100% found this document useful (2 votes)

479 views

Mar 101

msc marc guideline

Uploaded by

junee

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

100% found this document useful (2 votes)

479 views

Mar 101

msc marc guideline

Uploaded by

junee

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 568

Basic Nonlinear Analysis Using

Marc and Mentat

MAR101 Course Notes

November 2015

MA*V2015*Z*Z*Z*SM-MAR101-NT

Legal Information
MSC.Software Corporation reserves the right to make changes in specifications and other information contained in this
document without prior notice. The concepts, methods, and examples presented in this text are for illustrative and
educational purposes only, and are not intended to be exhaustive or to apply to any particular engineering problem or
design. MSC.Software Corporation assumes no liability or responsibility to any person or company for direct or indirect
damages resulting from the use of any information contained herein.
Copyright 2015 MSC.Software Corporation. All Rights Reserved. This notice shall be marked on any reproduction of
this documentation, in whole or in part. Any reproduction or distribution of this document, in whole or in part, without the
prior written consent of MSC.Software Corporation is prohibited.
The MSC.Software corporate logo, Adams, Dytran, Easy5, Fatigue, Laminate Modeler, Marc, Mentat, MD Nastran, Patran,
MSC, MSC Nastran, Mvision, Patran, SimDesigner, SimEnterprise, SimManager, SimXpert and Sofy are trademarks or
registered trademarks of the MSC.Software Corporation in the United States and/or other countries. NASTRAN is a
registered trademark of NASA. All other trademarks belong to their respective owners.

S0 - 2

CONTENTS
Section
1

Page
Course Overview
Course Objectives

1-2

SimCompanion -Where to go for Help

1-3

SimCompanion Access to Communities

1-6

Marc/Mentat Overview
MSC Software Overview

2-2

Marc/Mentat How do they work together?

2-6

Marc Overview

2-7

Summary of Marc Analysis Types

2-8

Marc Advanced Solution Features

2-9

Mentat Overview

2-19

Mentat Advanced Features

2-20

Marc/Mentat Documentation

2-23

Nonlinearity Overview
Section Contents

3-2

Some Basics of Nonlinearity

3-3

Comparison of Linear and Nonlinear Analylsis

3-4

Marc Basic Functionalities

3-5

General Considerations in Nonlinear Analysis

3-6

Sources of Nonlinearity

3-8

Geometric Nonlinearity

3-11

Comparison of Nonlinear and Linear Solutions

3-12

Stress-stiffening

3-13

MAR101, Section 0, November 2015

S0 - 3

CONTENTS
Section
3

Page
Nonlinear Overview (Continued)
Geometric Nonlinearity (Cont.)
Types of Geometric Nonlinearity

3-15

Examples

3-16

Large Strains

3-20

Non-Conservative Loading (Follower Force)

3-21

Contact Nonlinearity

3-22

Contact: Basics

3-23

Contact Bodies in Marc

3-24

Examples

3-25

Material Nonlinearity

3-42

Materials in Marc

3-43

Elastic

3-44

Plastic

3-45

Hyperelastic

3-46

Workshop 1 Analysis of Rubber Seal

Introduction to Mentat
Mentat Getting started

4-4

Pre-processing in Mentat

4-5

Communicating with Mentat

4-6

Mentat Layout

4-11

Model Navigator

4-24

Importing CAD Geometry into Mentat

4-33

MAR101, Section 0, November 2015

S0 - 4

CONTENTS
Section
4

Page
Introduction to Mentat (Continued)
Pre-processing in Mentat Continued
Mentat Hints and Shortcuts

4-38

Selecting in Mentat

4-40

Visibility in Mentat

4-43

Procedure Files in Mentat

4-44

Workshop 2 CAD Import and Repair

Worskhop 3A Linear Analysis of Cantilever Beam
Post-processing in Mentat

4-46

Post Processing Basics

4-47

Creating Plots

4-49

Animations

4-51

Animated GIF Movies

4-52

Workshop 3B Post-procesing of Linear Analylsis of Cantilever Beam

Geometric Nonlinearity
Geometric Nonlinearity Example

5-2

Geometric Nonlinearity: Objectivity

5-3

Geometric Nonlinearity Stiffness Breakdown

5-7

Strain Measures

5-8

Stress Measures

5-10

Classes of Large Deformation Problems

5-12

Large Strain Options

5-14

Geometric Nonlinearity Guidelines

5-17

MAR101, Section 0, November 2015

S0 - 5

CONTENTS
Section
5

Page
Geometric Nonlinearity (Continued)
Geometric Nonlinearity: Follower Forces

5-20

Shear Locking

5-21

Assumed Strain Option

5-24

Workshop 3C Nonlinear Analysis of a Cantilever Beam

Material Properties
Summary of Marc Materials

6-3

Capabilities Supported

6-4

Elasticity

6-7

Pasticity

6-9

Measure of Stress and Strain

6-14

Plastic Stress and Strain Data

6-18

Hardening Laws

6-22

Isotropic Hardening

6-23

Kinematic Hardening

6-26

Combined Hardening

6-29

Orthotropic Material Properties

6-33

Temperature Dependent Physical Quantities

6-34

Hyperelasticity

6-36

Characterizing Hyperelastic Materials

6-39

Hyperelastic Formulations

6-44

Stress Invariants in Mooney-Rivlin Formula

6-45

Neo-hookean and Mooney-Rivlin models

6-46

MAR101, Section 0, November 2015

S0 - 6

CONTENTS
Section
6

Page
Material Properties (Continued)
Hyperelastic Formulations Continued
3-Term Mooney Rivlin Model

6-47

Yeoh Model

6-49

Ogden Model and Arruda-Boyce Model

6-50

Gent Strain Energy Model

6-51

Marlow Model

6-53

Models for Large Volumetric Deformations and Foam Model

6-54

Storakers Model and Blatz-Ko Model

6-55

Model Summary

6-56

Composites: Laminated Shell and Solid Shell

6-57

Laminated Composite Shells

6-58

Orientation of Layers

6-63

Comparison of Modulus Ratios for Composites

6-67

Workshop 4 Plastic Deformation of Cantilever Beam

Workshop 5 Experimental Hyperelastic Analysis of Rubber Seal

Contact Analysis
Contact Basics With Examples

7-2

General Considerations in Contact

7-3

Contace Methods

7-4

Node-to-Segment Contact

7-7

Segment-to-Segment contact

7-9

Introduction to Contact Bodies

MAR101, Section 0, November 2015
Copyright 2015 MSC Software Corporation

7-17
S0 - 7

CONTENTS
Section
7

Page
Contact Analysis (Continued)
Deformable Bodies

7-19

Rigid (Geometric) Bodies

7-21

Rigid (Geometric) Bodies Allowing Heat Transfer

7-23

Example: Stress and Thermal Coupling Analysis

7-24

Workshop 6 Hertz Contact

7-28

Contact Detection in a Static Analysis

7-29

Possible Contact Situations

7-30

Distance Tolerance

7-35

Bias Factor

7-37

Contact Detection and Speed

7-39

Contact Detection for Shells

7-40

Beam-to-Beam Contact

7-41

Beam-to-Shell Contact

7-49

Pipe-in-Pipe and Exterior Contact

7-50

Shell Edge-to-Edge Contact

7-51

Contact and Element Types

7-52

Definitions of Contact Bodies

7-56

Defining Deformable Bodies in Mentat

7-57

Improved Boundary Description

7-58

Defining Rigid (Geometric) Bodies in Mentat

7-64

Control of Rigid (Geometric) Bodies

7-65

Orientation of Rigid (Geometric) Contact Bodies

7-73

MAR101, Section 0, November 2015

S0 - 8

CONTENTS
Section
7

Page
Contact Analysis (Continued)
Contact Tables

7-75

Contact Interaction

7-76

View Modes

7-83

Parameters Available in all View Modes

7-85

Workshop 7 Intereference Fit

Workshop 8 Hertz Contact Analysis with Friction
Special Topic Symmetry Using Contact Bodies

7-87

Symmetry Planes

7-88

Cyclic Symmetry

7-90

Setting up and Running the Analysis, Multi-Stepping, and Restarts

Analysis Overview

8-2

Job Submit

8-3

Multi-Step Analysis

8-5

User Subroutines

8-11

Meshing Adaptivity Global Adaptive Re-meshing

8-13

Meshing Adaptivity Local Adaptive Re-meshing

8-16

Running Large Jobs

8-17

Job Monitoring

8-19

Debugging the Input

8-21

Status File

8-22

Files Generated by Marc

8-24

Restarting a Job

8-26

MAR101, Section 0, November 2015

S0 - 9

CONTENTS
Section
8

Page
Setting up and Running the Analysis, Multi-Stepping, and Restarts (Continued)
Restart Form in Mentat Restart Data Mode

8-27

Restart From in Mentat Write Restart Data

8-30

Restart Form in Mentat Completion of Unfinished Loadcase

8-31

Workshop 9 Buckling Analysis

Numerical Analysis of Nonlinear Problems

Part I: Numerical Analysis of Nonlinear Behavior

9-3

Nonlinear Analysis: The Approach

9-4

Iterative Solution Methods

9-8

Newton Raphson Method

9-11

Modified Newton Raphson Method

9-14

Iterative Methods in Mentat

9-18

Analysis Convergence

9-19

Convergence Criteria

9-22

Convergence Testing

9-26

Relative Convergence Criteria Issues

9-27

Automatic Criteria Switching

9-29

Part II: Load Controls and GUI Controls

9-32

Load Incrementation Control and Automatic Time Stepping

9-33

The Time Scale

9-36

Load Increment Cut Back (Step Reduction)

9-39

MAR101, Section 0, November 2015

S0 - 10

CONTENTS
Section
9

Page
Numerical Analysis of Nonlinear Problems (Continued)
Part II: Load Controls and GUI Controls continued
Graphical User Interface Control

9-40

Mentat Load Increment Parameters Number of Cutbacks

9-42

Mentat Load Increment Parameters Out of Core Element Storage

9-43

Mentat Load Increment Parameters Loadcase Properties Form

9-44

Mentat Load Increment Parameters Fixed

9-45

Mentat Load Increment Parameters Adaptive Stepping Multi-Criteria

9-46

Mentat Load Increment Parameters Adaptive Stepping Arc Length

9-60

Mentat Iteration Parameters Solution Control

9-61

Automatic Time Stepping Example

9-63

Geometrically Nonlinear Framework

9-65

Total and Lagrangian Overview

9-66

How to choose the framework?

9-67

Total Lagrangian Usage

9-68

Updated Lagrangian Usage

9-69

Advanced Analysis Options

9-70

Cost of Nonlinear Analysis

9-72

Time Consuming Steps

9-73

Cost of Nonlinear Analysis: Example

9-74

Solvers

9-76

Iterative vs Direct Solver

9-78

Workshop 10 Pin Insertion and Extraction

MAR101, Section 0, November 2015
Copyright 2015 MSC Software Corporation

S0 - 11

CONTENTS
Section
10

Page
Resolving Convergence Problems
Overview

10-2

Nonlinear analysis guidelines

10-3

Information Available to Assist in Troubleshooting

10-8

Analysis Files

10-9

Analysis Messages

10-10

Status File

10-12

Troubleshooting Analysis Failure

10-13

General

10-14

Contact

10-17

Hyperelastic and Plastic Material Data

10-22

Include Reality

10-24

Elements

10-25

Convergence Criteria Behavior

10-27

Analysis Failure: Exit Numbers

10-34

Exit number 2004

10-35

Exit numbers 1005/1009

10-43

Workshop 11 Contact Analysis to Generate Force-deflection of a Spring

Appendix
A

Defining the Contact Constraints

MAR101, Section 0, November 2015

S0 - 12

SECTION 1
COURSE OVERVIEW

MAR101, Section 1, November 2015

S1 - 1

COURSE OBJECTIVES
Become knowledgeable with Nonlinear Finite Element Analysis:
Large deflection
Material nonlinearities
Contact

Learn capabilities of MARC nonlinear FEA software

Learn to use basic capabilities of MARC
Learn Mentat pre/post-processor
Pre-processing
Model generation
Loads definition

Post-processing
Deformation plots
Contour plots
History plots

MAR101, Section 1, November 2015

S1 - 2

SIMCOMPANION
One stop for full online
support
Find answers to your
questions
Search across ALL content
Subscribe to email notification
Single sign-on to ALL content
Access to other support
resources
Case Management Portal
Discussion Forums
Training Information

http://simcompanion.mscsoftware.com
MAR101, Section 1, November 2015
Copyright 2015 MSC.Software Corporation

S1 - 3

SIMCOMPANION
Personalized Support via the
following channels
Web
Submit a Case Online
Manage My Cases

Email
List of Addresses in Support Contact
Information

Phone
List of Phone Numbers in Support
Contact Information

MAR101, Section 1, November 2015

S1 - 4

SIMCOMPANION

MAR101, Section 1, November 2015

S1 - 5

SIMCOMPANION
Access to Communities
VPD Community Discussion Forums
Subscribe to discussion communities of interest

MAR101, Section 1, November 2015

S1 - 6

SECTION 2
MARC/MENTAT OVERVIEW

MAR101, Section 2, November 2015

S2 - 1

COMPANY OVERVIEW 50 YEARS OF MSC MILESTONES

1963 - Company founded by
Dr. Richard MacNeal and Mr. Robert
Schwendler. Developed first program
called SADSAM (for Structural Analysis by
Digital Simulation of Analog Methods) - the
forerunner to MSC Nastran.

1972 - MSC releases proprietary version

of NASTRAN, called MSC Nastran.

1965 - MSC participates in NASAsponsored project to develop a unified

approach to computerized structural
analysis. The program became known as
NASTRAN
(NASA Structural Analysis Program).

1994 - MSC merges with PDA

Engineering (Developer of PATRAN) to
become the largest single provider of
finite element analysis (FEA) software to
the CAE market.

1965 - A team of researchers at Brown

University initiated the development of the
technology leading to the MARC program.

1999 - MSC.Software merges with

MARC Analysis Research to lead both,
the linear and the nonlinear analysis
CAE market worldwide.

1971 - The MARC Analysis Research

Corporation was founded.

Continuous development with yearly

releases of Marc

MAR101, Section 2, November 2015

S2 - 2

1972 - MARC Corp releases the first

proprietary version of MARC, the first
commercial Nonlinear finite element
analysis program.

COMPANY OVERVIEW
The MSC.Software Corporation (formerly MacNeal-Schwendler
Corporation) has been supplying sophisticated computer-aided
engineering (CAE) tools since 1963
MSC.Software is the developer, distributor, and supporter of the
most complete and widely-used structural analysis program in
the world, MSC Nastran, as well as the first commercial nonlinear
analysis program in the world, Marc. In addition:
MSC Nastran
Patran
Adams
Marc
Mentat

MSC MVision
MSC Fatigue
MSC Laminate Modeler
And more

MARC101

MAR101, Section 2, November 2015

S2 - 3

COMPANY OVERVIEW - SOLVERS

MSC Nastran (implicit)

Linear Analysis
Vibration
Classic Dynamics
Basic Nonlinear

Marc (implicit)

Advanced Nonlinear
Spring back
Welding
Superplastic Forming

MARC101

Adams
Rigid Body Kinematics

MAR101, Section 2, November 2015

S2 - 4

COMPANY OVERVIEW PRE/POST PROCESSORS

Patran

Mentat

MARC101
SimXpert

MAR101, Section 2, November 2015

Adams/view

S2 - 5

MARC / MENTAT HOW DO THEY WORK TOGETHER?

Finite Element Analysis
Pre-processing Model Definition

Geometry
Material
Loads
Creates input file to solver, .dat file
Typically user interactive

MENTAT
input file

Processing - Solution

MARC

Compute intensive
Typically no user interaction
Creates results file .t16

results file

Post-processing - plots
Review of Solution results

Displacements
Stresses
Strains
Special requests

MENTAT

Typically user interactive

Creates images, animations, graphs
MAR101, Section 2, November 2015
Copyright 2015 MSC.Software Corporation

S2 - 6

MARC OVERVIEW
Analyzed and influenced final design
decisions on:

Automotive parts
Nuclear reactor housings
Biomedical equipment
Offshore platform components
Fiberglass fabric roof structures
Rocket motor casings
Ship hulls
Elastomeric mounts
Space vehicles
Electronic components
Steam-piping systems
Engine pistons
Tires
Jet engine rotors
Welding, casting, quenching
Large strain metal extrusions

MAR101, Section 2, November 2015

TOYOTA

S2 - 7

SUMMARY OF MARC ANALYSIS TYPES

Structural

Linear & Nonlinear Solutions

Static & Transient Analysis
Buckling & Post-Buckling
Fracture Mechanics
Time and Frequency based
(Classical) Dynamics
Frequency (and Modal) Extraction
Direct (Transient)
Modal (Transient)
Frequency Response (Steady
State)
Spectrum Response
Design Sensitivity and
Optimization

MAR101, Section 2, November 2015

Non-Structural

S2 - 8

Thermal
Electrostatic
Piezoelectric
Magnetostatic
Current/Thermal
Magnetodynamic
Diffusion
Acoustic
Fluid

MARC ADVANCED SOLUTION FEATURES

Overview
Fully implemented large
deformation and large strain
Robust Newton-Raphson and
Arc-length methods
Manual/automatic load
incremental procedures
Industry-leading parallel
processing
Global and local remeshing
Highly regarded contact
capability
Highly customizable via User
Subroutines

MAR101, Section 2, November 2015

S2 - 9

MARC ADVANCED SOLUTION FEATURES

Types of elements

Large Element Library

Sophisticated formulations to
account for large strain and its
associated numerical difficulties
All elements support large
deformations
Most elements support large
strains
Many large rotation-increment
shells/beams
Composite Elements
Incompressible Elements
All elements may be combined
User control over integration
methods

0D (Point) Elements
1D (Bar and Beam) Elements
2D Solid (Continuum) Elements
Axisymmetric Shell Elements
3D Solid (Continuum) Elements
3D Shell Elements
3D Solid Shell Element
Interface Elements
Semi-infinite Elements
Rebar

Example: Mode 1 for encastre beam with

combined element types
MAR101, Section 2, November 2015
Copyright 2015 MSC.Software Corporation

S2 - 10

MARC ADVANCED SOLUTION FEATURES

Materials
Large library of built-in material models, including:

Elastic (isotropic, orthotropic, anisotropic)

Ductile Failure (isotropic, orthotropic, anisotropic)
Progressive composite failure (shell and solid)
Brittle damage (post-yield softening)
Micro-void damage
Plastics
Hyperelastic (rubber, elastomer)
Creep
Viscoelastic/viscoplastic

Large library of built-in material

properties
Temperature dependent
Strain rate dependent

MAR101, Section 2, November 2015

S2 - 11

MARC ADVANCED SOLUTION FEATURES

Contact
Deformable and/or rigid bodies
Analytic or discrete contact
surfaces
Velocity, force, or displacement
control
Self contact
Complete control over permissible
contact, if required
Stick-slip or continuous friction
models

MAR101, Section 2, November 2015

S2 - 12

Video :

MARC ADVANCED SOLUTION FEATURES

Contact Continued
Marc tracks the motion of all contact bodies with respect to each other at all
times.
When two bodies come in contact, Marc automatically determines the area
of contact and calculates the contact normal and friction stresses.
Rigid body

Contact area

Contact stress
(including friction)
Calculated

Deformable
Body

Rigid body

MAR101, Section 2, November 2015

S2 - 13

MARC ADVANCED SOLUTION FEATURES

Local Adaptive Re-meshing

Automatic MPC

Penetration of Element Edge

Due to Mesh Discretization
MAR101, Section 2, November 2015
Copyright 2015 MSC.Software Corporation

Result of Local
Adaptive Re-meshing

S2 - 14

MARC ADVANCED SOLUTION FEATURES

Global Adaptive Re-meshing
For problems exhibiting very large strains and mesh distortion which could
prevent the analysis from continuing
Example:
Two rubber pieces coming into contact
The piece on the left undergoes self contact as the hole closes up

MAR101, Section 2, November 2015

S2 - 15

MARC ADVANCED SOLUTION FEATURES

Distributed Domain Decomposition
Parallel Processing
Automatic Subdivision based on Metis
Manual Decomposition based on Sets

MAR101, Section 2, November 2015

S2 - 16

MARC ADVANCED SOLUTION FEATURES

Distributed Domain Decomposition Continued
Coupled Thermal-Mechanical Analysis (V-16 Diesel Engine)

Uses 8-noded hexahedral elements

Total Number of Elements: 1,550,000
Total Number of Nodes: 2,170,000
Total Number of Degrees of Freedom: 6,510,000
Includes interference between head and block as well as bolt
loads
Decomposed into 11 Domains
Largest domain required 0.8GB of RAM

MAR101, Section 2, November 2015

S2 - 17

MARC ADVANCED SOLUTION FEATURES

Pre-state Axisymmetric to 3D

Example Data Transfer from Pre-State

Axisymmetric to 3-D Analysis

MAR101, Section 2, November 2015

S2 - 18

MENTAT OVERVIEW
Full support of all MARC
features
Modern Graphical User
Interface (GUI)
Intuitive easy to learn GUI
Consistent with popular
Windows standards

Advanced Geometry
importation / manipulation
Most graphics standards
supported

ACIS
Parasolid
SolidWorks
Pro/Engineer
CATIA V4/5
IGES
More (STL,STEP,Inventor,
Unigraphics,DXF)

MAR101, Section 2, November 2015

S2 - 19

MENTAT ADVANCED FEATURES

On-line documentation
Advanced meshing
capabilities
Direct input or import of
material properties
Graphic representation of
applied boundary conditions

MAR101, Section 2, November 2015

S2 - 20

MENTAT ADVANCED FEATURES

Extensive Post-Processing
Capabilities

Deformation plots
Contour plots
Vector plots
Time history plots

MAR101, Section 2, November 2015

S2 - 21

MENTAT ADVANCED FEATURES:

Fully Customizable

Custom menu support

Python support
Utilities to allow direct access to model databases and results files
Allows automation for repetitive requirements

MAR101, Section 2, November 2015

S2 - 22

MARC/MENTAT DOCUMENTATION
Volume A: Theory and User Information
Volume B: Element Library
Volume C: Program Input
Volume D: User Subroutines
Volume E: Demonstration Problems
Users Guide: Step by step example
problems (Mentat-based)

MAR101, Section 2, November 2015

S2 - 23

MARC Users Guide Example Problems

The User's Guide contains example problems of various capabilities of Marc/Mentat
that can be run with one button click as shown bellow

MAR101, Section 2, November 2015

S2 - 24

SECTION 3
NONLINEARITY OVERVIEW

MAR101, Section 3, November 2015

S3 - 1

SECTION CONTENTS
Some basics
Geometric nonlinearity
Contact nonlinearity
Material nonlinearity

MAR101, Section 3, November 2015

S3 - 2

SOME BASICS OF NONLINEARITY

MAR101, Section 3, November 2015

S3 - 3

COMPARISON OF LINEAR AND NONLINEAR ANALYSIS

Linear

Nonlinear

Stiffness unchanged with load

or deflection
Subset of nonlinear simulation
Less computationally intensive
Superposition applies

MAR101, Section 3, November 2015

Stiffness is a function of
loads/deformation
Stiffness recalculated many
times
Iterative solution subject to
convergence requirements
More computationally intensive
More representative of real
world
Superposition does not apply

S3 - 4

MARC BASIC FUNCTIONALITIES

Contact between structures
Automatic handling of load increments
Accurate material laws
Accurate and robust element technology
Robust solution techniques

MAR101, Section 3, November 2015

S3 - 5

GENERAL CONSIDERATIONS IN NONLINEAR ANALYSIS

Importance of nonlinearity
Classical analysis techniques are limited in their application to problems
that fit into linear assumptions.
Most Real World structures exhibit nonlinear behavior sometimes a
linear response is reasonable.

Questions to ask to help identify nonlinearity

Is the stiffness of the structure a function of load or deflection?
What is the source of nonlinearity?
Can this nonlinearity be ignored?

Working with nonlinear analysis

Analysis time increased

Debugging can be time consuming for large analyses
Use small test cases to understand the behavior first
Increased complexity leads to higher memory requirements

MAR101, Section 3, November 2015

S3 - 6

GENERAL CONSIDERATIONS IN NONLINEAR ANALYSIS

Nonlinear analysis involving large meshes, various contact
bodies and inelastic materials remain challenging.
As complexity increases so does solution time. This has a big
impact on the time it takes to obtain a stable nonlinear solution.
Simplify whenever possible
Break the problem into smaller problems, analyze these, and only if
necessary, analyze the whole problem
Reduce the problem to 2D

MAR101, Section 3, November 2015

S3 - 7

SOURCES OF NONLINEARITY
Geometric Nonlinearity

Large deflections
Large rotations
Preloads
Structural instability and collapse

Contact
Material Nonlinearity
Plasticity
Creep
Progressive failure

MAR101, Section 3, November 2015

S3 - 8

EXAMPLE OF 3-DIMENSIONAL DOOR SEAL SIMULATION

Wrinkle caused by the
contact between door and
the seal

Self Contacted area

Example
Lip bending due to the contact
between door and seal

MAR101, Section 3, November 2015

S3 - 9

A reinforced rubber seal,

as shown here combines
all three sources of
nonlinearity

MULTIPLE SOURCES OF NONLINEARITY

Marc was originally written as a nonlinear code
It provides nonlinear capability for all elements and materials
Almost all the individual nonlinear capabilities can be used
together
Hammer
Initial position Outline

Hyperelastic
(rubber) band

Stamp
Trigger
Impact Contact Pair

Sliding Contact

Plastic
Seal

Anvil
MAR101, Section 3, November 2015
Copyright 2015 MSC.Software Corporation

S3 - 10

GEOMETRIC NONLINEARITY

MAR101, Section 3, November 2015

S3 - 11

COMPARISON OF NONLINEAR AND LINEAR SOLUTIONS

Nonlinear stiffness matrix
assembled using:

[ K ] [ B ]T [ D][ B ] dV
Strain-displacement relationship
is defined by the [B] matrix. It
contains both, linear and
nonlinear terms.

Linear Solution

Cantilever beam example

If the deflection is large, the
beam develops axial strains
which absorb strain energy.
This results in a smaller vertical
deflection of the beam tip.
Geometric nonlinearity occurs
whenever the displacement
changes the effective stiffness of
the structure.
MAR101, Section 3, November 2015
Copyright 2015 MSC.Software Corporation

S3 - 12

Nonlinear Solution

EXAMPLE OF STRESS STIFFENING

Geometric nonlinearity occurs in stress stiffened structures such
as a violin string as shown below:

MAR101, Section 3, November 2015

S3 - 13

EXAMPLE OF STRESS STIFFENING

Eigenmodes and Eigenfrequencies of a blade on a jet engine are
much higher when the engine is rotating at operating speed than
when it is stationary.

MAR101, Section 3, November 2015

S3 - 14

TYPES OF GEOMETRIC NONLINEARITY

Large displacements and/or rotations
Buckling
snap through
snap back
collapse

Large strain

MAR101, Section 3, November 2015

S3 - 15

EXAMPLE OF SNAP THROUGH

Must be analyzed as nonlinear
Post buckling situation
Numerically difficult to
capture

MAR101, Section 3, November 2015

S3 - 16

LATERAL COLLAPSE OF FLAT BEAM

Buckling and collapse is the
opposite of stress stiffening
Structures under compressive load
can fail suddenly and collapse
Very important in many structural
design applications

MAR101, Section 3, November 2015

S3 - 17

BUCKLING AND COLLAPSE OF A SHELL

The shell collapses after the critical buckling point load is
exceeded
Without a nonlinear analysis, the critical load will be
overestimated this is dangerous.

