Multi Step Nonlinear
Multi Step Nonlinear
Multi-Step Nonlinear
User’s Guide
Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
Overview of nonlinear capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
Program architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
Rudiments of the user interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2
Case control section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-3
Bulk data section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-4
Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-4
Nonlinear characteristics and general recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-5
Figures
© 2014 Siemens Product Lifecycle Management Software Inc. All Rights Reserved.
This software and related documentation are proprietary to Siemens Product Lifecycle Management
Software Inc. Siemens and the Siemens logo are registered trademarks of Siemens AG. NX is a
trademark or registered trademark of Siemens Product Lifecycle Management Software Inc. or its
subsidiaries in the United States and in other countries.
NASTRAN is a registered trademark of the National Aeronautics and Space Administration. NX
Nastran is an enhanced proprietary version developed and maintained by Siemens Product Lifecycle
Management Software Inc.
MSC is a registered trademark of MSC.Software Corporation. MSC.Nastran and MSC.Patran are
trademarks of MSC.Software Corporation.
All other trademarks are the property of their respective owners.
TAUCS Copyright and License
TAUCS Version 2.0, November 29, 2001. Copyright (c) 2001, 2002, 2003 by Sivan Toledo, Tel-Aviv
University, [email protected]. All Rights Reserved.
TAUCS License:
Your use or distribution of TAUCS or any derivative code implies that you agree to this License.
THIS MATERIAL IS PROVIDED AS IS, WITH ABSOLUTELY NO WARRANTY EXPRESSED OR
IMPLIED. ANY USE IS AT YOUR OWN RISK.
Permission is hereby granted to use or copy this program, provided that the Copyright, this License,
and the Availability of the original version is retained on all copies. User documentation of any code
that uses this code or any derivative code must cite the Copyright, this License, the Availability note,
and "Used by permission." If this code or any derivative code is accessible from within MATLAB, then
typing "help taucs" must cite the Copyright, and "type taucs" must also cite this License and the
Availability note. Permission to modify the code and to distribute modified code is granted, provided
the Copyright, this License, and the Availability note are retained, and a notice that the code was
modified is included. This software is provided to you free of charge.
Availability (TAUCS)
As of version 2.1, we distribute the code in 4 formats: zip and tarred-gzipped (tgz), with or without
binaries for external libraries. The bundled external libraries should allow you to build the test
programs on Linux, Windows, and MacOS X without installing additional software. We recommend
that you download the full distributions, and then perhaps replace the bundled libraries by higher
performance ones (e.g., with a BLAS library that is specifically optimized for your machine). If you
want to conserve bandwidth and you want to install the required libraries yourself, download the
lean distributions. The zip and tgz files are identical, except that on Linux, Unix, and MacOS,
unpacking the tgz file ensures that the configure script is marked as executable (unpack with tar
zxvpf), otherwise you will have to change its permissions manually.
with the Executive System and with other modules through parameters, which may be input and/or
output variables of the module. Modules utilize main memory dynamically. If the size of the main
memory is insufficient to complete an operation, the module uses scratch files, which reside in the
secondary storage as an extension of the main memory. This is known as a spill operation.
DMAP is a kind of macro program using a data block oriented language. The solution sequence is a
collection of module statements written in the DMAP language tailored to process a sequential series
of operations, resulting in a specific type of structural analysis. A typical solution sequence consists
of three phases of functional operations: formation, assembly, and reduction of matrices; solution
of equations; and data recovery. Solution sequences that process superelements have built-in
superelement loops in the first and the last phases.
The nonlinear solution sequences have built-in loops in the second phase for subcase changes, load
increments, and stiffness matrix updates. Nested in this DMAP loop, nonlinear solution processes
comprise a number of internal iteration loops. Confining the discussion to SOL 401, the hierarchy
of the nonlinear looping is shown in the table below. Central to the nonlinear processes is module
NLTRD3. The module is self-contained to perform iterations for converged solutions.
• Element data
• Material data
• Solution methods
The first three classes of data may not be changed during the course of an analysis whereas the last
three classes of data may be changed in midcourse via subcases under the case control section.
1.3.3 Parameters
Parameters are used for requesting special features or specifying miscellaneous data. Parameters
are initialized in the MPL, which can be overridden by a DMAP initialization. Modules may change the
parameter values while the program is running.
There are two different types of parameters: user parameters (V,Y,name in the DMAP) and DMAP
(non-user) parameters. For the user parameters, users are authorized to change the initial values
by specifying PARAM data in the bulk data section or occasionally in the case control section. The
following table lists the parameters supported in SOL 401.
COLPHEXA OMAXR RGBEAMA
COUPMASS OMPT RGBEAME
• The size of the model should be determined based on the purpose of the analysis, the trade-offs
between accuracy and efficiency, and the scheduled deadline.
• Prior contemplation of the geometric modeling will increase efficiency in the long run. Factors
to be considered include selection of coordinate systems, symmetric considerations for
simplification, and systematic numbering of nodal points and elements for easy classification
of locality.
• Discretization should be based on the anticipated stress gradient, i.e., a finer mesh in the area of
stress concentrations.
• Element types and the mesh size should be judiciously chosen. For example, avoid highly
distorted and/or stretched elements (with high aspect ratio).
• The model should be verified prior to the analysis by some visual means, such as plots and
graphic displays.
Nonlinear analysis requires better insight into structural behavior. First of all, the type of nonlinearities
involved must be determined. The geometric nonlinearity is characterized by large rotations which
usually cause large displacements. Intuitively, geometric nonlinear effects should be significant if the
deformed shape of the structure appears distinctive from the original geometry without amplifying the
displacements. There is no distinct limit for large displacements because geometric nonlinear effects
are related to the dimensions of the structure and the boundary conditions. The key to this issue is to
know where the loading point is in the load-deflection curve of the critical area.
Additional recommendations are important for nonlinear analysis:
• PARAM,LGDISP,1 must be defined to turn on geometry nonlinearity.
• Material nonlinear effects can also be included. See Support for plasticity analysis and Support
for creep analysis.
• The nonlinear region usually requires a finer mesh. Use a finer mesh if severe element distortions
or stress concentrations are anticipated.
• The subcase structure should be utilized properly to divide the load or time history for
conveniences in data recovery, and database storage control, not to mention changing constraints
and loading paths.
• Many options are available in solution methods to be specified on the NLCNTL and the TSTEP1
bulk entries. The defaults should be used on all options before gaining experience.
• The axisymmetric elements CQUADX4, CQUADX8, CTRAX3, CTRAX6, the plane strain
elements CPLSTN3, CPLSTN4, CPLSTN6, CPLSTN8, and the plane stress elements CPLSTS3,
CPLSTS4, CPLSTS6, CPLSTS8 are supported for linear, geometric nonlinear, and material
nonlinear analysis.
The grid points on these elements must all lie in either the XZ plane, or all in the XY plane of the
basic coordinate system. The software automatically determines the orientation.
When axisymmetric elements are defined on the XZ plane, X is the radial direction, and Z is the
axial direction. The grid points defining these elements must have X ≥ 0.
When axisymmetric elements are defined on the XY plane, Y is the radial direction, and X is the
axial direction. The grid points defining these elements must have Y ≥ 0.
• A special, generalized plane strain formulation is available using the CPLSTN3, CPLSTN4,
CPLSTN6, and CPLSTN8 element types. See Generalized plane strain analysis.
• The RBE2 and RBAR rigid elements are supported with optional large displacement effects and
thermal expansion. The RBE3 rigid element is also supported, but it does not support the large
displacement effects or thermal expansion. See Rigid element support.
• The PSOLID or the PCOMPS bulk entries define the element properties. The PCOMPS is
optionally used to define a layered solid composite property.
The PGPLSN bulk entry is defines the special generalized plane strain element properties.
See Support for plasticity analysis and Support for creep analysis.
• You can request stress norm, stress error norm, strain energy norm, and strain energy error
norm output. The output is computed and stored on an individual element basis. The NX
Advanced Simulation product uses the output for adaptive meshing. See Error estimator for
mesh refinement.
• You can compute and output the j-integral in a crack simulation. The j-integral output can
be requested and used by third-party software like Zencrack to perform a fracture mechanics
analysis. The CHEXA bulk entry allows for a collapsed element definition. See Crack simulation.
The ANALYSIS case control command does not have a default in SOL 401. You must define it in
every subcase, and it cannot be defined above the subcases (globally). The first subcase can be
either a static or a modal subcase.
The modal subcase should include the METHOD case control command which selects the EIGRL
bulk entry. The EIGRL entry defines the data needed to perform the real eigenvalue analysis with the
Lanczos method. The modal subcase automatically includes the stress stiffening from the previous
static subcase, and can potentially include follower stiffness and spin softening depending on the type
of loading in the previous static subcase. The NLCNTL bulk entry has parameter inputs which allow
you to control the stiffness contributions for the modal subcase.
• PARAM,LGDISP,-1 (default) – Large displacement effects are turned off. Subcases which include
ANALYSIS=STATIC are linear static subcases.
• PARAM,LGDISP,1 – Large displacement effects are turned on. Subcases which include
ANALYSIS=STATIC are nonlinear static subcases.
PARAM,LGDISP,1 turns on large displacement effects, but small strains are assumed.
Material nonlinear effects can also be included. See Support for plasticity analysis and Support for
creep analysis.
-etc-
In this example, assuming a start time=0.0 for the subcase, the resulting time steps are as follows.
The time steps in which output occurs are highlighted. Output always occurs at the end time.
The 1st row has an end time of 10.0, 5 increments, and output at every 2nd time step.
Time Step 1 Time Step 2 Time Step 3 Time Step 4 Time Step 5
2.0 4.0 6.0 8.0 10.0
The 2nd row has an end time of 50.0, 4 increments, and output frequency at every 3rd time step.
The 3rd row has an end time of 100.0, 2 increments, and output at all time steps.
In the same example, assuming a start time=5.0 for the subcase, the resulting time steps for the
first row are as follows.
Time Step 1 Time Step 2 Time Step 3 Time Step 4 Time Step 5
6.0 7.0 8.0 9.0 10.0
• When Tend1=0.0,
o No other times are allowed. This is the only time for the associated subcase.
o Nouti is ignored.
• Nouti controls the frequency of results output. The table below summarizes the input options.
o For a sequentially dependent (SD) static subcase (default), the final Tendi from a previous SD
or NSD static subcase is the start time (Tstart) for the current SD subcase. If an SD subcase
has no previous SD or NSD static subcases, the start time is 0.0 for that SD subcase, and
Tend1=0.0 is permitted. Otherwise, Tend1 > Tstart for all other SD subcases.
• If a creep material is included, the software uses adaptive time stepping by default. The adaptive
time stepping can result in additional solution time steps which are not defined by the TSTEP1
entry. See Support for creep analysis.
Similarily, when running SOL 401 in the context of the NX Multiphysics environment, additional
solution time steps beyond what is defined on the TSTEP1 entries are possible.
For both of these cases, the Nout field on the TSTEP1 bulk entry still determines the frequency of
results output.
o Time-unassigned loads are selected with the LOAD case control command,
LOAD=n
o Time-assigned loads are selected with the DLOAD case control command,
DLOAD=n
where n points to a load set defined by a TLOAD1 bulk entry, or a DLOAD bulk entry if you
want to combine multiple TLOAD1 entries into a single load set.
o TABLEDi (i=1,2,3,4) - Table that defines the load variation with time.
o The supported static load inputs are the DAREA, FORCE, FORCE1, FORCE2, GRAV,
PLOAD, PLOAD4, RFORCE, RFORCE1, SLOAD, and SPCD bulk entries.
o The TYPE field on the TLOAD1 entry should be “0” for all load inputs selected by the
EXCITEID field, except for the SPCD entry. The SPCD entry requires “1” in the TYPE field.
o A real value is supported in the DELAY field on a TLOAD1 entry to optionally shift the time
steps used to compute the associated loads.
o A temperature load cannot be selected on the EXCITEID field. See Thermal Loads.
o The TID field selects a TABLEDi, which defines a load scaling versus time function.
o The figure below demonstrates how the DLOAD, TLOAD1, FORCE (for example), and
TABLEDi bulk entries relate to one another.
When there is more than one time-assigned load set, the DLOAD bulk entry is required:
$2345678$2345678$2345678$2345678$2345678$2345678$2345678$2345678$2345678
$
$ DLOAD COMBINES MULTIPLE TLOAD1 (102 AND 105)
DLOAD 17 1. 1. 102 1. 105
$
$ TIME-ASSIGNED FORCE, EXCITEID=125, TYPE=0 (DEFAULT), TIME FUNCTION TID=13
TLOAD1 102 125 13
$
FORCE 125 80 0 1. 3. 0. 0.
$
$ TIME FUNCTION 13 USED FOR FORCE LOAD
TABLED2 13 0. +
+ 0. 0. 1. 100. 2. 0. ENDT
$
$ TIME-ASSIGNED FORCE EXCITEID=3, TYPE=0 (DEFAULT), TIME FUNCTION TID=12
TLOAD1 105 3 1 12
$
FORCE 3 73 0 2. 8. 0. 0.
$
$ TIME FUNCTION 12 USED FOR FORCE LOAD
TABLED2 12 0. +
+ 0. 0. 2. 1. ENDT
When there is only one time-assigned load set, the DLOAD entry is not required:
$2345678$2345678$2345678$2345678$2345678$2345678$2345678$2345678$2345678
$
$ TIME-ASSIGNED FORCE, EXCITEID=125, TYPE=0 (DEFAULT), TIME FUNCTION TID=13
TLOAD1 102 125 13
$
FORCE 125 80 0 1. 3. 0. 0.
