t t ( )
n
| |dt =
S
ij
n
S
ij
n
c
n
cE
ij
--------- =

n
exp t
n
( )
n 1 =
N

+ =

o
n

0
exp t
n
( )
n 1 =
N

+ =
o
n
S
ij
t ( ) S
ij

t ( ) T
ij
n
n 1 =
N

t ( ) + =
S
ij
c

cE
ij
---------- 1 o
n
n 1 =
N

\ .
|
|
| |

c
0
cE
ij
--------- = =
T
ij
n
o
n
S
ij
0
t ( )exp t t ( )
n
| |dt
0
t
}
=
109
Model Generation
Material
The equations are reformulated in terms of the short time values of the energy function
(0-99)
(0-100)
It is assumed that the viscoelastic behavior in MSC Nastran Implicit Nonlinear acts only on the
deviatoricbehavior.
Viscoelastic Models in MSC Nastran
MSC Nastran Implicit Nonlinear has two models that represent viscoelastic materials. The first can be
defined as a Kelvin-Voigt model. The latter is a general hereditary integral approach.
Kelvin-Voigt Model
The Kelvin model allows the rate of change of the inelastic strain to be a function of the total stress and
previous strain.
The Kelvin material behavior (viscoelasticity) is modeled by assuming an additional creep strain ,
governed by
(0-101)
where and may be defined in the user subroutine CRPVIS and the total strain is
(0-102)
(0-103)
(0-104)
(0-105)
(0-106)
AS
ij
t
m
( ) 1 1 |
n
h ( )
n 1 =
N

o
n
\ .
|
|
| |
S
ij
0
t
m
( ) S
ij
0
t
m
h ( ) { } =
o
n
S
ij
n
t
m
h ( )
n 1 =
N

AS
ij
t
m
( ) |
n
h ( )o
n
S
ij
0
t
m
( ) S
ij
n
t
m
h ( ) | | o
n
h ( )S
ij
n
t
m
h ( ) =
c
ij
k
d
dt
-----c
ij
k
A
ijkl
o'
kl
B
ijkl
c
kl
k
=
A B
c
ij
c
ij
e
c
ij
p
c
ij
c
c
ij
k
c
ij
th
+ + + + =
c
ij
th
=thermal strain components
c
ij
e
=elastic strain components (instantaneous response)
c
ij
p
=plastic strain components
c
ij
c
=
creep strains defined via the CRPLAW and VSWELL user subroutines
Modeling Guide
Material
110
(0-107)
The CRPVIS user subroutine is called at each integration point of each element when the Kelvin model
is used.
Use the NLPARM option and set a nonzero time increment to define the time step and to set the tolerance
control for the maximum strain in any increment.
This option allows Maxwell models to be included in series with the Kelvinmodel.
Hereditary Integral Model
The stress-strain equations in viscoelasticity are not only dependent on the current stress and strain state
(as represented in the Kelvin model), but also on the entire history of development of these states. This
constitutive behavior is most readily expressed in terms of hereditary or Duhamel integrals. These
integrals are formed by considering the stress or strain build-up at successive times. Two equivalent
integral forms exist: the stress relaxation form and the creep function form. In MSC Nastran Implicit
Nonlinear, the stress relaxation form is used.
The viscoelasticity option in MSC Nastran Implicit Nonlinear can be used for both the small strain and
large strain Mooney, Ogden, Arruda Boyce, and Gent material stress-relaxation problems. A description
of these models is as follows:
Experimental Determination of Viscoelastic Material Parameters
The free energy function versus time data being used for large strain viscoelasticity can be generated by
fitting experimental data provided the following two tests are done:
1. Standard quasi-static tests (tensile, planar-shear, simple-shear, equi-biaxial tension, volumetric)
to determine the elastomer free energy constants.
2. Standard relaxation tests to obtain stress versus time.
Temperature Dependence of Viscoelastic Materials
The rate processes in many viscoelastic materials is known to be highly sensitive to temperature changes.
Such temperature-dependent properties cannot be neglected in the presence of any appreciable
temperature variation. For example, there is a large class of polymers which are adequately represented
by linear viscoelastic laws at uniform temperature. These polymers exhibit an approximate translational
shift of all the characteristic response functions with a change of temperature, along a logarithmic time
axis. This shift occurs without a change of shape. These temperature-sensitive viscoelastic materials are
characterized as Thermo-Rheologically Simple.
A reduced or pseudo time can be defined for the materials of this type and for a given temperature
field. This new parameter is a function of both time and space variables. The viscoelastic law has the
same form as one at constant temperature in real time. If the shifted time is used, however, the
transformed viscoelastic equilibrium and compatibility equations are not equivalent to the corresponding
elastic equations.
c
ij
k
=Kelvin model strain components as defined above
W
0
111
Model Generation
Material
In the case where the temperature varies with time, the extended constitutive law implies a nonlinear
dependence of the instantaneous stress state at each material point of the body upon the entire local
temperature history. In other words, the functionals are linear in the strains but nonlinear in
thetemperature.
The time scale of experimental data is extended for Thermo-Rheologically Simple materials. All
characteristic functions of the material must obey the same property. The shift function is a basic
property of the material and must be determined experimentally. Asa consequence of the shifting of the
mechanical properties data parallel to the time axis, the values of the zero and infinite frequency complex
moduli do not change due to shifting. Hence, elastic materials with temperature-dependent
characteristics neither belong to nor are consistent with the above hypothesis for the class of
Thermo-Rheologically Simple viscoelastic solids.
In addition to the Thermo-Rheologically Simple material behavior variations of initial stress-strain
moduli , the temperature of the other mechanical properties (coefficient of thermal expansion, etc.)
due to changes in temperature can be specified.
Note, however, that only the instantaneous moduli are effected. Hence, the long term moduli given by
(0-108)
can easily become negative if the temperature effects are not defined properly.
The effect of temperature, u, on the material behavior is introduced through the dependence of the elastic
modulus, G, on temperature, and through a reduced time concept:
(0-109)
where G=G(u), and xi(t) is the reduced time, defined by
(0-110)
where A(u(t)) is the shift function at time t. Often the shift function is approximated by the Williams
Landell Ferry (WLF) form:
(0-111)
where C
1
, C
2
and u
0
are constants (u
0
is the glassy transition temperature).
G
ijkl
0
G
ijkl

G
ijkl
0
t ( ) G
ijkl
n
n 1 =
N

=
t G g t ( ) s ( ) ( )