MAR101, Section 3, November 2015

S3 - 18

EXAMPLE OF A RUBBER BOOT

Geometric nonlinearity
Large strain and large deformation

Self-contact
Material nonlinearity

MAR101, Section 3, November 2015

S3 - 19

LARGE STRAINS
Large strains are used to:
Design rubber components
Design metal forming processes

Examples problems:

Engine mounts
Gaskets and seals
Solid propellant
Stretching of a thin sheet with
a hemispherical punch (demo
e8x52.dat)

Forming using a hemispherical punch

MAR101, Section 3, November 2015
Copyright 2015 MSC.Software Corporation

S3 - 20

NON-CONSERVATIVE LOADING

Geometric nonlinearity enables non-conservative loading.

The pressure stays normal to the deformed shape
Because the load now changes direction it can produce a significant change
in the reaction forces and moments
Changes in both direction and area are taken into account
Also sometimes called follower force

MAR101, Section 3, November 2015

S3 - 21

CONTACT NONLINEARITY

MAR101, Section 3, November 2015

S3 - 22

CONTACT: BASICS
Finite elements are based on the concept of local support
Nodes and elements usually communicate only with their nearest
neighbors

Elements not connected via a common node are not aware of

each other and pass through unrestrained

Standard finite element solutions are not sufficient for contact

problems
Marc addresses this matter with contact algorithms
MAR101, Section 3, November 2015
Copyright 2015 MSC.Software Corporation

S3 - 23

CONTACT BODIES IN MARC

Marc introduces the concept of Contact Bodies to determine
potential contact within a simulation
Deformable body
Body is deformable
Stress and temperature distribution

Rigid body
Body is not deformable (rigid)
No stress distribution

Continuously monitors relative locations of Contact Bodies

Automatically calculates forces between contact bodies
Automatically accounts for continuously varying contact area and
relative sliding
Works together with geometric and material nonlinearity
Applicable to static, dynamic, fluid, electrical, and thermal
analyses
MAR101, Section 3, November 2015
Copyright 2015 MSC.Software Corporation

S3 - 24

EXAMPLE OF FRICTION CLUTCH

Accounts for heat generation through friction
Plastic work may also be converted to heat
The heat produced by thermal expansion may affect the contact
conditions
Convection, conduction, and radiation between contact bodies
Varying heat transfer as bodies
approach
Other examples include
Forming process for metals and
plastics (rolling, extrusion)
Sealing processes
Impact analyses
Analysis of brakes and clutches
Car tire rolling

MAR101, Section 3, November 2015

S3 - 25

WORKING WITH CONTACT BODIES

Whole components can be
selected and declared as a
contact body
Contact between bodies is
determined by a (optional)
contact table
default is that all bodies see all
other contact bodies, including selfcontact

Seal after Mount step

It is not necessary to select outer

edges or surfaces
Neither is it a requirement to
define complicated master-slave
relationships

Door and Glass Closed

MAR101, Section 3, November 2015
Copyright 2015 MSC.Software Corporation

S3 - 26

EXAMPLE OF A GEAR RACK

Contact area is continuously changing

MAR101, Section 3, November 2015

S3 - 27

EXAMPLE OF BOLT LOADING

Each of the four parts is

meshed separately

Each of the four parts is

defined as a contact body

Contact possibilities are

specified in the Contact
Table, if required

Contact provides the

connectivity

MAR101, Section 3, November 2015

S3 - 28

EXAMPLE OF 3-D RIVETS UNDER CYCLIC LOADS

Each of the parts is meshed separately
Each of the parts is defined as a contact body
Contact possibilities are specified in the Contact Table, if required
Contact provides the connectivity

MAR101, Section 3, November 2015

S3 - 29

EXAMPLE OF 3-D MODEL UNDER CYCLIC LOADS

Magnified Displacement 4x

MAR101, Section 3, November 2015

S3 - 30

EXAMPLE OF DYNAMIC PIPE CRUSH

Dynamic impact (contact)
Large deformation
Large strain
Material Nonlinearity

MAR101, Section 3, November 2015

S3 - 31

EXAMPLE OF RUBBER SHOCK ABSORBER

Self contact may occur in addition to other contact

The floor, wall, and tool are represented by rigid surfaces

MAR101, Section 3, November 2015

S3 - 32

EXAMPLE OF STRESS EVALUATION

Peak stress is automatically

located

In some cases, larger stresses will

develop away from the region
experiencing contact

In other cases, there will be a

large concentration of stress in
the contact region

MAR101, Section 3, November 2015

S3 - 33

EXAMPLE OF CAN OPENING

Contact bodies may be given a failure criteria

MAR101, Section 3, November 2015

S3 - 34

EXAMPLE OF CHANGING CONTACT AREA

Crushing of a cylinder

Contact area is automatically evaluated

MAR101, Section 3, November 2015

S3 - 35

EXAMPLE OF RIGID CONTACT

Simulating rigid components

Some contacting parts are much stiffer than the others and may
be considered rigid

Marc allows the creation of rigid bodies and their interaction with
deformable bodies

The alternative is to specify a high stiffness for a deformable

body but too much stiffness causes ill-conditioned (or singular)
stiffness matrices

MAR101, Section 3, November 2015

S3 - 36

EXAMPLE OF RIGID CONTACT

Multi-pass Ring Rolling

Process

Rigid contact bodies very

widely used for gaskets, seals,
and metal forming

Ring
Rigid Surface
MAR101, Section 3, November 2015
Copyright 2015 MSC.Software Corporation

S3 - 37

EXAMPLE OF SURGICAL EYE IMPLANT

The solid body is completely free of BCs other than the contact
bodies.
Other codes require temporary constraints, while Marc can
resolve the problem even when the rigid surfaces are initially
separated.
Undeformed Shell

Undeformed Shell

Pod

MAR101, Section 3, November 2015

S3 - 38

EXAMPLE: CONTACT WITH LARGE DEFORMATION AND

PLASTICITY
The deformable structure may be experiencing local or global
buckling and collapse, as well as very large displacements
For shell elements, thickness can be taken into account in
determining contact

MAR101, Section 3, November 2015

S3 - 39

EXAMPLE: QUADRATIC CONTACT

Full support
for quadratic
elements

MAR101, Section 3, November 2015

S3 - 40

EXAMPLE: CONTACT AND RE-MESHING

MAR101, Section 3, November 2015

S3 - 41

MATERIAL NONLINEARITY

MAR101, Section 3, November 2015

S3 - 42

MATERIALS IN MARC
Marc supports 20+ material models including:

Elastic
Elastic-Plastic
Creep
Rubber (hyperelastic)
Foam
Composite
Mixtures
Rebar
Cohesive Interface
Gasket
Shape memory
Visco-plastic

Materials discussed in
the MAR101 course

All can be combined with other nonlinearities

MAR101, Section 3, November 2015

S3 - 43

MOST WIDELY USED MATERIAL MODELS

Elastic
An ideally elastic material has the
following properties:
Linear relationship between stress
and strain
Hookes Law

The deformation between any

elastic reference state and an
alternate state is reversible.
As a result, loads are said to be
fully recoverable.

Non-path dependent loading

Computationally simple

MAR101, Section 3, November 2015

S3 - 44

MOST WIDELY USED MATERIAL MODELS

Plastic
As stresses and strains are
increased beyond the yield point,
metals start to exhibit a nonlinear
behavior.
The yield point defines the shift
from elastic to plastic behavior
for a material.
Loading beyond the yield stress
includes permanent plastic
deformation.
A yielded ductile metal will
unload along a curve that is
parallel to the initial linear elastic
curve.

MAR101, Section 3, November 2015

S3 - 45

Plastic strain

MOST WIDELY USED MATERIALS

Hyperelastic
Hyperelastic materials, such as
rubber, exhibit highly nonlinear
elastic stress-strain behavior.
Hyperelastic materials can remain
elastic up to large strain values
(often up to 100% strain and
beyond).
Hyperelastic materials present a
different behavior (measured by the
relation between stresses and
strains) when subject to different
type of loadings.
MAR103 Course combines
laboratory tests and Marc analyses
for a thorough learning experience
on working with elastomers.

MAR101, Section 3, November 2015

S3 - 46

EXERCISE
Workshop 1 Analysis of a Rubber Seal
Be sure to ask for help if there is anything you do not understand

MAR101, Section 3, November 2015

S3 - 47

MAR101, Section 3, November 2015

S3 - 48

SECTION 4
INTRODUCTION TO MENTAT

MAR101, Section 4, November 2015

S4 - 1

MARC / MENTAT HOW DO THEY WORK TOGETHER?

FEA Analysis
Pre-processing Model Definition

Geometry
Material
Loads
Creates input file to solver, .dat file
Typically user interactive

MENTAT
input file

Solution
Compute intensive
Typically no user interaction
Creates results file .t16

MARC
results file

Post-processing
Review of Solution results

Displacements
Stresses
Strains
Special requests

MENTAT

Typically user interactive

Creates images, animations, graphs
MAR101, Section 4, November 2015
Copyright 2015 MSC.Software Corporation

S4 - 2

MENTAT
Where to get help:
MARC 2015 User Guide, Getting Started > Basics of Mentat and Finite
Element Modeling :
Mentat: a graphical user interface program that allows you to execute a finite
element analysis process from start to finish
Full description of the Mentat program and how to use it

MAR101 Introduction to Marc and Mentat course notes

Section 4 Introduction to Mentat
Workshops
Many step by step examples using Mentat

MAR102 course notes Advanced Nonlinear Analysis using Marc and

Mentat
Website: http://simcompanion.mscsoftware.com/

MAR101, Section 4, November 2015

S4 - 3

MENTAT GETTING STARTED

Starting Mentat
Windows
Double click Icon on Desktop OR
Start>All Programs>MSC.Software>Marc 2015.0.0>Marc Mentat 2015.0.0:

Linux/Unix
Type mentat at command line (may be different on your installation, consult with
your IT dept)
MAR101, Section 4, November 2015
Copyright 2015 MSC.Software Corporation

S4 - 4

PRE-PROCESSING IN MENTAT

MAR101, Section 4, November 2015

S4 - 5

HOW MENTAT COMMUNICATES WITH YOU

Mentat communicates with you via prompts and messages in the
dialog area and with other visual queues.
Mentat's prompts urge you to take action through the input of
data or commands.

MAR101, Section 4, November 2015

S4 - 6

HOW MENTAT COMMUNICATES WITH YOU

Some basic rules:
If you misspell a keyword or enter an incorrect response, Mentat warns you
through a message posted in the dialog area.
Mentat does not require that you complete every action you initiate.
For example, if you are prompted for a filename and change your mind, click
Enter instead of typing in the filename which will indicate to Mentat to abort the
action.

If the program is waiting for a list of items to operate on and instead you
enter a command that also requires a list of items or any additional data,
Mentat ignores the original request and processes the command. If the
command you enter does not request additional data, you are returned to
the original data request from before the interrupt.
The program assumes that you want to repeat the previous operation on a
new set of items and prompts you for a new list to operate on. This process
repeats itself until you indicate otherwise by entering a new command or a
click Enter.

MAR101, Section 4, November 2015

S4 - 7

HOW YOU COMMUNICATE WITH MENTAT

There are three methods available for the user to input data into
Mentat
Keyboard input
Can be used for virtually all types of input requirements
Required for names or dimensions
Not practical in some cases like lists (can be too long) or commands (must know
the command)
Can be input in dialog area or pop-up menu (click the box to be input in other
than dialog area)

Mouse
Most practical for command selection
Most practical for selection of entities
Widely used for graphic manipulation

Combination
In some cases a combination can be used - for example consider translating
entities where a dimension (keyboard input) must be input as well as identifying
which entities are to be moved (mouse selection)

MAR101, Section 4, November 2015

S4 - 8

COMMUNICATING WITH MENTAT - USING THE MOUSE

It is important to make a distinction between using the mouse in
the menu area versus the graphics area because the three mouse
buttons have very different functions in each area.
Menu area operation:
To select a menu item with the mouse, move the mouse arrow over the item
that you want to select and left-click
Hovering the mouse arrow over a menu item will provide a short description
of its function
Each menu Item has a help panel with a short description and explanation
of the function of that menu item. To activate the help feature, position the
mouse arrow over the menu item which you require help with and then click
the middle mouse button.

MAR101, Section 4, November 2015

S4 - 9

COMMUNICATING WITH MENTAT - USING THE MOUSE

GRAPHIC area operation:
To pick existing items (that is create a list of items)
move the mouse over the item to be identified followed by left-click.
You can undo that action by clicking the middle mouse button anywhere in the
graphics area.
At times, you will need to identify more than a single item. A list of items must be
terminated by a click of the right-click with the mouse arrow positioned anywhere
in the graphics area.

To select the location for new items

You can use a pre-defined grid space to easily locate items to be created.
If you left-click the mouse button while the mouse arrow is in the vicinity of a point
in the grid space, your item will be created at the coordinates of that point
In addition, you can also pick an existing node, point, or surface-grid-point to
specify a location.

MAR101, Section 4, November 2015

S4 - 10

Mentat Enhancements

MAR101, Section 4, November 2015

S4 - 11

MENTAT LAYOUT
Static Menu Area

Menu bar

Tool bar

Dynamic
Menu Areas

Graphics
Area
Dialog Area

Status Areas
MAR101, Section 4, November 2015
Copyright 2015 MSC.Software Corporation

S4 - 12

Main Menu

MENTAT LAYOUT
The Mentat Window is divided into three major areas:
Graphics area
used to display the current state of the database. When you start Mentat, the
graphics area is blank to indicate that the database is empty.

Menu area
Reserved to show the selectable menu-items
It is divided into two submenus types
Static: always present and contains items that are applicable and selectable at all times
Dynamic: contents of the dynamic menu area change as the static menu items are
selected

Dialog area
A scrollable area of about five visible lines where all program prompts, warnings,
and responses appear, and where the user can input data or commands.
Status area within the dialog area is reserved to communicate the state of the
program to the user. Either Working or Ready appears in the status area to
reflect the current state of the program. For intensive operations, an additional
progress widget will appear.

MAR101, Section 4, November 2015

S4 - 13

MENTAT LAYOUT A CLOSER LOOK

Window Header
Menu Bar
Toolbar

Main Menu

Select Bar
Dialog Area
Dynamics Menu
MAR101, Section 4, November 2015
Copyright 2015 MSC.Software Corporation

S4 - 14

MENTAT LAYOUT WINDOW HEADER

Useful Information:
Current database
If a save operation is performed this is the file that will be written

Mentat version
Current view

MAR101, Section 4, November 2015

S4 - 15

MENTAT LAYOUT MENU BAR

File
New: create new database
(purging current database)
Open: opens previously
existing database
Save: saves current database
Restore: retrieves last saved
version of the current database
undo of all changes since last
save

Import: imports CAD geometry

Covered in more detail later in
this section

Exit: leaves Mentat

Make sure to Save prior to
Exit if you want to save
changes made since last Save

MAR101, Section 4, November 2015

S4 - 16

MENTAT LAYOUT MENU BAR

Select
Useful or necessary when assigning
attributes (materials, geometric properties,
boundary conditions, etc.) to large groups
of entities
Covered in more detail later in this course

MAR101, Section 4, November 2015

S4 - 17

MENTAT LAYOUT MENU BAR

View
Controls how entities are plotted in the graphics area

MAR101, Section 4, November 2015

S4 - 18

MENTAT LAYOUT MENU BAR

Tools
Catch-all for items that have no other
logical home
Most often used for:
Animation creates animation of
simulation results
Distance measures distances between
entities
Calculations measures quantities
related to FE mesh such as element
volume, area, etc.

MAR101, Section 4, November 2015

S4 - 19

MENTAT LAYOUT MENU BAR

Window
Controls what type of plot is displayed
If multiple (windows displayed)in graphics area, organizes graphics
windows
Creates screen snapshot file

MAR101, Section 4, November 2015

S4 - 20

MENTAT LAYOUT MENU BAR

Help
Access on-line documentation

MAR101, Section 4, November 2015

S4 - 21

User Guide New and Improved - HTML

MAR101,
11/10/201 Section 4, November 2015
MSC Software Confidential
5 2015 MSC.Software Corporation
Copyright

S422
- 22

MENTAT LAYOUT TOOL BAR

Main Toolbar

Create: new database

Purges existing database

Open: existing database

Save: current database
Undo: last operation
Only a single level of undo is available

MAR101, Section 4, November 2015

S4 - 23

MENTAT LAYOUT TOOL BAR

View Toolbar

Fill View: visible geometry to graphics area

Dynamic mouse manipulation of graphics window

Toggled by:
Click Icon
Hold down Alt

Yellow background indicates

dynamic mode is active

If active, use the mouse buttons:

Left: Pan
Middle: 3D rotation
Right: Zoom

Mouse Zoom
Drag a rectangle around the area
of interest.

Incremental graphic motion

Relative to screen coordinates

MAR101, Section 4, November 2015

S4 - 24

MENTAT LAYOUT TOOL BAR

Plot Toolbar
Element display
Wireframe or solid
Grayed out if no elements exist

Surface display
Wireframe or solid(with/without internal
lines)
Grayed out if no surfaces exist

Solid display
Wireframe or solid
Grayed out if no solids exist

MAR101, Section 4, November 2015

S4 - 25

MODEL NAVIGATOR
Graphical representation of the model
Easy access to menus and properties forms
Filter entities

MAR101, Section 4, November 2015

S4 - 26

MODEL NAVIGATOR
Opening and closing branches of the Model Navigator

+
open close

MAR101, Section 4, November 2015

S4 - 27

MODEL NAVIGATOR
Right Mouse Button

Right-click in white
area of model
navigator for menu
MAR101, Section 4, November 2015
Copyright 2015 MSC.Software Corporation

S4 - 28

MODEL NAVIGATOR
Right Mouse Button
Identify contact bodies

MAR101, Section 4, November 2015

S4 - 29

MODEL NAVIGATOR
Right Mouse Button
Identify boundary conditions

MAR101, Section 4, November 2015

S4 - 30

MODEL NAVIGATOR
Right Mouse Button
Add a new entity

MAR101, Section 4, November 2015

S4 - 31

MODEL NAVIGATOR
Location and undocking the Model Navigator

MAR101, Section 4, November 2015

S4 - 32

MODEL NAVIGATOR
Toggle between modes
List/Model
Dynamic Menu/Model Navigator

MAR101, Section 4, November 2015

S4 - 33

MODEL NAVIGATOR
List Mode
Filter to reduce the number of visible entities.
Toggle entities on/off

MAR101, Section 4, November 2015

S4 - 34

Model Browser Drag and Drop

LBC show up in
Loadcase

Drag and Drop LBC

into Loadcase
MAR101, Section 4, November 2015
Copyright 2015 MSC.Software Corporation

S4 - 35

Copy Contact Table into Job

MAR101, Section 4, November 2015

S4 - 36

Model Browser Drag and Drop Operations

B.C into Loadcases
B.C into Jobs
Initial Conditions into Jobs
Contact Table into Loadcase
Contact Table into Job
Loadcase into Jobs
Mesh Adaptivity into Loadcases

MAR101, Section 4, November 2015

S4 - 37

IMPORTING CAD GEOMETRY INTO MENTAT

MAR101, Section 4, November 2015

S4 - 38

PARASOLID IMPORT
Import / Parasolid
This method requires the Feature:
Mentat_Parasolid_Modeling
This method imports the Parasolid
geometry directly into the kernel.
For 2014 and later, this is the least
expensive way to import Parasolid
Geometry into Mentat.

MAR101, Section 4, November 2015

S4 - 39

PARASOLID IMPORT
Import / General CAD as Solids
This method uses the ParaSolid Kernel and the
CT Translators.
This method requires the Features:
Mentat_Parasolid_Modeling
Mentat_Geometry_Translators

This method imports the Parasolid geometry into

the Parasolid kernel using the CT Libraries.
The benefit is that there are more geometry
clean-up tools and defeaturing capabilities.
The additional functionality that the user would
get is the ability to import the other geometry
formats supported by PID 10635 (ACIS, STEP,
and IGES)
Note that meshing is easy one step

MAR101, Section 4, November 2015

S4 - 40

Apply Boundary Conditions Directly on Solids

Vertices
Edges
Faces
Volumes

MAR101, Section 4, November 2015

S4 - 41

Apply Boundary Conditions Directly on Solids

MAR101, Section 4, November 2015

S4 - 42

Apply Boundary Conditions Directly on Solids

MAR101, Section 4, November 2015

S4 - 43

Solid Mesh Associativity - Generative

Apply BC to Geometry
Transferred to Mesh

MAR101, Section 4, November 2015

S4 - 44

Solid Mesh Associativity - Generative

Change Geometry
F.E. Mesh Automatically Changed
BC Automatically Applied

MAR101, Section 4, November 2015

S4 - 45

MENTAT HINTS AND SHORTCUTS

Mouse in graphics
Left-mouse click to pick
Right-mouse click to accept pick

Mouse in menu
Left-mouse click to pick another menu or function
Middle-mouse click (or F1) for help

Save your work (Ctrl+S)

Go to Files and select Save As and specify a file name. This will save the
current Mentat database to disk.
Use Save from then on

Dialog region
At the lower left of screen displays current activity and prompts for input
Check this region frequently to see if input is required

MAR101, Section 4, November 2015

S4 - 46

MENTAT HINTS AND SHORTCUTS

Dynamic Viewing
Identified by hand icon
Used to position the model in the graphics area
Activated by holding the Alt key

Active

Inactive

Left-click translates the model

Right-click zooms in/out,
Middle-click rotates in 3-D.

Zoom can also use Ctrl+Up/Down

Be sure to turn off Dynamic View before picking in the graphics area

Use Reset View

and Fill View

(Ctrl+F) to return to original view

Help
All of the workshop problems have Mentat procedure and data files.
They are located in a marc.ug directory under Mentats main directory. The
directory/file structure looks like: ~mentat/examples/marc.ug/s3/c3.9/ for Section
3, Chapter 3.9. Furthermore, you can click on the filename listed in the input files
table to download the files via the web.
MAR101, Section 4, November 2015
Copyright 2015 MSC.Software Corporation

S4 - 47

SELECTING IN MENTAT
Often it is required to identify objects or entities within Mentat that are to
be operated on. This is done using Mentats select capabilities
Example of selecting elements to delete
Click Geometry & Mesh under the Geometry & Mesh tab
Click the Rem button to the right of Elements in the pop-up form
Note the prompt in the dialog window Enter remove element list. It is now
up to the user to select the entities (elements in this example) to be deleted.
This can be done in numerous ways:
Left-Mouse button: With the mouse arrow over an item, left-click the mouse
button. Used for selection of relatively few items
Box Method: Left-click the mouse button and drag a rectangle to box in entities.
This allows picking of large numbers of items
Lasso Method: Hold down the Ctrl key and hold down the left mouse button to
drag an outline around desired entities. This is a free hand enclosure picking,
must finish at start point
Polygon Method: Hold down the Ctrl key and left-mouse click for the locations of
the end points of the vertices of a polygon that encloses the items you want to
select
Dialog Window: If you know the number, an element number can be typed into the
dialog window. Rarely used.
MAR101, Section 4, November 2015
Copyright 2015 MSC.Software Corporation

S4 - 48

SELECTING IN MENTAT
By default selections are cumulative
Middle-click of the mouse button rejects last selection pick
Right-click accepts all picks and sends list to Mentat
Using the Selection bar can streamline the picking process:

All Existing all entities of the type being selected that are
defined in the database

All Selected all previously selected entities

All Unselected anything not previously selected

All Visible anything visible in the graphics window

All Invisible anything not visible in the graphics window

Pick Set all entities in a previous defined and stored set

End List same as right-click, ends picking and sends list to

Mentat

MAR101, Section 4, November 2015

S4 - 49

SELECTING IN MENTAT
Selecting lists can be stored in Sets
Useful if you must select the same set of entities numerous times
Sets are defined using Select option in Menu Bar
Once a new set has been requested, Mentat will prompt the user to select
entities

MAR101, Section 4, November 2015

S4 - 50

VISIBILITY IN MENTAT
The Visibility form allows turning on or off specific items in the
graphics window

Removes items that are not of interest

Allows easier selecting
Focuses on a subset of the full model
Purely graphical no other modification to entities (items are not removed
from database)

Accessed from View in Menu Bar

MAR101, Section 4, November 2015

S4 - 51

PROCEDURE FILES
Procedure files playback previously recorded Mentat sessions
Interrupt recovery
Mentat automatically creates a procedure file of every session (*.proc filename)

Model modification
Procedure files are text files that can be edited with any text editor
Relatively easy to read
To make modifications the user can create a copy and change any input
parameter and then re-run the copy of the procedure file in a new Mentat session
Accessed from Menu Bar, Tools selection

MAR101, Section 4, November 2015

S4 - 52

EXERCISES
Workshop 2 CAD Import and Repair
Workshop 3A Linear Analysis of Cantilever Beam
Be sure to ask for help if there is anything you do not understand

MAR101, Section 4, November 2015

S4 - 53

POST-PROCESSING IN MENTAT

MAR101, Section 4, November 2015

S4 - 54

POST-PROCESSING WHAT IS IT?

The review of previously run simulations
Examination of primary degrees-of-freedom
Displacements/rotations for structural analyses

Examination of derived quantities

Stresses, strains, contact status, failure criteria
Vector
Plots

Deformation
Plots

MAR101, Section 4, November 2015

Contour Bands
Plots

S4 - 55

History
Plots

POST-PROCESSING BASICS
Post-processing in Mentat
Attach the results file to the Mentat session
Every Marc solution run will create a results file,
typically with the .t16 suffix
The .t16 is a binary results file containing the model
and the requested results

Create plots
Available under the Results Tab in the Main Menu

MAR101, Section 4, November 2015

S4 - 56

CREATING PLOTS
Most commonly used plots:

Model Plot
Plot of FEM distorted or color coded by requested
items
Can be combined that is exaggerated deformed
shape can have color contours based on stress

Path Plot
X-Y graph of the variation of requested item along
a defined geometric path

Defined path

Y deflection along path

MAR101, Section 4, November 2015
Copyright 2015 MSC.Software Corporation

S4 - 57

CREATING PLOTS
Most commonly used plots continued

History Plot
Plots the variation of an item vs. time
Time is often used in Mentat/Marc as a dimensionless tracking parameter
In this example, time, the X-axis, correlates to the percentage of the applied load
and the Y-axis represents the Displacement in the Y direction

MAR101, Section 4, November 2015

S4 - 58

ANIMATIONS
Animation form
Use Tools > animation from the Menu bar
Use results tab and click animation

MAR101, Section 4, November 2015

S4 - 59

ANIMATED GIF MOVIES

Easy to embed into Power Point
Available within the Movies selection of the results tab:

MAR101, Section 4, November 2015

S4 - 60

Clipping Planes for 3-d Solids

Easy definition of cutting plane
Make part of the model invisible
Create continuous contours on the
visible part of the structure

MAR101, Section 4, November 2015

S4 - 61

Clipping Planes

MAR101, Section 4, November 2015

S4 - 62

Clipping Plane showing Crack Propagation

MAR101, Section 4, November 2015

S4 - 63

Original Results E8x61

MAR101, Section 4, November 2015

S4 - 64

Clipping on Based upon Results

MAR101, Section 4, November 2015

S4 - 65

Display of other Quantity on Clipped Model

MAR101, Section 4, November 2015

S4 - 66

Non-uniform Scaling of Displacements

Non-uniform scale
factor of 25,1,1 in x,y,z
direction

Uniform Scale Factor

of 25

MAR101, Section 4, November 2015

S4 - 67

Control Over Display of Scaling Factor

MAR101, Section 4, November 2015

S4 - 68

EXERCISE
Workshop 3B Post-Processing of Linear Analysis of Cantilever
Beam
Be sure to ask for help if there is anything you do not understand

MAR101, Section 4, November 2015

S4 - 69

MAR101, Section 4, November 2015

S4 - 70

SECTION 5
GEOMETRIC NONLINEARITY

MAR101, Section 5, November 2015

S5 - 1

GEOMETRIC NONLINEARITY EXAMPLE

Linear Elastic Bar-spring
The solution to a linear elastic spring
problem is straightforward. The
displacement u is proportional to the
force P. We can write:

Pku

Spring

100 in.