PLOAD 125 100.0 21 30 18 10
PLOAD 125 100.0 10 18 22 25
$
$ TIME FUNCTION 13 USED FOR LOAD
TABLED2 13 0. +
+ 0. 0. 1. 100. 2. 0. ENDT
$
o The bolt preload capability allows you to model bolts with either 3D solid elements or
2D plane stress elements. A bolt preload can be combined with geometric nonlinear and
plasticity conditions. See Bolt preload.
o Loads in any subcase are total loads as opposed to incremental loads from the previous
subcase. In other words, the ending load from a previous subcase does not become the
initial loading for the consecutive subcase.
o LOAD=n or DLOAD=n defined at the global level is used in all statics subcases unless
a different LOAD=n or DLOAD=n is defined in a subcase.
o The TSTEP1 bulk entry defines the time step intervals in which a solution will be generated
and output in a static subcase. If your time steps defined by the TSTEP1 entry exceed the
time values defined in your TABLEDi entry, by default, the software will extrapolate the data
defined in the TABLEDi entry. The software will issue a warning if extrapolation occurs. If
you do not want the software to extrapolate the data, you can enter “1” in the EXTRAP
field on the TABLEDi entry.
o In SOL 401, when RFORCE or RFORCE1 entries are referenced by the EXCITEID field on
a TLOAD1 entry, the data on the associated TABLEDi, along with the scale factors S and
Si on a DLOAD entry (if defined), scale the angular velocity (ω) and acceleration (α), which
are used to compute an inertia force in the equation F = [m] [ω x (ω x r)) + α x r]. Since ω is
squared in the force computation, the resulting scaling is not linearly related to the computed
force (F). All other solutions scale the computed force (F).
• Tref is the initial temperature used when computing the temperature dependent coefficient of
thermal expansion, and is defined on the MATi entry. See Computing the coefficient of thermal
expansion.
• If the coefficient of thermal expansion is defined as temperature dependent with the MATTi
entries, αload is evaluated at Tload, and αinit is evaluated at Tinit .
If the coefficient of thermal expansion is not defined as temperature dependent, αload and αinit are
assigned the single value defined on the MATi entry.
• Tinit is defined using the TEMP(INIT) case control command, and must be the same for all
subcases. Typically, the TEMP(INIT) command is defined globally, and selects one of the
following.
o The TEMP(INIT) can select the TEMP and TEMPD entries in the bulk data.
For example,
...
TEMP(INIT) = 100
...
BEGIN BULK
...
$ Initial temperatures defined in the bulk data
TEMP,100,5,232.0,6,354.4,...etc
...
$ TEMPD defines a temperature for grid points not included on a TEMP entry
TEMPD,100,450.0
...
o The TEMP(INIT) command can select the TEMPEX and TEMPD bulk entries. The new
TEMPEX entry references an external BUN file using the unit number defined with an ASSIGN
statement. The unit number must be unique to other BUN files, and to other reserved unit
numbers. The BUN file used to define Tinit must only include a single set of temperature data.
If the BUN file only defines temperatures for a portion of the model (subset), the TEMPD
entry must be included in the bulk data to define a temperature for the grid points not
included in the BUN file.
For example,
...
ASSIGN BUN=‘temperature0.bun’ UNIT=21
...
TEMP(INIT) = 100
...
BEGIN BULK
...
$ Initial temperatures defined in the BUN file
TEMPEX,100,21
$ Temperature for grid points not in the BUN file
TEMPD,100,630.2
...
• There are a variety of options to define Tload. These options can be defined globally and in a
subcase. Any subcase definition will override any global definition. For example, if you define a
time-unassigned Tload globally using the TEMP(LOAD) command, and you define a time-assigned
Tload in a subcase using the DTEMP command, the time-assigned Tload is used for that subcase.
o You can define a time-unassigned Tload with all temperatures defined in the bulk data. The
TEMP(LOAD) case control command selects the TEMP and TEMPD entries in the bulk data.
For example,
...
SUBCASE 5
TEMP(LOAD) = 150
BEGIN BULK
...
$ time-unassigned grid point load temperatures for subcase 5
TEMP,150,74,232.0,23,354.4,...
$ TEMPD defines a temperature for grid points not included on a TEMP entry
TEMPD,150,450.0
...
o You can define a time-unassigned Tload with temperatures defined in an external BUN file.
The TEMP(LOAD) case control command selects the TEMPEX bulk entry and optionally the
TEMPD entry. The new TEMPEX entry references the external file using the unit number
defined with an ASSIGN statement. The unit number must be unique to other BUN files, and
to other reserved unit numbers. The BUN file selected with the TEMPEX bulk entry must
only include a single set of temperature data.
If the BUN file only defines temperatures for a portion of the model (subset), the TEMPD
entry must be included in the bulk data to define a temperature for the grid points not
included in the BUN file.
TEMPEX example:
...
ASSIGN BUN=‘temperature1.bun’ UNIT=22
$...
SUBCASE 10
TEMP(LOAD) = 200
BEGIN BULK
...
$ Time-unassigned load temperatures for subcase 10
TEMPEX, 200, 22
$ Temperature for grid points not in the BUN file
TEMPD,200,630.2
...
o You can define a time-assigned Tload with temperatures defined in the bulk data or in a BUN
file. The new DTEMP case control command selects the new DTEMP bulk entry, which
defines a list of time points versus set IDs. The set IDs are either the IDs of TEMP and
TEMPD entries in the bulk data, or the IDs of TEMPEX and TEMPD entries in the bulk data.
You cannot combine TEMP and TEMPEX entries with the same set ID.
Example with TEMP and TEMPD entries in the bulk data:
Note: This example assumes the TEMP entries for temperature sets 500 and 501 define temperatur
...
SUBCASE 15
DTEMP(LOAD) = 250
...
BEGIN BULK
...
$ DTEMP is a list of time points versus set IDs
DTEMP,250,,,,,,,,+
+,.2,500,.4,501,.6,502
...
$ Load temperatures at t=.2
TEMP,500,5,232.0,6,354.4,7,284.2
...
$ Load temperatures at t=.4
TEMP,501,5,234.1,6,356.3,7,287.8
...
$ Load temperatures at t=.6
TEMP,502,5,237.3,6,358.4,7,292.4
$ Temperature for grid points not defined with TEMP entry 502.
TEMPD,502,630.2...
DTEMP,250,,,,,,,,+
+,.2,500,.4,501,.6,502
...
$ Load temperatures at t=.2
TEMPEX,500,22
$ Temperature for grid points not defined in BUN file
TEMPD,500,345.4
...
$ Load temperatures at t=.4
TEMPEX,501,23
...
$ Load temperatures at t=.6
TEMPEX,502,24
...
$ If the BUN file for t=.2 and t=.4 includes data for all grid points, the TEMPD is not needed
o You can define a time-assigned Tload with temperatures defined in a single, external BUN
file. The new DTEMP case control command selects the new DTEMPEX bulk entry, which
references the external file using the unit number defined in the ASSIGN statement. The
unit number must be unique to other BUN files, and to other reserved unit numbers. The
single BUN file selected with the DTEMPEX bulk entry must include temperature data for
all grid points, and for multiple time points. The BUN file can include temperatures for grids
which are not in the model, but unlike the TEMPEX example above, the BUN file selected
with the DTEMPEX cannot define temperatures for only a portion of the model (subset). The
TEMPD entry cannot be combined with the DTEMPEX entry.
DTEMPEX example:
...
ASSIGN BUN=‘temperature.bun’ UNIT=23
$...
SUBCASE 20
DTEMP = 300
BEGIN BULK
...
$ Time-assigned load temperatures for subcase 20
DTEMPEX, 300, 23
...
Additional information:
• The specification of TEMP(MATERIAL) or TEMP(BOTH) are unsupported and will cause a
fatal error if defined.
• The TVAR parameter on the NLCNTL bulk entry controls if time-unassigned temperature loads
selected with the TEMP(LOAD) case control command are ramped, or not ramped for each
subcase.
o When TVAR=RAMP, the software ramps the load temperatures from the final Tload defined
for the previous static subcase to the Tload defined for the current subcase. The software
determines the load temperature increments using the total number of time increments
defined for that subcase. If Tload is not defined in the previous subcase, the software ramps
from Tinit to the current Tload.
The default is “RAMP” except when Tend1 = 0.0 is defined on the TSTEP1 entry in the first
static subcase. “STEP” occurs in this case.
• For the time-assigned temperature data, the software will interpolate the grid point temperatures
when times are defined between the time points in the data. Although, if a solution time is
outside the data range, the software will use the data at the closest time point, and a warning
will be written to the f06 file.
• You can turn off the thermal strain computation by defining the parameter setting THRMST=NO
(default=YES) on the NLCNTL bulk entry. This is useful for temperature dependent material
evaluation without thermal loading.
• When temperature dependent material properties are defined with the MATTi entries for a static
subcase, the properties are evaluated at Tload selected with either the TEMP(LOAD) or DTEMP
case control. Both Tload and Tinit must be defined when temperature dependent properties
are defined.
• A modal subcase which is not sequentially dependent (SEQDEP=NO) can include temperature
dependent material properties defined with the MATTi entries. The properties are evaluated at
Tload selected with the TEMP(LOAD) case control. The DTEMP case control command is not
supported in a modal subcase. Both Tload and Tinit must be defined when temperature dependent
properties are defined.
• The OTEMP case control command can be included to request solution temperature output.
advantage of using SPC to specify boundary conditions is that these boundary conditions can be
changed from subcase to subcase by selecting a different SPC set inside each subcase.
• The SPC=n case control command selects either the SPC, SPC1, or SPCADD bulk entry.
The SPC condition can change between subcases.
The SPC entry can optionally be used to define a time-unassigned enforced displacement. It
cannot be defined as time-assigned. That is, it cannot be selected with the EXCITEID on the
TLOAD1 entry. For a time-assigned enforced displacement, you can use the SPCD entry. See
Mechanical Loads for information on the SPCD entry.
The SPCFORCES case control command is supported to request the SPC force output.
• The MPC=n case control command selects either the MPC or MPCADD bulk entry.
MPCs can change between subcases, so the MPC=n case control command must be defined
globally.
MPCs do not update for large displacements (PARAM,LGDISP,1).
If multiple enforced displacement conditions are applied to the same DOF, the software uses the
following precedence.
• A time-assigned enforced displacement defined with the SPCD entry, which is referenced by
the EXCITEID on the TLOAD1 entry, will overwrite time-unassigned enforced displacements
defined with the SPCD or SPC entries.
• A time-unassigned enforced displacement defined with the SPCD entry, which is referenced by
the LOAD=n case control command, will overwrite a time-unassigned enforced displacement
defined with the SPC entry.
• Edge-to-edge glue
You can define edge-to-edge glue between the edges of the axisymmetric elements CTRAX3,
CQUADX4, CTRAX6, CQUADX8, the plane stress elements CPLSTS3, CPLSTS4, CPLSTS6,
CPLSTS8, and the plane strain elements CPLSTN3, CPLSTN4, CPLSTN6, CPLSTN8. You
create edge regions with the BEDGE bulk entry, then pair the regions using the source and
target fields on the BGSET bulk entry.
Additional information.
• The inputs to define contact conditions are consistent with other solutions (the BGSET case
control and the BGSET, BGPARM, BGADD, BSURFS, BCPROPS, BEDGE bulk entries).
• The BGSET case control command must be above the subcases. As a result, the glue conditions
apply to all static and modal subcases.
• The source side element characteristics are used to define the glue stiffness. Therefore it
is possible for differences depending on which element faces or edges are selected as the
source region.
• The axisymmetric, plane stress, and plane strain elements can be defined in either the XZ plane
or in the XY plane. Edge-to-edge glue is supported in either orientation. Glue conditions are
computed on the axisymmetric elements on a 2*PI section by default.
• Only GLUETYPE=2 is supported. If you request GLUETYPE=1, the software will continue the
solution using GLUETYPE=2. See the BGPARM entry in the Quick Reference Guide for a
definition of GLUETYPE.
• When large displacement occur with PARM,LGDISP,1, the glue stiffness orientation will update
as a result of large displacement effects. The glue stiffness is only computed once, at the
beginning of a solution.
• The BGRESULTS case control command is supported to request the glue forces and tractions in
SORT1 format.
• The generalized plane strain element is not supported by glue or contact regions.
• Surface-to-surface contact
You can define surface-to-surface contact between the faces of the solid elements CTETRA,
CHEXA, CPENTA and CPYRAM. You create solid element face regions with the BSURFS, or
BCPROPS bulk entries, then pair the regions using the source and target fields on the BCTSET
bulk entry.
• Edge-to-edge contact
You can define edge-to-edge contact between the edges of the axisymmetric elements CTRAX3,
CQUADX4, CTRAX6, CQUADX8, the plane stress elements CPLSTS3, CPLSTS4, CPLSTS6,
CPLSTS8, and the plane strain elements CPLSTN3, CPLSTN4, CPLSTN6, CPLSTN8. You
create edge regions with the BEDGE bulk entry, then pair the regions using the source and target
fields on the BCTSET bulk entry.
• The BCSET case control command must be above the subcases. As a result, the contact
definition applies to all static subcases.
A modal subcase which is sequentially dependent (default) uses the final stiffness from the
previous static subcase. When contact is defined, the final stiffness from the static subcase
includes the contact stiffness.
• The contact condition can update when large displacement effects are turned on with
PARAM,LGDISP,1. The active and inactive contact elements are updated along with any force
adjustments. When sliding occurs, the orientation of the current contact stiffness are updated. If
large enough sliding occurs within a user defined tolerance, contact elements are recreated in the
current deformed configuration, contact stiffness is recomputed, and contact tractions from the
previous iteration are applied to the newly created contact elements.
• The BCTPARM entry can be used to define SOL 401 contact parameters. The OFFSET region
parameter on the BCRPARA entry is supported.
• Contact results can be requested with the BCRESULTS case control command.
• The generalized plane strain element is not supported by glue or contact regions.