s ( )ds
0
t
}
+
\ .
|
|
| |
=
t ( )
ds
A u s ( ) ( )
-------------------
0
t
}
=
A ( ) log
C
1
u u
0
( )
C
2
u u
0
( ) +
-------------------------------- =
Modeling Guide
Material
112
Narayanaswamy Model
The annealing of flat glass requires that the residual stresses be of an acceptable magnitude, while the
specification for optical glass components usually includes a homogenous refractive index. The design
of heat treated processes can be accomplished using the Narayanaswamy model. This allows you to study
the time dependence of physical properties (for example, volumes) of glass subjected to a change in
temperature.
For more information pertaining to the Narayanaswamy Model, see Marc Volume A: Theory and User
Information, Chapter 7 Material Library.
Specifying Viscoelastic Material Entries
The viscoelastic MATVE and MATTVE material options are provided for cases where dissipative losses
caused by viscous (internal friction) effects in materials must be modeled. For time domain analysis,
this option is used with an elastic model to define classical linear, small strain, viscoelastic behavior, or
with hyperelastic or foam models to define finite linear, large deformation, viscoelastic behavior. As
described in the previous section, viscoelastic relaxation data can be fit using the experimental data fitting
(EDF) capability available in SimXpert.
References
MATVE in the MSC NASTRAN QRG.
MATTVE in the MSC NASTRAN QRG.
SimXpert Materials Application Input Data
To define a viscoelastic material in SimXpert:
1. Select Materials and Properties > Isotropic > Advanced > Add Constitutive Model > Time
Dependent.
2. Enter parameter values.
3. Click OK.
This input data creates the viscoelastic options. All inputs must have the same number of time points (at
the same times) in the referenced fields. The following equations may be useful when creating the Prony
series for the bulk and shear moduli: .
Entry Description
MATVE Specifies isotropic visco-elastic material properties to be used for quasi-static
or dynamic analysis in MSC Nastran Implicit Nonlinear (SOL 600) only.
MATTVE Specifies temperature-dependent visco-elastic material properties in terms of
Thermo-Rheologically Simple behavior to be used for quasi-static or transient
dynamic analysis in MSC Nastran Implicit Nonlinear (SOL 600) only.
K E 3 1 2v ( ) ( ) G E 2 1 v + ( ) ( ) = =
113
Model Generation
Material
Viscoelastic material models support the following material data.
Isotropic -Viscoelastic Description
Shift Function No Function
Thermal Exp. Solid Coeff. Solid coefficient of thermal expansion
Thermal Exp. Liquid Coeff. Liquid coefficient of thermal expansion
Deviatoric Terms of Prony Series Deviatoric Scale Factor -- Multiplier (scale factor) for deviatoric
behavior in Prony series (Real >0., Default =0).
Deviatoric Time Constant -- Defines time constants for deviatoric
behavior in Prony series (Real >0., Default =0).
Volumetric Terms of Prony Series Volumetric Scale Factor -- Multiplier (scale factor) for volumetric
behavior in Prony series (Real >0., Default =0.).
Volumetric Time Constant -- Defines time constants for
volumetric behavior in Prony series (Real >0., Default =1000.).
Shift Function Williams-Landel-Ferry
Thermal Exp. Solid Coeff. Solid coefficient of thermal expansion
Thermal Exp. Liquid Coeff. Liquid coefficient of thermal expansion
Deviatoric Terms of Prony Series Deviatoric Scale Factor -- Multiplier (scale factor) for deviatoric
behavior in Prony series (Real >0., Default =0).
Deviatoric Time Constant -- Defines time constants for deviatoric
behavior in Prony series (Real >0., Default =0).
Volumetric Terms of Prony Series Volumetric Scale Factor -- Multiplier (scale factor) for volumetric
behavior in Prony series (Real >0., Default =0.).
Volumetric Time Constant -- Defines time constants for
volumetric behavior in Prony series (Real >0., Default =1000.).
Reference Temperature Reference or glass transition temperature.
Shift Function Coeff. 1 Constant A1.
Shift Function Coeff. 2 Constant A2.
Shift Function Power Series Expansion
Thermal Exp. Solid Coeff. Solid coefficient of thermal expansion
Thermal Exp. Liquid Coeff. Liquid coefficient of thermal expansion
Deviatoric Terms of Prony Series Deviatoric Scale Factor -- Multiplier (scale factor) for deviatoric
behavior in Prony series (Real >0., Default =0).
Deviatoric Time Constant -- Defines time constants for deviatoric
behavior in Prony series (Real >0., Default =0).
Volumetric Terms of Prony Series Volumetric Scale Factor -- Multiplier (scale factor) for volumetric
behavior in Prony series (Real >0., Default =0.).
Volumetric Time Constant -- Defines time constants for
volumetric behavior in Prony series (Real >0., Default =1000.).
Modeling Guide
Material
114
Inelastic
Most materials of engineering interest initially respond elastically. Elastic behavior means that the
deformation is fully recoverable, so that, when the load is removed, the specimen returns to its original
shape. If the load exceeds some limit (the yield load), the deformation is no longer fully recoverable.
Some parts of the deformation will remain when the load is removed as, for example, when a paper clip
is bent too much, or when a billet of metal is rolled or forged in a manufacturing process. Plasticity
Reference Temperature Reference or glass transition temperature.
Number of Power Terms Number of coefficients in the power series representation.
Shift Function Coeff. 0 Coefficient C0 of the shift function.
Orthotropic-Viscoelastic Description
Shift Function No Function
Prony Terms Dev. Time Constant -- Defines time constants for deviatoric
behavior in Prony series (Real >0., Default =0)
Youngs Modulus i
Poissons Ratio ij
Shear Modulus ij
Shift Function Williams-Landel-Ferry
Prony Terms Dev. Time Constant -- Defines time constants for deviatoric
behavior in Prony series (Real >0., Default =0)
Youngs Modulus i
Poissons Ratio ij
Shear Modulus ij
Reference Temperature Reference or glass transition temperature.
Shift Function Coeff. 1 Constant A1.
Shift Function Coeff. 2 Constant A2.
Shift Function Power Series Expansion
Prony Terms Dev. Time Constant -- Defines time constants for deviatoric
behavior in Prony series (Real >0., Default =0)
Youngs Modulus i
Poissons Ratio ij
Shear Modulus ij
Reference Temperature Reference or glass transition temperature.
Number of Power Terms Number of coefficients in the power series representation.
Shift Function Coeff. 0 Coefficient C0 of the shift function.
Isotropic -Viscoelastic Description
115
Model Generation
Material
theories model the materials mechanical response as it undergoes such nonrecoverable deformation in
a ductile fashion. The theories have been developed most intensively for metals, but they are applied to
soils, concrete, rock, ice, and so on. These materials behave in very different ways (for example, even
large values of pure hydrostatic pressure cause very little inelastic deformation in metals, but quite small
hydrostatic pressure may cause a significant, non-recoverable volume change in a soil sample), but the
fundamental concepts of plasticity theories are sufficiently general that models based on these concepts
have been successfully developed for a wide range of materials. A number of these plasticity modes are
available in the MSC Nastran Implicit Nonlinear material library.
In nonlinear material behavior, the material parameters depend on the state of stress. Up to the
proportional limit (the point at which linearity in material behavior ceases) the linear elastic formulation
for the behavior can be used. Beyond that point, and especially after the onset of yield, nonlinear
formulations are required. In general, two ingredients are required to ascertain material behavior:
1. An initial yield criterion to determine the state of stress at which yielding is considered to begin
2. Mathematical rules to explain the post-yielding behavior.
There are two major theories of plastic behavior that address these criterion differently. In the first, called
deformation theory, the plastic strains are uniquely defined by the state of stress. The second theory,
called flow or incremental theory, expresses the increments of plastic strain (irrecoverable strains) as
functions of the current stress, the strain increments, and the stress increments. Incremental theory is
more general and can be adapted in its particulars to fit a variety of material behaviors. The plasticity
models in MSC Nastran Implicit Nonlinear are incremental theories, in which the mechanical strain
rate is decomposed into an elastic part and a plastic (inelastic) part through various assumed flow rules.
The incremental plasticity models are formulated in terms of:
A yield surface, which generalizes the concept of yield load into a test function which can be
used to determine if the material will respond purely elastically at a particular state of stress,
temperature, and so on.
A flow rule that defines the inelastic deformation that must occur if the material point is no
longer responding purely elastically.
and some evolution laws that define the hardening - the way in which the yield and/or flow
definitions change as inelastic deformation occurs.
The models also need an elasticity definition, to deal with the recoverable part of the strain models divide
into those that are rate-dependent and those that are rate-independent.
MSC Nastran Implicit Nonlinear includes the following models of inelastic behavior.
Metal Plasticity (von Mises or Hill)
ORNL (Oak Ridge National Laboratory) - characterizes creep behavior and cyclic loading
effects on stainless steel materials.
Porous Metal Plasticity (Gurson) - includes effects of hydrostatic pressure and failure processes
in ductile materials.
Pressure-Dependent models - models the behavior of granular (soil and rock) materials or
polymers, in which the yield behavior depends on the equivalent pressure stress.
Modeling Guide
Material
116
Linear Mohr-Coulomb
Parabolic Mohr-Coulomb
Buyukozturk Concrete
Yield Conditions
The yield stress of a material is a measured stress level that separates the elastic and inelastic behavior of
the material. The magnitude of the yield stress is generally obtained from a uniaxial test. However, the
stresses in a structure are usually multi-axial. A measurement of yielding for the multi-axial state of stress
is called the yield condition. Depending on how the multi-axial state of stress is represented, there can be
many forms of yield conditions. For example, the yield condition can be dependent on all stress
components, on shear components only, or on hydrostatic stress. A number of yield conditions are
available in MSC Nastran Implicit Nonlinear, and are discussed in thissection.
Metal Plasticity
The von Mises yield surface is widely used for plasticity in isotropic metals. It is assumed that the yield
and plastic flow describe isotropic metals at low temperatures where creep effects can be ignored.
Anisotropic metals and composite materials, can be treated by extensions of von Mises yield function, as
described in Hills yield function.
von Mises
The success of the von Mises criterion is due to the continuous nature of the function that defines this
criterion and its agreement with observed behavior for the commonly encountered ductile materials.
Thevon Mises criterion states that yield occurs when the effective (or equivalent) stress (o) equals
theyield stress (o
y
) as measured in a uniaxial test. Figure 0-25 shows the von Mises yield surface in two-
dimensional and three-dimensional stress space.
Figure 0-25 von Mises Yield Surface
For an isotropic material
Yield
Surface
Elastic
Region
(b) t-Plane
Yield
Surface
Elastic
Region
(a) Two-dimensional Stress Space
o'
3
o'
2
o'
1
o
1
o
2
117
Model Generation
Material
(0-112)
where o
1
, o
2
, and o
3
are the principal Cauchy stresses.
o can also be expressed in terms of non-principal Cauchy stresses.
(0-113)
The yield condition can also be expressed in terms of the deviatoric stressesas:
(0-114)
where is the deviatoric Cauchy stress expressed as
(0-115)
For isotropic material, the von Mises yield condition is the default condition in MSC Nastran
Implicit Nonlinear.
o o
1
o
2
( )
2
o
2
o
3
( )
2
o
3
o
1
( )
2
+ + | |
1 2
2 =
o o
x
o
y
( )
2
o
y
o
z
( )
2
o
z
o
x
( )
2
6 t
xy
2
t
yz
2
t
zx
2
+ + ( ) + + + | |
1 2
= ( ) 2
o
3
2
---o'
ij
o'
ij
=
o'
ij
o'
ij
o
ij
1
3
---o
kk
o
ij
=
Modeling Guide
Material
118
Hills Yield Function
Hills yield surface has been widely used both as a yield surface and as a failure surface for anisotropic
and composite materials. Hills yield function is a generalization of von Mises as expressed below.
(0-116)
Note the following points about Hills surface:
1. It degenerates into von Mises surface when all three direct yield stresses are equal
(Fx =Fy =Fz) and all three shear yield stresses are equal.
2. It is invariant with respect to hydrostatic stress, as is von Mises.
3. Hill's surface, unlike von Mises, is not always an ellipsoid in stress space. When it is not an
ellipsoid, it is not appropriate for use as a yield function (since it does not have an inside and an
outside, thereby dividing stress space into elastic and plastic regions).
Mohr-Coulomb Material (Hydrostatic Stress Dependence)
MSC Nastran Implicit Nonlinear includes options for elastic-plastic behavior based on a yield surface
that exhibits hydrostatic stress dependence. Such behavior is observed in a wide class of soil and rocklike
materials. These materials are generally classified as Mohr-Coulomb materials (generalized von Mises
materials). Ice is also thought to be a Mohr-Coulomb material. The generalized Mohr-Coulomb model
developed by Drucker and Prager is implemented in MSC Nastran Implicit Nonlinear. There aretwo
types of Mohr-Coulomb materials: linear and parabolic. Each is discussed on the following pages.
Linear Mohr-Coulomb Material
o
xx
Fx
--------
\ .
| |
2
o
yy
Fy
--------
\ .
| |
2
o
zz
Fz
-------
\ .
| |
2
+ +
1
F
x
2
------
1
F
y
2
------
1
F
z
2
------ +
\ .
|
| |
o
xx
o
yy