The deflection is related to the load

via the (constant) stiffness. It is
readily obtained from:

uP k
However, even if all the materials
are linear elastic and the strains are
small, large deformations would still
mandate a nonlinear analysis

MAR101, Section 5, November 2015

1
in.

S5 - 2

GEOMETRIC NONLINEARITY: OBJECTIVITY

Objective strain measure capability of elements to accommodate
large motions without inducing strain
Engineering strain is defined as:

ex du dx
It is also known as infinitesimal or small strain
It is not an objective strain measure, since it strains under rigid body rotation

MAR101, Section 5, November 2015

S5 - 3

GEOMETRIC NONLINEARITY: OBJECTIVITY

Consider small rigid body rotation
where:
2

For small (tan() )

For arbitrary rigid body rotation

2
2
Hence,

Rigid body rotation produces nonzero

strain not accurate!

MAR101, Section 5, November 2015

S5 - 4

GEOMETRIC NONLINEARITY: OBJECTIVITY

Green-Lagrange strains are:
2
2
2
1
u

v
w

u
2 x

x
x

2
2
2

u
v
w
1

2 y

y
y
y

yy w
2
2
2

1 u
v
w

u v 2 z
z
z

w
yz y x

v w
x y x y x y

zx z y
u u v v w w

w u y z y z y z
x z u u v v w w

z x z x z x

Green-Lagrange strain is an objective measure

Objectivity / Invariance provides large deformation capability by
handling rigid body motion correctly

MAR101, Section 5, November 2015

S5 - 5

GEOMETRIC NONLINEARITY: OBJECTIVITY

Using Green strain, an arbitrary rigid
body rotation gives:

e cos 1 1 2 (cos 1)2

e0
Rigid body rotation No Strain
accurate!
Second order terms give coupling
between strain axes (hence, stress
stiffening is accounted for)
Second order terms do not
automatically eliminate the
infinitesimal / small strain assumption

MAR101, Section 5, November 2015

S5 - 6

GEOMETRIC NONLINEARITY: THE STIFFNESS BREAKDOWN

Consider a simple truss
The stiffness (KT) can be
expressed as follows:

Linear
Stiffness

MAR101, Section 5, November 2015

Initial
Slope
Stiffness

Geometric
(Initial Stress or
Diffferential)
Stiffness

S5 - 7

Spring
Stiffness

STRAIN MEASURES
E = Engineering (infinitesimal) strain = (L L0)/L0
A measure preferred by structural engineers
Work conjugate to the Engineering stress measure
It is only applicable to small deformation and small strain analyses

L = Logarithmic (natural) strain = dL/L = ln(L/ L0)

A measure that is incremental in form and preferred by metallurgists
Work conjugate to the Cauchy stress measure
This measure is typically used for large deformation, large strain analyses

G = Green-Lagrange strain = (L2 L02)/2L02

This measure is typically used for large deformation, small strain analysis
Additional work in the stress-strain relationship extends it to large strain work
(Greens strain accommodates finite rotations but not finite strains)
It is work conjugate to the 2nd Piola-Kirchhoff stress measure

A = Almansi strain = (L2 L02)/2L2

This measure is also known as Eulerian strain
Work conjugate to the Almansi stress measure
Not used in Marc
MAR101, Section 5, November 2015
Copyright 2015 MSC.Software Corporation

S5 - 8

STRAIN MEASURES
The salient property of the last three tensors is that they are invariant
under rigid body rotation.
Additional strain measures include Stretch and Biot.
Provided the strains remain small (say <3-4%), the different strain/stress
measures will provide the same solutions.
Small strain examples include a fishing rod bending under the weight of
a large fish, helicopter rotor blades under static dead load, and the hair
spring of a mechanical watch.
Conversions between the strain measures are readily possible as
follows:
(1 2) Green-Lagrange and Engineering
G

1
L ln(1 E ) ln(1 2 G )
2

A 2 E E2 21 E

Logarithmic and Engineering

Almansi and Engineering

The strain measure used by Marc is determined mainly by the choice of a

total or updated Lagrangian solution framework (This will be discussed
later).
MAR101, Section 5, November 2015
Copyright 2015 MSC.Software Corporation

S5 - 9

STRESS MEASURES
E = Engineering (nominal, conventional) = F/Ao
Defined in terms of the original area and the original geometry.

C = Cauchy (true) stress = F/A

Defined in terms of current area and current deformed geometry (force per unit
deformed area).
As a result, it is the most naturally understood stress measure, that is, it most
naturally describes the material response.

1 = 1st Piola-Kirchhoff stress

Defined in terms of the original area and the current deformed geometry (current
force per unit undeformed area).

= 2nd Piola-Kirchhoff stress

Defined in terms of the initial area and current deformed geometry (transformed
current force per unit undeformed area).
It is work conjugate to the Green-Lagrange strain measure.
For small strain, the 2P-K stress can be interpreted as the Cauchy stress related to
(local) axes that rotate with the material.
Without additional knowledge concerning the deformations, the 2P-K stresses are
difficult to interpret.
2P-K stresses are not uncommonly transformed into Cauchy stress to give a true
stress of use to engineers.
MAR101, Section 5, November 2015
Copyright 2015 MSC.Software Corporation

S5 - 10

STRESS MEASURES
Stress conversions can be obtained via:

The stress measure used by Marc is determined mainly by the

choice of a total or updated Lagrangian solution framework
(discussed later).
In the analysis set up, your choice of framework will determine
whether Marc uses 2nd Piola Kirchhoff or Cauchy stress, then
Marc will carry out the necessary conversations.

MAR101, Section 5, November 2015

S5 - 11

CLASSES OF LARGE DEFORMATION PROBLEMS

Large displacement, small strain
Changes in the stress-strain law are neglected
Consider the material matrix for a thin, isotropic beam:
EA

EI xx

EI yy
GI xx
GI yy

Any changes in the geometric properties of the beam elements are ignored
For instance, the area of a beam being extended will not change

Large displacement, large strain

Changes in the stress-strain law are included
For instance, the changing area of a beam element is now included in the evaluation of
the material coefficients
Thus, large strain capability needs sophisticated material models as well
In addition, element formulations need to be:
insensitive to (strong) distortion
able to represent Non-dilatational (incompressible) deformation modes (locking)

MAR101, Section 5, November 2015

S5 - 12

THREE OPTIONS FOR LARGE STRAIN ANALYSIS

Automatic (default)
Recommended option - Marc chooses the best option based upon the
material model and the element type

Total Lagrange
Updated Lagrange

MAR101, Section 5, November 2015

S5 - 13

LARGE STRAIN OPTIONS

Total Lagrange
For Elasticity
Total Lagrange framework
Use Herrmann elements only

Updated Lagrange
For Elasticity
Only available for the Mooney-Rivlin, Ogden, Arruda-Boyce, Gent,
Bergstrom-Boyce, Marlow, and Foam models (including rubber damage)
Both Herrmann and displacement elements can be used
Not yet available for plane stress, membrane, and shell elements
A mixed formulation is used. The deformation is taken into account by
means of the left Cauchy-Green deformation tensor.

MAR101, Section 5, November 2015

S5 - 14

UPDATED LAGRANGE LARGE STRAIN OPTIONS

Additive Decomposition

For Plasticity
Updated Lagrange framework
A hypoelastic, rate-based formulation
Strain rate is decomposed into a sum of elastic and plastic terms (d = de + dp).
The Jaumann rate of Cauchy stress is used with this formulation.
This formulation does not accurately model the material response if the elastic strains
also become very large.

Multiplicative Decomposition
For Plasticity
Updated Lagrange framework
Hyperelastic based formulation with a multiplicative decomposition of the deformation
gradient (F = Fe F Fp).
This gives an accurate treatment of elastic deformations - important for spring-back.
Allows larger increments of strain to be used, with greater accuracy and better
convergence.
Only isotropic materials are supported at present.

MAR101, Section 5, November 2015

S5 - 15

MARC FORMS FOR LARGE DEFORMATIONS/LARGE STRAIN

All settings for Large Deformation/Large Strain are defined in the
Jobs tab

MAR101, Section 5, November 2015

S5 - 16

GEOMETRIC NONLINEARITY GUIDELINES

Clearly there are many options that are available to the user!
Which are appropriate?
Defaults are usually the best choice but user may override
see following page for recommended decision making process

Geometric nonlinearity often implies material nonlinearity

The Large Strain aspect of geometric nonlinearity implies that the use of
material models more sophisticated than Hookes Law are necessary
Advanced material models to be covered in later sections

When in doubt Large Deformation/Large Strain is the most

general option
Must supply appropriate material models
Longer run times due to higher nonlinearity level

MAR101, Section 5, November 2015

S5 - 17

GEOMETRIC NONLINEARITY GUIDELINES

Small/
large
strain
?

Small

Large

Elastic/
Plastic?

Elastic

Other
material
models

Plastic

Isotropic
?

Mooney
-Rivlin
or
Ogden?

Mooney/
Ogden

Updated
Lagrange
[2]

Proceed
with
solution

Total
Lagrange
[3]
Updated Lagrange
Additive
Decomposition
[1]

Non
Isotropic

Isotropic
Updated Lagrange
Multiplicative
Decomposition
Large
Deformation/Small
Strain Marc
Default

[1] This formulation does not accurately model the material response if the elastic
strains also become very large
[2] Not yet available for plane stress, membrane and shell elements
[3] Herrmann elements only
MAR101, Section 5, November 2015
Copyright 2015 MSC.Software Corporation

S5 - 18

GEOMETRIC NONLINEARITY: GUIDELINES

For any given geometric
nonlinear simulation:

P k (u ) u
The stiffness k is dependent on
the displacements u.
P can be calculated in terms of
u, but a direct solution for u in
terms of P is not possible.

u P k (u )
An iterative solution method
must be employed.
The default iterative method in
Marc is the Newton-Raphson
method. Iterative methods to be
discussed in future sections
MAR101, Section 5, November 2015
Copyright 2015 MSC.Software Corporation

Target: To ensure the displacement

increments do not result in errors in
the estimate of P beyond tolerance.

S5 - 19

GEOMETRIC NONLINEARITY: FOLLOWER FORCES

Used with Large Displacement/Large Strain option
Distributed loads (fluxes) are then based on the current geometry
The follower force contributes to a stiffness effect on the tangential
stiffness matrix
This effect can optionally be taken into account.
It affects the rate of convergence but not the final
result.
It may result in a asymmetric stiffness matrix, in
which case choose either the asymmetric or
symmetric solver. In the latter case, the matrix is
made symmetric.

The options for this command are:

No Follower Force: Default
Follower Force: No associated stiffness
Follower Force/Stiffness: Includes follower force
stiffness
Follower Force (Begin Inc): Uses displacements at
the beginning of the increment, as opposed to the
last iteration
MAR101, Section 5, November 2015
Copyright 2015 MSC.Software Corporation

S5 - 20

SHEAR LOCKING
Fundamental characteristics of bending for one element

Linear variation of axial strain, exx, through the thickness (y direction)

No strain in the thickness direction, eyy (if we take Poissons ratio as zero)
No membrane shear strain

MAR101, Section 5, November 2015

S5 - 21

SHEAR LOCKING
Fully integrated 1st order 4 node quad elements in bending:
Shear Locking
The axial strain can be viewed as the
change in length of the horizontal lines
through the Integration points. The
thickness strain is the change in length of
the vertical lines, and the shear strain is the
change in the angle between the horizontal
and vertical lines

The element detects shear strains that are physically non-existent

but are present solely because of the numerical formulation used
To control this problem there are special element types called Assumed
Strain (AS) and Reduced Integration (RI) elements
RI, AS, and quadratic elements will be discussed later

MAR101, Section 5, November 2015

S5 - 22

SHEAR LOCKING
The element cannot bend without shear
The negative consequencesignificant effort (strain energy) goes
into shearing the element rather than bending it
Leads to overly stiff behavior
Reduced integration elements will correct shear locking on a
problem like this:
P

Standard Orthogonal elements

However a new problem develops called, Hourglassing

(discussed later) which can be controlled by using assumed
strain elements
Thus shear-locking is controlled with RI elements, but RI
elements introduce hourglassing, which in turn is controlled by
AS elements.

MAR101, Section 5, November 2015

S5 - 23

ASSUMED STRAIN OPTION

Dont use first order fully integrated elements in regions
dominated by bending. Instead, try assumed strain elements.
Linear shape-function elements
Quad4 (2D Solid, 2D Shell)
Hex8

Assumed strain elements will produce inaccurate results on

skewed meshes.
The further away from 90 the edge angle, the worse the results
Both parallelogram or trapezoidal -shape are bad (see the figures below)

Skewed trapezoidal elements

Skewed Parallelogram Elements

MAR101, Section 5, November 2015

S5 - 24

ASSUMED STRAIN OPTION

The assumed strain includes additional shape functions that eliminate
the shear locking and provide improved accuracy in bending-dominated
problems
They are a compromise in cost between the first and second order
reduced integration elements. They have many of the advantages of
both.
They can model bending with only one element through the thickness.
They have no hourglass modes and can be used confidently with plasticity
and contact; however, they are sensitive to distortion and cannot capture
bending behavior well in meshes with alternating trapezoidal shaped
elements.

MAR101, Section 5, November 2015

S5 - 25

COMPARISON OF ASSUMED STRAIN AND STANDARD

ELEMENTS
Cantilever beam example
using conventional and
assumed strain fields
Result table shows the
normalized tip deflection
(FE:Theory)
Note the effectiveness of
the 3x1 and 6x1 mesh
when assumed strain is
invoked

Mesh

Quad4 (AS)

Quad4 (Standard)

3x1

0.932

0.025

6x1

0.952

0.093

12 x 2

0.983

0.291

24 x 4

0.994

0.621

48 x 8

0.998

0.868

96 x 16

1.004

0.963

a
b
Section
A-A
MAR101, Section 5, November 2015
Copyright 2015 MSC.Software Corporation

S5 - 26

ASSUMED STRAIN ELEMENTS

Advantages:
These are the most cost effective continuum elements for bending
dominated problems
Can model bending with only one element through the thickness
They have no hourglass modes
They can be used confidently with plasticity and contact

Disadvantages:
The substantially improved accuracy of the solution is at the expense of a
slightly increased computational costs during the stiffness assembly
They are sensitive to distortion and cannot capture bending behavior well in
distorted meshes

The most benefit of the assumed strain procedure is obtained for

coarse meshes. The effect decreases with mesh refinement
If a fine mesh is used, then the assumed strain option could be
left off (and providing a slightly faster analysis)
MAR101, Section 5, November 2015
Copyright 2015 MSC.Software Corporation

S5 - 27

Motivation for New Triangle and Tetrahedral Element

Reduce commonly known drawbacks of standard elements:
Shear locking (lack of flexibility in bending)
Volume locking for (nearly) incompressible material behavior

Shear locking
Iterative Solver
Cantilever beam; n = 0
Mesh

Error in tip displacement

88.5%

66.4%

33.2%

11.1%

An accurate solution requires a (very) fine mesh

MAR101, Section 5, November 2015

S5 - 28

New Element - Strain Smoothing

Cantilever Beam - 2D bending

Cantilever beam

B
C
D

Mesh

Error in tip displacement

(n = 0)
#6

# 155

# 239

88.5%

23.9%

66.4%

33.2%

11.1%

Mesh

Error in tip displacement

(n = 0.49)
#6

# 155

# 239

84.7%

81.7%

9.7%

4.0%

90.0%

54.8%

3.8%

33.2%

0.4%

72.6%

24.6%

1.4%

11.1%

0.04%

42.2%

8.1%

0.4%

MAR101, Section 5, November 2015

S5 - 29

New Element - Plasticity - Necking

Element Technology All rubber, plastics, metals (plasticity)
Triangular and Tetrahedral elements
No volume locking
Well-suited for global remeshing and rezoning

MAR101, Section 5, November 2015

S5 - 30

Element 241 Additive Plasticity

MAR101, Section 5, November 2015

S5 - 31

EXERCISE
Workshop 3C Nonlinear Analysis of a Cantilever Beam
Be sure to ask for help if there is anything you do not understand

MAR101, Section 5, November 2015

S5 - 32

SECTION 6
MATERIAL NONLINEARITY

MAR101, Section 6, November 2015

S6 - 1

OVERVIEW
Material types available in Marc
Material linearity and material nonlinearity
True stress and true strain versus engineering or nominal stress
and strain
Elastic-plastic material models
Hardening Laws

Hyper-elasticity material models

Laminated composite shells

MAR101, Section 6, November 2015

S6 - 2

SUMMARY OF MARC MATERIALS

Marc supports an extensive library of material types including:
Elastic-Plastic
Isotropic
Orthotropic
Anisotropic

Rigid-Plastic
Hypoelastic
Mooney
Ogden
Foam
Arruda-Boyce
Gent
Marlow
Bergstrom-Boyce
Parallel Rheological
Shape Memory
User defined

MAR101, Section 6, November 2015

These are the most commonly used

materials in Marc and will be
considered in detail in this course
Information on other materials can be
found in the on-line documentation as
well as the MARC102 Advanced
Marc Course

S6 - 3

MARC MATERIALS ELASTIC-PLASTIC

Capabilities supported:
Elastic
Linear relation between stress/strain

Plasticity
Nonlinear behavior past a specified yield point

Thermal Expansion
Strains due to temperature change

Viscoelasticity
Time dependent linear relation between stress/strain

Viscoplasticity
Time dependent plasticity

Creep
Time/Rate/Stress dependent inelastic behavior

Damage Effects
Methods to calculate cumulative damage

Cure Shrinkage
Volumetric shrinkage due to a curing process
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 4

MARC MATERIALS ELASTIC-PLASTIC

Capabilities supported continued
Damping
Material dependent damping

Forming Limit
Failure estimation of forming processes

Grain Size
Alternate hardening model

MAR101, Section 6, November 2015

S6 - 5

MATERIAL PROPERTIES IN MARC MENTAT

All of the constitutive models of a given material are entered under

the same material set name.
Example:
Plastic
Creep
Damping

MAR101, Section 6, November 2015

S6 - 6

ELASTICITY
An ideally elastic material
has the following properties:
A unique, natural, elastic
reference state to which it will
return when the deformation
causing forces are removed.
The deformation between this
elastic reference state and the
current state is reversible.
As a result, loads are said to be
fully recoverable

There is a one-to-one
relationship between stress and
strain

MAR101, Section 6, November 2015

Linear-Elastic Material
Representation

S6 - 7

DEFINING ELASTIC MATERIALS IN MENTAT

The Marc Mentat Materials form allows the modeling of:

Metals
Plastics
Rubbers
Composites
Foams

MAR101, Section 6, November 2015

S6 - 8

PLASTICITY
As stresses (and strains) are increased beyond the yield point,
metals start to exhibit a nonlinear behavior: Plasticity
The yield point defines the shift from elastic to plastic behavior
for a material

MAR101, Section 6, November 2015

S6 - 9

PLASTICITY
Loading beyond the yield stress induces permanent (plastic)
deformation
A yielded ductile metal will unload along a curve that is parallel to the initial
linear elastic curve
For metals, the yield stress usually occurs at .05% - .1% of the materials
Elastic Modulus

MAR101, Section 6, November 2015

S6 - 10

ELASTIC PERFECTLY PLASTIC MATERIAL

Simplest form of elastic-plastic behavior is elastic perfectly plastic
behavior
Assumes the stresses above yield are constant, as shown in
diagram below
Perfectly plastic is good for modeling first order material plasticity
The yield value may be a function of temperature

MAR101, Section 6, November 2015

S6 - 11

The plastic portion

requires selecting a
Hardening Law.

True stress

Yield Point

True Strain e

Plastic data should be

entered in terms of true
stresses and true plastic
strains rather than in
terms of engineering
stresses and engineering
strains. (Discussed in
detail later)

True stress

Marc expects data for the

elastic and plastic range to
be entered separately.

GENERAL PLASTICITY

True Plastic Strain

MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 12

DEFINING PLASTICITY OF A MATERIAL IN MENTAT

Note the empty
field here for
elastic-perfectly
plastic option

MAR101, Section 6, November 2015

S6 - 13

MEASURES OF STRESS AND STRAIN

As noted previously Marc expects data in the form of true
stresses vs. true plastic strains
It is imperative the analyst recognize that there is a difference
between stress and strain measures
Specifically the difference between engineering strain/stress and
true strain/stress is highly important
The differences are best observed by reviewing the phenomena of
Necking - describe on the following page

MAR101, Section 6, November 2015

S6 - 14

NECKING
At high strains, a metal may
experience highly localized
extension and thinning, usually
called necking.
The nominal stress of a metal as
it is necking is much lower than
its ultimate strength.
This behavior is due to the
following factors:
The geometry of the specimen
The nature of the test itself (that is
tension or compression)
The stress and strain measure used
(that is nominal stress and strain)

This behavior is best represented

by Isotropic Hardening

MAR101, Section 6, November 2015

S6 - 15

ENGINEERING STRESS AND STRAIN

Engineering stress and strain are computed using the un-deformed
shape dimensions as reference values
Engineering (nominal) stress:

nom

Where: nom is the engineering stress

F is the force
A0 is the undeformed or initial area

Engineering (nominal) strain:

nom (l l0 ) l0
Where:
l0 is the initial (undeformed) length
l is the length at the time the strain is
measured

Typically, laboratory measurements are

expressed in engineering stresses and
strains
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 16

TRUE STRESS AND TRUE (LOG) STRAIN

True stress and true strain are
defined as follows:
True stress:

A is the instantaneous area

True strain:

lo is the initial length

l is the instantaneous length

Plastic laws must be entered as

true stress and true strain
values in the table forms.
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 17

Marc uses true stress and

true strain to account for
changes in area during finite
deformations, which results
in a more accurate
mathematical model.

PLASTIC STRESS AND STRAIN DATA

Plasticity definition in Marc is defined as the post-yield, or plastic
portion of the stress-strain curve
Typical engineering data for stress-strain curves are defined as
total nominal strain

true nom (1 nom )

true ln(1 nom )
The first point of a plastic strain definition represents the yield
point, corresponding to a plastic strain value of zero. Thus, when
creating a material property table in Mentat, subtract the elastic
strain at each point from the total true strain values:

total elastic plastic

Thus:

or
alternatively
stated

plastic total true E

MAR101, Section 6, November 2015

S6 - 18

plastic total elastic

CONVERTING ENGINEERING MATERIAL DATA

Example of converting material test data to valid Marc inputs
Nominal
Stress
snom

Nominal
Strain
enom

True
Stress
snom(1+enom)

Total True
Strain
ln(1+enom)

Plastic
Strain
etot-strue/E

200E6

0.00095

200.2E6

0.00095

0.0000

240E6

0.0250

246.0E6

0.0247

0.0235

280E6

0.0500

294.0E6

0.0488

0.0474

340E6

0.1000

374.0E6

0.0953

0.0935

380E6

0.1500

437.0E6

0.1398

0.1377

400E6

0.2000

480.0E6

0.1823

0.1800

Extrapolated

Linear interpolation is used between data points

The materials response is extrapolated outside the specified
range
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 19

PLASTIC STRESS AND STRAIN DATA

Defining plasticity data within Mentat
Make all appropriate stress/strain
conversions
Create a table within Mentat that defines
true plastic strain vs. true stress

MAR101, Section 6, November 2015

S6 - 20

GENERAL ELASTIC-PLASTIC MATERIAL

In addition to yield stress and plastic strain vs. true stress curve
the user must specify:
Yield Criterion

Yield Criterion

Default Von Mises

Hardening Laws
Default - Isotropic

Strain Rate Method

Default - No strain rate
dependence

Hardening Rules

Strain Rate Method

MAR101, Section 6, November 2015

S6 - 21

HARDENING LAWS
Hardening laws define how the yield point changes after
initial yielding AND a reversal of loading
Ideally Plastic:
F ( ij ) y 0

Isotropic Hardening:
p

F ( ij ) y ( ) 0
Kinematic Hardening:

F ( ij y ) 0
Combined:
Starts as Isotropic and continues as
Kinematic
p

F ( ij ij ) y ( ) 0
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 22

ISOTROPIC HARDENING
Isotropic Hardening is good for modeling plasticity, where
material flow is the predominant effect that is being captured.
No shift in stresses - cannot be used to model hysteresis.
Commonly used to model drawing or other metal forming
operations.

Hardening Rule on the

Plasticity Properties Form
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 23

VON MISES YIELD AND ISOTROPIC HARDENING

For an isotropic material

[(1 2 )2 (2 3 )2 (3 1)2 ]1/ 2 / 2

where 1 , 2 and 3 are the principal

Cauchy stresses. It can also be
expressed in terms of non-principal
Cauchy stresses as follows:

[( x y ) 2 ( y z ) 2 ( z x ) 2 6 ( xy2 yz2 zx2 )]1 / 2 ) / 2

The yield condition can also be

expressed in terms of the deviatoric
stresses as:

Where

3 ' '
ij ij
2

is the deviatoric Cauchy

stress expressed as
ij

'
ij

MAR101, Section 6, November 2015

1
3

S6 - 24

VON MISES YIELD AND ISOTROPIC HARDENING

Isotropic Work Hardening assumes that the center of

the yield surface, Figure a, remains stationary in the
stress space, but that the size (radius) of the yield
surface expands due to work hardening.

Figure b depicts the load path of a uniaxial test with

loading and unloading of a specimen

It is first loaded from stress free, point 0, to initial yield at point 1

it is then continuously loaded to point 2
Then unloading from 2 to 3 following the elastic slope E (Youngs
modulus)
Elastic reloading from 3 to 2 takes place
The specimen is plastically loaded again from 2 to 4.
Elastically unloaded from 4 to 5
Reverse plastic loading occurs between 5 and 6

Figure a: Von Mises Yield Surface

Other comments for Figure b

Stress at 1 is the yield stress and stresses at points 2 and 4 are

larger than yield due to work hardening
During unloading, the stress state can remain elastic as in point 3,
or it can reach a subsequent (reversed) yield point (for example,
point 5)
The isotropic work hardening rule states that the reverse yield
occurs at current stress level in the reversed direction. Take the
stress level at point 4; then the reverse yield takes place at point
5 (negative of the stress at Point 4)

Figure b: Loading Path

MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 25

KINEMATIC HARDENING
Kinematic Hardening is good for simulating loading and unloading
effects where the compression yield is less than the tension yield
due to hardening, as depicted in the figure below.
The plastic deformation of a material will often increase its yield
stress for subsequent loadings.