1 2 3 4 5 6 7 8 9 10
NLCNTL ID Param1 Value1 Param2 Value2 Param3 Value3
Param4 Value4 Param5 Value5 -etc-
For example,
NLCNTL 1 EPSU 1E-3 EPSP 1E-3 EPSW 1E-7 +
+ CONV PW KSTEP 5 MAXITER 25
Name Description
CONV Specifies the convergence criteria. (Character = “U”, “P”, “W”, or any combination;
Default = “W”)
CREEP Include creep effects. (Character = “YES” or “NO”; Default = “YES”)
CRCERAT For the ratio of maximum creep increment to elastic strain method, the ratio of
maximum creep increment to elastic strain that is used to calculate the next time
step. Valid for creep analysis only. (Real ≥ 0.0; Default = 0.1)
CRCINC For the maximum creep increment method, the maximum creep increment that is
used to calculate the next time step. Valid for creep analysis only. (Real ≥ 0.0;
Default = 1.0E-4)
CRICOFF Creep strain increment below which the next time step is the product of the current
time step and CRMFMX. Valid for creep analysis only. (0.0 < Real < 1.0; Default
= 1.0E-6)
CRINFAC Integration factor used to calculate incremental creep strain. Valid for creep analysis
only. (0.0 ≤ Real ≤ 1.0; Default = 0.5)
CRINTS Initial time step or constant time step. Valid for creep analysis only. (Real > 0.0;
Default = 0.01)
Name Description
CRMFMN Minimum time step multiplying factor. If the next time step calculated by the
adaptive time stepping algorithm is smaller than the product of the current time step
and CRMFMN, the software halves the current time step, recalculates the current
creep strain increment, and reenters the adaptive time stepping algorithm at the
point the creep strain increment is compared to CRICOFF. Valid for creep analysis
only. (0 ≤ Real ≤ 1.0; Default = 0.1)
CRMFMX Maximum time step multiplying factor. See the CRICOFF parameter for additional
information. Valid for creep analysis only. (Real ≥ 1.0; Default = 5.0)
CRSBCDT Controls whether the first time step in a sequential subcase uses CRINTS or the
time step calculated at the end of the previous subcase. Valid for creep analysis
only. (Integer; Default = 1)
0: Use the time step calculated at the end of the previous subcase
1: Use CRINTS
CRTEABS Maximum absolute truncation error. Valid for creep analysis only. (0.0 ≤ Real <
1.0; Default = 1.0E-4)
CRTECO For the error truncation method, use CRTEABS to calculate the next time step if the
creep strain is less than CRTECO, and use CRTEREL to calculate the next time
step if the creep strain is greater than CRTECO. Valid for creep analysis only. (0.0
≤ Real < 1.0; Default = 0.01)
CRTEREL Maximum relative truncation error. Valid for creep analysis only. (0.0 ≤ Real <
1.0; Default = 0.01)
CRTSC Specifies the time stepping method. Valid for creep analysis only. (Integer or blank;
Default = 12)
0: Use constant time stepping
1: Use adaptive time stepping based on the error truncation method only
2: Use adaptive time stepping based on the ratio of maximum creep increment to
elastic strain method only
3: Use adaptive time stepping based on the maximum creep increment method only
12: Use adaptive time stepping based on both the error truncation method and the
ratio of maximum creep increment to elastic strain method
13: Use adaptive time stepping based on both the error truncation method and the
ratio of maximum creep increment method
23: Use adaptive time stepping based on both the ratio of maximum creep
increment to elastic strain method and the maximum creep increment method
123: Use adaptive time stepping based on the error truncation method, the ratio
of maximum creep increment to elastic strain method, and the maximum creep
increment method
Name Description
CRTSMN Minimum time step. If the next time step is larger than CRTSMN, the software
accepts the next time step. If the next time step is smaller than CRTSMN, the
software halves the current time step, recalculates the current creep strain
increment, and reenters the adaptive time stepping algorithm at the point the creep
strain increment is compared to CRICOFF. Valid for creep analysis only. (0.0 ≤ Real
≤ CRTSMX; Default = 0.001*CRINTS)
CRTSMX Maximum time step. If CRTSMX is set to 0.0 (default), the software accepts the next
time step. If CRTSMX is nonzero and the next time step is larger than CRTSMX,
the software uses CRTSMX as the next time step. Otherwise, the next time step is
compared to CRTSMN. Valid for creep analysis only. (Real ≥ 0.0; Default = 0.0)
EPSBOLT Bolt preload convergence tolerance. (Real > 0.0; Default = 1.0E-2)
EPSP Error tolerance on force. (Real > 0.0; Default = 1.0E-2)
EPSU Error tolerance on displacement. (Real > 0.0; Default = 1.0E-2)
EPSW Error tolerance on work. (Real > 0.0; Default = 1.0E-6)
FOLLOWK Include follower stiffness. Follower stiffness is computed from the follower loads
defined with the FORCE1, FORCE2, PLOAD, or PLOAD4 entries. (Character =
“YES” or “NO”; Statics default = “NO”; Modal default = “YES”)
ITRBOLT Maximum number of bolt iterations before the bolt preload calculation is considered
non-converged. (Integer > 0; Default = 20)
KUPDATE Stiffness update strategy. (Integer; Default = 0)
-1: Initial stiffness approach
0: Auto stiffness update
1: Full Newton-Raphson
> 1: Quasi Newton-Raphson, and KUPDATE is the number of iterations before
a stiffness update
MAXBIS Maximum number of bisections allowed. (0 < Integer < 10; Default = 5)
MAXDIV Number of divergences before solution is assumed to diverge. (Integer > 0; Default
= 3)
MAXITER Maximum number of iterations per time step. (Integer > 0; Default = 25)
Name Description
DISP Relative error in terms of displacements. See Error functions and weighted
normalization.
LOAD Relative error in terms of loads. See Error functions and weighted
normalization.
WORK Relative error in terms of work. See Error functions and weighted normalization.
NO. OF ITR DIV Number of occurrences of probable divergence during the iteration. See
Divergence criteria.
STIFFNESS Value for the current stiffness parameter for the current iteration.
PARAMETER
CURRENT
STIFFNESS % Change in the value for the current stiffness parameter between and prior
PARAMETER % iteration.
CHANGE
• The element iterative solver, which is already supported in SOL 101, performs particularly well
with solid element-dominated models. It may be a faster choice if lower accuracy is acceptable.
As in SOL 101, convergence tolerances and other options may be set by supplying an SMETHOD
card in case control and matching ITER card in bulk data.
• For problems involving contact and 3D solid elements, the element iterative solver is generally
faster as compared to the sparse direct solver.
• The PARDISO solver is a hybrid direct-iterative solver, potentially faster with larger numbers of
cores than the sparse solver but with slightly lower accuracy.
• The axisymmetric elements CQUADX4, CQUADX8, CTRAX3, CTRAX6, the plane strain
elements CPLSTN3, CPLSTN4, CPLSTN6, CPLSTN8, and the plane stress elements CPLSTS3,
CPLSTS4, CPLSTS6, CPLSTS8 are supported for linear, geometric nonlinear, and material
nonlinear analysis.
The grid points on these elements must all lie in either the XZ plane, or all in the XY plane of the
basic coordinate system. The software automatically determines the orientation.
When axisymmetric elements are defined on the XZ plane, X is the radial direction, and Z is the
axial direction. The grid points defining these elements must have X ≥ 0.
When axisymmetric elements are defined on the XY plane, Y is the radial direction, and X is the
axial direction. The grid points defining these elements must have Y ≥ 0.
• A special, generalized plane strain formulation is available using the CPLSTN3, CPLSTN4,
CPLSTN6, and CPLSTN8 element types. See Generalized plane strain analysis.
• The RBE2 and RBAR rigid elements are supported with optional large displacement effects and
thermal expansion. The RBE3 rigid element is also supported, but it does not support the large
displacement effects or thermal expansion. See Rigid element support.
• The PSOLID or the PCOMPS bulk entries define the element properties. The PCOMPS is
optionally used to define a layered solid composite property.
The PGPLSN bulk entry is defines the special generalized plane strain element properties.
• You can request stress norm, stress error norm, strain energy norm, and strain energy error
norm output. The output is computed and stored on an individual element basis. The NX
Advanced Simulation product uses the output for adaptive meshing. See Error estimator for
mesh refinement.
• You can compute and output the j-integral in a crack simulation. The j-integral output can
be requested and used by third-party software like Zencrack to perform a fracture mechanics
analysis. The CHEXA bulk entry allows for a collapsed element definition. See Crack simulation.
• The control grid point. The control grid point is the location where out-of-plane loads or enforced
displacements are applied to the set of elements that reference the PGPLSN bulk entry.
For the generalized plane strain analysis, NX Nastran calculates the standard in-plane plane strain
stiffness, but also calculates three net out-of-plane stiffness values relative to the displacement
coordinate system of the control grid point. Consequently, how you specify the displacement
coordinate system for the control grid point is very important. You should specify the displacement
coordinate system of the control grid point such that one axis is normal to the cross section and the
other two axes are parallel to the principal axes of the cross section. By so doing, the three net
out-of-plane stiffness values that NX Nastran calculates represent the normal stiffness of the cross
section, and the two bending stiffness for symmetrical bending of the cross section.
Because the CPLSTN3, CPLSTN4, CPLSTN6, and CPLSTN8 plane strain elements can only be
defined in the XY- or XZ-planes of the basic coordinate system, the direction normal to the cross
section is always in the Z- or Y-direction, respectively of the basic coordinate system. NX Nastran
checks that one of the axes of the displacement coordinate system of the control grid point coincides
with the correct normal direction and issues an error if one does not.
NX Nastran does not check the other two coordinate directions of the displacement coordinate
system for the control grid point. It is your responsibility to assure that these directions are parallel to
the principal axes of the cross section.
If you specify additive stiffness, the normal stiffness is added to the normal stiffness that NX Nastran
calculates for the cross section . The additive rotational stiffness values are added to the bending
stiffness values as follows:
• If the model lies in the XY-plane of the basic coordinate system, the KR1 value on the PGPLSN
bulk entry is added to the bending stiffness about the X-axis of the displacement coordinate
system of the control grid point. the KR2 value on the PGPLSN bulk entry is added to the bending
stiffness about the Y-axis of the displacement coordinate system of the control grid point.
• If the model lies in the XZ-plane of the basic coordinate system, the KR1 value on the PGPLSN
bulk entry is added to the bending stiffness about the X-axis of the displacement coordinate
system of the control grid point. the KR2 value on the PGPLSN bulk entry is added to the bending
stiffness about the Z-axis of the displacement coordinate system of the control grid point.
You can apply loads to the control grid point and to the generalized plane strain element mesh. At the
control grid point, you can account for the centrifugal force that is attributable to the portion of the
blade from the cross section you are modeling to the blade tip. To allow you to define a mechanically
equivalent loading at the control grid point, you can specify not only a force that acts normal to the
cross section, but also the bending moments that act on the cross section about axes parallel to the
principal axes of the cross section.
To the generalized plane strain mesh, apply surface tractions, body forces, and in-plane enforced
displacements that you want to include in the analysis. For example, you can apply aerodynamic
forces to the grid points that lie on the periphery of the mesh.
From the net out-of-plane stiffness values and the loads that are applied to the control grid point, NX
Nastran calculates the thickness change over the cross section. Similar to planes remaining plane in
pure bending of beams, NX Nastran enforces that the surface defined by the thickness change is
planar. From the thickness change over the cross section, NX Nastran calculates the out-of-plane
strain of the elements at the grid locations. During the solution of the finite element model, NX
Nastran uses the out-of-plane strain and any surface tractions, body forces, and in-plane enforced
displacements that you specified.
If an enforced displacement and enforced rotations are applied at the control grid point, the thickness
change of the cross section is directly specified. From the thickness change, NX Nastran calculates
the out-of-plane strain directly and the solution of the finite element model is as before.
Note the generalized plane strain element is not supported by glue or contact regions.
The following constitutive models are available with generalized plane strain elements:
• To model plasticity of an isotropic material, use the MAT1 and MATS1 bulk entries in combination.
• To model plasticity of an orthotropic material, use the MAT3 and MATS1 bulk entries in
combination. The elastic portion of the response is orthotropic, and the plastic portion of the
response is isotropic.
• To model creep of an isotropic material, use the MAT1 and MATCRP bulk entries in combination.
• To model creep of an isotropic material with temperature-dependent properties, use the MAT1,
MATT1, and MATCRP bulk entries in combination.
• To model creep of an orthotropic material, use the MAT3 and MATCRP bulk entries in combination.
• To model creep of an orthotropic material with temperature-dependent properties, use the MAT3,
MATT3, MATCRP, and TABLEM1 bulk entries in combination. The elastic portion of the response
is orthotropic, and the creep portion of the response is isotropic.
For additional information, see the PGPLSN bulk entry in the NX Nastran Quick Reference Guide.
• When RIGID=AUTO and PARAM,LGDISP,1, the software automatically applies the RIGID=STIFF
option. RBE2 and RBAR elements include large displacement effects and thermal expansion.
• When RIGID=STIFF and PARAM,LGDISP,-1, the large displacement effects are not included, in
general. RBE2 and RBAR elements include thermal expansion.
• When RIGID=STIFF and PARAM,LGDISP,1, the RBE2 and RBAR element behavior is the same
as RIGID=AUTO and PARAM,LGDISP,1. RBE2 and RBAR elements include large displacement
effects and thermal expansion.
• When RIGID=LINEAR, the RBE2 and RBAR elements do not include large displacement effects
or thermal expansion. This behavior is independent of the PARAM,LGDISP setting.
To compute large displacement effects and thermal expansion, the software internally replaces the
RBE2 and RBAR elements with either a stiff beam element, or a stiff spring element. A coincident grid
tolerance is used to determine if a beam or a spring is used. For the RBAR, if the distance between
the connecting grids is less than the tolerance, the stiff spring formulation is used. For the RBE2, if
the distance between the grid defined in the GN field on the RBE2 entry, and any of the grids defined
in the GM fields on the RBE2 entry, is less than the tolerance, the stiff spring formulation is used. You
can optionally define the coincident grid tolerance explicitly with the parameter RGLCRIT. By default,
it is automatically computed by the software:
Coincident Grid Tolerance = 1E-6 * LMODEL (units=length)
where LMODEL is the largest dimension of the model determined by the software.
You can optionally define the beam stiffness and area explicitly using the parameters RGBEAME and
RGBEAMA, respectively. By default, they are automatically computed by:
Beam Stiffness = 1e+2 * EMAX (units=force/length^2)
Beam Area = (LMODEL * 1e-2)^2 (units=length^2)
where EMAX is the largest Young’s modulus in the model. If no material is specified in the model,
EMAX is set to 1.0E12.