1
F
x
2
------
1
F
y
2
------
1
F
z
2
------ +
\ .
|
| |
o
xx
o
zz

1
F
x
2
------
1
F
y
2
------
1
F
z
2
------ + +
\ .
|
| |
o
yy
o
zz

+
t
xy
F
xy
--------
\ .
| |
2
t
yz
F
yz
-------
\ .
| |
2
t
zx
F
zx
-------
\ .
| |
2
+ + 1 =
119
Model Generation
Material
The deviatoric yield function, as shown in Figure 0-26, is assumed to be a linear function of the
hydrostatic stress.
(0-117)
where
(0-118)
(0-119)
The constants and can be related to and by
(0-120)
where is the cohesion and is the angle of friction.
Figure 0-26 Yield Envelope of Plane Strain (Linear Mohr-Coulomb Material)
Parabolic Mohr-Coulomb Material
The hydrostatic dependence is generalized to give a yield envelope which is parabolic in the case of plane
strain (see Figure 0-27).
(0-121)
f oJ
1
J
2
1 2
o
3
------- + 0 = =
J
1
o
ii
=
J
2
1
2
---o'
ij
o'
ij
=
o o c |
c
o
3 1 12o
2
( ) | |
1 2
------------------------------------------
3o
1 3o
2
( )
1 2
------------------------------- ; | sin = =
c |
Yield Envelope
c
o
|
R
o
x
+o
y
2
t
f 3J
2
3|oJ
1
+ ( )
1 2
o 0 = =
Modeling Guide
Material
120
(0-122)
where is the cohesion.
Figure 0-27 Resultant Yield Condition of Plane Strain (Parabolic Mohr-Coulomb Material
Buyukozturk Criterion (Hydrostatic Stress Dependence)
The Buyukozturk concrete plasticity model is a particular form of the generalized Drucker-Prager
plasticity model, which is developed specifically for plane stress cases by Buyukozturk. This yield
criterion, which originally has been proposed as a failure criterion, has the general form:
(0-123)
The Buyukozturk criterion reduces to the parabolic Mohr-Coulomb criterion if .
Oak Ridge National Laboratory Options
Oak Ridge National Laboratory (ORNL) has performed a large number of creep tests on stainless and
other alloy steels. It has also set certain rules that characterize creep behavior for application in the
nuclear structures. A summary of the ORNL rules on creep is discussed in Marc Volume A, Theory and
User Information. In MSC Nastran Implicit Nonlinear, the ORNL options are based on the definitions of
ORNL-TM- 3602 [1] for stainless steels and ORNL recommendations [2] for 2 1/4 Cr-1 Mo steel.
The initial yield stress should be used for the initial inelastic loading calculations for both the stainless
steels and 2 1/4 Cr-1 Mo steel. The 10th-cycle yield stress should be used for the hardened material. The
100th-cycle yield stress must be used in the following circumstances:
1. To accommodate cyclic softening of 2 1/4 Cr-1 Mo steel after many loadcycles.
2. After a long period of high temperature exposure.
3. After the occurrence of creep strain.
|
o
3 3c
2
o
2
( ) ( )
1 2
----------------------------------------- =
o
2
3 c
2 o
2
3
------
\ .
| |
=
c
c2
c
o
R
o
x
+o
y
2
o
t
f | 3oJ
1
J
1
2
3J
2
o
2
+ + =
0 =
121
Model Generation
Material
Work Hardening Rules
The work hardening rule defines the way the yield surface changes with plastic straining. A material is
said to be perfectly plastic if, upon the stress state touching the yield surface, an infinitesimal increase
in stress causes an arbitrarily large plastic strain. MSC Nastran Implicit Nonlinear models all materials
as work hardening, and treats perfectly plastic materials as a special case. Because the tangent stiffness
method is used, no difficulties arise in setting the work hardening slope equal to zero. Besides perfect
plasticity, three possibilities are provided: isotropic hardening and kinematic hardening.
The isotropic work hardening rule assumes that the center of the yield surface remains stationary in the
stress space, but that the size (radius) of the yield surface expands, due to work hardening. This type of
hardening is appropriate when the straining is the same in all directions.
For many materials, the isotropic work hardening model is inaccurate if unloading occurs (as in cyclic
loading problems). For these problems, the kinematic hardening model or the combined hardening
model represents the material better.
Isotropic, Kinematic, and Combined Hardening
The isotropic work hardening rule assumes that the center of the yield surface remains stationary in the
stress space, but that the size (radius) of the yield surface expands, due to work hardening. The change
of the von Mises yield surface is plotted in Figure 0-28b.
A review of the load path of a uniaxial test that involves both the loading and unloading of a specimen
will assist in describing the isotropic work hardening rule. The specimen is first loaded from stress free
(point 0) to initial yield at point 1, as shown in Figure 0-28a. It is then continuously loaded to point 2.
Then, unloading from 2 to 3 following the elastic slope E (Youngs modulus) and then elastic reloading
from 3 to 2 takes place. Finally, the specimen is plastically loaded again from 2 to 4 and elastically
unloaded from 4 to 5. Reverse plastic loading occurs between 5 and 6.
It is obvious that the stress at 1 is equal to the initial yield stress and stresses at points 2 and 4 are
larger than , due to work hardening. During unloading, the stress state can remain elastic (for
example, point 3), or it can reach a subsequent (reversed) yield point (for example, point 5). The isotropic
work hardening rule states that the reverse yield occurs at current stress level in the reversed direction.
Original
Hardened
Isotropic Hardening Kinematic Hardening
o
y
o
y
Modeling Guide
Material
122
Figure 0-28 Schematic of Isotropic Hardening Rule (Uniaxial Test)
Let be the stress level at point 4. Then, the reverse yield can only take place at a stress level of
(point 5).
For many materials, the isotropic work hardening model is inaccurate if unloading occurs (as in cyclic
loading problems). For these problems, the kinematic hardening model or the combined hardening model
represents the material better.
6
5
1
2
6
5
4
3
2
1
0
E
E
E
+o4
o
o
y
o4
4
3
0
(a) Loading Path
(b) von Mises Yield Surface
o'
3
o'
2
o'
1
o
4
o
4