MAR101, Section 6, November 2015

S6 - 26

KINEMATIC HARDENING

For many materials, the Kinematic Hardening model gives a better

representation of loading/unloading behavior than the isotropic
hardening model. For cyclic loading, however, the kinematic
hardening model cannot represent either cyclic hardening or
cyclic softening.

Kinematic Hardening is not applicable for metal forming

simulations in which there is significant plastic flow and straining
is generally monotonic.

MAR101, Section 6, November 2015

S6 - 27

VON MISES YIELD AND KINEMATIC HARDENING

Figure a illustrates the Kinematic Hardening rule. The von

Mises yield surface does not change in size or shape, but
the center of the yield surface can move in stress space.
The loading path of a uniaxial test is shown in Figure b.
The specimen is loaded in the following order

From stress free, point 0, to initial yield at 1

It is then continuously loaded to 2

Unloading is from 2 to 3

Elastic reloading takes place from 3 to 2

Plastic reloading takes place from 2 to 4

Elastic unloading occurs from 4 to 5

Reverse plastic occurs from 5 to 6.

Figure a: Von Mises

Yield Surface

Other comments for Figure b

As in isotropic hardening, stress at 1 is equal to the initial yield

stress, and stresses at 2 and 4 are higher than , due to work
hardening.

Point 3 is elastic, and reverse yield takes place at point 5

Under the kinematic hardening rule, the reverse yield occurs at the
level of 5 ( 4 2 y ) , rather than at the stress level of 4

Similarly, if the specimen is loaded to a higher stress level, point 7,

and then unloaded to the subsequent yield point 8, the stress at
point 8 is 8 ( 7 2 y )

If the specimen is unloaded from a tensile stress state (such as

point 4 and 7), the reverse yield can occur at a stress state in either
the reverse direction (point 5) or the same direction (point 8).

MAR101, Section 6, November 2015

S6 - 28

Figure b: Loading Path

COMBINED HARDENING
The initial hardening is assumed to be almost entirely isotropic,
but after some plastic straining, the elastic range attains an
essentially constant value, that is pure kinematic hardening.

MAR101, Section 6, November 2015

S6 - 29

COMBINED HARDENING
Combined Hardening is good for simulating the shift of the stress-strain
curve apparent in a cyclical loading (hysteresis), either for cyclic
hardening or cyclic softening.
This is called the Bauschinger effect, which in Marc requires using
Combined Hardening.
If there is a shift of stresses with neither hardening or softening (the
maximum stress in each cycle is the same), then the behavior is called
ratchetting.
A test often used to characterize the plastic behavior of metals is the
Plastic Shakedown, essentially by producing symmetric strain cycles.
Soft or annealed metals tend to harden towards a stable limit, and
initially hardened metals tend to soften. These types of things show up in
the Plastic Shakedown test.
Another test is the Relaxation of Mean Stress. It appears on an
asymmetric strain experiment, one in which the specimen is allowed to
strain well into the plastic range before starting the load cycling.
As cycles increases, the mean stress tends to zero
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 30

COMBINED HARDENING
Combined Hardening implies a constant shift of the center of the elastic
domain with a growth of elastic domain around this center until pure
kinematic hardening is attained. In this model, there is a variable
proportion between the isotropic and kinematic contributions that
depends on the extent of plastic deformation - as measured by the mean
plastic strain.
The basic assumption of the combined hardening model is that such
behavior is reasonably approximated by a classical constant kinematic
hardening constraint with the superposition of initial isotropic hardening.
The isotropic hardening rate eventually decays to zero as a function of
the equivalent plastic strain measured by

The work hardening data at small strains governs the isotropic behavior,
and the data at large strains governs the kinematic hardening behavior.
If the last work hardening slope is zero, the behavior is the same as the
isotropic hardening model.
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 31

ADDITIONAL TOPICS
Now we have a good idea of the basics of elastic and elasticplastic material specification, two of the most common material
types encountered.
Lets look at some other material capabilities:

Orthotropic materials
Temperature dependencies
Hyperelasticity rubber materials
Laminate simulation

MAR101, Section 6, November 2015

S6 - 32

RELATIONSHIPS FOR ORTHOTROPIC MATERIAL PROPERTIES

Plane Stress

Solid

To maintain symmetry

nyx = nxy * Ey/Ex

nzx = nxz * Ex/Ez
nzy = nyz * Ez/Ey

To obtain a valid material

nxy < (Ex/Ey)1/2

To obtain a valid material

Plane Strain

nxy < (Ex/Ey)

nxz < (Ex/Ez)
nyz < (Ey/Ez)

To maintain symmetry
Ey * (nxy*Ez + nyz*nxz*Ex) = Ex *
(nxy*Ez + nxz*nyz*Ey)

Shell

Axisymmetric

To maintain symmetry

nyx = nxy * Ey/Ex

nzx = nzx * Ez/Ex
nzy = nyz * Ez/Ey

To obtain a valid material

nxy < (Ex/Ey)

To obtain a valid material

nxy < (Ex/Ey)
nxz < (Ex/Ez)
nyz < (Ey/Ez)
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 33

TEMPERATURE DEPENDENT PHYSICAL QUANTITIES

T is temperature in the expressions below

MAR101, Section 6, November 2015

S6 - 34

TEMPERATURE DEPENDENT PHYSICAL QUANTITIES

Temperature

Plastic Strain

Stress

0.00000E+000

3.00000E+004

6.00000E+001

2.00000E-003

3.200000E+004

7.00000E+001

9.00000E-002

3.70000E+004

8.00000E+001

8.50000E-001

4.40000E+005

Slope-Break Point Data

Function-Variable Point Data

Slope

Break Point

Function

Variable

S1 = (F2 F1)/(T2 T1)

S2 = (F3 F2)/(T3 T2)

S3 = (F4 - F3)/(T4 T3)

MAR101, Section 6, November 2015

S6 - 35

HYPERELASTICITY
Hyperelastic materials, such
as rubber, exhibit highly
nonlinear elastic stress-strain
behavior.
Hyperelastic materials can
remain elastic up to large
strain values. Often up to
100% strain and beyond.
Marc excels in simulating
rubber made components.
It is possible to match
physical behavior as
illustrated by the model made
by an MSC customer shown to
the right.

MAR101, Section 6, November 2015

S6 - 36

Marc

Physical Test

HYPERELASTIC EXAMPLE: MODELING A RUBBER BOOT

The FEA of rubber boots presents large displacements, large strains,
incompressible material behavior, susceptibility to local buckling, and
varying boundary conditions caused by the 3-D contact between the
various parts of the boot.
Proper design should also consider
bellows shape optimization, fatigue
life, maintainability, replaceability,
and cost.
In this example, the boot is clamped
on one side to a rigid surface, and on
the other side to a translating and
rotating axis. Axial compression is
first applied, followed by bending.
Improved fatigue life was the design
goal, and nonlinear FEA was
successfully used to minimize time,
cost, and achieve a design with an
acceptable product life cycle.
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 37

Axial Compression, Bending,

Contact and Buckling in
Rubber Boot

HYPERELASTICITY
Marc makes the following assumptions about hyperelastic
materials:
The interpolation basis of Hermann elements uses independent pressure
stress fields in each element
Constant in first order elements
Linear in second order elements

Pressure as well as displacement interpolated independently

p a0 a1 X a2Y a3Z

p a0 (at center only)

(at corner nodes only)

u at nodes

Hermann Element
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 38

CHARACTERIZING HYPERELASTIC MATERIALS

Hyperelastic materials present a different behavior (measured by
the relation between stresses and strains) when subject to different
type of loadings.
Three types of experimental data may be required:
Simple Tension Test
Planar Tension Test
Biaxial Extension Test

MAR103 Course combines

laboratory tests and Marc
analyses for a thorough
learning experience on
working with elastomers.

MAR101, Section 6, November 2015

S6 - 39

MANY FORMULATIONS AVAILABLE FOR HYPERELASTICITY

Many material models

are available in Marc,
all of which are
supported by Mentat

Hyper-elastic
materials use the
strain energy potential,
W, to relate stresses
to strains

C10, C01, C20, .. are

the coefficients of the
Strain Energy
Function

MAR101, Section 6, November 2015

Multiple volumetric
representations
available

S6 - 40

CHARACTERIZING HYPERELASTIC MATERIALS

Calculation of coefficients based on curve fitting of physical test

data
General guidelines
Its just curve fitting!!
Phenomenological models - not
material law, no Polymer physics
as basis
Dont use too high order fit

Number of data points

Dont use too many (regularize
if necessary)
Add subtract points as needed

Range and scope of data

Check fit outside of range of data
Check fit in all modes of deformation
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 41

CHARACTERIZING HYPERELASTIC MATERIALS

General guidelines continued
Mooney-Rivlin model obtained by fitting tensile data is quite inadequate in
other modes of deformation, especially compression
Using only tensile data is dangerous!

Rubber Models allow for either Constant Bulk Modulus or Series

Representation to capture volumetric behavior
Extrapolations can be dangerous
Always check your models predicted response
Check to see if response is outside of data range

Unstable material model = numerical difficulties in FEA

See MARC103 course notes for details of curve fitting

MAR101, Section 6, November 2015

S6 - 42

CHARACTERIZING HYPERELASTIC MATERIALS

Definitions, Stretch Ratios, Engineering Strain:

Incompressibility

1 2 3 1
From thermodynamics and statistical mechanics, first order
approximation (Neo-Hookean):
W

1
G (12 22 23 3)
2

MAR101, Section 6, November 2015

(Strain energy potential)

S6 - 43

HYPERELASTIC FORMULATIONS
Many hyperelastic formulations are based on the generalized
Mooney-Rivlin polynomial function of strain energy, which leads
to a better agreement with test data for both unfilled as well as
filled rubbers.
It can be written as:
W

Cij ( I1 - 3) ( I2 - 3)
i

i, j 1

i 1

1
( I3 - 1)2 i
Di

where:

Strain energy potential

I3
I1, I2
cij

Elastic volume ratio or third strain invariant

Strain Invariants - measure of distortion in the material

Coefficients related to shear behavior of the material

Di
N

=
=

Compressibility behavior of the material

Order of the polynomial
If we take the
derivative of W with
respect to strain, we
obtain the stress.

MAR101, Section 6, November 2015

S6 - 44

HYPERELASTIC FORMULATIONS
W

Cij ( I1 - 3) ( I2 - 3)
i

i, j 1

i 1

1
( I3 - 1)2 i
Di

The strain invariants present in the generalized Mooney-Rivlin

formula are functions of the stretch ratios as follows:
I 1 12 22 32
I 2 12 22 22 32 32 12
I 3 12 22 32

Different equations result from the generalized formula according

to the order of the polynomial.
In case of perfectly incompressible material, I3 =1, reducing the
strain energy potential to:

MAR101, Section 6, November 2015

S6 - 45

HYPERELASTIC FORMULATIONS
The simplest model of rubber elasticity is the Neo-Hookean
model, represented as:

W C10 ( I1 3)
This first-order model exhibits a constant shear modulus and gives a good
correlation with the experimental data - up to 40% strain in uniaxial
tension and up to 90% strains in simple shear

The earliest phenomenological theory of nonlinear elasticity is

now known as the Mooney-Rivlin model written as:

W C 10 ( I 1 3 ) C 01 ( I 2 3 )
Although it shows good agreement with tensile test data up to 100% strains,
it has been found inadequate in describing the compression mode of
deformation. Moreover, this model fails to account for the stiffening of the
material at large strains.

MAR101, Section 6, November 2015

S6 - 46

HYPERELASTIC FORMULATIONS
The 3-term Mooney-Rivlin model
Sometimes produces a better fit on vulcanized rubber when loading is not
simple tension

W C10 ( I1 3) C01 ( I 2 3)
C11 ( I1 3)(I 2 3)

However, Tschoegls investigations, published in 1971,

underscored the fact that the retention of higher than 1 order
terms in the generalized Mooney-Rivlin polynomial function of
strain energy lead to a better agreement with test data for both
unfilled as well as filled rubbers.
Signorini
W C10 ( I1 3) C01 ( I 2 3)
C20 ( I1 3) 2

MAR101, Section 6, November 2015

Second Order
Invariant

James-GreenSimpson
(Third Order)

W C10 ( I1 3) C01 ( I 2 3)

C11 ( I1 3)( I 2 3)

C20 ( I1 3) 2

C20 ( I1 3) 2 C30 ( I1 3)3

S6 - 47

HYPERELASTIC FORMULATIONS
All these models account for non-constant shear modulus.
However, caution needs to be exercised on inclusion of higher
order terms to fit the data, since this may result in unstable energy
functions yielding nonphysical results outside the range of the
experimental data.
Three Term Mooney-Rivlin:

Second Order Invariant:

W C10 ( I1 3) C01 ( I 2 3)

C11 ( I1 3)( I 2 3)

Signorini:

C11 ( I1 3)( I 2 3) C20 ( I1 3)2

James Green-Simpson (Third Order):

W C10 ( I1 3) C01 ( I 2 3)

W C10 ( I1 3) C01 ( I 2 3)
C11 ( I1 3)( I 2 3)

C20 ( I1 3) 2

MAR101, Section 6, November 2015

C20 ( I1 3)2 C30 ( I1 3)3

S6 - 48

HYPERELASTIC FORMULATIONS
Yeoh Model
This model differs from the other
higher order polynomial models in
that it depends on the first strain
invariant only:
W C10 ( I1 3) C20 ( I1 3)2 C30 ( I1 3)3

It fits various modes of deformation

using the data obtained from a
uniaxial tension test only. This leads
to reduced requirements on material
testing. However, it behaves poorly
for low strains.
W

EI m
log m *
6
I m I1

From Yeoh [1990].

where :
I1* I1 3
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

Yeoh Model: Tensile and

Compressive Data Replotted
Against (I13).

S6 - 49

HYPERELASTIC FORMULATIONS
Arruda-Boyce Model

Ogden Model
Proposed the energy function as
separable functions of principal
stretches:
1

n 3 n a n
an
an
W J (1 2 3 3) 4.5K ( J 3 1) 2
n 1 n
N

Where J is the dilatancy

Is unique since the standard

tensile test data provides accuracy
for multiple modes of deformation
at all strain levels.
Like the Gent model, it accounts
for the physics of network
deformation:
1
1
W nK ( I1 3)
( I12 9)
20 N
2
11
( I13 27)
2
1050 N
19
4
(
I
81)
1
3
7000 N
519
( I15 243)
673750 N 4

MAR101, Section 6, November 2015

S6 - 50

HYPERELASTIC FORMULATIONS
Gent Strain Energy Model
Accounts for the underlying molecular structure of elastomer, simulating the
non-Gaussian behavior of individual chains in the network.
Similar to the Arruda-Boyce model but based upon limiting the chain
extensibility.
The limit of the first invariant is Im which represents locking tendency due to
finite extensibility of polymer chains

MAR101, Section 6, November 2015

S6 - 51

HYPERELASTIC FORMULATIONS
The figure shows how a 3-term Ogden model compares with
Treloars data [Treloar, 1975] in simple tension, pure shear, and
biaxial tension. The Ogden constants in this case were
calculated:

In practice, more than a 3-term Ogden model is rarely used

MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 52

HYPERELASTIC FORMULATIONS
Marlow Model
Simplified model
Enter engineering stress-strain data
May be temperature dependent

MAR101, Section 6, November 2015

S6 - 53

HYPERELASTIC FORMULATIONS
There are several models for materials going through large volumetric
deformations.
Blatz-Kos
Penns
Storakers

Marc has adopted the foam model

for compressible materials with
the following representation:

n n

1 2 n 3 n 3
n 1 n
N

W
N

n 1

1-J n
n

Where , , and are material

constants
The second term represents
volumetric change
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 54

HYPERELASTIC FORMULATIONS
This model [Hill-1978, Storakers-1986] with n = 2 provides good
correspondence with data in uniaxial and equibiaxial tension.
The Blatz-Ko model [Blatz and Ko, 1968] proposed for polymers
and compressible foam-like materials.

MAR101, Section 6, November 2015

S6 - 55

HYPERELASTIC SUMMARY
Neo-Hookean model: Good correlation with the experimental data - up to
40% strain in uniaxial tension and up to 90% strains in simple shear
Mooney-Rivlin model: Although it shows good agreement with tensile test
data up to 100% strains, it has been found inadequate in describing the
compression mode of deformation. Moreover, this model fails to account for
the stiffening of the material at large strains.
Yeoh model: Behaves poorly for low strains
Gent model: Strain energy model accounts for the underlying molecular
structure of elastomer
Arruda-Boyce model: Is unique since the standard tensile test data provides
accuracy for multiple modes of deformation at all strain levels
Ogden Model: In practice, more than a 3-term formulation is rarely used
Blatz-Kos, Penns, and Storakers models: For large volumetric
deformations, Marc has adopted the foam model
Hill Storakers model: Provides good correspondence with data in uniaxial
and equibiaxial tension
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 56

COMPOSITES: LAMINATED SHELL AND SOLID SHELL

Laminated Shell

MAR101, Section 6, November 2015

Stacked Brick or Solid Shell

S6 - 57

LAMINATED COMPOSITE SHELLS

The shells in Marc support
multi-layer construction
The user must define:
A number of layers or lamina
Independent material and
orientation in each layer
Different number of integration
points used for Simpsons
integration through each layer

Classical Lamination Theory

is the most commonly used
material model for composites
One example is to
macroscopically simulate
high-stiffness fibers along
different orientations in a
matrix material
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 58

Marc uses the following strain

distributions for the composite
shells
Membrane strains vary linearly
which results in good membrane
and bending behavior
Trapezoidal rule for integration
through layer thickness

CREATING COMPOSITE SHELLS IN MENTAT

Layer numbers are defined by the element normal
Layer 1 is on the top surface of the element
Layer numbers grow sequentially in the direction of the element normal

Note: these
materials must
be previously
defined

Layer 1
Layer 2
Layer 3

MAR101, Section 6, November 2015

S6 - 59

LAMINATED COMPOSITE SHELLS

Transverse shear strains are assumed constant through the shell
thickness.
This results in the violation of the equilibrium condition that 13 and
should be zero at the shell surface and that stresses between layers
should be continuous.
Therefore, the computed
transverse shear strains
and stresses cannot be
used to determine
interlaminar shear failure.
Optionally, for Element
types 22, 45, 75, 140, 149,
150 and 185, a parabolic
distribution of transverse
shear strains can be
included, allowing
improved inter-laminar
shear stresses.
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 60

LAMINATED COMPOSITE SHELLS

Composite failure is computed on a layer by layer basis
Modify an individual layers failure criteria

MAR101, Section 6, November 2015

S6 - 61

LAMINATED COMPOSITE SHELLS

If the composite structure is thin, anisotropic layered shell elements
provide a viable option.
The rebar element has recently gained popularity due to its
computational economy
Designed originally for concrete reinforced with steel rods and then extended for
cord-rubber composites

MAR101, Section 6, November 2015

S6 - 62

ORIENTATION OF LAYERS
Options for defining orientation:

Local element system for shells and solid shell

Global system for continuum composite elements
Rebar directions
Redefined by orientation option
Redefined by composite ply system

MAR101, Section 6, November 2015

S6 - 63

LAYING A LAMINATE FOR COMPOSITE SHELLS

Laying a Laminate on a surface in Marc Mentat:
Mentat uses a coordinate system (such as Coord 2 in the figure) to create a Marc
UUPLANE, defined as the plane formed by axes X and Y of that system (or any
plane parallel to the original one).
At any point on the surface,
Intersection
there should be a UUPLANE
(Reference
Line)
that is not coplanar with the
surfaces tangent plane at the
Point on
Surface
X
point. Thus, a straight curve
intersection between the
Coord 2
Z
UUPLANE and the tangent
Y
Normal
plane can be found. This defines
a reference line.
Normal
The laminate is laid by rotating
X
the material in X direction at a
given angle about the normal to
Intersection
the surface (material Z) at the
Reference
Angle(reference
point from the reference line. In
line)
this example, if that angle is
constant for all elements, the
laminate will follow a helicoidal
pattern around the cylinder.
MAR101, Section 6, November 2015
Copyright 2015 MSC.Software Corporation

S6 - 64

ORIENTATION ALLIGNED WITH CURVE

Composite solid element
Results agree via curve and
element edge orientation
curve

element
orientation

MAR101, Section 6, November 2015

S6 - 65

ORIENTATION VIA 3D LOCAL

Use vectors between face centroids to define preferred directions
for 3D Hex elements
1st direction (Red)
4-1-5-8 face to 3-2-6-7 face
2nd direction (Green)
1-2-6-5 face to 4-3-7-8 face
3rd direction (Blue)
1st direction x 2nd direction

MAR101, Section 6, November 2015

S6 - 66

COMPARISON OF MODULUS RATIOS FOR COMPOSITES

MODULUS RATIO COMPARISONS FOR RIGID AND FLEXIBLE COMPOSITES

Filamentary
composite
system

Reinforcement
modulus Ec
(Gpa)

Matrix modulus
Er
(Gpa)

Longitudinal ply
modulus, E1
(Gpa)

Transverse ply
modulus, E2
(Gpa)

Modulus ratio,
Ec/Er

Anisotropy
E1/E2

Glass- epoxy

75.0

3.4000

50.0

18.000

22.0

2.8

Graphite-epoxy

250.0

3.4000

200.0

5.200

74.0

38.0

Nylon-rubber

3.5

0.0055

1.1

0.014

640.0

79.0

Rayon-rubber

5.1

0.0055

1.7

0.014

930.0

120.0

Steel-rubber

83.0

0.0140

18.0

0.021

5,900.0

860.0

E1 and E2 are calculated at volume fractions typical of use for the different composites

MAR101, Section 6, November 2015

S6 - 67

EXERCISE
Workshop 4 Plastic Deformation of Cantilever Beam
Workshop 5 Experimental Hyperelastic Analysis of Rubber Seal
Be sure to ask for help if there is anything you do not understand

MAR101, Section 6, November 2015

S6 - 68

MAR101, Section 6, November 2015

S6 - 69

MAR101, Section 6, November 2015

S6 - 70

SECTION 7
CONTACT ANALYSIS

MAR101, Section 7, November 2015

S7 - 1

CONTACT: BASICS
Continuously monitors relative node locations
Evaluates contact normal forces
If required, evaluates frictional forces
Automatically handles transfers of force between contacting
bodies
Automatically accounts for
continuously varying contact
area and enormous relative
sliding
Works together with
geometric and material
nonlinearity
Applicable to static, dynamic,
fluid, electrical, and thermal
analyses
MAR101, Section 7, November 2015
Copyright 2015 MSC.Software Corporation

S7 - 2

GENERAL CONSIDERATIONS IN CONTACT

Nonlinear analysis involving large meshes, various contact
bodies, and inelastic materials remain challenging.
As complexity increases so does solution time. This has a big
impact on the time it takes to obtain a stable nonlinear solution.
Simplify whenever possible
Break the problem into smaller problems. Analyze these and only if
necessary analyze the whole problem
Whenever possible, reduce the problem to 2D

MAR101, Section 7, November 2015

S7 - 3

CONTACT METHODS

MAR101, Section 7, November 2015

S7 - 4

CONTACT METHODS
Mathematically, the contact problem occurs as a constrained
optimization problem where contact conditions occur as
inequalities described as Kuhn-Tucker conditions.
Among the several approaches within the finite element
framework that have been used to model the frictional contact
and impose the non-penetration constraint, the most popular
ones are:

Penalty Methods [Peric and Owen, 1992],

Lagrange Multiplier [Chaudhary and Bathe, 1986],
Augmented Lagrangian [Laursen and Simo, 1993],
Perturbed Lagrangian [Simo, Wriggers, and Taylor, 1985],
Hybrid Methods [Wunderlich, 1981],
Gap Elements, Interface Elements,
Direct application of contact forces
Solver Constraints

MAR101, Section 7, November 2015

S7 - 5

CONTACT METHODS
These methods are derived from three numerical techniques:
Solver constraints

No additional degrees of freedom

Increased bandwidth of system matrix
No special elements required
Used by Marc for node-to-segment contact

Penalty functions

Number of equations remains the same

Increase of bandwidth of system matrix
Choice of penalty function is problem dependent
Used by Marc for segment-to-segment contact

Lagrange multipliers
Increased number of equations
Zero terms on main diagonal of system matrix
Used by Marc for gap elements

MAR101, Section 7, November 2015

S7 - 6

NODE-TO-SEGMENT CONTACT
Marc Solver Constraint Method allows for automated solution of
problems involving contact
The algorithm automatically detects nodes entering contact and
generates the appropriate constraints to ensure no penetration
occurs
It also maintains compatibility of displacements across touching
surfaces
Both, deformable-to-rigid and deformable-to-deformable contact
situations are allowed
The user need only identify bodies which are potential candidates
for contact during the analysis
Self-contact, common in rubber problems, is also permitted
Additional kinematic restraints are not permitted (springs, etc.)
on nodes which contact rigid bodies

MAR101, Section 7, November 2015

S7 - 7

NODE-TO-SEGMENT CONTACT
The nonlinear element stiffness matrix is assembled using:

K BT D B dV
Modification of [B] only leads to geometric nonlinearly (small strain)
Modification of [D] only leads to material nonlinearity (small strain)
Modification of both [B] and [D] provides large strain capability
Contact nonlinearity in Marc is imposed to the global stiffness matrix at
the solution stage through additional constraint equations to ensure that
penetration does not occur.

K aa
K
ba

K ab ua f a

K bb ub f b

Where:
Kaa represents the nodes in contact
Kbb represents the nodes not in contact

These additional constraints cause a change in stiffness, not of the

constituent element but of the overall stiffness
Contact is considered a nonlinear boundary value problem

MAR101, Section 7, November 2015

S7 - 8

SEGMENT-TO-SEGMENT CONTACT
Improved accuracy
Easy to use
No master-slave concept
Supports double-sided shell contact
Improved accuracy

Full model

MAR101, Section 7, November 2015

Node-to-Segment
Small Plate

S7 - 9

Segment-to-Segment
Small plate

SEGMENT-TO-SEGMENT CONTACT
Segment-to-segment contact identifies unique segments that
interact using a penalty method (and augmentation procedure if
requested)
Segments are defined as unique, overlapping, regions of
element edges or faces
A 2D example is shown here, but Marc operates the same for 3D
with elements and faces

First polyline with

3 polyline points

MAR101, Section 7, November 2015

S7 - 10

SEGMENT-TO-SEGMENT CONTACT
Result of first pass (loop
over auxiliary points to
find segments defining
polylines):

MAR101, Section 7, November 2015

S7 - 11

SEGMENT-TO-SEGMENT CONTACT
The equation reflecting the normal contact conditions to be taken
into account is:

G ( u, u) ng n d Tt g t d 0

Where:
G is the contact body boundary being in contact
gn is the gap function
gt is the tangential gap vector
ln is the Lagrange multiplier (contact pressure)
lt is the tangent stress vector

MAR101, Section 7, November 2015

S7 - 12

SEGMENT-TO-SEGMENT CONTACT
Gap function scenarios
Gn > 0 - point is not in contact
Gn = 0 - point is on boundary in contact
Gn < 0 - point is penetrating which is not physically admissible

Contact contribution to the equations of equilibrium

Implemented with an Augmented Lagrangian Method

MAR101, Section 7, November 2015

S7 - 13

SEGMENT-TO-SEGMENT CONTACT, LIMITATIONS

The method currently cannot be used with any of the following:

DDM (domain decomposition method).