You can optionally define the spring stiffness explicitly using the parameter RGSPRGK. By default, it
is automatically computed by:
Spring Stiffness = EMAX * LMODEL (units = force/length)
Additional information:
• MPCFORCE and GPFORCE output are supported with all of the rigid elements. Since the
software internally replaces an RBAR or RBE2 with a stiff beam or spring element when
RIGID=STIFF, these elements are no longer represented as MPC equations. As a result,
MPCFORCE output is not applicable to these elements. GPFORCE and FORCE output is
applicable.
• GPFORCE output will correctly account for large displacements, except for DOF which are
included in MPC equations.
• The TEMP(LOAD) and TEMP(INIT) value used on RBAR elements is an average calculated
from the grid point values. On RBE2 elements, an average TEMP(LOAD) and TEMP(INIT) is
calculated for each leg of the element using the values on the independent/dependent grid pairs
such that each leg can have a different thermal strain if the temperatures vary at the grids.
The rigid element thermal strains are calculated from
εthermal = α(AVGTEMP(LOAD) – AVGTEMP(INIT))
If TEMP(LOAD) or TEMP(INIT) are not defined, they are assumed to be zero.
where
Ω is the element volume,
σunaveraged is the unaveraged stress vector,
σaveraged is an averaged stress vector computed at a grid point using the stress vectors from elements
connected to the grid point,
D matrix is the constitutive relation.
When computing σaveraged, stress values are not averaged across different element families, material
properties, material coordinate systems, orientation angles in 2D solid elements, and thicknesses in
plane stress elements.
You use the STRESS, STNERGY and STEP describers on the new ADAPTERR case control
command to request the output.
• The ADAPTERR case control command must be defined above the subcases (globally).
• The STRESS describer requests the stress norm and the stress error norm.
• The STNERGY describer requests the strain energy norm and strain energy error norm.
• You can specify both the STRESS and STNERGY describers to request stress norm, stress error
normal, strain energy norm and strain energy error norm output.
• The software always outputs the maximum value on each element, for each output type
requested, by comparing the values from all solution steps. In addition, if you specify the STEP
describer, the software will output what you have requested at the output increment steps defined
with the TSTEP1 entries.
• The creation of the CRAKTP bulk entry. You can use the CRAKTP bulk entry to specify
information related to the crack tip.
For additional information, see the CRAKTP bulk entry in the NX Nastran Quick Reference Guide.
• The creation of the VCEV bulk entry. You can use the VCEV bulk entry to define virtual crack
tip extension vectors.
For additional information, see the VCEV bulk entry in the NX Nastran Quick Reference Guide.
• The modification of the CHEXA bulk entry to allow for collapsed CHEXA element definition. Note
that the collapsed CHEXA element is not supported in a glue or contact region.
• The creation of the COLPHEXA parameter. You can allow collapsed CHEXA elements to bypass
input file checks with the new COLPHEXA parameter. To do so, specify PARAM,COLPHEXA,YES
in the bulk section of the input file.
1 2 3 4 5 6 7 8 9 10
G14 G16 G17 G6 G19 G20
where the same grid ID is entered in the G2, G3, and G10 fields, another grid ID is entered in
both the G14 and G15 fields, and another grid ID is entered in the G6, G7, and G18 fields.
• In Format 2, 20 unique grid IDs are specified in the 20 grid ID fields of the CHEXA bulk entry.
However, eight of the grid IDs do not have unique coordinates. Format 2 is typically used for
elasto-plastic material models. With Format 2, mid-side grids remain at the mid-span locations.
For the collapsed CHEXA element shown in Figure 3-2, the Format 2 specification is as follows:
1 2 3 4 5 6 7 8 9 10
CHEXA EID PID G1 G2 G3 G4 G5 G6
G7 G8 G9 G10 G11 G12 G13 G14
G15 G16 G17 G18 G19 G20
where the grids entered in the G2, G3, and G10 fields would share the same coordinates, the
grids entered in the G14 and G15 fields would share the same coordinates, and the grids entered
in the G6, G7, and G18 fields would share the same coordinates. Unlike Format 1 where grids in
the CHEXA element connectivity are merged, Format 2 does not merge coincident grids in the
CHEXA element connectivity. Thus, these grids can move independently of one another.
as shown in Figure 3-3. The relation between volume and Cartesian coordinates can be established
for a linear tetrahedron (without midside nodes) as follows:
Equation 3-4.
The strains are calculated as
Equation 3-5.
where strain vector
Equation 3-6.
element matrix
Equation 3-7.
nodal displacement vector
Equation 3-8.
with
Equation 3-9.
Equation 3-10.
Because the shape functions are defined in terms of local coordinates, a coordinate transformation is
required to obtain global derivatives. Volume coordinates involve four coordinates (one of which is
dependent), and the Jacobian matrix will become rectangular. To avoid this difficulty, let us introduce
local coordinates (ξ,η,ζ) as follows:
Equation 3-11.
Derivatives with respect to the local coordinates can be expressed in terms of global derivatives
using the chain rule, i.e.,
Equation 3-12.
where the Jacobian matrix is expressed in terms of shape functions as
Equation 3-13.
This Jacobian matrix must be inverted to obtain global derivatives, from which the element matrix is
formed. Notice that the determinant of the Jacobian matrix is called Jacobian which represents a
volume change, i.e.,
Equation 3-14.
Derivatives of the shape functions with respect to the local coordinates are obtained as follows:
These shape functions and derivatives may be reduced to those for 4-noded and 10-noded
tetrahedron elements. It can be verified that, for any combination of deleted midside nodes, there
exist a condition
where (xij = xi - xj) is used for convenience. Upon inverting the Jacobian matrix, we have
where
Equation 3-15.
While analytical integration is possible, there are some advantages in using numerical integration.
The Gaussian quadrature is used for the tetrahedron as usual. The minimum number of intergration
points required for non-singular stiffness matrix may be determined based on
Equation 3-16.
These numbers are found to be 1 and 4 for linear and quadratic tetrahedron, respectively. Using
the integration formulas, shown in Figure 3-4, the element stiffness matrices for linear and quadratic
tetrahedron may be computed as
Equation 3-18.
where [Ň] = [N1I N2I ..... N10I], with I being (3x3) identity matrix. Again the Gaussian quadrature
4-point formula is used to find
Equation 3-19.
Notice that 4-point scheme is to be used even for 4-noded tetrahedron. For computational
convenience, the consistent mass may be converted to the lumped mass. One way to achieve this is
to take the diagonal terms and scale them so that the total mass is preserved.
Equation 3-20.
Equation 3-22.
where {εo}T = αΔT < 1 1 1 0 0 0 >. With anisotropic thermal properties, {εo}T becomes ΔT < α1,
α2 ... α6 > in general.
The pressure load applied to any surface of the tetrahedron may be distributed to the grid points using
the shape functions applicable to the 6-noded triangular element with removable midside nodes, i.e.,
Equation 3-23.
where {n} is a unit direction vector associated with a pressure (p) and
with I being (3 x 3) identity matrix. Shape functions (Ni’) are shown in Figure 3-5. Area integration
should be performed using Gaussian quadrature formulas shown in Figure 3-6, i.e.,
Equation 3-24.
During the nonlinear iteration process, the equilibrium is not reached until convergence is achieved.
The equilibrium is sought in the global level when the residual load R approaches zero. The residual
load vector is defined as
Equation 3-25.
where {P} is the applied load vector including the thermal load and Σ implies assemblage in the global
coordinate system (coordinate transformation required).
Equation 3-26.
Equation 3-27.
Then, the transformation from element coordinates to the basic coordinates is simply
Equation 3-28.
where
Equation 3-29.
and < xe, ye, ze > Tbasic is the position vector of the element coordinate system with respect to the
basic coordinate system. Notice that [Tbe] is an orthogonal matrix, i.e.,
Equation 3-30.
with
where {σ}, [B], {u} and {α} are defined in the element coordinate system, and the shape function
{Ni} interpolates Gauss point temperatures from the nodal temperatures. In case of nonlinear
analysis, stresses are computed again upon convergence, starting from the last converged state (last
converged solution of σ and u), i.e.,
Equation 3-31.
with
the nonlinear material routine computes updated stresses {σnew}, which are stored in ESTNL.
To output grid point stresses, the stresses, the integration points must be extrapolated to the nodal
points. A linear extrapolation can be applied using stresses at the element c.g. and the corner Gauss
point associated with a grid point. For a tetrahedron, referring to Figure 3-4,
Equation 3-32.
where
This yields
Equation 3-33.
In matrix form for all the grid points in the element
Equation 3-34.
where
This process must be operated on every component of stress vector. No extrapolation is required
for 1-point integration.
• The MAT3 and MATT3 (temperature dependent) bulk entries define isotropic materials.
• The MAT9 and MATT9 (temperature dependent) bulk entries define anisotropic materials.
• The MAT11 and MATT11 (temperature dependent) bulk entries define orthotropic materials.
• Plastic and creep materials can optionally be assigned to the 3D solid elements, axisymmetric
elements, the plane stress elements, and the plane strain elements. You can enable one or both
plasticity/creep in all subcases, or in specific subcases.
User interface
• The MATNL parameter. With the MATNL parameter, you can globally switch the plasticity
analysis capability on or off.
• The PLASTIC parameter is available on the NLCNTL bulk entry to optionally turn off the plasticity
capability in a subcase.
• The MATS1 bulk entry allows you to define stress versus plastic strain tabular data.
2. Specify PARAM,MATNL,1.
3. Include a NLCNTL case control command that points to a NLCNTL bulk entry.
If your input file contains subcases, and you want to include the effects of plasticity in specific
subcases, but not others, you have two options.
Option 1: Use a global NLCNTL case control command
1. Specify PARAM,MATNL,1.
2. Include a NLCNTL case control command above the subcases that points to a NLCNTL bulk entry.
3. On the NLCNTL bulk entry pointed to by the global NLCNTL case control command, specify
any applicable parameters.
4. In the subcases that you want to disable the plasticity analysis capability, include a NLCNTL case
control command that points to a NLCNTL bulk entry.
5. On the NLCNTL bulk entry pointed to by the NLCNTL case control commands in the subcases,
specify “PLASTIC” in a PARAMi field and “NO” in the corresponding VALUEi field.
2. Include NLCNTL case control commands in each subcase. Multiple NLCNTL case control
commands can point to a single NLCNTL bulk entry.
3. In subcases that you want to enable the plasticity analysis capability, have the NLCNTL case
control command point to a NLCNTL bulk entry with any applicable parameters specified.
4. In subcases that you want to disable the plasticity analysis capability, have the NLCNTL case
control command point to a NLCNTL bulk entry that has “PLASTIC” specified in a PARAMi field
and “NO” specified in the corresponding VALUEi field.
In a SOL 401 plasticity analysis, the property bulk entry referenced by all non-rigid elements must
reference a MAT1 bulk entry and a MATS1 bulk entry that have the same material identification
number. If the properties on the MAT1 bulk entry are temperature-dependent, include a MATT1 bulk
entry with the same material identification number.
On the MATS1 bulk entry, specify TYPE = “PLASTIC” or “PLSTRN” to select the strain type in tabular
data used to describe a multilinear stress-strain curve. Specify TYPE = “PLASTIC” if you want to
use total strains. Specify TYPE = “PLSTRN” if you want to use plastic strains. Total and plastic
strains are related as follows:
where
To describe a bilinear stress-strain curve, specify either TYPE = “PLASTIC” or “PLSTRN” and enter
the work hardening slope, H, directly.
For additional information, see the MATS1 bulk entry in the NX Nastran Quick Reference Guide.
The von Mises yield function for an isotropic hardening material may be stated as:
J2 is the second invariant of the deviatoric stress tensor. The deviatoric stresses, sij, are given by:
where σm is the mean or average normal stress, and δij is the Kronecker delta.
Isotropic hardening assumes that the uniaxial yield stress, σy, is a function of plastic straining.
For kinematic or combined hardening, the von Mises yield function is
J ́2 is the second invariant of the shift stress that is defined as the deviatoric stress minus the
back stress. The back stress may also be thought of as the position of the center of the yield
surface in the six-dimensional stress space. For initially isotropic materials, the back stress
tensor components are initially zero.
2. A flow rule determines the relative magnitudes of the components of the plastic strain increment
tensor.
The software uses an associated flow rule in which the plastic potential function, g, is the same
as the yield function, f, and the components of the plastic strain increment are given by:
In vector form, for the von Mises yield surface with kinematic hardening, the plastic flow vector is
given by:
For isotropic hardening, this equation reduces to the well-known Prandtl-Reuss equations.
3. A hardening rule defines the changes in the yield function as a result of plastic straining.
Isotropic Hardening
When you select isotropic hardening, the software uses a piece-wise linear stress-strain curve.
The isotropic hardening assumption isn't very realistic for most materials subjected to cyclic
loading. However, it is relatively simple and efficient.
Isotropic hardening assumes that the yield surface expands uniformly as a result of plastic
straining. This assumption is achieved by making the yield stress a function of the integrated
effective plastic strain increments, which for a von Mises material is:
The slope of the stress plastic strain curve, Ep, is called the plastic modulus. It can be obtained from
the uniaxial stress-strain curve and is defined by:
Kinematic Hardening
When you select kinematic hardening, the software assumes a bilinear stress-strain curve. If the
material database contains a multilinear representation, only the yield point and the tangent modulus
of the first segment beyond the yield point are used to characterize the stress-strain behavior.
Kinematic hardening assumes that the yield surface translates in the stress space but doesn't change
size or shape. The yield stress, σy, doesn't change, but the back stress, αij, is a function of plastic
straining. The Ziegler-Prager kinematic hardening is one of the most widely used models. This model
assumes that the back stress increment is in the direction of the stress minus the back stress.
where,
The reduced effective plastic strain associated with isotropic hardening is related to the effective
plastic strain by the following:
where De is the elastic modulus matrix, εp is the plastic strain, is the effective plastic strain, and
TYF is the von Mises yield function.
The software will compute a consistent tangent modulus for use in generating the tangent stiffness
matrix. When this procedure is used with a full Newton-Raphson iteration, quadratic convergence can
be obtained. A detailed description of this procedure may be found in Crisfield, 1991.