123
Model Generation
Material
Kinematic Hardening
Under the kinematic hardening rule, the von Mises yield surface does not change in size or shape, but
the center of the yield surface can move in stress space. Figure 0-28d illustrates this condition. Zieglers
law is used to define the translation of the yield surface in the stress space.
The loading path of a uniaxial test is shown in Figure 0-28c. The specimen is loaded in the following
order: from stress free (point 0) to initial yield (point 1), 2 (loading), 3 (unloading), 2 (reloading), 4
(loading), 5 and 6 (unloading). As in isotropic hardening, stress at 1 is equal to the initial yield stress ,
and stresses at 2 and 4 are higher than , due to work hardening. Point 3 is elastic, and reverse yield
takes place at point 5. Under the kinematic hardening rule, the reverse yield occurs at the level of
, rather than at the stress level of . Similarly, if the specimen is loaded to a
higher stress level (point 7), and then unloaded to the subsequent yield point 8, the stress at point 8
is . If the specimen is unloaded from a (tensile) stress state (such as point 4 and 7),
the reverse yield can occur at a stress state in either the reverse (point 5) or the same (point 8) direction.
For many materials, the kinematic hardening model gives a better representation of loading/unloading
behavior than the isotropic hardening model. For cyclic loading, however, the kinematic hardening
model can represent neither cyclic hardening nor cyclic softening.
Combined Hardening
Figure 0-30 shows a material with highly nonlinear hardening. Here, the initial hardening is assumed to
be almost entirely isotropic, but after some plastic straining, the elastic range attains an essentially
constant value (that is, pure kinematic hardening). The basic assumption of the combined hardening
model is that such behavior is reasonably approximated by a classical constant kinematic hardening
constraint, with the superposition of initial isotropic hardening. The isotropic hardening rate eventually
decays to zero as a function of the equivalent plastic strain measured by
(0-124)
o
y
o
y
o
5
o
4
2o
y
( ) = o
4

o
7
o
8
o
7
2o
y
( ) =
c

p
c
p
dt
}
2
3
---
c

ij
p
c

ij
p
\ .
| |
1 2
dt
}
= =
FullyHardened
Pure Kinematic
Range
Combined
Hardening
Range
Initial
Elastic
Range
Initial
Yield
One-half Current
Elastic Range
o
c
3
2
do
dc
p
Kinematic Slope,
Stress
Strain
Modeling Guide
Material
124
Figure 0-29 Basic Uniaxial Tension Behavior of the Combined Hardening Model
This implies a constant shift of the center of the elastic domain, with a growth of elastic domain around
this center until pure kinematic hardening is attained. In this model, there is a variable proportion between
the isotropic and kinematic contributions that depends on the extent of plastic deformation (as measured
by ).
The work hardening data at small strains governs the isotropic behavior, and the data at large strains
( ) governs the kinematic hardening behavior. If the last work hardening slope is zero, the
behavior is the same as the isotropic hardeningmodel.
Experimental Determination of Work Hardening Slope
In a uniaxial test, the work hardening slope is defined as the slope of the stress-plastic strain curve. The
work hardening slope relates the incremental stress to incremental plastic strain in the inelastic region
and dictates the conditions of subsequent yielding. A number of work hardening rules (isotropic,
kinematic, and combined) are available in MSC Nastran Implicit Nonlinear. A description of these work
hardening rules is given below. The uniaxial stress-plastic strain curve can be represented by a piecewise
linear function or through the user subroutine WKSLP. This requires the use of MARCIN to specify the
MARC WORKHARD option.
Figure 0-30 Work hardening Slopes
c
p
c
p
1000 >
E E E E
Strain
Stress
Ac
1
p
o
Ao3
Ao2
Ao1
Ac
2
p
Ac
3
p
125
Model Generation
Material
You enter a table of yield stress, plastic strain points.
The yield stress and the work hardening data must be compatible with the procedure used in the analysis.
For small strain analyses, the engineering stress and engineering strain are appropriate. If only
PARAM,LGDISP is used, the yield stress should be entered as the second Piola-Kirchhoff stress, and
the workhard data be given with respect to plastic Green-Lagrange strains. If PARAM,LGDISP,1 or 2
are used, the yield stress must be defined as a true or Cauchy stress, and the work hardening data with
respect to logarithmic plastic strains. Engineering stress and strain may be defined and Bulk Data
parameter MRTABLS1 used to provide the program with rules to convert to the proper stress and strain
measures. See MRTABLS1 in the MMC NASTRAN QRG.
Flow Rules
Yield stress and work hardening rules are two experimentally related phenomena that characterize plastic
material behavior. The flow rule is also essential in establishing the incremental stress-strain relations for
plastic material. The flow rule describes the differential changes in the plastic strain components
as a function of the current stress state. So long as a material point is elastic, Hookes law provides a
relationship between total stress and strain. After a material becomes plastic, however, there is no longer
a unique relationship between total stress and strain. The problem then is usually solved incrementally,
following the exact loading path.
For points which are plastic, a flow rule is used to relate increments of stress to plastic strain. MSC
Nastran Implicit Nonlinear uses an associated flow rule, which prescribes that increments of plastic
strain are computed as a constant times the gradient of the yield function.
Ao
1
Ac
1
p
----------
Ao
2
Ac
2
p
----------
Ao
3
Ac
3
p
----------
0.0
Ac
1
p
Ac
1
p
Ac
2
p
+
Slope Breakpoint
Note: The data points should be based on a plot of the stress versus plastic strain for a tensile test.
The elastic strain components should not be included.
dc
p
Modeling Guide
Material
126
In other words, considering the yield function as a surface in stress space, the plastic strain increment is
a vector in the direction of the outward normal to the surface at the point where it is touched by the
stresses on the loading path.
The equation representing this is:
(0-125)
where is a constant. Writing the six equations explicitly:
(0-126)
These stress versus plastic strain equations are analogous to the stress versus total strain equations of
elasticity, where elastic strains can be computed as the gradient of a strain energy potential function,
namely;
(0-127)
Thus, the yield function F plays the role of a plastic potential. If a theory of plasticity uses something
other than the yield function as a plastic potential, a so called non-associated flow rule results. Non-
associated flow rules are not available in MSC Nastran Implicit Nonlinear.
dc
ij
p

cF
co
ij
--------- =
dc
xx
p
cF
co
xx
----------- =
dc
yy
p
cF
co
yy
----------- =
dc
zz
p
cF
co
zz
---------- =
dc
xy
p
cF
ct
xy
---------- =
dc
yz
p
cF
ct
yz
---------- =
dc
zx
p
cF
ct
xz
---------- =
dc
ij
cU
co
ij
--------- =
127
Model Generation
Material
For the von Mises and modified Hill yield functions programmed in MSC Nastran Implicit Nonlinear,
the derivatives in the yield function are obtained simply by differentiating with respect to individual
components of stress. For example, for the modified Hill function, we have:
(0-128)
The constant in these flow rule equations is evaluated automatically by MSC Nastran Implicit Nonlinear
on the basis of material stability during plastic flow (by the requirement that the stress state remain on
the yield surface during plastic straining).
The Prandtl-Reuss representation of the flow rule is available in MSC Nastran Implicit Nonlinear. In
conjunction with the von Mises yield function, this can be represented as:
(0-129)
where and are equivalent plastic strain increment and equivalent stress, respectively.
The significance of this representation is illustrated in Figure 0-31. This figure illustrates the
stress-space for the two-dimensional case. The solid curve gives theyield surface (locus of all stress
states causing yield) as defined by the vonMisescriterion.
Equation (0-139) expresses the condition that the direction of inelastic straining is normal to the yield
surface. This condition is called either the normality condition or the associated flow rule.
If the von Mises yield surface is used, then the normal is equal to the deviatoric stress.
dc
xx
p