Brake Squeal
Anisotropic friction
Wear
Pore pressure
Piezoelectric analysis

MAR101, Section 7, November 2015

S7 - 14

SEGMENT-TO-SEGMENT WITH SHELL SELF CONTACT

MAR101, Section 7, November 2015

S7 - 15

SEGMENT-TO-SEGMENT, SHIP BUMPER WITH SELF

CONTACT

Notice contact
status on both
sides

MAR101, Section 7, November 2015

S7 - 16

INTRODUCTION TO CONTACT BODIES

MAR101, Section 7, November 2015

S7 - 17

TYPES OF CONTACT BODIES

Deformable body
Body is deformable
Stress and temperature distribution

Rigid (Geometric) body

Body is not deformable (rigid)
No stress distribution
Constant temperature

Rigid (Meshed) body allowing heat transfer

Body is not deformable (rigid)
No stress distribution
Temperature distribution

A contact analysis requires at

least one deformable body

MAR101, Section 7, November 2015

S7 - 18

DEFORMABLE BODIES
Each deformable body consists of one or more finite elements
A deformable body does not need to completely correspond with
a physical body
Shared Nodes at the interface

Include all elements in the contact body in a coupled analysis if

heat transfer to the environment is taken into account
Nodes or elements must belong to NO MORE than one
deformable body
MAR101, Section 7, November 2015
Copyright 2015 MSC.Software Corporation

S7 - 19

DEFORMABLE BODIES
Internally, finite element data is transferred into segments and
nodal points defining the boundary of the deformable body
2D model: a segment corresponds to an element edge
3D model: a segment corresponds to an element face

The Marc element type chosen must be a stress element

MAR101, Section 7, November 2015

S7 - 20

RIGID (GEOMETRIC) BODIES

A rigid body is defined by means of a number of geometrical
entities
Discrete description
Straight line, circular arc, spline
Surface of revolution, Bezier surface, ruled surface, 4-point patch, poly-surface

Analytical description
NURBS curve or surface
Cone surface
Sphere surface

Mentat uses the analytical NURBS

description
Bezier Surface

Ruled Surface
MAR101, Section 7, November 2015
Copyright 2015 MSC.Software Corporation

S7 - 21

RIGID (GEOMETRIC) BODIES

Internally, the geometry of a rigid body is stored:
Piecewise linearly - for each discrete entity
Exactly - for analytical entities

Analytical entities (NURBS) are more accurate for curved

geometries, as they can provide a continuously varying slope as
well as continuity of the normal vector along the surface.
The number of subdivisions for analytical entities is used for
searching purposes (it might influence the amount of memory
allocated).
Each rigid body may have a prescribed motion:

Velocity
Position
Force or moment
Scaling

MAR101, Section 7, November 2015

S7 - 22

RIGID (GEOMETRIC) BODIES ALLOWING HEAT TRANSFER

Each rigid body allowing heat transfer consists of one or more
finite elements
The motion of a rigid body allowing heat transfer is defined
similarly to rigid bodies without heat transfer except that defining
a force or moment is not possible
It is not necessary that a rigid body allowing heat transfer
completely corresponds to the physical body
Include all elements if:
Heat transfer to the environment is taken into account
The body has a non-zero velocity

Nodes or elements must belong to NO MORE than one rigid body

allowing heat transfer
Internally, finite element data is transferred into segments and
nodal points, similar to deformable bodies
The Marc element type chosen must be a heat transfer element
MAR101, Section 7, November 2015
Copyright 2015 MSC.Software Corporation

S7 - 23

EXAMPLE: EXTRUSION PROBLEM

50
20o

Before Analysis

R =6
billet

4.75

4
35

25
channel

After Analysis
MAR101, Section 7, November 2015
Copyright 2015 MSC.Software Corporation

S7 - 24

MODEL 1: DEFORMABLE-TO-RIGID CONTACT, COUPLED

ANALYSIS
Rigid (Geometric) Body
No stress distribution
Constant temperature
Geometrical entities
circular arc
straight lines

Marc element 10
(full integration axisymmetric)

MAR101, Section 7, November 2015

S7 - 25

MODEL 2: DEFORMABLE-TO-RIGID CONTACT, COUPLED

ANALYSIS
Rigid (Meshed) Body
No stress distribution
Temperature distribution
Marc element 40
(full integration axisymmetric thermal)

Marc element 10
(full integration axisymmetric)

MAR101, Section 7, November 2015

S7 - 26

MODEL 3: DEFORMABLE-TO-DEFORMABLE CONTACT,

COUPLED ANALYSIS
Deformable (Meshed) Body
Stress distribution
Temperature distribution
Marc element 10
(full integration axisymmetric)

Marc element 10
(full integration axisymmetric)

MAR101, Section 7, November 2015

S7 - 27

EXERCISES
Workshop 6 Hertz Contact
Be sure to ask for help if there is anything you do not understand

MAR101, Section 7, November 2015

S7 - 28

CONTACT DETECTION IN A STATIC

ANALYSIS

MAR101, Section 7, November 2015

S7 - 29

POSSIBLE CONTACT SITUATIONS

Contacting
(touching) body

Distance tolerance
1
2

Contacted (touched)
body

3
4

1) Node outside element, outside distance tolerance

2) Node outside element, inside distance tolerance
3) Node inside element, inside distance tolerance
4) Node inside element, outside distance tolerance

MAR101, Section 7, November 2015

S7 - 30

POSSIBLE CONTACT SITUATIONS

1) Node outside element, outside distance tolerance
Bodies are not in contact
Contacting node remains in current position

MAR101, Section 7, November 2015

S7 - 31

POSSIBLE CONTACT SITUATIONS

2) Node outside element, inside distance tolerance
Contacting node is projected onto segment of contacted body
According to internal equilibrium (mass preservation)
Remains in contact if necessary force is less than separation force

MAR101, Section 7, November 2015

S7 - 32

POSSIBLE CONTACT SITUATIONS

3) Node inside element, inside distance tolerance
Contacting node is pushed back onto segment of contacted body
According to internal equilibrium

MAR101, Section 7, November 2015

S7 - 33

POSSIBLE CONTACT SITUATIONS

4) Node inside element, outside distance tolerance
Node penetrated
Increment will be recycled with modified time step
If this situation occurs at the beginning of analysis, contact will not be found

MAR101, Section 7, November 2015

S7 - 34

DISTANCE TOLERANCE
The size of the contact tolerance has a significant impact on the
computational costs and the accuracy of the solution.
Contact tolerance too small:
Detection of contact is difficult, leading to higher costs.
More nodes are likely to be considered penetrating leading to increase in
increment splitting, therefore, increasing the computational costs.

Contact tolerance too large:

Nodes are considered in contact prematurely, resulting in a loss of accuracy.
A large amount nodes might penetrate
the surface.

MAR101, Section 7, November 2015

S7 - 35

DISTANCE TOLERANCE
Measured normal to the contacted body
May be user-defined
By default, this tolerance is evaluated as:
1/20x smallest element edge for continuum elements
1/4x smallest thickness for beam and shell elements

MAR101, Section 7, November 2015

S7 - 36

BIAS FACTOR
Without bias factor, the contact tolerance is equally applied to
both sides of a segment.
Can be changed by introducing bias factor 0 < B < 1 (default: 0.95)
disttol

disttol (1-B)

disttol

disttol (1+B)

Improves accuracy, since the distance below which a node comes

into contact is reduced
Reduces increment splitting since the distance to cause
penetration is increased.
The recommended value is B = 0.95 for most contact analyses.
For analyses involving frictional contact, a bias into a contact
body is recommended (0.95 - 0.99)

MAR101, Section 7, November 2015

S7 - 37

DISTANCE TOLERANCE
Recommended usage is to leave the tolerance blank and let Marc
evaluate this
If necessary, specify a tolerance in the contact table for a
specific contact pair

MAR101, Section 7, November 2015

S7 - 38

CONTACT DETECTION AND SPEED

Fast and effective linear element loop (i.e. not quadratic) used to
determine the boundaries of deformable contact bodies
Fast boxing algorithm using multi-level boxing
A super box is created for every 200 segments
If a contacting node is outside the super box, all the segments
inside the super box can be skipped

Box

Segment

Box per segment

MAR101, Section 7, November 2015
Copyright 2015 MSC.Software Corporation

S7 - 39

Super box

CONTACT DETECTION FOR SHELLS

Shell Elements Check for Contact
Checks for contact be made for on:

Both, the top and bottom surfaces (default)

Only the top surface
Only the bottom surface
Resolve contact at the center of shell (ignore thickness)

MAR101, Section 7, November 2015

S7 - 40

BEAM-TO-BEAM CONTACT SEGMENT-TO-SEGMENT

Can be used for both closed and open cross-sections
Geometric nature of beams exactly represented
Contact occurs between segments

MAR101, Section 7, November 2015

S7 - 41

BEAM -TO-BEAM CONTACT

Mentat Beam Contact Control Form

MAR101, Section 7, November 2015

S7 - 42

BEAM-TO-SHELL CONTACT
Beam-to-Shell contact
includes:

Shell offsets
Beam offsets
Beam cross-section
Beam orientation

Patent Pending

MAR101, Section 7, November 2015

S7 - 43

PIPE-IN-PIPE AND EXTERIOR CONTACT

Patent Pending

MAR101, Section 7, November 2015

S7 - 44

SHELL EDGE-TO-EDGE CONTACT

MAR101, Section 7, November 2015

S7 - 45

CONTACT AND ELEMENT TYPES

Linear Elements
All linear elements are available for use with contact (continuum, beams,
shells)
No special treatment is required
In many manufacturing and rubber analyses, lower-order elements behave
better than their higher-order counterparts because of their ability to handle
large distortions as well as large strain materially nonlinear analyses.

Quadratic Elements
Quadratic elements are fully supported
Quadratic contact takes into account the curved geometry and shape
functions of such elements and takes account of both corner and midside
nodes
They are especially useful in small strain, elastic (or mildly inelastic)
analyses, for example automotive engines

MAR101, Section 7, November 2015

S7 - 46

CONTACT AND ELEMENT TYPES

Example: Pin Connection
A number of pin connections are used to mount a thick polymer insulation
layer on a perforated steel plate

MAR101, Section 7, November 2015

S7 - 47

CONTACT AND ELEMENT TYPES

Quadratic Elements
Higher-order isoparametric elements use shape functions which, when the
elements are loaded by a uniform pressure, lead to equivalent nodal loads
that oscillate between the corner and midside nodes.
Two procedures have been implemented to eliminate this problem:
Linearized Contact
In this case, the midside nodes on the exterior surface are automatically tied to the corner
nodes
This effectively results in a linear variation of both, the geometry and the displacement on
the exterior element edges
All elements in the interior of the body behave in the conventional higher order manner
The constraints on the exterior can cause the behavior of the complete structure to be too
stiff; whilst in the area of contact, the stress distribution might be irregular

True Quadratic Contact (genuine)

In this case, no special constraints introduced on the exterior surface other than that
coming from contact
Both the midside and the corner nodes may come into contact and when contact is
established with another deformable body consisting of quadratic elements, a constraint
equation corresponding to the complete quadratic shape function is automatically
incorporated.
Since the above mentioned oscillating nodal loads cannot be used for separation, the
decision whether or not a node should separate is based on the contact normal stress
rather than the contact normal force
MAR101, Section 7, November 2015
Copyright 2015 MSC.Software Corporation

S7 - 48

CONTACT AND ELEMENT TYPES

Node to
Segment
Only

MAR101, Section 7, November 2015

S7 - 49

DEFINITIONS OF CONTACT BODIES

MAR101, Section 7, November 2015

S7 - 50

DEFINING DEFORMABLE BODIES IN MENTAT

Rigid Body
(Curves)

The Contact form

supports Marcs
contact capability.
Contact Bodies are
selected as a set of
elements in the
application region
panel.
Deformable Body
(2D Plain Strain)

MAR101, Section 7, November 2015

S7 - 51

IMPROVED BOUNDARY DESCRIPTION

In the (default) discrete description, the boundary of the
contacted body is described by the finite elements that the body
comprises.
This can cause problems due to the fact that the normals of the
body are not continuous for a curved boundary described with
lower-order elements.
Remedy: deformable bodies may also be defined with an
analytical description

contacting body

actual geometry
4
3

2
1
contacted body
MAR101, Section 7, November 2015
Copyright 2015 MSC.Software Corporation

S7 - 52

finite element
approximation

IMPROVED BOUNDARY DESCRIPTION

Analytical definition of the contact surface improves the accuracy by
representing the geometry with a continuous analytical function
For 2D, local spline curves are defined through the boundary nodes
For 3D, local Coons surfaces are defined through neighboring surface
segments
The nodes of the contacting body now touch these analytical entities
instead of the actual finite elements
These analytical entities are updated as the body is deformed
This is important for concentric shafts or rolling simulation
actual geometry

4
3

contacting body

2
1
contacted body
MAR101, Section 7, November 2015
Copyright 2015 MSC.Software Corporation

S7 - 53

finite element
approximation

IMPROVED BOUNDARY DESCRIPTION

2D Case
Define cubic spline for segment 2-3 based on
tangent at node 2 and node 3
position of node 2 and node 3

C1 continuity is obtained by defining the tangent at node 2 and node 3

based on the position of nodes 1, 2, 3 and nodes 2, 3, 4, respectively.
The user should indicate
nodes/points where the outer
boundary description is
discontinuous, that is, corners
and edges that are to remain
unsmoothed should be
identified.

actual
geometry
4
3
2
contacted
body

MAR101, Section 7, November 2015

S7 - 54

IMPROVED BOUNDARY DESCRIPTION

3D case
Define Coons surface for contacted segment 1-2-3-4 based on
Position of nodes 1, 2, 3 and 4
Tangent vectors at node 1, 2, 3 and 4
Assume twist vectors to be zero

C1 continuity at the nodes is obtained by defining the normal vectors at the

nodes based on weighted contributions of the adjacent segments
The user should indicate edges where the outer boundary description is
discontinuous

4
1

MAR101, Section 7, November 2015

S7 - 55

IMPROVED BOUNDARY DESCRIPTION

Angular Break

No Angular Break

code will create a new

surface, analytical
representation

one large continuous

change in slope, rounded
corners

MAR101, Section 7, November 2015

S7 - 56

IMPROVED BOUNDARY DESCRIPTION

Select analytic definition
Specify the discontinuities via corner points/nodes (2D)
Specify the discontinuities via edges (3D)
Request a visual check of the analytical surface in Mentat

MAR101, Section 7, November 2015

S7 - 57

DEFINING RIGID (GEOMETRIC) BODIES IN MENTAT

Contact form supports Marcs rigid contact capability.

Contact Bodies are selected as a set of curves or surfaces.
MAR101, Section 7, November 2015
Copyright 2015 MSC.Software Corporation

S7 - 58

CONTROL OF RIGID (GEOMETRIC) BODIES

Rigid bodies can be controlled via:
Velocity
Position
Load

For prescribe translational and/or

rotational control as a function of
time you can use a time table
Prescribe force and/or moment on
rigid body as a function of time
These controls may be used for all
structural rigid bodies

MAR101, Section 7, November 2015

S7 - 59

VELOCITY CONTROLLED RIGID (GEOMETRIC) BODY

The motion of a rigid body at any point A is defined by the
translational velocity at the center of rotation and the angular
velocity (radians/time) about the rotation axis.

v A vC rCA
Where
vA = velocity of point A of rigid body
vC = velocity of centre of rotation C
w = angular velocity of center of rotation C
rCA = radius vector from A to C

Rigid bodies provide rotational freedom when continuum

elements are being used
If a rigid body has a non-zero velocity during increment 0, Marc
tries to find initial contact using at most 1,000 trials

MAR101, Section 7, November 2015

S7 - 60

VELOCITY CONTROLLED RIGID (GEOMETRIC) BODY

MAR101, Section 7, November 2015

S7 - 61

POSITION CONTROLLED RIGID (GEOMETRIC) BODIES

Definition is similar to velocity control
Position and angle are used to define the motion of the center of
rotation of the rigid body
In the current release of Marc, position will be interpreted as
displacement of the center of rotation
Similar to velocity controlled motion, the motion of any point on
the rigid body will be a combination of the motion of the center of
rotation uC and the rotation

u A uC rCA

MAR101, Section 7, November 2015

S7 - 62

POSITION CONTROLLED RIGID (GEOMETRIC) BODIES

MAR101, Section 7, November 2015

S7 - 63

LOAD CONTROLLED RIGID (GEOMETRIC) BODIES

Forces and moments can be applied to rigid bodies
Translational Forces translations are applied to the control node
Moments are applied to the auxiliary node
The control and auxiliary nodes are separate with respect to the structure
The control node also defines the center of rotation
These nodes are used to store the
forces and moments acting on
rigid bodies
To apply forces:
Create a control node at the center of
rotation
Define point load boundary condition
Assign boundary condition to control
node

Control Node
Defines center of
rotation

Auxiliary Node

MAR101, Section 7, November 2015

S7 - 64

applied forces

applied moments

LOAD CONTROLLED RIGID (GEOMETRIC) BODIES

To apply moments:
Create an auxiliary node.
Define point load boundary condition.
Assign boundary condition to auxiliary node.

For easier handling, it should be placed in the

vicinity of the control node used for the forces
MX My, Mz should be specified in the first field of
the point load form - usually indicating Fx Fy, Fz.
Rotations applied to this node are with respect
to the coordinates of the control node
Note: Unloaded degrees of freedom will be
unrestrained and may cause rigid body motion
unless additional fixed boundary conditions are
applied

MAR101, Section 7, November 2015

S7 - 65

LOAD CONTROLLED RIGID (GEOMETRIC) BODIES

Select Load for Body Control on the Contact Body Properties
form
Click Parameters to enter necessary load control information

MAR101, Section 7, November 2015

S7 - 66

ORIENTATION OF RIGID (GEOMETRIC) CONTACT BODIES

Each rigid body has an
inside and an outside
surface
Rigid (Geometric) body
contact can only be
detected with the outside
Tick marks (2D) or
vectors (3D) indicate the
inside of the rigid body

This ensures that the

normals to the rigid surface
point outwards from the
rigid body

Hole
MAR101, Section 7, November 2015
Copyright 2015 MSC.Software Corporation

S7 - 67

ORIENTATION OF RIGID (GEOMETRIC) CONTACT BODIES

Change
orientation
(Flip Curves)

MAR101, Section 7, November 2015

S7 - 68

CONTACT TABLE

MAR101, Section 7, November 2015

S7 - 69

CONTACT INTERACTION
It is simply a way to define the behavior between a pair of contact
bodies
Default is all bodies can potentially touch all other bodies
Mentat divides bodies into types

Meshed (Deformable)
Meshed (Rigid)
Geometric
Geometric (with Nodes)
Symmetry

MAR101, Section 7, November 2015

S7 - 70

CONTACT INTERACTION
Initially

Populated

MAR101, Section 7, November 2015

S7 - 71

CONTACT INTERACTION
To activate
potential contact
between paired
bodies, click on
square between
them

Check the
Active check
box

? means
activated but
no interaction
is defined
MAR101, Section 7, November 2015
Copyright 2015 MSC.Software Corporation

S7 - 72

CONTACT INTERACTION
Contact Interaction Form

Defines behavior between bodies

Displays created interactions
Can create a new interaction
Can use an existing interaction
Can copy an existing interaction

MAR101, Section 7, November 2015

S7 - 73

CONTACT INTERACTION
Contact Interaction Properties Form
Click edit on the Contact Table Entry Properties form to edit the contact
interactions of a contact body pair

MAR101, Section 7, November 2015

S7 - 74

CONTACT INTERACTION SUB MENUS

Contact interaction Parameters

Separation
Friction
Wear
Augmentation

MAR101, Section 7, November 2015

S7 - 75

Cohesive Contact
Provide a user controllable soft contact mechanism
Reduce computational costs and improve accuracy of assembly analyses
Simplified Analysis of Bonded Joints Replace Detail Model with
Simplified Model
Cohesive glue implemented for small sliding segment to segment contact
User specifies either stiffness or contact stress vs. relative displacement
on contact interaction menu.
Infinitely stiff fasteners
50%

50%

Infinitely soft fasteners

33%

Real fasteners
35%

MAR101, Section 7, November 2015

30%

35%

S7 - 76

Cohesive Glue Implementation

Cohesive glue implemented for small sliding segment to segment
contact
User to specify cohesive material property data and refer material
on contact interaction menu (GLUE ).

MAR101, Section 7, November 2015

S7 - 77

Initial and Deformed

MAR101, Section 7, November 2015

S7 - 78

Stress vs Displacement Constant K

User Defined Constant K with cut-off at 0.005

User Defined Constant K

MAR101, Section 7, November 2015

S7 - 79

Separating Parallel Plate Model

MAR101, Section 7, November 2015

S7 - 80

User Defined Normal Stress vs. Displacement

User Defined Stress

MAR101, Section 7, November 2015

Resultant Normal Stress for different Points

S7 - 81

Contact Stress at different Times

MAR101, Section 7, November 2015

S7 - 82

Cohesive Menu

MAR101, Section 7, November 2015

S7 - 83

Cohesive Example

MAR101, Section 7, November 2015

S7 - 84

Soft Contact Penetration will occur

MAR101, Section 7, November 2015

S7 - 85

ENTRY LIST MODE

Easily visualize all of the contact pairs
Easily check pairs to activate contact interaction

MAR101, Section 7, November 2015

S7 - 86

BODY VIEW MODE

Featured body
Identifies all bodies potentially in contact
Graphically pick the body

MAR101, Section 7, November 2015

S7 - 87

ADD/REPLACE ENTRIES
Click the Add/Replace entries button
Opens new form to edit the table

MAR101, Section 7, November 2015

S7 - 88

REDUCE COMPLEXITY VIA BODY VISIBILITY

Select bodies to be visible
Use to control size of lists, matrix, and filter

MAR101, Section 7, November 2015

S7 - 89

EXERCISE
Workshop 7 Interference Fit
Workshop 8 Hertz Contact Analysis with Friction
Be sure to ask for help if there is anything you do not understand

MAR101, Section 7, November 2015

S7 - 90

SPECIAL TOPIC
SYMMETRY USING CONTACT BODIES

MAR101, Section 7, November 2015

S7 - 91

SYMMETRY PLANES
Symmetry planes often
provide an easy way for
imposing symmetry
conditions
May be used instead of more
traditional fixed boundary
conditions
Automatically sets very high
separation forces to disable
separation
Permits only the motion
tangential to the contact
segment
Only for rigid curves or
surfaces
MAR101, Section 7, November 2015
Copyright 2015 MSC.Software Corporation

S7 - 92

SYMMETRY PLANES

Automatic symmetry line extension to avoid exit 2400

MAR101, Section 7, November 2015

S7 - 93

CYCLIC SYMMETRY
If a structure has a periodically varying geometry and loading,
only a sector needs to be modeled.
A special set of constraint equations can be used:

MAR101, Section 7, November 2015

S7 - 94

CYCLIC SYMMETRY
Appropriate constraints on a mesh are automatically provided
A nodal point on the symmetry axis is automatically constrained in the
plane perpendicular to the symmetry axis
For continuum elements only
Input consists of
Direction vector of the axis of symmetry
Point on the axis of symmetry
Sector angle a in degrees

The option can be used together with contact

Any shape of the sector sides is allowed, provided that rotating the sector
360/a times about the symmetry axis results in the complete model
The option can be combined with global re-meshing/rezoning
In a coupled thermo-mechanical analysis, the temperature field is forced to
be cyclic symmetric
The possible rigid body rotation about the symmetry axis can be
automatically suppressed
The mesh does not need to line up on both sides of the segment
MAR101, Section 7, November 2015
Copyright 2015 MSC.Software Corporation

S7 - 95

MAR101, Section 7, November 2015

S7 - 96

MAR101, Section 7, November 2015

S7 - 97

MAR101, Section 7, November 2015

S7 - 98

MAR101, Section 7, November 2015

S7 - 99

MAR101, Section 7, November 2015

S7 - 100

SECTION 8
SETTING UP AND RUNNING THE ANALYSIS,
MULTI-STEPPING, AND RESTARTS

MAR101, Section 8, November 2015

S8 - 1

ANALYSIS OVERVIEW
How to setup and run an analysis

The Job/Analysis form in detail

Multi-step analysis
User subroutines
Mesh adaptivity
Running large jobs
Job monitoring
Re-starting jobs

MAR101, Section 8, November 2015

S8 - 2

JOB SUBMIT
On Job Properties form, click Run
On the Run Job form, click Submit
When the status indicates
Complete, click Open Post File

MAR101, Section 8, November 2015

S8 - 3

JOB SUBMIT: ADVANCED

Full Run Method
Write Input File: the Marc input file is created but no Marc job is launched
Execute 1 (Full Run Method): The check run is done and if there are no
fatal errors, the steps are executed in sequence
Submit 1: writes and executes the file and overwrites the previously edited
input file

MAR101, Section 8, November 2015

S8 - 4

MULTI-STEP ANALYSIS
Many Marc analyses require several steps. Usually, these will be
nonlinear loadcases or steps.
In a series of analysis steps or loadcases, the starting condition
for each step is the ending condition from the previous step
In a step, loads are applied as
total values
Example: Cup Forming problem
Step 1: Closes Blank-holder and
Pressurizes it
Step 2: Moves Punch
Step 3: Releases Punch
Step 4: Releases Holder
Step 5: Releases Die

MAR101, Section 8, November 2015

S8 - 5

MULTI-STEP ANALYSIS
Create the loadcases on the Loadcase Properties form to define the load
history
The loadcases selection order on the Job Properties form specifies the
order in which loads and boundary conditions are applied
Note that you are specifying total load, not incremental load

MAR101, Section 8, November 2015

S8 - 6

MULTI-STEP ANALYSIS
Occasionally, these steps may be punctuated by perturbation
steps
Linear analysis steps are perturbations about a base state
The base state is the ending condition of the last nonlinear step prior to the
linear perturbation
In a linear analysis step, the loads are defined as the magnitudes of the
load perturbations only
If an analysis step follows a linear perturbation step, any perturbation
response is ignored, that is steps 2,3,5,6 do not affect steps 1 and 4.
Example: Preloaded Cantilever Beam

Step 1: Preload P1 (Nonlinear Static)

Step 2: Natural Frequency Extraction
Step 3: Response Spectrum Analysis (Earthquake)
Step 4: Preload P2 > P1 (Nonlinear Static)
Step 5: Natural Frequency Extraction
Step 6: Response Spectrum Analysis (Earthquake)
preload

MAR101, Section 8, November 2015

S8 - 7

Many Marc analysis require several

steps. Usually, there will be several
general analysis steps.
Occasionally, these may be
punctuated by perturbation steps.

True Stress

MULTI-STEP ANALYSIS

In a series of general analysis

steps, the starting condition for
each step is the ending condition
from the previous step.