References:
• Crisfield, M. A., Non-linear Finite Element Analysis of Solids and Structures (Chichester: John
Wiley & Sons, 1991).
• Chakrabarty, J., Theory Of Plasticity (New York: McGraw-Hill Book Company, 1987), 55-119.
• Chen, W. F. and Han, D. J., Plasticity for Structural Engineers, (New York: Springer-Verlag,
1988), 239-281.
Equation 4-1.
where εec is the effective creep strain, σe is effective stress, t is time, and A, B, and D are user-defined
coefficients. Because the model uses effective stress and effective creep strain, the values for the
coefficients are directly relatable to results from uniaxial testing.
In the Bailey-Norton model, temperature is not accounted for explicitly. To account for
temperature-dependence, you can define the coefficients as tabular functions of temperature.
For some very simple cases, you can use Equation 4-1 directly to calculate the effective creep strain
as a function of time. However, for the general case where temperature and stress vary, and computer
simulation is required, Equation 4-1 is applied incrementally over a finite number of time steps.
During the creep analysis, the incremental creep strain for each time step is calculated by numerically
integrating the instantaneous creep strain rate. The formula for creep strain rate is obtained from
the following flow rule:
Equation 4-2.
where έijc are the components of the creep strain rate tensor, έec is the effective creep strain rate, and
Sij are the components of the deviatoric stress tensor.
The effective creep strain rate is obtained by differentiating Equation 4-1 with respect to time.
To evaluate each increment of creep strain, the software performs a numerical integration based
on the generalized trapezoidal rule as follows:
Equation 4-3.
where Δtn = tn – tn-1 is the duration of the subinterval, and β is a user-defined numerical integration
parameter. Generally, the default value for β of 0.5 is appropriate.
User interface
• With the MATNL parameter, you can globally switch the creep analysis capability on or off.
For more information, see the MATNL parameter.
• Parameters are available for use with the NLCNTL bulk entry. These parameters allow you to
turn off the creep capability in subcases, control adaptive time stepping or define a constant time
step, and define the integration factor in Equation 4-3.
For more information, see the NLCNTL bulk entry in the NX Nastran Quick Reference Guide.
• With the MATCRP bulk entry, you define parameters related to the creep constitutive model.
For more information, see the MATCRP bulk entry in the NX Nastran Quick Reference Guide.
2. Specify PARAM,MATNL,1.
3. Include a NLCNTL case control command that points to a NLCNTL bulk entry.
If your input file contains subcases, and you want to include the effects of creep in specific subcases,
but not others, you have two options.
Option 1: Use a global NLCNTL case control command
1. Specify PARAM,MATNL,1.
2. Include a NLCNTL case control command above the subcases that points to a NLCNTL bulk entry.
3. On the NLCNTL bulk entry pointed to by the global NLCNTL case control command, specify
any applicable parameters.
4. In the subcases that you want to disable the creep analysis capability, include a NLCNTL case
control command that points to a NLCNTL bulk entry.
5. On the NLCNTL bulk entry pointed to by the NLCNTL case control commands in the subcases,
specify “CREEP” in a PARAMi field and “NO” in the corresponding VALUEi field.
2. Include NLCNTL case control commands in each subcase. Multiple NLCNTL case control
commands can point to a single NLCNTL bulk entry.
3. In subcases that you want to enable the creep analysis capability, have the NLCNTL case control
command point to a NLCNTL bulk entry with any applicable parameters specified.
4. In subcases that you want to disable the creep analysis capability, have the NLCNTL case control
command point to a NLCNTL bulk entry that has “CREEP” specified in a PARAMi field and “NO”
specified in the corresponding VALUEi field.
To directly define solution times for the creep analysis, include a TSTEP case control command in
your input file that points to a TSTEP1 bulk entry. On the TSTEP1 bulk entry, you can specify the
solution times and specify which solution times you want results output.
The solution times you specify on the TSTEP1 bulk entry may result in time steps that are either too
coarse to produce accurate results, or too fine to produce results efficiently. To assist you in avoiding
such problems, the software uses adaptive time stepping by default. You can tweak the adaptive
time stepping algorithm or override adaptive time stepping altogether with new parameters for the
NLCNTL bulk entry. For more information on adaptive time stepping, see Time step control.
In a SOL 401 creep analysis, the property bulk entry referenced by all non-rigid elements must
reference a MAT1 bulk entry and a MATCRP bulk entry that have the same material identification
number. If the properties on the MAT1 bulk entry are temperature-dependent, include a MATT1 bulk
entry with the same material identification number.
You use the MATCRP bulk entry to specify:
• The stress threshold below which creep does not occur.
You can specify the coefficients in the Bailey-Norton creep model as either constant or as a function
of temperature. To specify a coefficient as temperature-dependent, enter the identification number
of a TABLEM1 bulk entry in the corresponding A, B, or D field of the MATCRP bulk entry. On the
TABLEM1 bulk entry, enter tabular data that describes how the coefficient varies with temperature. At
present, a MATCRP bulk entry cannot reference a TABLEM2, TABLEM3, or TABLEM4 bulk entry.
The time steps that result from the TSTEP1 bulk entry specification may be too coarse to produce
accurate results, or too fine to produce results efficiently. By default, the software uses an adaptive
time stepping algorithm to avoid such problems.
You can tweak the adaptive time stepping algorithm or override adaptive time stepping altogether and
use a constant time step with new parameters for the NLCNTL bulk entry. The new parameters are:
CRCERAT Ratio of maximum creep increment to elastic strain that is used to adaptively
specify the next time step.
CRCINC Maximum creep increment that is used to adaptively specify the next time step.
CRICOFF Creep strain increment below which the next time step is the product of the
current time step and the maximum time step multiplying factor.
• Override adaptive time stepping altogether and have the software use the value of the CRINTS
parameter as a constant time step.
• The next time step is based on the ratio of maximum creep increment to elastic strain criterion.
You specify the ratio of maximum creep increment to elastic strain with the CRCERAT parameter.
• The next time step is based on the maximum truncation error criterion. For this option, you
have three sub-options.
o Use the maximum absolute truncation error. You specify the maximum absolute truncation
error with the CRTEABS parameter.
o Use the maximum relative truncation error. You specify the maximum relative truncation
error with the CRTEREL parameter.
o Use the maximum absolute truncation error if the creep strain is less than the value specified
by the CRTECO parameter, and use the maximum relative truncation error if the creep strain
is greater than the value specified by the CRTECO parameter.
• The next time step is the shortest time step calculated by any combination of the maximum creep
strain increment, ratio of maximum creep increment to elastic strain, and maximum truncation
error criteria.
When the creep simulation begins, the value of the CRINTS parameter is always used as the first
time step. If adaptive time stepping is overridden, the value of the CRINTS parameter is used as a
constant time step throughout the simulation.
If adaptive time stepping is not overridden, after each time step the software compares the calculated
creep strain increment to the value of the CRICOFF parameter. If the creep strain increment is
greater than the value of the CRICOFF parameter, the software uses the adaptive time stepping
algorithm to calculate the next time step. If the creep strain increment is less than the value of
the CRICOFF parameter, the software uses the product of the current time step and the value of
the CRMFMX parameter as the next time step.
If the software uses the adaptive time stepping algorithm to calculate the next time step, the next time
step is compared to the product of the current time step and the value for the CRMFMN parameter. If
the next time step is smaller than the product of the current time step and the value for the CRMFMN
parameter, the software halves the current time step, recalculates the current creep strain increment,
and reenters the algorithm at the point the creep strain increment is compared to the value of the
CRICOFF parameter. If the next time step is larger than the product of the current time step and the
value of the CRMFMN parameter, the software keeps the next time step.
The next time step is then compared against the values of the CRTSMX and CRTSMN parameters.
First, the software checks to see if the value of the CRTSMX parameter is 0.0. If so, the software
accepts the value for the next time step and uses it to compute the next creep strain increment. If the
value of the CRTSMX parameter is not set to 0.0, the next time step is compared to the value of the
CRTSMX parameter. If the next time step is larger than the value of the CRTSMX parameter, the
software uses the value of the CRTSMX parameter as the next time step and uses it to compute
the next creep strain increment. If the next time step is smaller than the value of the CRTSMX
parameter, the next time step is compared to the value of the CRTSMN parameter. If the next time
step is smaller than the value for the CRTSMN parameter, the software halves the current time step,
recalculates the current creep strain increment, and reenters the algorithm at the point the creep
strain increment is compared to the CRICOFF parameter. If the next time step is larger than the
value of the CRTSMN parameter, the software accepts the value for the next time step and uses
it to compute the next creep strain increment.
The adaptive time stepping algorithm is summarized by the following flowchart. In the flowchart, the
notation for the value of a parameter is Vparameter name.
The effective stress in terms of the stress tensor components ij and deviatoric stress tensor
components sij is given by:
• A flow rule determines the creep strain rate tensor components for a general state of stress.
• A hardening rule determines creep strain rates from the uniaxial rate equation under changing
stress and temperature.
Strain Hardening
The strain hardening rule assumes that the material state is determined by the accumulated effective
creep strain. When stress or temperature changes, the shift from one creep curve to another is based
on the accumulated effective creep strain. This shift is illustrated in the following figure.
The strain hardening rule is generally considered to be superior to the time hardening rule and is,
therefore, the default for the software's creep equations. Under conditions of constant stress and
temperature, the time hardening and strain hardening assumptions produce identical results.
Time Hardening
The time hardening rule uses time as a material state variable. It assumes that the material state is
determined by the length of time the material has been creeping.
When stress or temperature changes, the shift from one creep curve to another is based on the
accumulated creep time. This shift is illustrated in the following figure.
The time hardening rule is usually not very realistic when very large changes in temperature or
stress occur over the interval of interest.
With the maximum truncation error criterion, NX Nastran also calculates the next time step by scaling
the current time step. The value of the CRTECO parameter on the NLCNTL bulk entry determines
whether the maximum absolute truncation error or the maximum relative truncation error is used to
adaptively specify the next time step.
• If the effective creep strain is less than the value of the CRTECO parameter, the maximum
absolute truncation error is used. The maximum absolute truncation error is specified with the
CRTEABS parameter on the NLCNTL bulk entry.
• If the effective creep strain is greater than the value of the CRTECO parameter, the maximum
relative truncation error is used. The maximum relative truncation error is specified with the
CRTEREL parameter on the NLCNTL bulk entry.
The next time step is calculated by scaling the current time. When the maximum absolute truncation
error is applied, the software computes the next time step as:
When the maximum relative truncation error is applied, the software computes the next time step as:
where e is the calculated truncation error, and is the current effective creep strain increment.
There are two approaches to calculate the truncation error e corresponding to the two different
scenarios:
1. At the first time step or if β=0 or β=1, the two point rule truncation error is used.
2. From the second time step and 0 < β< 1, the three point rule truncation error is used.
With the maximum creep strain increment criterion, NX Nastran calculates the next time step by
scaling the current time step as follows:
Equation 4-4.
where CRCINC is the value of the CRCINC parameter and Δεec is the current effective creep strain
increment.
With the ratio of maximum creep increment to elastic strain criterion, NX Nastran calculates the next
time step by scaling the current time as follows:
Equation 4-5.
where CRCERAT is the value of the CRCERAT parameter, Δεec is the current effective creep strain
increment, and εeE is the current total effective elastic strain.
• Use TYPE = “ELEM” to reference a GROUP bulk entry that includes a list of elements. The
MATOVR specification applies to the listed elements.
• Use TYPE = “PROP” to reference a GROUP bulk entry that includes a list of properties. The
MATOVR specification applies to all elements that reference the properties listed in the GROUP
bulk entry.
For more information, see the MATOVR bulk entry in the NX Nastran Quick Reference Guide.
Surface-to-surface contact conditions allow the solution to search and detect when a pair of element
faces come into contact. The contact conditions prevent the faces from penetrating and allow finite
sliding with optional friction effects. The surface-to-surface contact source and target regions
consist of solid element faces. From element faces in the source region, a top and bottom normal is
projected. The software creates a contact element if:
• Any of the source element normals intersect with an element in the target region.
• The distance between the two faces is equal to or less than the defined separation distance.
Edge-to-edge contact conditions allow the solution to search and detect when a pair of element edges
come into contact. The software creates contact elements if the distance between the edges is equal
to or less than the search distance which you specify. The contact elements prevent the edges from
penetrating and allow finite sliding with optional friction effects.
The axisymmetric, plane stress, and plane strain elements can be defined in either the XZ plane or in
the XY plane. Edge-to-edge contact is supported in either orientation. The generalized plane strain
element is not supported by glue or contact regions.
Subcase control
The BCSET case control command selects the contact conditions defined in the bulk data. The
BCSET command must be above the subcases. As a result, a contact definition applies to all static
subcases. A modal subcase which is sequentially dependent (default) uses the final stiffness from
the previous static subcase. When contact is defined, the final stiffness from the static subcase
includes the contact stiffness.
Solver options with contact
For models with 3D solid elements and contact, the element iterative solver is generally recommended
over the default sparse solver for better performance. The SOL 401 element iterative solver is select
by setting the SOLVER parameter on the NLCNTL bulk entry to “ELEMITER”.
• The BCPROPS entry is defined by its own unique ID and is a list of solid element property IDs.
The free faces of the solid elements selected with a property ID are automatically determined
by the software.
A contact surface can be defined with any face of a solid element. Although parabolic faces with
omitted midside nodes are permitted, their use could affect accuracy.
All region IDs defined with the BEDGE, BSURFS, or BCPROPS entries must be unique.
• For surface-to-surface contact pairs, SID and TID are contact regions defined with the BCPROPS
and BSURFS entries.
The minimum and maximum search distance fields (MINDi and MAXDi) define a range in which the
solver can initially determine if the distance between element edges or faces in a particular pair are
within the threshold for creating contact elements. These values are used to determine where contact
elements are created. The minimum distance can be negative if there is an interference condition
modeled as overlapping regions.