2o
xx
F
x
2
-----------
o
yy
F
x
F
y
------------
o
zz
F
x
F
z
----------- =
dc
yy
p

o
xx
F
x
F
y
------------
2o
yy
F
y
2
-----------
o
zz
F
y
F
z
----------- + =
dc
zz
p

o
xx
F
x
F
z
-----------
o
yy
F
y
F
z
-----------
2o
zz
F
z
2
---------- + =
dc
xy
p

o
xy
F
xy
2
-------- =
dc
yz
2

o
xz
F
yz
2
-------- =
dc
zx
p

o
yz
F
zx
2
-------- =
dc
ij
p
dc
p
co
co
ij
'
----------- =
dc
p
o
Modeling Guide
Material
128
Figure 0-31 Yield Surface and Normality Criterion 2-D Stress Space
Rate Dependent Yield
Strain rate effects cause the structural response of a body to change because they influence the material
properties of the body. These material changes lead to an instantaneous change in the strength of the
material. Strain rate effects become more pronounced for temperatures greater than half the melting
temperature ( ), but are sometimes present even at room temperature. The following discussion
explains the effect of strain rate on the size of the yield surface.
Using the von Mises yield condition and normality rule, we obtain an expression for the stress rate of
theform
For elastic-plastic response
(0-130)
and
(0-131)
where
(0-132)
Yield Surface
o
2
dc2
dcp
dcp
1
o
1
p
T
m
o

ij
L
ijkl
c

kl
r
ij
c

p
+ =
L
ijkl
C
ijkl
C
ijmn

co
co
mn
------------
co
co
pq
----------- C
pqkl
\ .
| |
D =
r
ij
C
ijmn

co
co
mn
------------
2
3
--- o
co
cc
p
-------- D =
D
4
9
--- o
2

co
cc
p
--------
co
co
ij
--------- C
ijkl

co
co
kl
---------- + =
129
Model Generation
Material
As strain rates increase, many materials show an increase in yield strength. The model provided in MSC
Nastran Implicit Nonlinear for this purpose is
where:
Yield stress variation with strain rate is given using one of the following options:
1. The breakpoints and slopes for a piecewise linear approximation to the yield stress strain rate
curve are given. The strain rate breakpoints should be in ascending order, or
2. The Cowper and Symonds model is used. The yield behavior is assumed to be completely
determined by one stress-strain curve and a scale factor depending on the strain rate.
Perfectly Plastic
A material is said to be perfectly plastic if, upon the stress state touching the yield surface, an
infinitesimal increase in stress causes an arbitrarily large plastic strain. The uniaxial stress-strain diagram
for an elastic-perfectly plastic material is shown in Figure 0-32. Some materials, such as mild steel,
behave in a manner which is close to perfectly plastic.
= the uniaxial equivalent plastic strain rate
= the effective yield stress at a non-zero strain rate
=
the static yield stress, which may depend on the equivalent plastic strain,
, via isotropic hardening, or on the temperature, .
=
are material parameters that may be functions of temperature. and are
defined on the input forms. This model is effective in both static and
dynamic procedures.
c

D
o
o
0
------ 1
\ .
| |
P
for o o
0
> =
c

pl
o
o
0
c
pl
T , ( )
c
pl
T
D T ( ) p T ( ) ,
D p
Note: If multiple material models are used, they must all be expressed as piecewise linear, or as
Cowper and Symonds model.
Modeling Guide
Material
130
.
Figure 0-32 Perfectly Plastic Material Stress-Strain Relationship
Experimental Stress-Strain Curves
Metals
In uniaxial tension tests of most metals (and many other materials), the following phenomena can be
observed. If the stress in the specimen is below the yield stress of the material, the material behaves
elastically and the stress in the specimen is proportional to the strain. If the stress in the specimen is
greater than the yield stress, the material no longer exhibits elastic behavior, and the stress-strain
relationship becomes nonlinear. Figure 0-33 shows a typical uniaxial stress-strain curve. Both the elastic
and inelastic regions are indicated.
Figure 0-33 Typical Uniaxial Stress-Strain Curve (Uniaxial Test)
x x
oxx
YS
E
1
Note: Stress and strain are total quantities.
Stress Inelastic
Region
Elastic Region
Yield
Stress
Strain
131
Model Generation
Material
Within the elastic region, the stress-strain relationship is unique. As illustrated in, if the stress in the
specimen is increased (loading) from zero (point 0) to (point 1), and then decreased (unloading) to
zero, the strain in the specimen is also increased from zero to , and then returned to zero. The elastic
strain is completely recovered upon the release of stress in thespecimen.
The loading-unloading situation in the inelastic region is different from the elastic behavior. If the
specimen is loaded beyond yield to point 2, where the stress in the specimen is and the total strain
is , upon release of the stress in the specimen the elastic strain, , is completely recovered.
However, the inelastic (plastic) strain, , remains in the specimen. Figure 0-34 illustrates this
relationship. Similarly, if the specimen is loaded to point 3 and then unloaded to zero stress state, the
plastic strain remains in the specimen. It is obvious that is not equal to . We can conclude that
in the inelasticregion:
Plastic strain permanently remains in the specimen upon removal of stress.
The amount of plastic strain remaining in the specimen is dependent upon the stress level at
which the unloading starts (path-dependent behavior).
The uniaxial stress-strain curve is usually plotted for total quantities (total stress versus total strain). The
total stress-strain curve shown in Figure 0-34 can be replotted as a total stress versus plastic strain curve,
as shown in Figure 0-35. The slope of the total stress versus plastic strain curve is defined as the work
hardening slope (H) of the material. The work hardening slope is a function of plastic strain.

o
1
c
1
o
2
c
2
c
2
e
c
2
p
c
3
p
c
2
p
c
3
p
Stress
Total Strain =Strain and Elastic Strain
Yield Stress
Strain
c
2
c
2
p
c
2
e
+ =
c
3
c
3
p
c
3
e
+ =
o
y
o
3
o
2
o
1
c
3
c
2
c
1
c
3
p
c
3
e
c
2
e
c
2
p
0
1
2
3
Modeling Guide
Material
132
Figure 0-34 Schematic of Simple Loading - Unloading (Uniaxial Test)
Figure 0-35 Definition of Work hardening Slope (Uniaxial Test)
The stress-strain curve shown in Figure 0-34 is directly plotted from experimental data. It can be
simplified for the purpose of numerical modeling. A few simplifications are shown in Figure 0-36 and
are listed below:
1. Bilinear representation constant work hardening slope.
2. Elastic perfectly-plastic material no work hardening.
3. Perfectly-plastic material no work hardening and no elastic response.
4. Piecewise linear representation multiple constant work hardening slopes.
5. Strain-softening material negative work hardening slope.
In addition to elastic material constants (Youngs modulus and Poissons ratio), it is essential to include
yield stress and work hardening slopes when dealing with inelastic (plastic) material behavior. These
quantities can vary with parameters such as temperature and strain rate. Since the yield stress is generally
measured from uniaxial tests, and the stresses in real structures are usually multi-axial, the yield
condition of a multi-axial stress state must be considered. The conditions of subsequent yield (work
hardening rules) must also be studied.
Plastic Strain
H =tan u (Workhardening Slope)
=do/dc
p
u
c
p
o
Total Stress
133
Model Generation
Material
Figure 0-36 Simplified Stress-Strain Curves (Uniaxial Test)
Geological Materials
Data for geological materials are most commonly available from tri-axial compression testing. In such a
test, the specimen is confined by pressure and an additional compression stress is superposed in one
direction. Thus, the principal stresses are all negative, with .
Figure 0-37 Tri-axial Compression and Tension
o
c
(1) Bilinear Representation
o
c
(3) Perfectly Plastic
o
c
(5) Strain Softening
o
c
(2) Elastic-Perfectly Plastic
o
c
(4) Piecewise Linear Representation
0 o
1
> o
2
o
3
> =
-o
3
o
2
o
1
=o
2
>o
3
-o
1
o
3
o
1
o
1
>o
2
=o
3
o
2
Modeling Guide
Material
134
The values of the stress invariants in a uniaxial compression experiment are:
p=-{1/3}(2o
1
+o
3
)
q=o
1
-o
3
r
3
=-(o
1
-o
3
)
3

so that t=q=o
1
-o
3

The tri-axial results may thus be plotted in the t-p plane shown above. Fitting the best straight line through
the results then provides | and d.
Tri-axial tension data are also needed to define K. Under tri-axial tension, the specimen is again confined
by pressure, then the pressure in one direction is reduced. In this case, the principal stresses
are
The stress invariants are now:
p=-{1/3}(o
1
+2o
3
),
q=o
1
-o
3,
r
3
=(o
1
-o
3
)
3
,
so that t={q/K}={1/K}(o
1
-o
3
)
K may thus be found by plotting these test results as q versus p and again fitting the best straight line.
The tri-axial compression and tension lines must intercept the p-axis at the same point, and the ratio of
values of q for tri-axial tension and compression at the same value of p then gives K as shown in
Figure 0-38.
Figure 0-38 Tri-axial Compression and Tension Data
o
1
o
2
> o
3
=
q
p