Log Strain

The most common use of multistepping is for load history control

In a general step, loads are applied
as total values

blank
blankholder
die

Example: Modified Olson Cup Test

Step 1: Close and pressurize blankholder

Step 2: Move punch up
The Modified Olson Cup Test is often used to determine
Step 3: Release punch
the material properties of a metal for the purpose of
Step 4: Release blankholder
stretch forming
Step 5: Release die

MAR101, Section 8, November 2015

S8 - 8

SETTING UP MULTI-STEP JOBS

Why should we run a multi-step
analysis?
Step 1: Insertion (Load factor = 1)
Step 2: Insertion (Load factor = 0)
Might want to change options:

Initial Model
MAR101, Section 8, November 2015
Copyright 2015 MSC.Software Corporation

Loading
S8 - 9

Extraction

MULTI-STAGE FORMING EXAMPLE

Initial
Sheet
First
stage

Stamping of a Metal Can

Each Stage is performed in a separate stage, the
results are reviewed, before the subsequent stage is
performed
MAR101, Section 8, November 2015
Copyright 2015 MSC.Software Corporation

S8 - 10

Second
stage

USER-SUBROUTINES
User-subroutines are a powerful
way to input new capabilities by
the user for specific needs
Use user-subroutines for:

Input file option activation

Material / LBC flag activation
Analysis flag activation
Material models
Work-hardening varying as a function
of temperature
Damage models, etc.
Shape memory alloy material
models

Boundary conditions
Heat flux varying spatially or with
other BCs
Friction varying as a function of
temperature
MAR101, Section 8, November 2015
Copyright 2015 MSC.Software Corporation

S8 - 11

USER-SUBROUTINES

MAR101, Section 8, November 2015

S8 - 12

MESH ADAPTIVITY
Global Adaptive Re-meshing

MAR101, Section 8, November 2015

S8 - 13

MESH ADAPTIVITY
Global Adaptive Re-meshing Continued
Full Mesh Density Control: desire nice mesh at crack tip

MAR101, Section 8, November 2015

S8 - 14

MESH ADAPTIVITY
Global Adaptive Re-meshing Continued
Full Density Control

MAR101, Section 8, November 2015

S8 - 15

MESH ADAPTIVITY
Local Adaptive Re-meshing
Splits the mesh at locations where nodes are in contact

MAR101, Section 8, November 2015

S8 - 16

RUNNING LARGE JOBS

Solver selection for running
large jobs
Use Pardiso Direct Sparse for
solid finite element mesh

MAR101, Section 8, November 2015

S8 - 17

RUNNING LARGE JOBS

Selecting DDM

MAR101, Section 8, November 2015

S8 - 18

JOB MONITORING
Monitor jobs while they are running:

Status
Increment
Singularity Ratio
Convergence Ratio
Exit Number

View Files:
Output file (.out)
Log file (.log )
Status file (.sts)

MAR101, Section 8, November 2015

S8 - 19

JOB MONITORING
View output file (jobname.out)
Shows all convergence controls used
Defaults overridden by controlsnot
usually needed

Shows all the details of the iterations

Messages from the solver
Numerical singularities, zero pivots, and
negative eigenvalues

Useful in pinpointing difficulties and

troubleshooting
locations of highest residuals
locations of excessive deformation
locations of contact changes

MAR101, Section 8, November 2015

S8 - 20

DEBUGGING THE INPUT

Check
for Input
Errors

Load case number

Increment number
Number of recycles (per increment and accumulated)
Number of separations in a contact analysis (per increment and
accumulated)
Number of time step cut backs (per increment and accumulated)
Number of times global re-meshing occurred
Time step
Total time
Wall time
CPU time
Max and Min Displacement

MAR101, Section 8, November 2015

S8 - 22

STATUS FILE

MAR101, Section 8, November 2015

S8 - 23

FILES GENERATED BY MARC

Marc Post file

Marc Restart File

Binary (optionally text) file

named jobname.t16
Contains results from analysis
and also contains all model
information
It can be attached to either
Patran, Marc Preference, or
Marc Mentat for job monitoring
and for post-processing
If it contains results from a
thermal analysis, it can be used
as a temperature-data file for a
subsequent structural analysis to
generate thermal loads.

MAR101, Section 8, November 2015

S8 - 24

Binary file named jobname.t08

Contains the entire Marc process
at the end of the last or
(optionally) selected converged
increment(s).
It can be used to continue the
analysis on a subsequent job as
originally defined or to continue it
with a new set of parameters and
load steps.

Marc Scratch Files

These are binary files (such as
jobname.t23) transferring
information between internal
libraries in the software.
Normally, deleted at the end of
the job. Should be deleted by
user if job crashes.

FILES GENERATED DURING ANALYSIS

Relevant files:
job_id. dat - The Marc input file - a text file generated by Marc which defines
the model (pre-processing analysis data, model setup, and job information)
which is documented in Volume C - Program Input of the Marc
documentation set (an important document to have).
job_id. out - The print file - detailed messages and output written by Marc
during processing the analysis. Often a large file.
job_id. log - The log file - summary messages written by Marc
documenting the progress of the analysis.
job_id. sts - The status file - a small terse quick-look text-editable job
summary file containing job statistics about the time/load increments.
job_id. t16/19 - The results file. This is often a very large file containing the
whole FEM model and the requested output results. A binary file by default
(.t16), it can optionally be written as an ASCII (.t19) file, or both.
job_id. t08 - The restart file. This is a binary file containing an image of
the state of the model at requested increment-counted intervals, or at the
last successfully completed increment in a previous job. It is created only by
request in the Set Restart Parameters form.
MAR101, Section 8, November 2015
Copyright 2015 MSC.Software Corporation

S8 - 25

RESTARTING A JOB
Why would we restart a job?
Unforeseen power or computer failure
Job goes well then fails to converge nice to reuse converged increments
Do not know when instability might hit and will have to switch to arc length
method
Use Restart to modify loading, procedure, or output controls of a step.
Useful to continue the analysis from some intermediate point, either to run an
analysis alternative or to correct the analysis.
Useful to compute additional eigenvalues in a Natural Frequency analysis.

MAR101, Section 8, November 2015

S8 - 26

RESTART FORM IN MENTAT

Restart Data Mode
Write: use if this is the first job for which you
are creating a restart file
Very useful for large jobs due to the excessive
disk storage requirements
Useful for examining the results before continuing
the analysis
Creates a restart file named jobname.t08

MAR101, Section 8, November 2015

S8 - 27

RESTART FORM IN MENTAT

Restart Data Mode
Read: use if you want to continue the
analysis from a previous job, in
which case:
The restart job name has to be given.
This is the name of the previous job
(the restart file you will read)
The increment number at which the
analysis should be restarted. If
omitted, the analysis will be restarted
at the last available step in the restart
file

MAR101, Section 8, November 2015

S8 - 28

RESTART FORM IN MENTAT

Restart Data Mode
Read and Write: use if you want
to continue the analysis from a
previous job and you also want to
create a restart file for the current
job.
Combination of Write and Read
types

MAR101, Section 8, November 2015

S8 - 29

RESTART FORM IN MENTAT

Write Restart Data
Single Increment File: the restart
data is written to a single file named
jobname.t08
Last converged only option: only the
last converged increment is saved
Last Converged & Periodic option:
Will allow a subsequent restart from
any saved increment, and not
necessarily from the last converged
increment. The value for the
parameter Increment Frequency
sets the interval between increments
for saving results for Restart.

Multiple Increment File: the restart

data is written to multiple files name
jobname.i_n.t08
n is the increment number
The number of increments between
writing the restart data is entered
using increment frequency
MAR101, Section 8, November 2015
Copyright 2015 MSC.Software Corporation

S8 - 30

RESTART FORM IN MENTAT

Completion of Unfinished Loadcase
Unmodified
Modified
Immediate

MAR101, Section 8, November 2015

S8 - 31

EXERCISE
Workshop 9 Buckling Analysis
Be sure to ask for help if there is anything you do not understand

MAR101, Section 8, November 2015

S8 - 32

SECTION 9
NUMERICAL ANALYSIS OF NONLINEAR
PROBLEMS

MAR101, Section 9, November 2015

S9 - 1

OVERVIEW
Part I: Numerical analysis of nonlinear behavior
Nonlinear analysis: the approach
Iterative solution methods (Newton Raphson)
Analysis convergence

Part II: Load controls and GUI controls

Load increment and automatic time stepping

Graphical user interface controls
Geometrical nonlinear framework
Cost of nonlinear analysis

MAR101, Section 9, November 2015

S9 - 2

PART I: NUMERICAL ANALYSIS OF NONLINEAR

BEHAVIOR
Nonlinear Analysis: The Approach
Iterative Solution Methods (Newton Raphson)
Analysis Convergence

MAR101, Section 9, November 2015

S9 - 3

NONLINEAR ANALYSIS: THE APPROACH

MAR101, Section 9, November 2015

S9 - 4

THE APPROACH
Finite Element Analysis is an approximate technique in which
real life is interpreted by an engineer and simulated by
appropriate numerical algorithms
Interpretation + Simulation = idealization
Good idealization requires:
Understanding of the structure and its general behavior under load
Understanding of the behavior that is of interest (and those that are not)
Understanding of the most appropriate FE features to use

This exacting process requires an inquisitive engineer with a good

understanding of FEA

MAR101, Section 9, November 2015

S9 - 5

THE APPROACH
The Finite Element Program must never be used as a black box
The reliability of the results MUST always be questioned
The FE results should NEVER be permitted to override the results
anticipated from engineering judgment and experience

MAR101, Section 9, November 2015

S9 - 6

THE APPROACH

MAR101, Section 9, November 2015

S9 - 7

ITERATIVE SOLUTION METHODS

MAR101, Section 9, November 2015

S9 - 8

ITERATIVE SOLUTION METHODS

Pure Incremental Schemes:
Load applied incrementally
No iterative correction
Residual eliminated on an incremental
basis
Path history
Stiffness updated incrementally
Drift from true equilibrium

Incremental-Iterative Schemes:
Load applied incrementally
Iterative correction to restore incremental
equilibrium
Path history
Stiffness updated incrementally or
iteratively
Newton-Raphson methods
MAR101, Section 9, November 2015
Copyright 2015 MSC.Software Corporation

S9 - 9

ITERATIVE SOLUTION METHODS

For a linear analysis, the solution of equilibrium equations (for example
Gaussian elimination) can be applied directly in one step
For nonlinear solutions both the stiffness and external forces may be
functions of the nodal displacements
The aim is to attain equilibrium between the internal forces {I} and the
external forces {F}
For the solution step, we must solve the equations:
{I} {F} = Out of balance force vector = 0
or
[K] {u} - {F} = Out of balance force vector = 0

It is not practical to have a zero out of balance force, therefore it is

limited to a small user specified value called the residual force, {R}.

K u F Residual force vector R

To solve such a nonlinear set of equations, Marc uses a Newton Raphson
procedure (by default). This is an incremental-iterative method.
MAR101, Section 9, November 2015
Copyright 2015 MSC.Software Corporation

S9 - 10

THE NEWTON RAPHSON METHOD

l(u) F(u) R

Where R is a small user specified out of balance load called the Residual Load

Steps 1- 5 are the predictive stages

Steps 6 - 9 are the corrective (iterative) stages

MAR101, Section 9, November 2015

S9 - 11

THE NEWTON RAPHSON METHOD

The implementation of the Newton Raphson method typically
uses a Taylor series expansion about the current position to
estimate the displacement correction direction and magnitude.
Drop the quadratic term and compute displacement correction
{u}
Define tangent stiffness matrix:
It relates small changes in load to
changes in displacement

The procedure involves the

computation of {u} for a given
{F}
Stop iterations when the
residual ({I} {F}) is equal to
the user specified tolerance

MAR101, Section 9, November 2015

F
2 F
2
u

{ F } {0}

u
2
u
u
F
[ KT ] Tangent Stiffness
u
{ F } [ K T ] u

u [ KT ]1F
S9 - 12

THE NEWTON RAPHSON METHOD (NR)

Quadratic rate of convergence
Needs a few iterations to converge
Evaluation and maybe inversion of tangent stiffness matrix at
each iteration
Expensive for large systems
Less likely to converge to an
unstable solution
Recommended default method
Recommended for geometric
nonlinearity (GNL)
May fail under extreme material
nonlinearity (for example, brittle cracking)

MAR101, Section 9, November 2015

S9 - 13

THE MODIFIED NEWTON RAPHSON METHOD (MNR)

Evaluation and inversion of tangent stiffness matrix only at start
of each increment
Slow convergence behavior
More iterations to converge
Computationally inexpensive
per iteration
May be essential for extreme
material nonlinearity
May be assisted using
additional iterative
acceleration techniques
Suitable for mildly nonlinear
problems

MAR101, Section 9, November 2015

S9 - 14

MODIFIED VS. FULL NEWTON CONVERGENCE RATES

The graph shows the
residual and displacement
convergence norms during
a nonlinear increment for
NR and MNR.
MNR has decreasing
convergence rate behavior
For a highly nonlinear
application, MNR will be
much slower than NR

MAR101, Section 9, November 2015

S9 - 15

MAR101, Section 9, November 2015

S9 - 16

ITERATIVE METHODS IN MENTAT

MAR101, Section 9, November 2015

S9 - 17

ANALYSIS CONVERGENCE

MAR101, Section 9, November 2015

S9 - 18

ANALYSIS CONVERGENCE
Two aspects to the convergence control
1. The convergence criteria
Residual Norm
Displacement Norm
Strain Energy Norm

2. The convergence tolerance

Threshold value below which convergence is deemed to have occurred.

Convergence
Monitor

Convergence
Criteria

MAR101, Section 9, November 2015

Convergence
Tolerance

S9 - 19

ANALYSIS CONVERGENCE
The convergence criteria monitor the extent to which the iterative procedure has
reached equilibrium state
Termination of the iterative process occurs when the convergence ratio is less
than the specified tolerance (default 0.1 for residual and displacement)
Too slack a tolerance gives a false state of equilibrium. In this case, the
reference state used during the iterative procedure can drift from equilibrium
and may cause the material response to differ from the true response.
Too tight a tolerance results in unnecessary iterations

Residual
Force
Criteria

Residuals jump at
start of a new increment
Residuals decrease as the
iterative procedure works

Numerical
Convergence
Prescribed value
below which
convergence
is assumed

True Equilibrium
Increment

Convergence
Tolerance
Iterations

MAR101, Section 9, November 2015

S9 - 20

CONVERGENCE CRITERIA
Relative Residual Criteria
The aim is to ensure that the out-of-balance force is much smaller than the
external forces

Where:
||Fresidual|| is the internal (out of balance) force
vector and represents the component with the
highest absolute value
||Freaction|| is the external reaction force vector
and represents the component with the highest
absolute value
TOL1 is the user specified convergence
tolerance (default 0.1)

Moments can be included in a similar way

according to:

where
||M|| represents the moment vectors
TOL2 is the user specified convergence
tolerance (0.1)

MAR101, Section 9, November 2015

S9 - 21

CONVERGENCE CRITERIA
Relative Displacement Criteria
The aim is to ensure that the maximum
displacement of the last iteration is small compared
to the maximum displacement of the increment

Where:
u|| is the maximum correction to the incremental displacement vector (iterative
displacement).
||u|| is the maximum incremental displacement.
TOL1 is the user specified convergence tolerance (default 0.1).

Rotations can be included in a similar way according to:

Where:
|||| represents the rotations
TOL2 is the user specified convergence tolerance (0.1).

A disadvantage of this approach is that it results in at least one iteration,

regardless of the accuracy of the solution
MAR101, Section 9, November 2015
Copyright 2015 MSC.Software Corporation

S9 - 22

CONVERGENCE CRITERIA
Relative Strain Energy Criteria
The aim is to ensure that the iterative strain energy is small compared to the
strain energy of the increment
With this method, the entire model is checked since the energies are the
total energies integrated over the whole volume

Where:
E is the iterative change in strain energy
E is the incremental strain energy
TOL1 is the user specified convergence tolerance (default 0.1)

A disadvantage of this approach is that it results in at least one iteration,

regardless of the accuracy of the solution
The advantage of this method is that it evaluates the global accuracy as
opposed to the local accuracy associated with a single node

MAR101, Section 9, November 2015

S9 - 23

CONVERGENCE CRITERIA
Convergence Checking
Relative
The residuals, displacements, or energy is scaled
appropriately so that a relative tolerance is given.

Absolute
Convergence tolerances are given as absolute
values of residuals or displacements.
Not available for strain energy.

Relative/Absolute
Relative tolerances are used unless reactions or
incremental displacements are below a specified cutoff value, in which case, absolute testing is done.
Useful for contact problems in which contact bodies
are not initially in contact, giving zero reaction forces.
Not available for strain energy.

MAR101, Section 9, November 2015

S9 - 24

CONVERGENCE TESTING
Combination of residual and displacement checking
Accepts solution if either:
Residuals and displacements fulfill the criteria
Residuals or displacements fulfill the criteria

Recommended settings (initial)

The convergence criteria
Relative Residual AND Relative Displacement

The convergence tolerance

0.1 for both

Automatic Switching ON

MAR101, Section 9, November 2015

S9 - 25

RELATIVE CONVERGENCE CRITERIA ISSUES

Some analyses cause the relative convergence criteria to become
meaningless:
Maximum displacement increment becomes too small.
maximum displacement change at node 3 degree of freedom 1 is equal to 2.3E-13
maximum displacement increment at node 2 degree of freedom 1 is equal to 6.9E-9
displacement convergence ratio 3.462

Maximum reaction force becomes too small

maximum residual force at node 2311 degree of freedom 1 is equal to 6.058E-08
maximum reaction force at node 2294 degree of freedom 2 is equal to 1.460E-08
residual convergence ratio 4.148

Extremely small strain energy density

Each of these scenarios represent the denominator of the convergence

test becoming numerically too small to handle. That is, dividing any
numerator by 1e-10 will cause a large convergence norm.
This gives the wrong impression that the analysis is not converging
Remedies:

Specify Auto-Switch (see later)

Remove offending criteria
Tighten remaining criteria
Consider relative/absolute testing

MAR101, Section 9, November 2015

S9 - 26

RELATIVE CONVERGENCE CRITERIA ISSUES

MAR101, Section 9, November 2015

S9 - 27

AUTOMATIC CRITERIA SWITCHING

Remedy to meaningless convergence criteria
Automatically detects small values in convergence calculations, and does
one of the following:
Switches to residual checking, if displacement increments become very small
(Max_Disp_Inc./Smallest_Elem_Size < 1e-6)
Switches to displacement checking, if reaction
forces become very small (<1e-8).
Switches to energy checking, if structure is
free of stress and deformation (strain energy
density < 1e-15).

Automatic switching is only available

for relative testing
If any kind of absolute testing is used
Automatic Switching will be deactivated

Auto
Step
MAR101, Section 9, November 2015
Copyright 2015 MSC.Software Corporation

Relative
Criteria

S9 - 28

Auto
Switch

Robust
Solution

AUTOMATIC CRITERIA SWITCHING

This example uses displacement convergence testing throughout
the analysis of both draw bending and spring-back.
At the last stages of the spring-back analysis, the convergence
testing is automatically switched to strain energy testing due to
extremely small displacement increment.

Draw Bending and Spring-back

S9 - 29

AUTOMATIC CRITERIA SWITCHING

This example uses residual
force convergence testing as
the default throughout the
analysis.
However, at the beginning of
the analysis the key has rigid
body motion which results in
very small reaction forces.
Therefore the convergence
testing is automatically
switched to displacement
testing and reset to residual
testing as soon as the key
comes in contact with the lock.

MAR101, Section 9, November 2015

S9 - 30

RBE2, RBE3 Convergence Testing

In large displacement large rotation simulations
and additional convergence control is made on
the rotations of the rigid links.
The default value is on an absolute value and too
small, resulting in unnecessary iterations for
many problems.
Either turn off or set value to between 0.02 and
0.1

MAR101, Section 9, November 2015

S9 - 31

NONLINEAR HEAT TRANSFER ANALYSIS

Nonlinearity can be caused by
Temperature dependent conductivity and/or specific heat
Radiation boundary condition
Temperature dependent convective (film) coefficient or heat flux

Consequently, heat capacity matrix, conductivity matrix, film

matrix and equivalent nodal flux vector may be temperature
dependent.
Steady state: applicable in case of mild nonlinearity; e.g. if
conductivity is slightly temperature dependent

MAR101, Section 9, November 2015

S9 - 32

NONLINEAR HEAT TRANSFER ANALYSIS

Solution can be obtained iteratively:
Start with KT0=Q.
Next approximations are obtained by successive substitution

n 1

K T Q T
n

continue until

T n 1 T n

max

Ttol3

Transient analysis: necessary in case of severe nonlinearity; e.g.

radiation boundary condition
Solution can be obtained using nonlinear backward difference scheme:

MAR101, Section 9, November 2015

S9 - 33

ESTIMATED TEMPERATURE AT END OF INCREMENT

Material properties are based upon
*

T
~

1 n 1(est)
n
T
T
~
2 ~

for the first iteration next iterations ,

T
~

est

n
n
n -1
T T T * t n 1 / t n
~
~
~

iterations are stopped if

Ti*-1 Ti*

MAR101, Section 9, November 2015

max

Ttol3

S9 - 34

CONTROL VALUES
Control values heat transfer analysis
Ttol1 : maximum incremental temperature change at node (def = 20); if the
automatic time stepping scheme is selected, the time step will be
automatically increased or decreased if necessary.
Ttol2 : maximum nodal temperature change before properties are
reevaluated and matrices reassembled (def = 100).
Ttol3 : maximum error in estimated nodal temperature; used for property
reevaluation (def = 0, test is by passed)

MAR101, Section 9, November 2015

S9 - 35

INITIAL TIME STEP ESTIMATE

Sometimes oscillatory behavior if time step is too small; e.g. heat
penetration in an originally iso-thermal block
Better approximation can be obtained if:
Time step is INCREASED
Mesh is refined
Heat capacity matrix is lumped (linear elements)

MAR101, Section 9, November 2015

S9 - 36

PART II: LOAD CONTROLS AND GUI CONTROLS

Load increment control and automatic time stepping
Graphical user interface controls
Geometrical nonlinear framework
Cost of nonlinear analysis

MAR101, Section 9, November 2015

S9 - 37

TIME STEP CONTROL

LOAD INCREMENTATION AND AUTOMATIC TIME
STEPPING
GRAPHICAL USER INTERFACE CONTROLS

MAR101, Section 9, November 2015

S9 - 38

LOAD INCREMENTATION CONTROL

There are three principle ways to apply loading in
Marc:
Constant load
incrementation
(Fixed)

F
F
Ninc

Variable load
incrementation
(adaptive)

F i F i 1

Variable load
incrementation
(adaptive/arclength)

The total load is divided equally

into a number of specified solution
increments.

The next load increment is

evaluated from the previous load
multiplied by a growth/reduction
factor.
The next load increment is
evaluated from the previous load
increment multiplied by the arclength magnitude.

F i F

Another very important application of load incrementation is to rigid body

contact bodies (discussed in contact section)
MAR101, Section 9, November 2015
Copyright 2015 MSC.Software Corporation

S9 - 39

LOAD INCREMENTAL CONTROL

Fixed
Gives more user control
Good where nonlinearity is mild
or where the response is smooth.

Adaptive

Load

Recommended method
Automatically locates and
handles sudden nonlinearity

THE TIME SCALE

Marc uses the concept of time to monitor load increments even in static
analyses (typically, a time period of unity)
It is invaluable in providing a scale over which a history of events is described
when using load tables
For example, both the number of increments required and the total time of the
solution are specified for fixed increment load increments
For adaptive load increments, only time is needed
Easy to determine how far the analysis has progressed from the time output. The
increment number does not give this information in an adaptive solution (50th/100
increments is not necessarily half way through)
1.0

Relative
load
magnitude

0.0

1.0

Time period

MAR101, Section 9, November 2015

S9 - 41

THE TIME SCALE

For both Fixed and Adaptive Load increments with table-driven
Input, Marc will automatically detect all points in the table
There is no need to break up the loadcase
where there is a change in slope on the
boundary condition history curve
Example:
Correct results will be obtained with just one
loadcase between times 0 and t3.

MAR101, Section 9, November 2015

S9 - 42

THE TIME SCALE

For adaptive load increments care must
be taken to appropriately define the
loading history in each loadcase
Select all points for loadcase results

MAR101, Section 9, November 2015

S9 - 43

LOAD INCREMENTATION CUT BACK (STEP REDUCTION)

If convergence is not achievable at a
specified load level, a step cut back is
carried out.

Load

Initial Load Level

It saves having to restart the analysis at

a specified load level.

Unconverged Solution after

maximum number of
allowable iterations
Reduced Load Level

A cut back is where an increment is

restarted but at a reduced load level.

Converged Solution at
Reduced Load Level

The incrementation number is reset for

adaptive incrementation.
Cut backs will continue until
convergence is obtained or until the
maximum number of reductions have
reached.
Additional sub-incrementation is
carried out with fixed incrementation to
catch up.
The maximum number of reductions
may be specified by the user.
MAR101, Section 9, November 2015
Copyright 2015 MSC.Software Corporation

S9 - 44

Displacement

Following convergence at a
reduced level, the reduced step
size may remain constant (fixed
incrementation) or increase again
(adaptive incrementation).

Note , if the previous increment did not

converge to a tight tolerance, cut backs
may not result in a converged solution

GRAPHICAL USER INTERFACE CONTROL

MAR101, Section 9, November 2015

S9 - 45

MENTAT LOAD INCREMENT PARAMETERS

MAR101, Section 9, November 2015

S9 - 46

LOAD INCREMENT PARAMETERS

The form to control the load incrementation
The data required depends upon the type of load stepping procedure
specified:
Constant
Or
Adaptive Multi-Criteria
Or
Adaptive Arc length
Or
Adaptive Temperature
It is possible to mix-and-match
different incrementation
procedures in an analysis.

MAR101, Section 9, November 2015

S9 - 47

MENTAT LOAD INCREMENT PARAMETERS

Number of Cutbacks:
Optionally invoked
The increments (solution) will stop if the
maximum number of cutbacks is
reached
Upon detection of exit numbers 3002,
1009, or 2400, the increment is cut
back.
Typical text in output file:
If the number of cutbacks is greater than 4 there
is a problem with your choice of time steps
0th
1.00
1st
0.50
2nd
0.25
3rd
0.125
4th
0.0625
MAR101, Section 9, November 2015
Copyright 2015 MSC.Software Corporation

S9 - 48

MENTAT LOAD INCREMENT PARAMETERS

Out-Of-Core Element Storage
Element quantities are stored out-of-core
This reduces the amount of main memory required in
the analysis, but may increase the amount of I/O
Marc Parameter Command is ELSTO

Out-Of-Core Incremental Backup

During the Newton-Raphson iteration process, Marc
makes a second copy of the solution space
Marc normally does this in memory unless sufficient
memory is not available; in which case, it uses auxiliary
disk space.
This option can be used to force Marc to use computer
hard drive space
Marc Parameter command is IBOOC
The slow-down related to incremental backup out-of-core is much less than that of
out-of-core element storage
Often useful for large problems when the multi-frontal sparse solver is used, as this
back-up copy is allocated before the decomposition memory is allocated
If element data is out-of-core, the incremental backup is automatically out-of-core as
well.
MAR101, Section 9, November 2015
Copyright 2015 MSC.Software Corporation

S9 - 49

MENTAT LOAD INCREMENT PARAMETERS

Loadcase Properties Form
The data required depends upon the type of load stepping procedure
specified
Fixed number of steps and total time
Adaptive approaches (Multi-Criteria, Arc Length,
Temperature) click parameters and enter
information on the Adapting Stepping
parameters form

It is possible to mix-and-match different

incremental procedures for different loadcases
in an analysis

MAR101, Section 9, November 2015

S9 - 50

MENTAT LOAD INCREMENT PARAMETERS

Fixed
Fixed number of increments
User defined number of load
increments, within which the total
load will be applied in equal steps

Total time
The user defined total time
duration of the loadcase.