The contact condition can update when large displacement effects are turned on with
PARAM,LGDISP,1. The active and inactive contact elements are updated along with any force
adjustments. When sliding occurs, the orientation of the current contact stiffness are updated. If large
enough sliding occurs within a user defined tolerance, contact elements are recreated in the current
deformed configuration, contact stiffness is recomputed, and contact tractions from the previous
iteration are applied to the newly created contact elements.
See the section “Contact Control Parameters - BCTPARM” for more information on contact
parameters.
Contact with Composite Solid Faces
Defining contact regions and pairs on composite solid faces which are perpendicular to the stack
direction (edge faces) may produce poor stress continuity. If the contact definition is between edge
faces belonging to different PCOMPS definitions, and if the number of plies on each PCOMPS
definition is small and the same, and the ply thicknesses are similar, the stress continuity should be
fairly smooth. This also applies to the results requested with the BCRESULTS case control command.
Additional Recommendations
When defining contact regions and pairs on geometry which is not tangent continuous, creating single
contact regions which cross corner transitions can result in non-uniform stress results around the
corners. It is recommended to break these areas into multiple regions and pairs as shown below.
When defining contact regions and pairs, it is recommended to not include the same element face in
multiple regions. In “A” below, an element is repeated in regions 2 and 3. In “B”, the same element
only exists in region 2. “B” is recommended. Repeating element faces multiple times in the same or
different regions can significantly increase memory requirements and degrade performance.
• For surface-to-surface contact definitions, all of the parameters listed below are supported.
• For edge-to-edge contact definitions, all of the parameters listed below are supported except
for REFINE and INTORD.
You can optionally define multiple BCTSET/BCTPARM bulk entry sets, each set with unique
contact set IDs (CSID), and then combine them with a single BCTADD bulk entry. The multiple
BCTSET/BCTPARM bulk entry sets are created to adjust certain contact parameters locally. Contact
parameters can also be adjusted globally with a BCTPARM bulk entry having the same CSID as the
BCSET case control command. Also see the section Contact Pairs which includes an input example.
Global and local contact parameters have the following definition and rules:
Any of the parameters on the BCTPARM bulk entry can be defined globally. A parameter’s default
value is used if it is not defined globally or locally.
Following is a description of the BCTPARM input parameters. Local contact parameters have “*”
next to their names.
Name Description
CTOL Contact augmentation traction convergence. The augmentation loop convergence
criteria can be based on traction convergence. The contact force ratio FRAT is
determined as:
FRAT = (λ k –λ k-1) * (λ k –λ k-1) / (λ k * λ k)
where k is the augmentation loop id. If FRAT < CTOL, the contact augmentation
loop is considered converged. (Default = 0.05)
PTOL* Contact penetration tolerance. If the contact penetrations exceed the penetration
tolerance, an extra augmentation loop is performed. If the penetrations are below
this tolerance, the augmentation loop is considered converged. In addition, if the
global solution convergence criteria is satisfied, then the time step is considered
converged. PTOL only applies when CNTCONV=0, 1, or 3. If CNTCONV=2,
PTOL is ignored and the contact convergence criteria is only based on CTOL.
(Default = 1.0E-2 *characteristic length)
CNTCONV Contact convergence criteria.
0 – The contact convergence criteria is based on the first of PTOL or CTOL.
(Default)
1 – The contact convergence criteria is based on PTOL.
2 – The convergence criteria is based on CTOL.
3 – The convergence criteria is based on both CTOL and PTOL.
MAXS Maximum number of augmentation (outer) loops. If the augmentation loop has not
converged in MAXS number of iterations, the solution will proceed to the next step
if the global convergence criteria has been met. (Default = 20)
Name Description
INIPENE* Use when the goal is for a pair of contact regions to be touching without
interference, but due to the faceted nature of finite elements around curved
geometry, some of the element edges or faces may have a slight gap or
penetration.
0 or 1 - Contact is evaluated exactly as the geometry is modeled. No corrections
will occur for gaps or penetrations (Default).
2 - Penetrations will be reset to a new initial condition in which there is no
interference.
3 - Gaps and penetrations are both reset to a new initial condition in which there
is no interference.
INIPENE is applied when contact elements are initially created, and if they are
recreated as a result of large displacement effects when PARAM,LGDISP,1 is
defined.
OPNSTF* Open contact stiffness scale factor. The open contact stiffness is computed by
OPNSTF * closed stiffness. (OPNSTF default = 1.0E-6)
OPNTOL* Open gap tolerance scale factor. The open contact stiffness (OPNSTF * closed
stiffness) is applied to the contact elements that have a gap value less than or
equal to OPNTOL * characteristic length, but greater or equal than GAPTOL *
characteristic length. The contact element stiffness is 0.0 if the gap is greater than
OPNTOL * characteristic length. (OPNTOL default = 1.0)
GAPTOL* Closed gap tolerance scale factor. The closed contact stiffness is applied to the
contact elements that have a gap less than GAPTOL * characteristic length.
(Default = 1.0E-6)
NOSEP* No separation contact option.
NOSEP=0 (default): When contact stiffness is recomputed in a consecutive
nonlinear iteration, contact elements which are inactive as a result of normal
tractions=0.0 and no penetration, and which have a gap greater than GAPTOL *
characteristic length will remain inactive in the consecutive iteration.
NOSEP=1: The open contact stiffness (OPNSTF * closed stiffness) is applied to
the inactive contact elements that have a gap value less than or equal to OPNTOL
* characteristic length, but greater or equal than GAPTOL * characteristic length..
The contact elements with a gap greater than OPNTOL * characteristic length
remain inactive. While sliding is permitted with this option, the magnitude of the
sliding can be controlled by the tangential penalty factor. To define frictionless
sliding, set the coefficient of friction=0.0 or tangential penalty factor (PENT)=0.0.
(Default=0)
Name Description
GUPDATE Geometry update flag
0 – Contact geometry updates will not be done during the analysis.
1 – Geometry update will be done for large deflection analysis whenever the
relative tangential sliding between source and target regions in a pair exceeds the
tolerance set by the GUPTOL parameter (Default for geometric nonlinear analysis).
GUPTOL* Geometry update tolerance. If the relative sliding distance between the source and
target regions exceeds this tolerance, a geometry update will be initiated with large
displacement. (Default = 0.1 * characteristic length)
DISCAL Displacement scaling option
0 – No scaling will be done.
1 – Scaling will be done if required during every iteration. A check will be performed
after every displacement increment to see if the incremental displacements would
cause penetration between the source and target regions. If the penetrations
exceed DISTOL, the entire incremental displacements will be scaled back to limit
the penetrations in the model. (Default)
DISTOL Tolerance for displacement scaling feature. (Default = 0.5* characteristic length)
KSTAB Stiffness stabilization for contact.
0 – Stiffness stabilization is off. (Default)
1 – The stiffness matrix is stabilized when it is singular due to inactive contact
constraints. The stabilization adds a factor (1.0) to the diagonal terms of the
stiffness matrix. KSTAB=1 is only supported with the sparse solver, and will
disable any open contact stiffness specified through the OPNSTF parameter.
Name Description
PENN* Penalty factor for normal direction. PENN and PENT are automatically calculated
by default. When PENT is defined but PENN is undefined, PENN = 10 * PENT.
PENT* Penalty factor for transverse direction. PENN and PENT are automatically
calculated by default. When PENN is defined but PENT is undefined, PENT
= PENN / 10.
Name Description
Equation 5-1.
where
Kc= Contact stiffness assembled for all active contact elements.
Fc= Contact forces resulting from normal and tangential components of the tractions.
Equation 5-2.
where
Ns = number of grid points on an element source face
Nt = number of grid points on an element target face
= normal and tangential base vectors evaluated at the source locations Sj.
The normal traction λni is evaluated for iteration i as
where
εn = normal penalty stiffness
where
If slipping:
The equivalent grid point forces resulting from an active contact element are then computed from
equation 6-2.
1. The stiffness matrix contribution of an active sticking contact element is:
For slipping contact elements, the second stiffness term is unsymmetric. If the normal traction is
assumed to be fixed during the Newton-Raphson iterations, the second term vanishes, and a
symmetric stiffness matrix is retained.
b. Compute k and kc
2. Do Newton-Raphson iterations (i=1 to MAXITER). MAXITER is defined on the NLCNTL bulk entry.
a. Compute internal forces (Fint)
Assuming that the normal traction is fixed during Newton-Raphson iterations, λnk is used as
opposed to Tnk resulting in symmetrization.
If (sticking)
else; (slipping)
Compute contact forces Fc using Tnk and Ttk.
d. Solve
f. Check for the solution convergence on the residual. The solution convergence criteria is
defined by the CONV parameter on the NLCNTL bulk entry.
(i) If the solution convergence criteria is satisfied; augment tractions by step 3.
(ii) If the solution convergence criteria is not satisfied; do more iterations, go to 2a, update
stiffness if NEWK=TRUE, and go to step 1b.
3.
Check for maximum penetration convergence and/or force convergence. The contact
convergence tolerances are defined by the PTOL and CTOL parameters on the BCTPARM bulk
entry. The penetration convergence is satisfied if the maximum penetration (Pmax) is smaller than
the penetration tolerance (PTOL). The force convergence is satisfied if the contact force ratio
(FRAT) is smaller than the contact force tolerance (CTOL), where
FRAT = (λ k –λ k-1) * (λ k –λ k-1) / (λ k * λ k).
The CNTCONV parameter on the BCTPARM bulk entry detemines which tolerance (PTOL,
CTOL) is used to satisfied contact convergence.
4. End.
• You use the BOLTFOR bulk entry to define the bolt preload force.
• You use the BOLTLD bulk entry to optionally combine and scale bolt preload forces.
• You use the BOLTLD case control command in your bolt preload subcase to invoke the capability.
The BOLTLD command references a BOLTLD bulk entry or BOLTFOR bulk entries. Bolt preload
can only be invoked in a single subcase.
Bolts in SOL 401 are defined with the ETYPE = 3 format on the BOLT bulk entry. This format requires
that you list all of the elements that are used to model the bolt shaft on the BOLT entry. The grid point
you enter in the GP field on the BOLT entry indicates to the software where to calculate the cross
sectional area of the bolt. For this calculation, NX Nastran uses the direction that you define with the
CSID and IDIR fields on the BOLT entry as the bolt axis. To avoid any cross sectional effects at the
bolt ends, it is best to select a grid point in the GP field closer to the middle of bolt length.
NX Nastran uses the cross sectional area and the bolt preload force to estimate the initial bolt stress
and strain. To account for compliance in the structure that is being bolted together, the software
iterates on the bolt strain until convergence is satified.
You can adjust the bolt preload convergence tolerance with the EPSBOLT parameter (default=1.0E-3),
which is defined on the NLCNTL bulk entry. For each bolt preload iteration, the software computes
the difference between the current bolt preload and the user-defined preload. If the difference is less
than value of the EPSBOLT parameter, the bolt preload calculation is considered converged. If
the difference is greater than EPSBOLT, the preload strain is recomputed for the next bolt preload
iteration. The iterations continue until either convergence is satisfied, or the number of iterations
reaches the value of the ITRBOLT parameter (default=20). The ITRBOLT parameter is also defined
on the NLCNTL bulk entry.
You must define TEND1 = 0.0 on the TSTEP1 bulk entry selected for the bolt preload subcase. You
can optionally specify the number of increments on the TSTEP1 entry even though the duration of
the time steps are zero. If you do not specify the number of increments, the software applies the
preload force in a single increment.
The bolt preload subcase includes the bolt preload force and can optionally include a temperature
load. No other loads are supported in the bolt preload subcase. The consecutive sequentially
dependent subcases can then include service loads. The resulting bolt strain is included in the
consecutive sequentially dependent subcases until a non-sequentially dependent static subcase
occurs.
If your input file includes globally-defined loads, you must include DLOAD = 0 in the bolt preload
subcase. Doing so instructs the software to ignore the global loads during the bolt preload solve.
Otherwise, a fatal error is issued and the run terminates.
Often, the bolt preload subcase is the first subcase, although this is not required. In addition, a bolt
preload subcase which is not the first can be defined as sequentially dependent. For example,
you can define an initial static subcase in order to resolve an interference condition using contact.
Then a sequentially dependent bolt preload subcase can follow. The final stiffness from the initial
subcase, which includes the contact stiffness, along with the contact forces, all are used in the bolt
preload subcase. Note that loads defined in the initial subcase cannot be included in the bolt preload
subcase. Also note that the TSTEP1 in the initial subcase must also have TEND1 = 0.0.
If the bolt preload subcase is the first, you do not need to specify SEQDEP = NO since the first
subcase is always non-sequentially dependent.
The convergence information related to bolt preload is listed in the f06 file. For example,
...
PRELOAD BOLT CALCULATIONS FOR BOLT ID 1
-------------------------------------
ITERATION = 3
USER PRELOAD = 200000000.000000
INTERNAL FORCE = 199999616.790662
PERCENTAGE DIFFERENCE = 1.916046692281961E-004
BOLT PRELOAD CONVERGENCE ATTAINED
---------------------------------
...
The software issues a fatal error message if the bolt preload iterative solution fails to converge.
See the BOLT bulk entry.
Equation 7-1.
requires where σij are stress tensor components, bi are body forces, and xj are space coordinates.
• The constitutive relations represented by stress-strain relations, e.g., for linear elasticity,
Equation 7-2.
where εkl are strain tensor components and Dijkl are elastic constants.
Equation 7-3.
where ui are displacements.
These systems of governing differential equations must be satisfied for every infinitesimal element
throughout the domain of the continuum. The complete set of state variables, namely displacements,
may be determined by solving these systems of equations supplemented by boundary conditions,
and in dynamic situations by initial conditions as well. For the nonlinear problems, the governing
equations should be satisfied throughout the history of load application. The material nonlinearity
is manifested in the constitutive relations. The geometric nonlinearity is pronounced in the
strain-displacement relations, but it also affects the equilibrium equation by changing applied loads.
Changes in constraints affect the boundary conditions, which constitute contact problems.
Most of the known solutions for the solid mechanics problems are based on ideal geometry and linear
approximations. However, the real nature is more complicated and inherently nonlinear. The linear
system is a very particular case of a general problem. Even the nonlinear solutions that we seek deal
with only a small subset of special cases in a general category of nonlinear problems. When the
nonlinear system is confronted, no general mathematical solutions exist and superposition no longer
applies. The system may even be non-conservative.