ht hc
d
Best fit to tri-axial
tension data
Best fit to tri-axial
compression data
|
135
Model Generation
Material
Matching Mohr-Coulomb Parameters
Sometimes, experimental data are not directly available. Instead, the user is provided with the friction
angle and cohesion values for the Mohr-Coulomb model. We, therefore, need to calculate values
for theparameters of the Drucker-Prager model to provide a reasonable match to the
Mohr-Coulombparameters.
The Mohr-Coulomb failure model is based on plotting Mohrs circle for states of stress at failure in the
plane of the maximum and minimum principal stresses. The failure line is the best straight line that
touches these Mohrs circles.
The Mohr-Coulomb model is thus
s+o
m
sin-c cos=0,
where s={1/2}(o
1
-o
3
)
is half of the difference between the maximum and minimum principal stresses (and is, therefore, the
maximum shear stress), and
o
m
={1/2}(o
1
+o
3
)
is the average of the maximum and minimum principal stresses.
We see that the Mohr-Coulomb model assumes that failure is independent of the value of the
intermediate principal stress. The Drucker-Prager model does not. The failure of typical geotechnical
materials generally includes some small dependence on the intermediate principal stress.
Matching Tri-axial Test Response
One approach to matching Mohr-Coulomb and Drucker-Prager model parameters is to make the two
models provide the same failure definition in tri-axial compression and tension. For this purpose, we can
rewrite the Mohr-Coulomb model in terms of principal stresses.
(0-133)
Using the results above (for the stress invariants p, q, and r), in tri-axial compression and tension, allows
the Drucker-Prager model to be written for tri-axial compression as
(0-134)
and, for tri-axial tension, as
o
1
o
3
o
1
o
3
+ ( ) | ( ) 2c | cos sin + 0 =
o
1
o
3

| tan
2
1
3
--- | tan +
\ .
| |
------------------------------ o
1
o
3
+ ( )
1
1
3
--- | tan
1
1
6
-- - | tan +
------------------------o
c
0
+ + 0 =
Modeling Guide
Material
136
(0-135)
We wish to make the equations for tri-axial compression and bi-axial tension identical to the general
Mohr-Coulomb equation for all values of (o
1
,o
3
).
Comparing the equations for tri-axial compression and tri-axial tension requires that:
(0-136)
so that
(0-137)
Comparing the coefficients of (o
1
+o
3
) in the equation for tri-axial compression and that for tri-axial
tension provides:
(0-138)
and hence, from the derived equation for K:
(0-139)
Finally, comparing the last terms in the general expression for the Mohr-Coulomb model and the equation
for tri-axial compression and using the expression for tan| provides:
(0-140)
The expression for tan|, K, and this last expression and thus provide Drucker-Prager parameters that
match the Mohr-Coulomb model in tri-axial compression and tension.
The value of K in the Drucker-Prager model is restricted to for the yield surface to remain
convex. Rewriting the expression for K as:
(0-141)
shows that this implies . Many real materials have a larger Mohr-Coulomb friction angle than
this value. In such circumstances, one approach is to choose K =0.778 and then to use the expression for
o
1
o
3

| tan
2
K
----
1
3
--- | tan
\ .
| |
---------------------------- o
1
o
3
+ ( )
1
1
3
--- | tan
1
K
----
1
6
--- | tan
------------------------o
c
0
+ + 0 =
1
1
6
--- | tan +
1
K
----
1
6
-- - | tan =
K
1
1
1
3
--- | tan +
------------------------ =
|
6 | sin
3 | sin
-------------------- tan
K
3 | sin
3 u sin +
--------------------- =
o
c
0
2c u cos
1 u sin
--------------------- =
K 0.778 >
u sin 3
1 K
1 k +
-------------
\ .
| |
=
| 22 s
137
Model Generation
Material
tan | to define | and the expression for to define , ignoring the expression for K. This matches
the models for tri-axial compression only, while providing the closest approximation that the model can
provide to failure being independent of the intermediate principal stress. If is significantly larger than
22, this approach may provide a poor Drucker-Prager match of the Mohr-Coulomb parameters. MSC
Nastran Implicit Nonlinear uses K=1 by default.
Matching Plane Strain Response
Plane strain problems are often encountered in geotechnical analysis: examples are long tunnels,
footings, and embankments. For this reason, the constitutive model parameters are often matched to
provide the same flow and failure response in plane strain.
The Drucker-Prager flow potential defines the plastic strain increment as:
(0-142)
where is the equivalent plastic strain increment.
Since we only wish to match the behavior in one plane we can assume K=1, which implies that t=q. Then:
(0-143)
Writing this expression in terms of principal stresses provides:
(0-144)
with similar expressions for and .
Assume plane strain in the 1-direction. Then, at limit load, we must have =0. From the above
expression, this provides the constraint:
(0-145)
so that:
o
c
0
o
c
0
dc
pl
dc
pl
1
1
1
3
--- tan
-------------------------
o c
c
t p tan ( )
\ .
|
|
| |
=
dc
pl
dc
pl
dc
pl
1
1
1
3
--- tan
-------------------------
\ .
|
|
| |
o c
c
q
o c
cp
tan
\ .
| |
=
dc
1
pl
dc
pl
1
1
1
3
--- tan
-------------------------
\ .
|
|
| |
1
2q
------ 2o
1
o
2
o
3
( )
1
3
--- tan +
\ .
| |
=
dc
2
pl
dc
3
pl
dc
1
pl
1
2q
------ 2o
1
o
2
o
3
( )
1
3
--- tan + 0 =
Modeling Guide
Material
138
(0-146)
Using this constraint, we can rewrite q and p in terms of the principal stresses in the plane of deformation,
(0-147)
and
(0-148)
With these expressions, the Drucker-Prager yield surface can be written in terms of o
2
and o
3
as
(0-149)
The Mohr-Coulomb yield surface in the (2,3) plane is:
(0-150)
By comparison,
(0-151)
(0-152)
o
1
1
2
--- o
2
o
3
+ ( )
1
3
--- q tan =
q
3 3
2 9 tan ( )
2

------------------------------------- o
2
o
3
( ) =
p
1
2
--- o
2
o
3
+ ( )
tan
2 3 9 tan ( )
2
( )
--------------------------------------------- o
2
o
3
( ) + =
9 | tan tan
2 3 9 tan ( )
2
( )
--------------------------------------------- o
2
o
3
( )
1
2
--- | o
2
o
3
+ ( ) d tan + 0 =
o
2
o
3
o
2
o
3
+ ( ) 2c cos sin + 0 =
sin
| 3 9 tan ( )
2
( ) tan
9 | tan tan
----------------------------------------------------- =
c cos
3 9 tan ( )
2
( )
9 | tan tan
-----------------------------------------d =
139
Model Generation
Material
Now consider the two extreme cases of flow definition: associated flow, =|, and non-dilatant flow,
when =0.
Assuming associated flow, the last two equations provide:
(0-153)
and
(0-154)
while for non-dilatant flow they give and
In either case, is immediately available as:
(0-155)
The difference between these two approaches increases with the friction angle but, for typical friction
angles, the results are not very different, as illustrated in the table below.
Plane strain matching of Drucker-Prager and Mohr-Coulomb models.
As strain rates increase, many materials show an increase in yield strength. This effect often becomes
important when the strain rates are in the range of -0.1 to one per second, and can be very important if
the strain rates are in the range of 10 to 100 per second, as commonly occurs in high energy dynamic
events or in manufacturing processes.
Mohr-Coulomb
Friction Angle, u Associated Flow Nondilatant Flow
Drucker-Prager
friction angle, | d/c
Drucker-Prager friction
angle, | d/c
10 16.7 1.70 16.7 1.70
20 30.2 1.60 30.6 1.63
30 39.8 1.44 40.9 1.50
40 46.2 1.24 48.1 1.33
50 50.5 1.02 53.0 1.11
| tan
3 sin
1
1
3
--- sin ( )
2
+
------------------------------------ =
d
c
---
3 cos
1
1
3
--- sin ( )
2
+
------------------------------------ =
| tan 3 sin =
d
c
--- cos =
o
c
0
o
c
0 1
1
1
3
-- | tan
-----------------------d =
Modeling Guide
Material
140
Temperature-Dependent Behavior
This section discusses the effects of temperature-dependent plasticity on the constitutiverelation.
The following constitutive relations for thermo-plasticity were developed by Naghdi. Temperature
effects are discussed using the isotropic hardening model and the von Mises yield condition.
The stress rate can be expressed in the form
(0-156)
For elastic-plastic behavior, the moduli are
(0-157)
and for purely elastic response
(0-158)
The term that relates the stress increment to the increment of temperature for elastic-plastic behavior is
(0-159)
and for purely elastic response
(0-160)
where
(0-161)
and
(0-162)
and are the coefficients of thermal expansion.
Temperature-Dependent Stress Strain Curves
MSC Nastran 2005, SOL 600 offers the capability of stress-strain curve dependence as a function of
temperature. The user specifies these stress strain curves at different temperatures and then specifies the
temperature to use for each subcase. Linear interpolation between the supplied curves is used to
o
ij