MAR101, Section 9, November 2015

S9 - 51

LOAD INCREMENT PARAMETERS: ADAPTIVE

Trial Time Step Size
Mentat: Initial Fraction of Loadcase Time
User defined time increment
The starting value for the load stepping
procedure.
For example, given a total solution time of unity, a
starting step size of 0.01 would represent 1% of
the total load to be applied.
Default is good

Time Step Scale Factor

Mentat: Time Step Increase Factor
Use default value
Discussed later

MAR101, Section 9, November 2015

S9 - 52

LOAD INCREMENT PARAMETERS: ADAPTIVE

Minimum Time Step
Mentat: Minimum Fraction of
Loadcase Time
The algorithm is not allowed to use a
time step smaller than this.
The analysis will not stop if it is
reached.
The default value is the smaller of the
current time increment or 1e-5 times
the total time. Which is too small

Maximum Time Step (optional)

Mentat: Maximum Fraction of
Loadcase Time
The algorithm is not allowed to use a
value larger than this.
The analysis will not stop if it is
reached.
Default =0.5 is too large use 0.05
MAR101, Section 9, November 2015
Copyright 2015 MSC.Software Corporation

S9 - 53

LOAD INCREMENT PARAMETERS: ADAPTIVE

Total Time

Max # of Steps
Mentat: Maximum # Steps
The maximum number of
increments in the current loadcase.
The analysis will stop if it is
reached.

Mentat: Total Loadcase Time

The user defined total time
duration of this loadcase

MAR101, Section 9, November 2015

S9 - 54

LOAD INCREMENT PARAMETERS: ADAPTIVE

# of Steps of Output
Mentat: Results at Fixed Time Intervals
Results can be requested at specific times rather than at specific increment
frequency.
For adaptive increment type, this is important for highly nonlinear analyses for which
results are desired at uniform times during the analysis (i.e. not bunched up)

MAR101, Section 9, November 2015

S9 - 55

ITERATION PARAMETERS
Desired # of Iterations per Increment:
Mentat: Desired # Recycles/Increment
Specified by user (desired)
Number of actual recycles to converge
(actual)
If actual < desired = easy increment
increase time step in next increment.
If actual = desired = target increment
continue with the same time step in next
increment.
If actual > desired = hard increment
decrease time step in next increment.
In a contact analysis, contact related recycles
(e.g. body contact or separation) are not
counted.
Increase value for lightly nonlinear analyses
In addition, the time step will be decreased if
Elements inside out (exit 1005 or 1009)
Nodes slide off rigid bodies (exit 2400)

MAR101, Section 9, November 2015

S9 - 56

ITERATION PARAMETERS
Time Step Scale Factor
Factor used to increase/decrease the time
step is user-defined (default is 1.2).
No increase of the time step during the
current increment.
Scale factor used is bounded by the user
defined minimum and maximum factors.

Exception
If there is consistent convergence
(convergence ratio reducing in 3 previous
consecutive increments) and the number of
actual recycles exceeds the desired number,
the number of recycles is allowed to go
beyond the desired number until
convergence or up to the user specified
maximum number. The time step is then
decreased for the next increment.
MAR101, Section 9, November 2015
Copyright 2015 MSC.Software Corporation

S9 - 57

ITERATION PARAMETERS
Max # of Iterations per Increment
Mentat: Max # Recycles
Relates to the Newton Raphson
iterations.
If exceeded, a cut back may follow.

Proceed when not converged

Marc should proceed with the next
increment even if convergence is not
obtained within the current increment.
If the maximum number of recycles is
reached without convergence, a warning
is given and the analysis is continued.
This option is, in general, not
recommended.

MAR101, Section 9, November 2015

S9 - 58

ITERATION PARAMETERS
Contribution of Initial Stress Stiffness:
Controls the contribution of the initial
stress matrix (geometric stiffness matrix)
to the tangent stiffness matrix.
Influences the convergence behavior,
but not the accuracy.
FULL: Full contribution (default)
NONE: No contribution
DEVIATORIC STRESS:
Only for deviatoric stress contribution
Often useful in rubber analysis, when the volumetric stresses are large.
The contribution of volumetric stresses can be varied linearly (NUMERICAL
PREFERENCES) using a factor (1=no cont., 0=full cont.).

BEGIN INCREMENT STRESS: Uses stresses at the beginning of increment

and not the stresses at the last iteration.
TENSILE STRESS: Only tensile stresses contributions are considered.

MAR101, Section 9, November 2015

S9 - 59

AUTOMATIC CRITERIA FOR LOAD INCREMENTATION

Additional option allows for
automatic physical criteria to be
used.
These automatic criteria serve as
upper-bound limits to prevent runaway Newton-Raphson iterations.
Criteria are only added in the
analysis if there are no competing
explicitly defined user-criteria
found.
Criteria are only used as limits;
they are used to control the time
step for the current increment but
not for the next increment.

MAR101, Section 9, November 2015

S9 - 60

ADDITIONAL AUTOMATIC CRITERIA FOR LOAD

INCREMENTATION
Five additional criteria

Total strain criterion for large displacement analyses maximum equivalent

total strain increment set to 50%.
Plastic strain criterion for large displacement, finite strain analyses maximum equivalent plastic strain increment set to10%.
Relative creep strain criterion for explicit creep analyses - maximum creep
strain change/elastic strain set to 0.5.
Relative stress change criterion for explicit creep analyses maximum
equivalent stress change/equivalent stress set to 0.5.
State variable criterion for large displacement analyses - maximum
temperature increment is such that the equivalent stress increment
associated with the change in thermal properties of the materials does not
exceed 50% of the total equivalent stress.

MAR101, Section 9, November 2015

S9 - 61

PHYSICAL CRITERIA FOR LOAD INCREMENTATION

User defined range of values for one
or more of the available criteria.
Specify a maximum time step
permissible within each range.
Can be used in conjunction with
standard adaptive incrementation
control.

MAR101, Section 9, November 2015

S9 - 62

PHYSICAL CRITERIA FOR LOAD INCREMENTATION

Criteria Range
A number of physical criteria can be selected (more than one if required).
Consider the Plastic Strain criterion.
The range of plastic strains together with the maximum permitted strain
increment are required.
If a plastic strain increment is detected to be beyond that permitted, a step
reduction will occur to ensure that the strain remains within the specified
bounds.

MAR101, Section 9, November 2015

S9 - 63

PHYSICAL CRITERIA FOR LOAD INCREMENTATION

Limits (default)
Decrease current time step if a physical
criterion is violated.
Desired-Actual control is still used to
determine the next increment step size.

Targets
Decrease current time if a physical criterion is
violated.
Physical criterion is used to determine the
next increment step size (not the number of
iterations - desired-actual control).
If the calculated values of the criteria are
higher than the user-defined values in any
iteration, the time step is scaled down and the
current increment is repeated.
The scale factor used for reduction (increase)
is the ratio between the actual value and the
target value and used for the next increment.
This factor is limited by user-specified
minimum and maximum factors (defaults to
0.1 and 10 respectively).

MAR101, Section 9, November 2015

S9 - 64

INERTIAL DAMPING: EXAMPLE

In a static analysis, artificial numerical damping can be activated.
If the time step becomes smaller than the user-defined minimum
value, a factored lumped mass matrix is added to the stiffness
matrix (the right hand side is modified consistently).
A time step equal to 0.001 * user-defined minimum time step is
used.
If the time step becomes larger than the user-defined minimum,
the option is switched off automatically.

MAR101, Section 9, November 2015

S9 - 65

MENTAT LOAD INCREMENT PARAMETERS

Adaptive Stepping Multi-Criteria
Inertial Damping
In a static analysis, artificial numerical damping can be activated
If the time step becomes smaller than the user-defined minimum value, a factored
lumped mass matrix is added to the stiffness matrix (the right hand side is
modified consistently)
A time step equal to 0.001 * user-defined minimum time step is used
If the time step becomes larger than the user-defined minimum, the option is
switched off automatically

MAR101, Section 9, November 2015

S9 - 66

Instabilities
Inertial Damping

Nonlinear Elasticity
Plasticity
Material Damage (Progressive Failure, Cracking)
Fracture
Buckling

Arc Length Method

Nonlinear Elasticity
Buckling

MAR101, Section 9, November 2015

S9 - 67

MENTAT LOAD INCREMENT PARAMETERS

Adaptive Stepping Multi-Criteria
Inertia Damping Example
Consider two truss structures. One loaded with a force and having quasi-static
inertial damping and one loaded by a displacement

Force

Displacement

MAR101, Section 9, November 2015

S9 - 68

MENTAT LOAD INCREMENT PARAMETERS

Adaptive Stepping Multi-Criteria
Inertial damping example continued
Response of
displacement Loaded
members

Effect of quasi-static
damping on Force
Loaded Node

MAR101, Section 9, November 2015

S9 - 69

Incrementation Procedure Which Procedure to Use?

MAR101, Section 9, November 2015

S9 - 70

Auto Time Multi-Criteria Critical Data

MAR101, Section 9, November 2015

S9 - 71

Load Case Selection What is correct

MAR101, Section 9, November 2015

S9 - 72

Iteration history

MAR101, Section 9, November 2015

S9 - 73

Equivalent Plastic Strain History

What is Going On?

Formation of a
Shear Band
Clearly an Instability

MAR101, Section 9, November 2015

S9 - 74

Total Work Strain Energy = Damping Loss

Damping Energy
Rate= 5.e-4
MAR101, Section 9, November 2015
Copyright 2015 MSC.Software Corporation

Damping Energy
Rate= 1.e-7
S9 - 75

Damping Effect
Effect of Damping Energy Rate on Calculated Plastic
Strain
0.5
0.45
0.4

Plastic Strain

0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
-8

-7

-6

-5

-4

-3

-2

Logarithmic Damping Energy RateTitle

Be Careful Damping Adds Stability, but

Absorbs Energy, Hence Less Deformation
(Plastic Strain)

MAR101, Section 9, November 2015

S9 - 76

-1

Refined Mesh

Plastic Strain
about the Same
Shear Band
Moved
Instability when
plasticity occurs
across the
region through
full element

MAR101, Section 9, November 2015

S9 - 77

MENTAT LOAD INCREMENT PARAMETERS

Adaptive Stepping Multi-Criteria
Maximum # Steps
The maximum number of increments in the current loadcase
The analysis will stop if it is reached

MAR101, Section 9, November 2015

S9 - 78

MENTAT ITERATION PARAMETERS

Adaptive Stepping Multi-Criteria
Desired # Recycles/Increment
Desired: specified by user
Actual: number of actual recycles to converge
If actual < desired = easy increment
increase time step in next increment.
If actual = desired = target increment
continue with the same time step in next
increment.
If actual > desired = hard increment
decrease time step in next increment.
In a contact analysis, contact related recycles
(for example, body contact or separation) are
not counted.
Increase value for lightly nonlinear analyses
In addition, the time step will be decreased if
Elements inside out (exit 1005 or 1009)
Nodes slide off rigid bodies (exit 2400)

MAR101, Section 9, November 2015

S9 - 79

MENTAT ITERATION PARAMETERS

Adaptive Stepping Multi-Criteria
Time Step Increase Factor
Factor used to increase/decrease the time
step is user-defined (default is 1.2)
No increase of the time step during the
current increment
Scale factor used is bounded by the user
defined minimum and maximum factors
Exception
If there is consistent convergence
(convergence ratio reducing in 3 previous
consecutive increments) and the number of
actual recycles exceeds the desired number,
the number of recycles is allowed to go
beyond the desired number until convergence
or up to the user specified maximum number.
The time step is then decreased for the next
increment.

MAR101, Section 9, November 2015

S9 - 80

MENTAT LOAD INCREMENT PARAMETERS

Adaptive Stepping Multi-Criteria
User Defined Criteria
User defined range of values for one or
more of the available criteria
Specify a maximum time step permissible
within each range
Can be used in conjunction with adaptive
multi-criteria incremental control

MAR101, Section 9, November 2015

S9 - 81

MENTAT LOAD INCREMENT PARAMETERS

Adaptive Stepping Multi-Criteria
User-Defined Criteria Continued
Limits (default)
Decrease current time step if a physical
criterion is violated
Desired-Actual control is still used to
determine the next increment step size

Targets
Decrease current time if a physical criterion is
violated
Physical criterion is used to determine the
next increment step size (not the desiredactual control)
If the calculated values of the criteria are
higher (lower) than the user-defined values in
any iteration, the time step is scaled down
(up) and the current increment is repeated.
The scale factor used for reduction (increase)
is the ratio between the actual value and the
target value
This factor is limited by user-specified
minimum and maximum factors (defaults to
0.1 and 10 respectively)
MAR101, Section 9, November 2015
Copyright 2015 MSC.Software Corporation

S9 - 82

MENTAT LOAD INCREMENT PARAMETERS

Adaptive Stepping Multi-Criteria
User Defined Physical Criteria Range
A number of physical criteria can be selected (more than one if required).
Consider the Plastic Strain criterion
The range of plastic strains together with the maximum permitted strain increment are
required
If a plastic strain increment is detected to be beyond that permitted, a step reduction will
occur to ensure that the strain remains within the specified bounds

MAR101, Section 9, November 2015

S9 - 83

MENTAT LOAD INCREMENT PARAMETERS

Adaptive Stepping Multi-Criteria
User Defined Additional Criteria
Total strain criterion for large displacement analyses maximum equivalent total
strain increment set to 50%
Plastic strain criterion for large displacement, finite strain analyses - maximum
equivalent plastic strain increment set to10%
Relative creep strain criterion for explicit creep analyses - maximum creep strain
change/elastic strain set to 0.5
Relative stress change criterion for explicit creep analyses maximum equivalent
stress change/equivalent stress set to 0.5
State variable criterion for large
displacement analyses - maximum
temperature increment is such that the
equivalent stress increment associated
with the change in thermal properties
of the materials does not exceed 50%
of the total equivalent stress.

MAR101, Section 9, November 2015

S9 - 84

MENTAT LOAD INCREMENT PARAMETERS

Adaptive Stepping Multi-Criteria
Automatic Criteria
Additional option allows for automatic
physical criteria to be used
These automatic criteria serve as
upper-bound limits to prevent run-away
Newton-Raphson iterations
Criteria are only added in the analysis if
there are no competing explicitly
defined user-criteria found
Criteria are only used as limits; they are
used to control the time step for the
current increment but not for the next
increment

MAR101, Section 9, November 2015

S9 - 85

MENTAT LOAD INCREMENT PARAMETERS

Adaptive Stepping Multi-Criteria
Results at Fixed Time Intervals
Results can be requested at specific times rather than at specific increment
frequency
For adaptive increment type, this is important for highly nonlinear analyses for which
results are desired at uniform times during the analysis (that is not bunched up)

MAR101, Section 9, November 2015

S9 - 86

ARC LENGTH METHODS

Arc Length Methods are often
called continuation methods
Objective is to modify the Load
such that the simulation can
continue within certain changes
in deflection.
Buckling and Post Buckling is
never easy
The following methods exist:

Crisfield
Riks-Ramm
Modified Riks / Ramm
Crisfield / Modified Riks-Ramm
Scaled Riks-Ramm
Advanced Crisfield

MAR101, Section 9, November 2015

S9 - 87

LOAD INCREMENT PARAMETERS: ADAPTIVE/ARC LENGTH

Arc Length Methods are often called
continuation methods
Objective is to modify the Load such
that the simulation can continue
within certain changes in deflection.
The method controls the evolution of
the arc length (l)
The following methods exist:

Crisfield
Riks-Ramm
Modified Riks / Ramm
Crisfield / Modified Riks-Ramm
Scaled Riks-Ramm
Advanced Crisfield

MAR101, Section 9, November 2015

S9 - 88

LOAD INCREMENT PARAMETERS: ADAPTIVE/ARC LENGTH

Initial Arclength lin= DU1TDU1
So initial step is critical
For Single Loadcase Load Applied=
P=l*Puser
Max Ratio Arc Length/Initial Arc
Length:
Keeps the arc length step size from
growing too large.

Min Ratio Arc Length/Initial Arc

Length:
Keeps the arc length step size from
reducing too much.

MAR101, Section 9, November 2015

S9 - 89

ARC LENGTH METHOD E11x6x7

MAR101, Section 9, November 2015

S9 - 90

ARC LENGTH METHOD E11x6x7

MAR101, Section 9, November 2015

S9 - 91

ARC LENGTH METHOD E4x7

MAR101, Section 9, November 2015

S9 - 92

MENTAT LOAD INCREMENT PARAMETERS

Adaptive Stepping Arc Length
Max Ratio Arc Length/Initial Arc Length:
keeps the arc length step size from
growing too large
Min Ratio Arc Length/Initial Arc Length:
keeps the arc length step size from
reducing too much

MAR101, Section 9, November 2015

S9 - 93

MENTAT ITERATION PARAMETERS

Solution Control
Max # Recycles
Relates to the Newton Raphson iterations
If exceeded, a cut back may follow

Non-Positive Definite
If turned on, the solution of a non-positive
definite system is forced.

Proceed when not converged

Marc should proceed with the next increment
even if convergence is not obtained within the
current increment
If the maximum number of recycles is reached
without convergence, a warning is given and
the analysis is continued
This option is, in general, not recommended

MAR101, Section 9, November 2015

S9 - 94

MENTAT ITERATION PARAMETERS

Solution Control
Contribution of Initial Stress Stiffness:
Controls the contribution of the initial stress
matrix (geometric stiffness matrix)
to the tangent stiffness matrix.
Influences the convergence behavior, but not
the accuracy
Select from the following:
Full: Full contribution (default)
None: No contribution
Deviatoric Stress:
Only for deviatoric stress contribution
Often useful in rubber analysis, when the volumetric stresses are large
The contribution of volumetric stresses can be varied linearly (numerical preferences)
using a factor (1=no cont., 0=full cont.)

Begin Increment Stress: Uses stresses at the beginning of increment and not the
stresses at the last iteration
Tensile Stress: Only tensile stresses contributions are considered

MAR101, Section 9, November 2015

S9 - 95

AUTOMATIC TIME STEPPING: EXAMPLE

Key insertion problem

3 deformable bodies
2 rigid bodies
Glued contact between rigid and deformable bodies
Elastic Strain physical criteria used
Quasi-Static Inertial Damping specified

Rigid Body
Key

Rigid Body

MAR101, Section 9, November 2015

3 Deformable
Bodies

S9 - 96

AUTOMATIC TIME STEPPING: EXAMPLE

Results

Total number of time step cut backs

versus increment

Time versus increment number

Step cut-back
MAR101, Section 9, November 2015
Copyright 2015 MSC.Software Corporation

Step cut-back
S9 - 97

GEOMETRICAL NONLINEAR FRAMEWORK

MAR101, Section 9, November 2015

S9 - 98

GEOMETRICAL NONLINEAR FRAMEWORK

Total Lagrangian
All integrals are evaluated with
respect to the initial undeformed
configuration.

0
1
2
3

Updated Lagrangian
All integrals are evaluated with
respect to the last completed
iteration of the current increment.

0
1
2
3

MAR101, Section 9, November 2015

S9 - 99

GEOMETRICAL NONLINEAR FRAMEWORK

How to choose the framework?
The choice affects the stress and strain measures used and access to
some capabilities that are only sensibly formulated in one or the other (for
example, updated Lagrangian for remeshing)
Both procedures can be used for linear or nonlinear materials, in
conjunction with static or dynamic analysis
Theoretically and numerically, the two formulations (TL & UL) yield exactly
the same results
The choice of one over another is dictated by such things as convenience of
modeling the physics of the problem, rezoning requirements, and
integration of constitutive equations

MAR101, Section 9, November 2015

S9 - 100

GEOMETRICAL NONLINEAR FRAMEWORK

Total Lagrangian Usage
Naturally formulated in terms of Green-Lagrange strains and 2nd Piola
Kirchhoff stress
Although this formulation is based on the initial element geometry, the
incremental stiffness matrices are formed to account for previously
developed stress and changes in geometry.
Integration of the constitutive equations for certain types of material
behavior (for example, plasticity) make the implementation of the total
Lagrange formulation inconvenient
Particularly suitable for the analysis of nonlinear hyperelastic problems
(Mooney/Ogden material or user subroutine HYPELA).
Useful for problems with plasticity and creep, where moderately large
rotations but small strains occur - a case typical in problems with beam or
shell bending. However, this is only due to the element approximations
involved.

MAR101, Section 9, November 2015

S9 - 101

GEOMETRICAL NONLINEAR FRAMEWORK

Updated Lagrangian Usage
Naturally formulated in terms of Cauchy stress and logarithmic strain since the
current configuration is the reference configuration.
Used for large strain plasticity analyses
Flexibility in accounting for elasticity and history effects
Residual stresses can be accurately calculated

For deformations involving excessive distortions, ease of rezoning favours the

updated Lagrangian formulation.
Mapping is to the current configuration and not to the undistorted mesh.
Analysis of shell and beam structures in which the total rotations are sufficiently
large so that the nonlinear terms in the curvature expressions may no longer be
neglected.
Many beam and shell elements assume that the out-of-plane derivatives are
related to the rotation in radians. This theoretically breaks down when rotations
approach 1 radian in practice, the limiting rotation is much more severe.

MAR101, Section 9, November 2015

S9 - 102

ADVANCED ANALYSIS OPTIONS

MAR101, Section 9, November 2015

S9 - 103

ADVANCED ANALYSIS OPTIONS

MAR101, Section 9, November 2015

S9 - 104

COST OF NONLINEAR ANALYSIS

MAR101, Section 9, November 2015

S9 - 105

COST OF NONLINEAR ANALYSIS

Time Consuming stages
Assemble stiffness matrix
Solve
Recover stresses
Loop over load increments
Loop over iterations
Loop over elements
Loop over Gauss points
end
end
end
end

Linear analyses have 1 pass of the load increment and

iterative loop, a much less expensive solution
Reduced integration elements give 1 pass of the GP
loop, a much less expensive solution

MAR101, Section 9, November 2015

S9 - 106

COST OF NONLINEAR ANALYSIS: EXAMPLE

A nonlinear increment took 12 recycles, giving a total time of
2504.60 seconds.
An entire linear analysis (single-step) would only take 372.77
seconds!

11 iteration

MAR101, Section 9, November 2015

S9 - 107

COST OF NONLINEAR ANALYSIS

Nonlinear analyses have significantly greater storage
requirements
Sufficient physical memory is a major consideration to obtain inmemory solutions (100% CPU)
Use iterative Sparse (incomplete Choleski)

MAR101, Section 9, November 2015

S9 - 108

SOLVERS
Each iteration of the NewtonRaphson Method requires solving the
system of equations
This can be done with a Direct Solver
or with an Iterative Solver
With recent advances in solver
technology, the time spent in
assembly and recovery now exceed
the time spent in the solver

MAR101, Section 9, November 2015

S9 - 109

SOLVERS AT A GLANCE
CASI Iterative (Solver 9)
is recommended for
bulky models

Pardiso (Solver 11)

is recommended for SMP

MUMPS (Solver 12)

is recommended for DMP

Multi-Frontal Sparse (Solver 8)

MAR101, Section 9, November 2015

S9 - 110

is Default

ITERATIVE VS. DIRECT SOLVER

Use the Iterative Sparse solver wherever possible (no beams,
shells or Hermann elements).
Otherwise, use the Multi-Frontal direct sparse solver.
Example: f(x) = sqrt(x) -1 = 0.
An iterative solver
requires more
iterations to achieve
solution.
Each cycle in an
iterative method takes
less time to compute
than the time needed
for a cycle in a direct
method.

MAR101, Section 9, November 2015

S9 - 111

THE ITERATIVE SOLVER

The iterative solver carries out a number of iterations during the
solution
These iterations are different than those of the Newton-Raphson
iterations.
For EXIT 2020, the iterative solver fails to converge within the
required number of iterations. It refers to the number of iterations
in the iterative solver and means one of the following:
The number of iterations permitted for the iterative solver must be increased
The iterative solver is not appropriate for use in the analysis

MAR101, Section 9, November 2015

S9 - 112

EXERCISE
Workshop 10 Pin Insertion and Extraction
Be sure to ask for help if there is anything you do not understand

MAR101, Section 9, November 2015

S9 - 113

MAR101, Section 9, November 2015

S9 - 114

SECTION 10
RESOLVING
CONVERGENCE PROBLEMS

MAR101, Section 10, November 2015

S10 - 1

OVERVIEW
Nonlinear Analysis Guidelines
Information Available For Help
Troubleshooting Analysis Failure
Analysis Failure : EXIT Numbers

MAR101, Section 10, November 2015

S10 - 2

NONLINEAR ANALYSIS GUIDELINES

Convergence of a nonlinear problem mostly is not just about the
convergence tolerance values or the criteria specified; it is an
overall issue of model integrity and representation of reality.
It is strongly recommended that small tests be performed to gain
experience of unknown (to you) element and solution types:
To understand its limitations.
To ensure that it provides the required behaviour for the actual simulation to
be carried out.
To prevent expensive surprises at the end of a project.

Small element tests are preferable (where possible), since it is so

much quicker and easier to verify the input and to evaluate the
response with only a few degrees of freedom.

MAR101, Section 10, November 2015

S10 - 3

NONLINEAR ANALYSIS GUIDELINES

There are a number of sources for examples and benchmarks
available that may help in this regard.
The MARC User Guide manual - contains an increasing amount of worked
examples written with the express intention of clearly demonstrating the use
of the facilities.
The MARC Demonstration Problem manual (volume E) - is in Marc data file
format only. The data files associated with this manual can be located in the
Marc installation directory.
The NAFEMS suite of examples. Further information is available on their
website (www.nafems.org).

MAR101, Section 10, November 2015

S10 - 4

NONLINEAR ANALYSIS GUIDELINES

Perform and scrutinize the results from a static linear analysis to
check the integrity and behaviour of the basic model. If the
nonlinear model already exists,
For contact analyses, this would mean changing all contact conditions to
GLUED.
For material nonlinearity, simply increase the failure criteria so that it cannot
be reached.
For geometrically nonlinear analyses, turn off Large Displacement / Large
Strain

Add each of the nonlinearities, one by one, to determine their

effect on the solution and its convergence behaviour. For
instance, start with contact, next add any geometric nonlinearity,
and then finally, any material nonlinearity, etc.
For contact analyses, contact conditions can be set to GLUED.
The next step would be TOUCHING, but with a separation force of 1e20.

MAR101, Section 10, November 2015

S10 - 5

NONLINEAR ANALYSIS GUIDELINES

If buckling is expected, a linear eigenvalue buckling analysis
should be performed to obtain the linear buckling load. This will
act as both, a benchmark value to compare against as well as a
useful aid in determining the load magnitude to be applied in the
subsequent geometrically nonlinear analysis.
Always use engineering common sense and verify the plausibility
of the results before making design decisions based upon them.

MAR101, Section 10, November 2015

S10 - 6

INFORMATION AVAILABLE FOR HELP

MAR101, Section 10, November 2015

S10 - 7

ANALYSIS MESSAGES
The problem begins with error messages like:

Failure to converge to tolerance (EXIT 3002)

Error encountered in stress recovery (EXIT 1009/1005)
The time step has become too small due to too many step cut-backs (EXIT 3009)
Unable to reduce the time step below the minimum value allowed (EXIT 3015)
Node on the boundary of a deformable body tried to slide out of surface definition in a
contact analysis (EXIT 2400)

The main place to look at is the end of the OUTPUT file (.out)
A successful analysis looks like:
**************************************************************************
This is a successful completion to a Marc simulation,
indicating that no additional incremental data was
found and that the analysis is complete
**************************************************************************
Marc Exit number 3004

An unsuccessful analysis looks like:

**************************************************************************
Analysis has failed to converge to required convergence
tolerances. One of several error conditions has been
detected and the run aborted. The output will specify
additional messages
**************************************************************************
Marc Exit number 3002

MAR101, Section 10, November 2015

S10 - 8

Associated message
for EXIT number
EXIT Number

ANALYSIS MESSAGES

The output file (.out) contains

all messages and potentially
the results.