The first phase of the structural analysis is the idealization of a physical system into a simpler and
more manageable engineering problem. The idealization process involves simplifications of the
geometry, boundary and joint conditions, and loading conditions, etc. using engineering intuitions,
experimental data, empirical observations, and classical solutions. If the idealized structural system
renders a problem that cannot be resorted to a classical method of analysis, further idealization is
required, namely discretization, for numerical analysis.
Finite elements represent spatial discretization of a continuum. As such, however, they do
not immediately impose nonlinearity. When nonlinearity has to be taken into account for large
displacements and/or stresses, a numerical model poses new dimensions to the discretization in
addition to the n-dimensional Euclidean space. That is, the discretization is applied to time, load,
and material properties by using piece wise linear curves. While discretization allows approximate
solutions by numerical methods, it introduces numerous mathematical singularities which may
complicate computational processes. Fortunately, the efficiency of modern digital computers makes
it feasible to apply complicated computational procedures to the complex systems of engineering
problems.
For the discrete system, governing differential equations are converted to algebraic equations. The
finite element model represents a structure by an assemblage of finite elements interconnected at
nodal points. State variables are the displacements (displacement method or stiffness approach.) of
the nodal points which carry fictitious forces representing distributed stresses actually acting on the
element boundaries. The equilibrium requirements are satisfied at nodal points by the nodal force
balance. The material constitutive laws are satisfied at the integration points of the element. The
compatibility is ensured by the displacement continuity between elements. It is noted, however, that
the compatibility of the nonconforming elements is ensured by a patch test.
Equation 7-4.
where U is the strain energy of the system and W is the potential energy of the external loads. The
equilibrium equations can be obtained by invoking the principle of virtual work or the Ritz method, i.e.,
Equation 7-5.
which implies that the total potential of the system must be stationary with respect to the state
variables (displacement) for equilibrium to be ensured. The functional IT is so called because it
involves the integral of implicit functions of the state variables, {u}.
Considering a three-dimensional continuum for a nonlinear problem, the stationarity condition results
in
Equation 7-6.
where the dots and δ denote infinitesimal increments and arbitrary variations, respectively. The
left-hand side represents variations in the strain energy increment and the right-hand side represents
variations in the external work which consists of body forces bi (such as a gravity load), traction forces
ti at the boundary surface (such as pressure loads), and concentrated forces pi. Now it remains
to determine admissible functions expressing the arguments of the functional II in terms of state
variables {u}, which are valid throughout the whole region and satisfy the boundary conditions.
The finite element method can be characterized by the following features distinguished from the
conventional Ritz methods or the matrix method for frame structures:
• The whole region of the system is divided into numerous subdomains, called finite elements,
which have simple geometrical shapes.
• The variational process is limited to each finite element, which aggregates into a whole region
when assembled.
• The admissible displacement field within each element, , can be expressed in terms of nodal
displacements using interpolation functions known as shape functions, N, i.e.,
Equation 7-7.
where {u} is a displacement vector consisting of all nodal points of the element.
The strain-displacement relations for the element can then be established in terms of nodal
displacements using the shape functions in Equation 7-7, i.e.,
Equation 7-8.
where
Equation 7-9.
and the element matrix [B] consists of derivatives of the shape functions, evaluated at the current
deformed geometry. Notice that the geometric linear problem requires that the element matrix be
evaluated only at the initial geometry. The software employs an approximate updated Lagrangian
approach for geometric nonlinear problems, by which linear strains are computed in the updated
element coordinate system in order to eliminate the effects of the rigid body rotation but the
equilibrium is established at the final position in the stationary coordinate system. This method does
not require reevaluation of the element matrix [B] (constant in the absence of large strains) while the
element coordinates are reevaluated continuously.
Equilibrium equations for an element may be obtained by reducing Equation 7-6 after the substitution
of Equations 7-7 and 7-8, based on the small deformation theory. Then the element boundary
stresses are statically equivalent to the nodal forces which balance the applied external loads, i.e.,
Equation 7-10.
with
Equation 7-11.
and
Equation 7-12.
where [Ns] is an appropriate interpolation function for the traction force. Notice that the equilibrium
equation for an incremental load may be expressed as
Equation 7-13.
where should be components of co-rotational stress which is independent of a rigid body rotation.
The element stiffness matrix can be obtained by substituting the constitutive relations into Equation
7-13, i.e.,
Equation 7-14.
where
Equation 7-15.
and [D] is the material tangent matrix. The nodal forces of an element can then be expressed as
Equation 7-16.
where the element stiffness is
Equation 7-17.
Notice that this expression represents an element stiffness due to the material stiffness without
geometric nonlinear effects. As will be shown later, an additional stiffness [J(d)] due to initial stresses
should be included for an incremental process because the initial stresses exist from the second
increment.
The equilibrium must be satisfied in the whole region throughout the complete history of load
application. Equilibrium equations for the global discrete system are obtained when all the elements
are assembled, i.e.,
Equation 7-18.
where Σ over m denotes a summation over all elements. For the incremental process, the equilibrium
equation may be rewritten as
Equation 7-19.
with
Equation 7-20.
where {σ0} represents an initial stress or the stress state at the preceding load step.
Because of the approximations involved in the interpolation functions, the finite element model
provides an approximate solution even if the equilibrium Equation 7-18 is satisfied exactly.
Consequently, the differential equations of equilibrium are not satisfied exactly even for linear
problems, but the error decreases as the finite element mesh is refined. This convergence condition
is required and ensured by element formulations with regard to the element convergence criteria.
• Local coordinate system: defined by the user in the Bulk Data, which may include special
coordinates such as cylindrical and spherical coordinate systems.
• Global coordinate system: a collective coordinate system which comprises all the local coordinate
systems specified for output quantities.
• Displaced element coordinate system: similar to element coordinate system but defined in the
displaced position.
• Material coordinate system: a Cartesian coordinate system used to orient anisotropic material
properties.
• Modal coordinate system: a generalized coordinate system defined for each eigenmode.
It is noted that the global system is a Cartesian coordinate system, although non-Cartesian coordinate
systems are adopted to orient the local Cartesian coordinates for output quantities. In the software,
all the displacements and forces, hence the system matrices, such as the stiffness matrix, are
expressed in the global coordinates. This implies that all the major computations involved in the
analysis are processed in Cartesian coordinates. Element and material coordinate systems are
defined in the element connectivity description. Now we only have to consider linear transformations
between Cartesian coordinate systems.
Let us consider a coordinate transformation between the primed and unprimed systems which are
right-handed Cartesian coordinates. The transformation matrix T consists of direction cosines of unit
vectors of the unprimed coordinate system, i.e.,
Equation 7-21.
where
Equation 7-22.
Notice that T is an orthogonal matrix and thus
Equation 7-23.
Because the work and energy are invariants with respect to coordinate transformation, i.e.,
Equation 7-24.
it follows that
Equation 7-25.
Then the equilibrium equation,
Equation 7-26.
may be expressed in the unprimed coordinate system by
Equation 7-27.
with
Equation 7-28.
It is noted that the modal matrix Φ is used as a transformation matrix for a modal transformation
which is not elaborated here.
Equation 7-29.
The forces and displacements are transformed from element to global coordinates and vice versa, i.e.
where Tbe transforms from element to basic coordinates and Tbg transforms from global to basic
coordinates.
It is noted that Tbe is identical for all the nodes of an element but Tbg may vary from node to node in
the same element. The element stiffness matrix is transformed into global coordinates by
Equation 7-30.
for which the building blocks of Tbe and Tbg are (3x3) matrices formed for each nodal point and have
to be assembled for an entire element, e.g. for a three-noded triangular shell element
Equation 7-31.
where the superscript is used to associate each (3x3) matrix with the nodal point and is repeated
for the rotational degrees of freedom.
Equation 7-32.
where
The module MCE1 partitions [Rmg] and solves for a transformation matrix [Gmn], i.e.,
Equation 7-33.
where
Then the module MCE2 partitions the global stiffness matrix, [Kgg], and reduce it to the n-set, i.e.,
Equation 7-34.
from which the system is reduced to
Equation 7-35.
where
and
The primes are used in K’nn, P’n, and Q’n to distinguish from Knn, Pn, and Qn, which are resulting
matrices after the reduction.
Equations in the n-set can be further reduced by eliminating single-point constraints, i.e.,
Equation 7-36.
which is reduced to
Equation 7-37.
where
with
Notice that the effects of constraint forces (Qs and Qm) are not visible in Equation 7-37. The
single-point constraint forces are recovered by
Equation 7-38.
Further reduction of equations in the f-set is performed by an elimination of the o-set, known as static
condensation. The f-set is partitioned by the UPARTN module as follows:
Equation 7-39.
from which
Equation 7-40.
where
and
Equation 7-41.
where
and
There are some rules to remember regarding the displacement sets in SOL 401. They are:
• The r-set is not supported. Do not use the SUPORT bulk entry.
• Rigid elements are formulated with linear multipoint constraint equations and do not have large
displacement capability. Consequently, erroneous results will be obtained if the rigid element
undergoes a large rotation. To avoid this, stiff elements should be used in place of rigid elements
for large displacement analysis.
• The TSTEP1 bulk entry for the load increment (time based). It is selected by the TSTEPNL
case control command.
• The EIGRL bulk entry for the modal analysis. It is selected by the METHOD case control
command.
The increment size can vary from subcase to subcase by specifying different TSTEPNL. It is
recommended to define separate TSTEPNL for every subcase even if the same values are specified,
so that changes can be accommodated in the subcase level as needed.
With values of 1 or 2 for LGDISP, all the potentially nonlinear elements become actively nonlinear
elements unlike the material nonlinear model.
This solver has a distinct approach to the large rotation, for which the element coordinates are
continuously updated to the current configuration during the iteration. The equilibrium is sought in the
deformed position. Consider the internal force computation as follows:
Equation 8-1.
Equation 8-2.
in which could be divided into two parts (linear and nonlinear), i.e.,
Equation 8-3.
Upon differentiation of Equation 8-1, we have
Equation 8-4.
where {σ} represents stresses with reference to the original coordinates. Substituting Equation
8-3 and
Equation 8-5.
Equation 8-4 becomes
Equation 8-6.
with
Equation 8-7.
and
Equation 8-8.
in which KL represents the usual linear stiffness matrix, KR a stiffness due to large rotation, and Kg a
geometric stiffness dependent on the initial stress level.
Now it remains to define the nonlinear part of the element matrix (BN). The definition of finite strains
based on the Lagrangian formulation (referred to the initial configuration) is as follows:
Equation 8-9.
Equation 8-10.
with other components obtained similarly. In matrix notation
where {εL} is the usual infinitesimal strain vector and {εN} is the nonlinear strain vector consisting of
the second order terms, i.e.,
Equation 8-11.
where
Equation 8-12.
and
Equation 8-13.
Introducing shape functions (Ni) and nodal displacements {u} (using an example of a 10-noded
tetrahedron), displacement derivatives are expressed by
Equation 8-14.
and
Equation 8-15.
where
Equation 8-16.
and
Equation 8-17.
From the properties of matrices A and θ, it can be shown that
from which
Equation 8-18.
The initial stress stiffness [Kσ] can be derived as follows:
in which
where
Equation 8-19.
with I being the (3x3) identity matrix. Finally the geometric stiffness is
Equation 8-20.
It has been found that stiffness matrices caused by geometric nonlinearity (KR and Kσ) can be
computed from the matrices [A], [G], and [M] with the following observations:
• [G] is dependent upon the initial geometry, hence stays constant unless the geometry is updated.
This matrix is used in forming [KR] and [Kσ].
• [A] is used in forming [KR]. [A] is dependent on the rotations and should be updated continuously.
• [M] is used in forming [Kσ]. [M] is dependent on the stresses and should be updated continuously.
• The matrix [KR] takes into account the effects of large rotations. The large displacement effects,
due to rigid body translation and rotation, are treated effectively in the absence of large strains by
updating element coordinates in the software.
• Geometric stiffness matrix [Kσ] takes into account the effects of the initial stresses. This effect
becomes important with geometric stiffening, and is used for instability analysis. The geometric
stiffness matrix [Kσ] is equivalent to the differential stiffness [Kd] in the software.
Equation 8-21.
where the superscript e denotes an elemental operation and the subscript d denotes the vectors in
the displaced element coordinate system. Then the element forces should be transformed into the
common coordinate system (namely global coordinate system denoted by a subscript g) before
assembly for global operations, i.e.,
Equation 8-22.
where the summation sign implies an assembly operation, and Tbd and Tbg are transformation
matrices from displaced to basic and from global to basic coordinate systems, respectively.
Equation 8-23.
where the position vector (Xeb in Figure 8-2) of the element coordinate system with respect to
the basic coordinate system is denoted by < xe, ye, ze > Tbasic and transformation matrix [Tbe] is
composed of direction cosines of unit vectors of the element coordinate system with respect to the
basic coordinate system, i.e.,
Equation 8-24.
As the element deforms or displaces, the element coordinate system moves and this is defined as a
displaced coordinate system. The displaced coordinate system is established in the same manner as
the element coordinate system. Again the transformation should be performed similarly, i.e.,
Equation 8-25.
where < xd, yd, zd > Tbasic is the position vector of the displaced element coordinate system with
respect to the basic coordinate system (Xdb in Figure 8-2) and [Tbd] is formed simliarly to [Tbe].
In order to isolate the deformation from the rigid body displacements, nodal displacements are
computed in the displaced element coordinate system by overlaying the original element as shown in
Figure 8-3.
The net displacements can be computed by subtracting the original nodal coordinates in the element
coordinate system from the displaced nodal coordinates in the displaced element coordinate system,
i.e.,
Equation 8-26.
in which the nodal coordinates in the element and displaced element coordinate systems can be
computed by the following transformations:
Equation 8-27.
and
Equation 8-28.