L
ijkl
c

kl
h
ij
T

+ =
L
ijkl
L
ijkl
C
ijkl
C
ijmn

co
co
mn
------------
co
co
pq
----------- C
pqkl
\ .
| |
D =
L
ijkl
C
ijkl
=
h
ij
X
ij
C
ijkl
o
kl
C
ijkl
co
co
kl
---------- o
pq
X
pq
2
3
---o
co
cT
------
\ .
| |
\ .
| |
D =
H
ij
X
ij
C
ijkl
o
kl
=
D
4
9
--- o
2

co
cc
p
--------
co
co
ij
--------- C
ijkl

co
co
kl
---------- + =
X
ij
cC
ijkl
cT
--------------
c
kl
e
=
o
kl
141
Model Generation
Material
determine the appropriate curve at the temperature specified for a particular subcase. Marcs AF-
Flowmat capability is used for this capability; therefore, user subroutines do not have to be supplied. This
capability is best explained with an example (this example can be obtained from MSC Nastran
development. The name of the file is mattep20.dat).
SOL 600, NLSTATI C pat h=1 st op=1
TI ME 10000
CEND
ECHO = NONE
DI SPLACEMENT( pl ot ) = ALL
SPCFORCE( PLOT) = ALL
St r ess( PLOT) = ALL
St r ai n( PLOT) = ALL
SPC = 1
NLPARM = 2
t emp( i ni t ) =10
subcase 1
t emp( l oad) =11
LOAD = 100
subcase 2
t emp( l oad) =12
LOAD = 200
subcase 3
t emp( l oad) =13
LOAD = 300
BEGI N BULK
par am, mr af f l ow, mymat 0
par am, mr t abl s1, 4
par am, mr t abl s2, 1
NLPARM 2 10 AUTO 1 20 P
PARAM, LGDI SP, 1
t empd, 10, 70.
t empd, 11, 110.
t empd, 12, 700.
t empd, 13, 1100.
$LOAD, 20, 1. 0, 2. 0, 1, 1. 0, 2
l oad, 100, 1. , 1. , 1
l oad, 200, 1. , - . 5, 1
l oad, 300, 1. , 1. 1, 1
PLOAD4 1 1 - 15.
.
.
.
$ Const r ai nt Set 1 : Unt i t l ed
SPC 1 1 123456 0.
SPC 1 8 123456 0.
SPC 1 15 123456 0.
SPC 1 22 123456 0.
SPC 1 29 123456 0.
$ Pr oper t y 1 : Unt i t l ed
PSHELL 1 1 0. 125 1 1 0.
$ Mat er i al 1 : AI SI 4340 St eel
MATEP, 1, TABLE, 35000. , 2, CAUCHY, I SOTROP, ADDMEAN
MAT1 1 2. 9E+7 0. 327. 331E- 4 6. 6E- 6 70. +MT 1
+MT 1 215000. 240000. 156000.
MAT4 14. 861E- 4 38. 647. 331E- 4
$ 1 2 3 4 5 6 7 8 9
$2345678 2345678 2345678 2345678 2345678 2345678 2345678 2345678 2345678
MATTEP 1 21
Modeling Guide
Material
142
MATT1 1 7
TABLEM1 7
+ 70. 0 6. 6E- 6 1000. 6. 5E- 6 1200. 6. 4E- 6 1500. 6. 3E- 6
+ 2000. 6. 2E- 6 ENDT
$2345678 2345678 2345678 2345678 2345678 2345678 2345678 2345678 2345678
TABLEST 21
+ 70. 0 31 1000. 32 1200. 33 1500. 34
+ 2000. 35 ENDT
TABLES1, 31
, 0. , 15000. , 1. 0, 16000. , 10. , 25000. , 100. , 30000. ,
, 99999. , 40000. , ENDT
TABLES1, 32
, 0. , 13000. , 1. 0, 14000. , 10. , 23000. , 100. , 28000. ,
, 99999. , 28000. , ENDT
TABLES1, 33
, 0. , 11000. , 1. 0, 12000. , 10. , 21000. , 100. , 26000. ,
, 99999. , 25000. , ENDT
TABLES1, 34
, 0. , 9000. , 1. 0, 10000. , 10. , 19000. , 100. , 22000. ,
, 99999. , 24000. , ENDT
TABLES1, 35
, 0. , 5000. , 1. 0, 7000. , 10. , 9000. , 100. , 13000. ,
, 99999. , 15000. , ENDT
GRI D 1 0 0. 0. 0. 0
.
.
.
CQUAD4
.
.
.
ENDDATA
In this input, the stress strain curves are specified by TABLES1 entries. The collection of stress-strain
curves to be used is specified in the TABLEST entry and the corresponding temperatures at which they
apply is specified in the TABLEM1 entry. The TABLEM1 I D is called out in field seven of the MATT1 entry
and the TABLEST I D is called out in field five of the MATTEP entry. TABLEST must list the stress strain
TABLES1 I Ds in order of increasing temperature and the first ID must be at the lowest temperature
specified anywhere in the analysis. In this example, it is a temperature of 70 corresponding to
temp(init)=10 in the Case Control. Similarly, the temperatures in the TABLEM1 entry must be in
increasing order. The stress-strain curves should cover the entire range of temperatures for the analysis
so that no extrapolation is needed. The actual temperatures for each subcase are given by the temp(load)
specifications for each subcase.
There is one parameter that is critical to this analysis:
143
Model Generation
Material
Specifying Elastoplastic Material Entries
Each of the elastoplastic models described in this section can be selected with the MATEP bulk
dataentry.
References
MATEP in the MSC NASTRAN QRG.
MATTEP in the MSC NASTRAN QRG.
SimXpert Materials Application Input Data
To define an inelastic material in SimXpert:
1. Select Materials and Properties > Isotropic, Orthotropic, or Anisotropic > Advanced > Add
Constitutive Model > Elasto Plastic.
2. Enter parameter values.
3. Click OK.
The required properties for describing elastic-plastic behavior vary based on material type, dimension,
type of nonlinear data input, hardening rule, yield criteria, strain rate method, and thermal dependencies.
The table below shows the various input options and criteria available to you for defining
elastoplasticbehavior.
param,mrafflow,mymat0 Name of the file containing temperature dependent stress versus plastic
strain curves in Marcs AF_flowmat format. This file can be generated
from the current MSC Nastran run using TABLEST and TABLES1 entries
or a pre-existing file can be used depending on the value of
PARAM,MRAFFLOR. The extension .mat will be added toName. If
this is a new file, it will be saved in the directory from which the MSC
Nastran execution is submitted. If a pre-existing file isto be used, it can
either be located in the directory where the MSC Nastran execution is
submitted and run or in the Marc AF_flowmat directory.
Entry Description
MATEP Specifies elasto plastic material properties to be used for large
deformationanalysis.
MATTEP Specifies temperature-dependent elasto-plastic material properties to be used
for static, quasi-static, or transient dynamic analysis.
Modeling Guide
Material
144
Elasto Plastic Model Summary
Constitutive
Model
Nonlinear Data
Input
Hardening
Rule Yield Criteria
Strain Rate
Method
Plastic Stress/Strain
Curve
Isotropic
Kinematic
Combined
von Mises
Tresca
Mohr-Coulomb
Drucker-Prager
Parabolic Mohr-
Coulomb
Buyukozturk Concrete
Oak Ridge National
Lab
2-1/4 Cr-Mo ORNL
Reversed Plasticity
ORNL
Full Alpha Reset
ORNL
Piecewise
Linear
Cowper-
Symonds
Hardening Slope Isotropic
Kinematic
Combined
von Mises
Tresca
Mohr-Coulomb
Drucker-Prager
None
145
Model Generation
Material
Nonlinear Data Input
The type of nonlinear data input you choose to use to define elasto plastic material behavior determines
the input data required.
Stress/Strain Curve - All stress-strain curves are input as piecewise linear. SimXpert transfers
the stress-strain curve input on the material property field directly to the TABLES1 entry.
The number of linear segments used to define the stress-strain curve may be different from one
material to another. The same strain breakpoints need not be used for all of the different
materials stress-strain curves. It is recommended to define the stress-strain curves throughout
the range of strains which the analysis is likely to predict. If the analysis predicts a plastic strain
greater than the last point defined by the user, MSC Nastran Implicit Nonlinear continues the
analysis after shifting the last strain breakpoint on that curve to match the predicted value,
thereby changing (reducing) the work hardening slope for the last segment of the curve.