Substitution of Equations 8-27 and 8-28 into Equation 8-26 results in
Equation 8-29.
where {ug} is a total displacement (translational components only) in the global coordinates. In the
absence of the large displacement effect, the net displacement ud in Equation 8-29 is reduced to:
Equation 8-30.
if the material is linear or
Equation 8-31.
if nonlinear material is involved. Consequently, the tangent stiffness matrix is formed in the global
system by assembling the element stiffness matrices transformed into the global coordinate system
from the displaced coordinate system, i.e.,
Equation 8-32.
The update process is performed at every iteration and the updated nodal displacements ud are used
whenever strains and stresses are computed. Effectively, the second order effect due to large rigid
body motion is eliminated. However, the displacement output shows the total displacements in
the global coordinates, i.e.,
Equation 8-33.
where the subscript i. denotes operations on each nodal point. The transformation matrix [Tbd] is
computed for each element after each iteration and stored in the ESTNL data block for stiffness matrix
update when requiredby the stiffness matrix update strategy. On the other hand, the transformation
[Tbg] is computed for each nodal point and it is not stored but recomputed whenever it is needed. The
nodal coordinates in the undeformed geometry, Xb, are available from the data block BGPDT.
This approach can be interpreted as approximate updated Lagrangian method, since the motion of
the body follows Lagrangian description. Stresses are computed in the deformed geometry just like
Cauchy stress. However, this method of displaced coordinate system is a unique and salient feature
in the software. The referential geometry in the updated Lagrangian method is brought up-to-date at
every incremental step upon convergence but fixed during the iterative process, which is inherently
different from the current method of updating the coordinate system.
Equation 8-34.
where p is the magnitude of the pressure on the surface A, interpolated by a shape function N, and
ñdA changes as a function of u. They generally occur with fluid pressures such as the pressurized
balloon, inflated tire, or the lift load on the airplane wing. Other physical applications involve
kinematics such as the classical fire hose instability problem or inertia loads on spinning bodies. In
the software, the term applies to specific load inputs as defined below.
The first category includes simple forces, and enforced displacements. The second, follower force
category, includes the following Bulk Data inputs:
FORCE1, FORCE2 The direction changes with displacements of the referenced GRID points. The
magnitudes of these concentrated loads are constant.
PLOAD, PLOAD4 The pressure loads follow the surface of the solid elements (HEXA, PENTA,
PYRAM and TETRA).
RFORCE Centrifugal loads change in magnitude and direction with motion of the masses
attached to the GRID points. The effect may be destabilizing if large motions
occur. It is recommended that lumped masses be used with these loads.
Also note that upstream superelements are assumed to be linear and therefore the upstream loads
will remain fixed in magnitude and direction. In addition, forces on omitted degrees of freedom (when
ASET or OMIT data are present) should not be follower forces.
8.3.2 Implementation
The follower forces depend on the GRID displacements and therefore must be recalculated for each
nonlinear iteration and line search. The basic equation for residual error, defined by Equation 9-2,
becomes:
Equation 8-35.
where the applied load vector {Pal is now a variable. Corrective Loads are computed based on the
updated geometry and added to the initially applied loads to account for the follower forces, i.e.,
Equation 8-36.
where
Equation 8-37.
Note that thermal effects are included in the vector {F}.
In turn, the tangent matrix could be calculated from Equation 9-3 using derivatives of the loads, which
is termed follower matrix. However, the nonlinear solution process ignores the stiffness effects of
the changing loads and use the approximation:
Equation 8-38.
The effect of the approximation is minor in most cases. However, it could become a major concern in
thin shell models with pressure loads causing large rotations, where the converged solutions will be
correct but the rate of convergence may be slow or cause divergence. Also the buckling solutions or
modal analysis on preloaded structure with pressure load may not be correct due to the approximate
tangent matrix if the effect of the follower matrix is significant.
The follower force effects in the analysis can be controlled by the parameter LGDISP. Three options
are available in PARAM LGDISP:
• = 0 for no geometric nonlinearity
3. tn=tn-1+∆t
4. Obtain the external force vector Pn1 for the first iteration of current time step.
a. If i>1: Include follower force effects to obtain Pn1.
7. If i>1:
a. Compute: E2=∆Uni-1. R
b. Compute Er=E2/E1
9. Compute E1=∆Uni.R
{Pg} + {Qg} - {Fg} = {0} where {Pg}, {Qg}, and {Fg} represent vectors of applied loads, constraint forces,
and element nodal forces, respectively. Element nodal forces are nonlinear functions of displacements
for nonlinear elements. Since the equilibrium condition is not immediately attained in the presence of
nonlinear elements, an iterative scheme such as the Newton-Raphson method is required. Since the
error vanishes at constrained points and the constraint forces vanish at free points, the unbalanced
forces acting at nodal points at any iteration step are conveniently defined as an error vector by
Equation 9-1.
and
{Δui} = {ui} – {ui-1},
{Ri} = {P} – {F(ui)}.
The iteration continues until the residual error {R} and the incremental displacements {Δu} become
negligible, which is signified by the convergence criteria.
The tangential stiffness consists of the geometric stiffness in addition to the material stiffness, i.e.,
without regard to the coordinate transformation,
Equation 9-4.
where [Km] and [Kd] refer to the material and the differential stiffness, respectively. The material
stiffness is given in Equation 4-17 with a material tangential matrix for [D]. The differential stiffness,
which is caused by the initial stress, is defined as follows:
Equation 9-5.
where [BN] represents the second order effects in the strain-displacement relations, [G] consists of
derivatives of shape functions and [M] is a function of stresses. Notice that the initial displacement
stiffness is not included in [KT] because its effects are already eliminated in the element formulation.
Newton's procedure was previously implemented using a corrective force. Recalling that the element
forces for linear elements are expressed as
{F} = [K] {u},
a corrective force vector may be defined as
2. If KUPDATE=0 (Auto method, default), stiffness is updated based on the change in the value
of current stiffness parameter since the last stiffness update.
3. If KUPDATE=-1 (Initial stiffness method), no stiffness update is made during the solution.
Current stiffness parameter approach for the automatic stiffness update method (KUPDATE=0)
The current stiffness parameter approach proposed by Bergan and Crisfield [*] is used for automatic
stiffness update method. The current stiffness parameter gives a scalar measure for the stiffness of
the structure at the current loading condition.
After the first iteration of the first step, S11=1, that is, the initial value of the current stiffness parameter
is one. After each iteration, the value of current stiffness parameter is recomputed. The change in
current stiffness parameter is computed for each iteration as:
Where, Sref is the reference value for current stiffness parameter. At the start of the solution, Sref is
set to 1.0. Stiffness is updated if ∆S≥α, where α=5.0 for a problem with structural loading only, α=20.0
for a problem with pure thermal load, and α=10.0 for a problem with a combination of structural and
thermal load. The value for α can be defined by the user on the NLCNTL card through parameter
CSTFPAR. Valid input for CSTFPAR is a real number.
Sref is updated upon updating stiffness to correspond to the value of Sni.
1. Bergan. P, Horrigmoe. G, Krakeland . B, and Soreide T., SOLUTION TECHNIQUES FOR
NON-LINEAR FINITE ELEMENT PROBLEMS, International Journal for Numerical Methods
in Engineering, Vol. 12, 1677-1696 (1978)
Equation 9-10.
Equation 9-11.
Equation 9-12.
where,
Rni-1=Pni-1-Fni-1 (Uni-2) is the residual for iteration i-1 for time step n,
ΔUni=K-1 Rni-1 is the displacement increment computed in iteration i-1 for time step n,
Rni=Pni-Fni (Uni-1) is the residual for iteration i for time step n,
εpi is the computed error for the force norm for iteration i for time step n,
εpi-1 is the computed error for the force norm for iteration i-1 for time step n,
εwi is the computed error for the energy norm for iteration i for time step n,
εwi-1 is the computed error for the energy norm for iteration i-1 for time step n.
For the first iteration of every new time step, the value for λpi and λwi is set to 0.9999. The variable
NDIV is initiated to 0 in the first iteration of each time step. NDIV is incremented if:
1. If Er>1 or Er<-100, NDIV=NDIV+1
NDIV is reset to 0 if neither of the two conditions above are met in any given iteration. The solution
is considered to have diverged if NDIV>MAXDIV. MAXDIV is a NLCNTL parameter which controls
maximum allowable divergences in a time step (Default=3).
Upon divergence, the software will:
1. Revert the solution to the last converged point.
2. Reform stiffness at the last converged configuration (unless KUPDATE=-1, in that case no
stiffness update is performed).
3. Perform bisection, and attempt to solve the time step with a reduced size from the last converged
point.
For a timestep of size Δt, on bisection the time step is reduced to:
Δt1=1/2 Δt
Bisection continues until the solution converges, that is,
Δtk = 1/2 Δtk-1 = 1/2k Δt
where k is a bisection count. Once the bisection is successful (rendering a converged solution),
the integration proceeds to the next time step. If k = 1, the same time step size is used for the
next time step.
If k > 1, an effort is made to accelerate the solution process by increasing the time step size for
the next step. The time step size for the next time step is influenced by all of the following factors:
a. The number of iterations (i) that were required to reach convergence with time step size of Δtk.
c. Remaining time left (δt) to complete the original time step of size Δt.
The time step size used for the next time step is m Δtk, where 1 ≤ m ≤ k – 1. The largest
value of m that satisfied the following conditions is used:
A. I ≤ 3 or I ≤ MAXITR/2m, where MAXITR is the maximum permissible iterations for a given
timestep.
B. δt/m Δtk is an integer greater than 0.
Although the goal of increasing time step size is to reduce the number of time steps till
completion of solution, for some problems, further bisection may be required in subsequent
steps.
The maximum number of bisections is limited by the parameter MAXBIS (default=5). The
bisection process is activated on an as-needed basis. You also have an option to suppress
bisection by specifying MAXBIS=0.
d. If the solution fails to converge even after attempting maximum permissible bisections, the
solution is terminated with a fatal message. Results corresponding to the last converged user
requested output time are printed.
Both aspects must be defined properly for the criteria to be effective, for the solution scheme to
be efficient, and for the solution to be accurate.
There are no universally accepted convergence criteria to date in the field of finite element analysis.
Conditions to be met by ideal convergence criteria for a general-purpose finite element analysis have
been contemplated. The convergence criteria should:
• be satisfied for linear cases at all times.
• be able to handle all the loading cases including constant loading, unloading, and no external
loading (applicable to creep analysis).
• have smooth transitions after the stiffness updates and loading changes.
• excessively loose tolerances cause not only inaccuracy but convergence difficulties in the
subsequent steps due to cumulative errors.
The fundamental difficulty of the convergence tests for a structural analysis lies in the fact that
the base vectors (forces and displacements) involve inconsistent units, namely, combinations of
forces and moments or translations and rotations. Indiscriminate use of these vectors will cause
unit-dependent convergence criteria. For example, while an error in forces is dominant when the
model is expressed in newton-meter, the error would be dominated by moments if the same model is
described in newton-millimeter.
The most natural and reasonable criterion for the convergence test is formulated in terms of an
energy error. The energy error is the logical choice because both the out-of-balance forces {R}
and the change in displacements {Δu} should be minimized by the iteration process. Furthermore,
energy quantities do not pose problems of inconsistent units due to mixed units associated with
translations and rotations.
Although the convergence test in terms of energy errors is usually adequate, some distinct errors
are not detected with this criterion; i.e., displacements are in gross error while the residual load
error is negligible, or vice versa. This would be the case if the degrees-of-freedom in error have a
very small or a very large stiffness. Such cases compel the need for criteria in terms of loads and
displacements. Nominally, by visualizing the load-deflection curve for a one-dimensional case, it can
be noticed that the convergence criterion in terms of loads governs the stiffening structure and the
criterion in terms of displacements governs the softening structure. Scalar error functions for these
criteria are formulated to be dimensionless by introducing the weighted normalization.
to displacement, force and energy (work) are computed in every iteration. Error in energy (work) is
used as the default convergence criterion. These criteria are computed as follows:
• EPSU (displacement criterion)
Equation 9-13.
where,
Uni is the displacement vector for iteration i of time step n,
Uni-1 is the displacement vector for iteration i-1 of time step n, and
Un-1 is the displacement vector at convergence for time step n-1.
The displacement error computation can be summarized as the ratio of L2 norm of the
incremental displacement in the current iteration and the L2 norm of the incremental displacement
in the current time step. The error in displacement in the first iteration for every time step should
be 1.000.
where,
Pni is the external force vector for iteration i of time step n,
Fni is the internal force vector for iteration i of time step n,
Rni is the residual force vector for iteration i of time step n, and
Rn1 is the residual force vector for iteration 1 of time step n.
Equation 9-14.
Equation 9-15.
Where,
ΔUniis the incremental displacement vector computed for iteration i of time step n,
Rni is the residual force vector for iteration i of time step n,
Uni is the total displacement vector for iteration i of time step n,
9.5.4 Implementation
The convergence tolerance determines the efficiency of the solution scheme as well as the accuracy
of the solution. The tolerance should be realistic, not too tight nor too loose. It is difficult to choose
optimal default values for the convergence tolerances. However, efforts have been made to set the
default values to provide reliable solutions to the general class of problems. Thus, default tolerances
should be adhered to until good reasons are found to change them.
The following three error functions (in terms of displacements, loads, and energy) are computed
and compared to tolerances.
Eu < EPSU (=10-2 by default)
Ep < EPSP (=10-2 by default)
Ew < EPSW (=10-6 by default)
where EPSU, EPSP, and EPSW are tolerances specified in the NLCNTL entry. However, only
those criteria chosen by the user (combinations of U,P, and/or W) are designed to be satisfied for
convergence.
It is noted that divergence conditions are established independent of convergence criteria.
Headquarters
Europe
Granite Park One
Stephenson House
5800 Granite Parkway
Sir William Siemens Square
Suite 600
Frimley, Camberley
Plano, TX 75024
Surrey, GU16 8QD
USA
+44 (0) 1276 413200
+1 972 987 3000
Asia-Pacific
Americas
Suites 4301-4302, 43/F
Granite Park One
AIA Kowloon Tower, Landmark East
5800 Granite Parkway
100 How Ming Street
Suite 600
Kwun Tong, Kowloon
Plano, TX 75024
Hong Kong
USA
+852 2230 3308
+1 314 264 8499