Hardening Slope - The hardening slope and the yield point are required with this Nonlinear Data
Input option.
Perfectly Plastic - Perfect plasticity is described by simply specifying the yield point.
Constitutive
Model Type
Hardening
Rule Yield Criteria
Strain Rate
Method
Plastic Perfectly Plastic None von Mises
Linear Mohr-Coulomb
Parabolic Mohr-
Coulomb
Buyukozturk Concrete
Oak Ridge National
Lab
2-1/4 Cr-Mo ORNL
Reversed Plasticity
ORNL
Full Alpha Reset
ORNL
Piecewise
Linear
Cowper-
Symonds
Power Law
Rate Power Law
J ohnson-Cook
Kumar
Piecewise
Linear
None Piecewise
Linear
Cowper-
Symonds
Elasto Plastic Model Summary
Modeling Guide
Material
146
The tables below provide descriptions for the input data for each of the four types of nonlinear input.
Isotropic - Stress/Strain Curve or Perfectly Plastic:
All Yield Functions
Property Name Description
Stress /Strain Curve
or
Yield Stress
Defines the Cauchy stress versus logarithmic strain (also called
equivalent tensile stress versus total equivalent strain) by reference to a
table. For Perfectly Plastic models, only a Yield Stress needs to be
entered.
Can also be strain rate dependent if Strain Rate Method is Piecewise
Linear. Accepts table of yield stress versus strain rate.
10th Cycle Yield Stress
versus Plastic Strain
or
10th Cycle Yield Stress
When set to ORNL, accepts field of 10th cycle yield stress versus
plastic strain. Can be temperature dependent also. For Perfectly Plastic
models, only a 10th Cycle Yield Stress needs to be entered.
Coefficient C Visible if Strain Rate Method is Cowper-Symonds.
Inverse Exponent P Visible if Strain Rate Method is Cowper-Symonds.
Alpha When set to Linear Mohr-Coulomb, defines the slope of the yield
surface in square root J 2 versus J 1 space. This property is required.
Beta When set to Parabolic Mohr-Coulomb, defines the beta parameter in
the equation that defines the parabolic yield surface in square root J 2
versus J 1 space. This property is required.
Note: 2 1/4 Cr-Mo ORNL, Reversed Plasticity ORNL, Full Alpha Reset ORNL are the same as
Oak Ridge National Labs. Generalized Plasticity is the same as von Mises.
Perfectly Plastic is identical to Stress/Strain except that no hardening rules apply.
147
Model Generation
Material
Hardening Slope - Nonlinear Data Input
Failure and Damage Models
One of the nonlinear features of a material's behavior is failure. When a certain criterion (failure
criterion) is met, the material fails and no longer sustains its loading and breaks. In a finite-element
method, this means that the element, where the material reaches the failure limit, cannot carry any
stresses anymore. The stress tensor is effectively zero. The element is flagged for failure, and, essentially,
is no longer part of the structure.
Failure criteria can be defined for a range of materials and element types. The failure models are
referenced from the material definition entries.
Anisotropic/Orthotropic - Stress/Strain Curve or Perfectly Plastic:
All Yield Functions
Description
Stress versus Strain
or
Tensile Yield Stress
Same as description for Isotropic Elastic-Plastic. If Strain Rate Method
is Piecewise Linear, accepts field of yield stress versus strain rate.
Or defines an isotropic yield stress. It is a required property when the
plasticity type is Perfectly Plastic.
Stress 11/22/33 Yield Ratios Defines the ratios of direct yield stresses to the isotropic yield stress in
the elements coordinate system.
Stress 12/23/31 Yield Ratios Defines the ratios of shear yield stresses to the isotropic shear yield
stress (yield divided by square root three) in the elements
coordinatesystem.
Note: Perfectly Plastic is identical to Elastic-Plastic except that no hardening rules apply. Stress
verus Plastic Strain is replaced with Yield Stress data only as is 10th Cycle Yield versus
Strain replaced with 10th Cycle Yield Stress data. Thus no tabular data is necessary.
Isotropic/Orthotropic/Anisotropic - Hardening Slope
Property Name Description
Hardening Slope Slope of the stress-strain curve once yielding has started.
Yield Point Defines the stress level at which plastic strain begins to develop.
Internal Friction Angle When yield function is set to Mohr-Coulomb or Drucker-Prager this
gives the parameter describing the effect of hydrostatic pressure on the
yield stress.
Modeling Guide
Material
148
Isotropic/Orthotropic/Anisotropic Failure Models
For isotropic, 2-D orthotropic, and 2-D anisotropic materials, you can implement one of five failure
models in MSC Nastran Implicit Nonlinear (SOL 600). Failure models are based on maximum stress
criteria, maximum strain criteria, or one of three composite stress/strain failure theories.
Maximum Stress Criterion
At each integration point, MSC Nastran Implicit Nonlinear calculates six quantities:
(0-163)
(0-164)
(0-165)
(0-166)
Failure Model Applicable Material Type
Maximum Stress Isotropic, 2-D Orthotropic, 2-D Anisotropic
Maximum Strain 2-D Orthotropic
Hill Isotropic, 2-D Orthotropic (stress or strain based), 2-D Anisotropic
Hoffman Isotropic, 2-D Orthotropic (stress or strain based), 2-D Anisotropic
Tsai-Wu Isotropic, 2-D Orthotropic (stress or strain based), 2-D Anisotropic
o
1
X
t
------
\ .
| |
F
o
1
X
c
------
\ .
| |
F
o
1
0 >
o
1
0 <
if
if
1.
o
2
Y
t
------
\ .
| |
F
o
2
Y
c
------
\ .
| |
F
o
2
0 >
o
2
0 <
if
if
2.
o
3
Z
t
------
\ .
| |
F
o
3
Z
c
------
\ .
| |
F
o
3
0 >
o
3
0 <
if
if
3.
o
12
S
12
--------
\ .
| |
F
4.
149
Model Generation
Material
(0-167)
(0-168)
where
Maximum Strain Failure Criterion
At each integration point, calculates six quantities:
(0-169)
(0-170)
is the failure index (F =1.0).
are the maximum allowable stresses in the 1-direction in tension and compression.
are maximum allowable stresses in the 2-direction in tension andcompression.
are maximum allowed stresses in the 3-direction in tension andcompression.
is the maximum allowable in-plane shear stress.
is the maximum allowable 23 shear stress.
is the maximum allowable 31 shear stress.
o
23
S
23
--------
\ .
| |
F
5.
o
31
S
31
--------
\ .
| |
F
6.
F
X
t
X
c
,
Y
t
Y
c
,
Z
t
Z
c
,
S
12
S
23
S
31
c
1
e
1t
------
\ .
| |
F
c
1
e
1c
-------
\ .
| |
F
c
1
0 >
c
1
0 <
if
if
1.
c
2
e
2t
------
\ .
| |
F
c
2
e
2c
-------
\ .
| |
F
c
2
0 >
c
2
0 <
if
if
2.
Modeling Guide
Material
150
(0-171)
(0-172)
(0-173)
(0-174)
where
Hill Failure Criterion
Assumptions:
Orthotropic materials only
Incompressibility during plastic deformation
Tensile and compressive behavior are identical
At each integration point, MSC Nastran Implicit Nonlinear calculates:
is the failure index (F=1.0).
are the maximum allowable strains in the 1 direction in tension andcompression.
are the maximum allowable strains in the 2 direction in tension and compression.
are the maximum allowable strains in the 3 direction in tension andcompression.
is the maximum allowable shear strain in the 12 plane.
is the maximum allowable shear strain in the 23 plane.
is the maximum allowable shear strain in the 31 plane.
c
3
e
3t
------
\ .
| |
F
c
3
e
3c
-------
\ .
| |
F
c
3
0 >
c
3
0 <
if
if
3.