The information on the solver

is dependent on the solver
choice

The log file (.log) mainly

contains convergence
information.

The status file (.sts) contains

a summary.

A typical nonlinear output file section:

start of assembly cycle number is
wall time =
12.00

solver workspace needed for out-of-core matrix storage = 7612

solver workspace needed for in-core matrix storage = 10114
matrix solution will be in-core
start of matrix solution
singularity ratio

3.4185E-12

end of matrix solution

maximum residual force at node 7 degree of freedom 1 is equal to 1.156E-04

maximum reaction force at node 35 degree of freedom 2 is equal to 2.809E-01
residual convergence ratio 4.117E-04
maximum displacement change at node 3 degree of freedom 1 is equal to 1.013E-02
maximum displacement increment at node 3 degree of freedom 1 is equal to 1.013E-02
displacement convergence ratio 1.000E+00
failure to converge to tolerance
increment will be recycled
MAR101, Section 10, November 2015
Copyright 2015 MSC.Software Corporation

S10 - 9

THE STATUS FILE

Summary of the convergence
behavior of the analysis

Things to look out for:

Sudden jumps in the number of cycles
(what happened?)
Large number of separations throughout
(a general contact issue)
Large number of separations part way
(contact lost? Contact causing local
failure? Frictional forces overcome?)
Note: Using Segment-to-Segment,
separations are not shown in the status
file.
Large number of cut-backs throughout
(load increment too large)
Large number of cut-backs part-way (what
happened?)

MAR101, Section 10, November 2015

S10 - 10

TROUBLESHOOTING ANALYSIS
FAILURE

MAR101, Section 10, November 2015

S10 - 11

GETTING CLUES

Post-process what there is.

The deformed shape often gives

obvious clues as to why the
simulation is not converging.

Exaggerating the deformation is a

good way to pick up cracks in the
model or localized effects from
incorrect contact definition.

Animating with a reasonable

deformation exaggeration can also
be of help.

If convergence is not achieved in

the first increment, it can be very
helpful to specify that Marc
continues with the proceed when
not converged option.

This means that a POST results file

will be generated.
Even if an increment fails to
converge, it may provide a pointer to
the problem.

MAR101, Section 10, November 2015

S10 - 12

The figure shows contact occurring

on only two nodes before nonconvergence.

Perhaps the other part coming into

contact slips away afterwards?

Perhaps a more refined mesh on

the contacted area is needed?

GENERAL
Ensure consistent units are used throughout the model
Note: N, mm, Kg are not consistent.

Reduce the time step. There may be significant nonlinearity

occurring at the beginning.

Usual for contact

May suggest an incorrect yield value for material nonlinearity.
Buckling or significant rotation may have occurred.
Large element distortion.

Make sure that the automatic cut-back capability is invoked.

If using the fixed load incrementation, change to the adaptive
scheme and include the automatic criteria.
Check that each component of the
structure is restrained against rigid
body motion.
Boundary conditions are the interface
between the model and the rest of
the world.

MAR101, Section 10, November 2015

S10 - 13

Unstable

Stable

CONTACT
Set Contact Tolerance Bias to 0.9 (particularly for shell contact)
Set Contact Tolerance to 0.0
The rigid surface markers should always point towards the interior of the
rigid body. If it does not, Marc may not detect contact between the rigid
surface and the deformable body.
Contact can be lost or not found because of too large a load increment.
Refine the mesh in the area of contact.
Coarse meshes can produce single point contact and promote instability.
It is necessary to capture the contact interaction accurately if contact distribution is of
importance.

Analytical surface definition may be incorrect and causing bulbous

corner/edge contact surfaces.
Consider smoothing the surfaces in contact
if there are sharp features, e.g. insert a
radius instead of a sharp corner for corner
contact.
Rigid
Rigid
Initial indeterminate contact state
can lead to chatter model components
in contact where possible.
Deformable
Deformable
Remove friction
Incorrect

MAR101, Section 10, November 2015

S10 - 14

Correct

CONTACT
Review and reconcile any initial contact over-closures and
openings.
Nodes initially penetrating significantly past the contact zone will
be ignored.
If this situation occurs at the beginning of the analysis, node will
not be found.
If it occurs later, the increment will be recycled with a modified
time step.

MAR101, Section 10, November 2015

S10 - 15

CONTACT
If using the stress-free check, make sure that the resulting
elements are not distorted when the slave nodes are moved by
the program to lie on the master surface.

MAR101, Section 10, November 2015

S10 - 16

CONTACT using SEGMENT-to-SEGMENT

Penetration is too large increase the penalty or add
augmentation
Note default penalty given on a body-by-body basis in output, typically
multiply default penalty by 10

If the amount of penetration is growing during sliding contact and

the augmentation procedure isnt used, then a modified
calculation of the contact normal pressure can be activated using
FEATURE,8401
If there is sliding contact, but the amount of friction is too small increase the penalty factor to simulate sticking or activate the
augmentation procedure for friction. Typically, this penalty factor
should be increased by a factor of 10.
Alternatively, the penalty factor used to simulate sticking can be
derived from the current contact normal pressure using
FEATURE,11701.

MAR101, Section 10, November 2015

S10 - 17

CONTACT using SEGMENT-to-SEGMENT

If there is little to no penetration, but a very slow convergence
behavior is observed- the penalty factor is too large and should
be decreased. Typically, the penalty factor should be decreased
by a factor of 10.
If there is contact between deformable bodies with a large
difference in stiffness or if there are shell elements involved in the
calculation and they are mostly loaded in bending, then
FEATURE,12001 can be activated. This causes the default penalty
factor to be based on the softest of two contacting bodies,
without taking into account the shell thickness when determining
the characteristic length
If the augmentation procedure is activated and the number of
iterations is large for every increment - then the distance beyond
which an augmentation will be applied can be significantly
increased. If this is done, the augmentation procedure is still
applied, but will not cause extra iterations if the convergence
criteria are fulfilled.
MAR101, Section 10, November 2015
Copyright 2015 MSC.Software Corporation

S10 - 18

HYPERELASTIC MATERIAL DATA

Check the material stability for elastomer
materials (in experimental data fitting).
Check that the material data covers the
entire strain range:
This can cause elements inside out errors.
The analysis may not converge if any part of
the model experiences strains beyond the
stability limits of the material.

Revert to a lower order polynomial fit

(e.g. Single-term Mooney) in the
experimental data fitting.
When fitting experimental data,
engineering stress/strain data is
expected.
Any material model in which the tangent
stiffness is zero or negative will most
often cause convergence problems.

MAR101, Section 10, November 2015

S10 - 19

PLASTIC MATERIAL DATA

For elasto-plastic materials, Cauchy stress and

log plastic strain data is expected.
At large strains, there are significant
differences in how the stresses and strains are
defined.

Beware of data extrapolation: extend the work

hardening data sufficiently to cover the entire
strain range.
The large stress value in the table is the one
that is used if the specified range is exceeded.

If a perfect plasticity model experiences

convergence difficulties, use a more realistic
plasticity model with non-zero work hardening.

Any material model in which the tangent

stiffness is zero or negative will often cause
convergence problems.

MAR101, Section 10, November 2015

S10 - 20

INCLUDE REALITY
Make sure all appropriate nonlinearities are
included.
Some structures rely on stress stiffening
effects for stability and would require a large
displacement analysis.
Is geometric nonlinearity required? Large
deformations/rotations may be causing large
non-physical strains.
If large strains are present in the analysis, it is
likely, for many materials, that failure is also
present (e.g. plasticity).
It is possible for many elasto-plastic analyses to
be in the small strain environment but the
addition of a robust large strain option is
recommended.

MAR101, Section 10, November 2015

S10 - 21

ELEMENTS
Be aware of the element mechanisms
associated with reduced integration
elements.
Always specify assumed strain option for
fully integrated 2D and 3D solid elements to
eliminate over-stiff solutions in the presence
of bending.
Always specify the constant dilatation
option for fully integrated 2D and 3D solid
elements in large-strain plasticity to avoid
volumetric locking.
This is due to over-constraints resulting from the
incompressible nature of plastic deformation.
Alternatively, use reduced integration or Herrmann Breathing or Hourglass Mode
elements.

Marc uses the global X-axis as the axis of

symmetry (z) for axisymmetric elements.

MAR101, Section 10, November 2015

S10 - 22

ELEMENTS
Specify a mesh so that the
shape of the elements is
reasonable throughout the
entire analysis.
Anticipate how the mesh will
deform.
For example, make element
sides shorter in the direction
that will be elongated the most.

The figure shows the mesh before (top)

and after (bottom) deformation. Elements
on the left stretched more readily due to
plastic necking. The analyst anticipated this
and refined the mesh towards the left.
A uniform mesh would have produced a
poorer simulation.
MAR101, Section 10, November 2015
Copyright 2015 MSC.Software Corporation

S10 - 23

GENERAL
If the analysis is still recalcitrant, remove nonlinearities to try and
isolate the cause of the problem.
For contact analyses, this would mean changing all contact conditions to
GLUED.
For material nonlinearity, simply increase the failure criteria so that it cannot
be reached.
Turn on non-positive definite.
Turn on quasi-static inertial damping.

Some specific clues can be found by looking at the behavior of

the convergence ratios during the solution.
Some specific clues can be found by looking at the change in
singularity ratio during the solution. If it suddenly reduces, it is
likely that an instability is occurring

MAR101, Section 10, November 2015

S10 - 24

CONVERGENCE CRITERIA BEHAVIOUR

Monotonic Divergence:
Material failure, e.g. point loads/supports causing massive localised failure
Contact lost because of too large a load increment or wrong contact settings.
Refine the mesh in the area of contact definitions. Coarse meshes can produce
single point contact and promote instability.
Buckling has occurred without arc-length methods requested.
Reduce load step to reduce the amount of nonlinearity occurring in an increment.
Convergence criteria too slack?
Tighten the convergence criteria,
particularly for geometric nonlinearity the Convergence
Actual variation of
Criteria
convergence criteria
solution may be drifting too far from the
true equilibrium position.
Ideal variation of
convergence criteria

Convergence Tolerance

Number of iterations

MAR101, Section 10, November 2015

S10 - 25

CONVERGENCE CRITERIA BEHAVIOUR

Slow Convergence:
Not uncommon in contact analyses whilst contact is being established.
Convergence tolerance too tight?
Use full Newton-Raphson to obtain full quadratic convergence
characteristics.
Friction issues
Check the relative sliding velocity is an appropriate value (1-10%).
Use bilinear model.
Unfeasibly large friction coefficients (tangential chatter).

Elements (bars, beams, springs) that are simulating

stiff members round-off issues if their
stiffness's are arbitrarily large.
Evaluate stiffnesses from real
geometry and materials.
Follower force with the stiffness
contribution gives a better convergence
rate and may help in the presence
of large rotations.
Convergence
Criteria

Actual variation of
convergence criteria

Ideal variation of
convergence criteria

Convergence Tolerance

MAR101, Section 10, November 2015

S10 - 26

Number of iterations

CONVERGENCE CRITERIA BEHAVIOUR

Slow Convergence
a. Gap elements can produce slow
displacement norm convergence behaviour.
Both, the iterative and incremental
displacements associated with a high
stiffness spring are tiny.
This causes the displacement norm
calculation of:
(iterative displacement ) / (displacement
increment)
to produce extremely small numbers.
The changes occurring in the
displacement values are lost because of
machine precision.
b. Analytical contact surface definitions give a
continuous normal and better convergence
and would be better than a
discrete surface definition for a coarse
mesh.
c. A poorly conditioned system leads to
consistently slow convergence.
Large: Small element sizes
Stiff: Soft materials
Poor quality element shapes

MAR101, Section 10, November 2015

S10 - 27

Convergence
Criteria

Actual variation of
convergence criteria

Ideal variation of
convergence criteria

Convergence Tolerance

Number of iterations

CONVERGENCE CRITERIA BEHAVIOUR

Oscillating Convergence
Typical behavior
Friction oscillating between stick and slip , use bilinear friction, take larger
steps
Plasticity is moving on and off yield surface, take larger steps
Contact is oscillating between contact and separation increase separation
force / stress
Local Buckling is occurring, add damping
Convergence
Criteria

Actual variation of
convergence criteria

Ideal variation of
convergence criteria

Convergence Tolerance

Number of iterations

MAR101, Section 10, November 2015

S10 - 28

CONVERGENCE CRITERIA BEHAVIOUR

Oscillating Divergence
Buckling is occurring
Catastrophic material failure
Extreme friction parameters with a large
load increment.
An element mechanism may have been
excited in a reduced integration element.
The remedy is to use a fully integrated
element.

Convergence
Criteria

Actual variation of
convergence criteria

Ideal variation of
convergence criteria

Convergence Tolerance

Number of iterations

MAR101, Section 10, November 2015

S10 - 29

CONVERGENCE CRITERIA BEHAVIOUR

Fixed Non-Convergence:
If the concrete (or similar softening) material model is being used, slacken the
residual convergence norm.
Locking can occur in highly constrained structures in a materially nonlinear analysis
in which massive plastic strain is developed (plasticity is assumed incompressible).
The effective plastic strain magnitude should be displayed to check this.
Traditional linear tri/tet displacement elements are notoriously guilty of this
numerical phenomenon.
The Hermann and the reduced integration elements were designed to eliminate
such behaviour.
The very large pressures associated with hyperelastic analyses can cause illconditioning (and, hence, slow convergence) if the full initial stress stiffness matrix
Convergence
is selected.
Critieria
Actual variation of
convergence criteria
In this case, either the none or tensile only
options would be recommended.
The initial stress selection does not influence
the accuracy, just the rate of convergence.

Ideal variation of
convergence criteria

Convergence Tolerance

Number of iterations

MAR101, Section 10, November 2015

S10 - 30

CONVERGENCE CRITERIA BEHAVIOUR

Fixed Non-Convergence :
Elements that are of poor quality will give problems.
Significant twist, skewness or distortion represents a serious degrading of
performance in most cases.
Aspect ratios greater than 1:10 (not so crucial).

This is particularly so if the elements are subject to large stress gradients of

sufficient magnitude to cause material failure or large deformations.
Convergence criteria too tight for the analysis.

Convergence
Critieria
Actual variation of
convergence criteria

Ideal variation of
convergence criteria
Convergence Tolerance

Number of iterations

MAR101, Section 10, November 2015

S10 - 31

ANALYSIS FAILURE:
EXIT NUMBERS

MAR101, Section 10, November 2015

S10 - 32

NON-POSITIVE DEFINITE : EXIT 2004

The associated error message is:

**************************************************************************
The determinant of the stiffness matrix becomes zero or
negative when indicated node has been reached during the
Gaussian elimination phase of the solution process. This
means that the stiffness matrix is non-positive definite
**************************************************************************
Marc Exit number 2004

During forward elimination, the diagonal stiffness coefficients are

recorded to evaluate the singularity ratio:
R = min(Kkk(k) / Kkk(k-1))

Seen in the output file as

Singularity Ratio

1.7495E-01

There are three results:

All Kkk(k) And Kkk(k-1) Are Positive
Any Kkk(k) Are Zero
Any Kkk(k) Are Negative

MAR101, Section 10, November 2015

S10 - 33

NON-POSITIVE DEFINITE : EXIT 2004

All Kkk(k) And Kkk(k-1) Are Positive (Positive Pivot)
Indicates a positive definite matrix.
The singularity ratio indicates the extent of the loss of accuracy during the elimination
process.
The number of digits lost during the elimination process is approximately equal to log10R, so the smallest value that would be acceptable is around 1e-6.
As the singularity ratio becomes smaller, the stiffness matrix is said to be poorly
conditioned (ill-conditioned).
It is checked by monitoring the Singularity Ratio:
Singularity Ratio
Singularity Ratio

1.7495E-01
4.3578E-23

Poor conditioning occurs because of large variations in the magnitude of diagonal

stiffness terms, for example:
Large stiff elements being connected to small less stiff elements (stiff beam being used to
transfer load into the structure)
Elements with highly disparate stiffness. For example, a beam element may have a bending
stiffness that, in order of magnitude, is less than its axial stiffness.
Cantilever beam has axial (~EA/L) and shear/rotational stiffness (~12EI/L3).

MAR101, Section 10, November 2015

S10 - 34

NON-POSITIVE DEFINITE : EXIT 2004

Any Kkk(k) Are Zero (Zero Pivot)
Indicates a singular non-positive definite matrix, i.e. rigid body modes are
present in the structure.
It is triggered by a Singularity Ratio = 0.0 and indicated with:
Error: non-positive definite system at node 15 degree of freedom 2

The multifrontal direct sparse solver automatically detects zero stiffness

coefficients and adds in small, artificial terms to permit the solution to
continue.
useful to see the deformation and obtain some clues as to the reason
for the singularity but, if the singularity is not spotted, the results may be
significantly inaccurate.

MAR101, Section 10, November 2015

S10 - 35

NON-POSITIVE DEFINITE : EXIT 2004

Any Kkk(k) Are Negative (Negative Pivot)
Indicates a non-singular, non-positive definite matrix.
It is triggered by a negative Singularity Ratio and indicated with:
Error: non-positive definite system at node 15 degree of freedom 2

The instability of a well-conditioned stiffness matrix originates from bifurcation and limit point, for
example:
Load
Load

Cstiff>0
Pivmin<0

limit point

Bifurcation point
Cstiff>0
Pivmin>0

Cstiff>0
Pivmin>0

Displacement

Cstiff<0
Pivmin<0

Displacement

Such situations can indicate real buckling and mandate the use of arc-length methods.
For example, an axially loaded straight strut will generate one negative pivot if loaded in a
geometrically nonlinear analysis to just beyond the first buckling load. Without additional side
perturbation, the strut will remain straight and resist the buckling path offered at this point.
Further increase of the load to just beyond the second buckling load will generate two negative
pivots and so on.
However, they can just as easily be the result of errors in modelling.

MAR101, Section 10, November 2015

S10 - 36

NON-POSITIVE DEFINITE : EXIT 2004

Remedies: Meshing Aspect
The aspect ratio of some elements is greater than the recommended limits.
Some element shapes are too distorted. This refers particularly to the curvature of element sides
and the central positioning of the mid-side nodes for higher order elements.
Coarse mesh refinement in areas remote from structural support conditions in significantly flexible
structures can cause conditioning problems.
Element mechanisms may have been excited by the loading patterns that may be eliminated by
invoking the full integration rule for these elements - thin shell elements are known to be prone to
such mechanisms in the case of very thin, curved surface analyses. Also membrane elements.
Rigid body (particularly torsional) motion may occur when connecting beam and shell elements to
continuum elements due to insufficient additional restraint.
Multiple bar elements used independently and without the use of a geometrically nonlinear
analysis to generate stress stiffening are prone to zero pivots, for example, a simple cantilever
beam.
Bars have no transverse (shear) stiffness and are particularly useful for modeling reinforcement
bars or tie linkages where there is no moment connectivity. These elements will not present any
difficulties when used in conjunction with other plane elements (shells, plates, etc.) since the
transverse stiffness required to prevent a numerical mechanism will be contributed from the
underlying surface elements.

MAR101, Section 10, November 2015

S10 - 37

NON-POSITIVE DEFINITE : EXIT 2004

Remedies: Geometric Properties Aspects
Zero magnitude for shear area or torsional constant parameters in the geometric
properties for beams.
Defining incompatible 1st and 2nd moment section properties for beams.

Remedies: Material Aspects

Ensure the same set of units are used to define the nodal coordinates and material
properties.
Incorrect nonlinear material parameters such as a zero yield stress or significant
variations in the magnitude of multi-hardening curves.
Ill-conditioning may occur in highly contained, large strain analyses using hyperelastic
material models turn off the initial stress-stiffness contributions.
In hyperelastic analyses, the strain state may be in a region where the input data for
the strain energy function is invalid.
It may also be caused by incorrect material properties (for example, negative yield
stress ). This situation may arise through temperature dependence of properties.

MAR101, Section 10, November 2015

S10 - 38

NON-POSITIVE DEFINITE : EXIT 2004

Remedies: Support Nodes
Are supports defined and assigned? The structure must be restrained
against free body translation and rotation (except for dynamic analyses).
For beams, the problem could be with unrestrained torsional motion.
All nodes of axisymmetric elements lying on the axis of symmetry must be
restrained to prevent any radial displacement across the symmetry axis.
Mentat and Patran do not create such center line supports automatically.
The axis of symmetry for Marc is assumed to be the global X-axis.

Remedies: Modelling Integrity

The mesh contains gaps or has discontinuities in the connection of the
elements.

Remedies: Contact
Is there any other restraint to body motion other than contact conditions?

MAR101, Section 10, November 2015

S10 - 39

ELEMENT ERRORS: EXIT 1005/1009

Errors during stiffness or mass matrix generation. The output
will reveal which element has a particular problem
Error encountered in stress recovery
Examples of these specific errors and the corresponding element
numbers are as follows:

*** Error: Element 10 Inside Out at Point 3

*** Error Element 10 has zero length
*** Error - Shell element collapsed to a line
*** Error Element 10 has zero area
*** Error - Element 10 is too distorted

Indicates an element shape problem that will prevent the element

stiffness or mass matrix from being evaluated at all.
It can be helpfully discussed in terms of the Jacobian matrix.

MAR101, Section 10, November 2015

S10 - 40

ELEMENT ERRORS : EXIT 1005

The Jacobian Matrix
Important component of the isoparametric element formulation - commonly
denoted as [J].
Consider a 3-noded bar element with the following natural coordinate
system:

The coordinate variable x (xi) thus varies between -1 and +1 and represents
a normalized coordinate system.
The local coordinate system for the same element, however, is specified as
x
x=0

x=L/2

x=L

The coordinate variable now varies in the range 0 to L.

A correspondence between the two is needed in order to carry out
numerical integration effectively.
MAR101, Section 10, November 2015
Copyright 2015 MSC.Software Corporation

S10 - 41

ELEMENT ERRORS: EXIT 1005

The Jacobian Matrix
When formulating finite elements, it is necessary to obtain the results for the
derivative:

d d d

dx dx d

dx
J
d

The Jacobian is then defined as

d
1 d

dx J d

Giving the correspondence

The Jacobian thus relates the natural coordinate derivatives to the corresponding
local coordinate derivatives.
The isoparametric formulation relies on there being a one-to one relationship
between the natural and the local coordinate systems (and their derivatives).
That is, there is a unique mapping between the two systems.
Determinants of this matrix equal or close to zero imply a collapsing element and
automatically invoke an error message.
MAR101, Section 10, November 2015
Copyright 2015 MSC.Software Corporation

S10 - 42

ELEMENT ERROR: EXIT 1005

The Jacobian Matrix
Unacceptable element shapes:

Interior Angle Greater than 180o

Overlapping Area of Element Geometry

Incorrect Location of Midside Node

Extremely Distorted Quadrilateral

For non-zero, positive values, the Jacobian determinant represents a measure of the
Gauss point volume (or length/area for 1D and 2D element types respectively):

dV dx dy dz [ J ] d d d
MAR101, Section 10, November 2015
Copyright 2015 MSC.Software Corporation

S10 - 43

ELEMENT ERRORS: EXIT 1005

The Jacobian Matrix
If this error occurs during the first assembly of the stiffness matrix, it is most likely
due to data input errors related to element shape or node ordering.
During later increments, it is likely to be due to excessive deformation in an element
(for geometrically nonlinear analyses only) as a result of load increments that are too
large.
The remedy is to reduce the load that is applied in each increment and to ensure that the
automatic cut-back facility is invoked.
Note that these errors can occur during the iterative process so that it is not always possible to
visualise the excessive deformation. In this case, it is necessary to check the overall material
behaviour and magnitude of the incremental loads.

Incorrect definition of the two/three dimensional continuum and plate elements.

These element types require an counter-clockwise node numbering sequence. The remedy is
to reverse the ordering of the elements (Mentat: Mesh Generation> Check> Upside Down)
Note that the local axis system of the elements may be viewed using Mentat: Plot> Draw>
Faces (with Elements> Solids unselected. Alternatively, Mesh Generation> Check> ID
Backfaces) will indicate the same information with filled contours.

When performing an elastomer (rubber) analysis, it is possible to obtain this error if

specified magnitude of the bulk modulus is too low and/or the incremental load level
is too high. In both cases, the incompressibility of the material may be insufficient to
resist the applied load level and, because geometric nonlinearity is typically required
with such materials, the elements can be inverted.
MAR101, Section 10, November 2015
Copyright 2015 MSC.Software Corporation

S10 - 44

EXERCISE
Perform Workshop 11: Contact Analysis to Generate ForceDeflection of a Spring
Be sure to ask for help if there is anything you do not understand.

MAR101, Section 10, November 2015

S10 - 45

MAR101, Section 10, November 2015

S10 - 46

APPENDIX A
DEFINING THE CONTACT CONSTRAINTS

MAR101, Appendix A, November 2015

A-1

DEFORMABLE 2D STRESS ANALYSIS

Set up tying relation:

DnA = 1/2 (1-xA) DnB + 1/2 (1+xA) DnC d
where
n= displacement component in local y direction
(normal to segment BC)
x= natural coordinate along segment BC
MAR101, Appendix A, November 2015
Copyright 2015 MSC.Software Corporation

A-2

DEFORMABLE 2D HEAT TRANSFER ANALYSIS

Define heat fluxes between A and A and B and B:

qAA = hAA (TA TA) areaAA
qBB = hBB (TB TB) areaBB
where
h= film coefficient
T= temperature
MAR101, Appendix A, November 2015
Copyright 2015 MSC.Software Corporation

A-3

DEFORMABLE 2D HEAT TRANSFER ANALYSIS

For plane stress and plane strain elements:

For axisymmetric elements:

Eliminate temperatures TA and TB:

= 1/2 (1

where
x= natural coordinates along segments CD and DE

MAR101, Appendix A, November 2015

A-4

DEFORMABLE 3D STRESS ANALYSIS

Set up tying relation

where
w = displacement component in local z direction, corresponding to the normal
of segment BCDE

x, h = natural coordinates on segment BCDE

MAR101, Appendix A, November 2015

A-5

DEFORMABLE 3D HEAT TRANSFER ANALYSIS

Define heat fluxes between A and A, B and B, C and C, D and D:

MAR101, Appendix A, November 2015

A-6

where
h = film coefficient
T = temperature

DEFORMABLE 3D HEAT TRANSFER ANALYSIS

Eliminate temperatures TA, TB, TC and TD using:=

x, h = natural coordinates on segments EFML, etc.

MAR101, Appendix A, November 2015

A-7

RIGID 2D STRESS ANALYSIS

Transformation to local system (x, y),

x is tangential and y is normal to the contacted segment or contacted curve
(discrete or analytical description).

Define prescribed displacement component

Dnrigidbody = displacement component in local y direction following

rigid body movement
MAR101, Appendix A, November 2015
Copyright 2015 MSC.Software Corporation

A-8

RIGID 2D HEAT TRANSFER ANALYSIS

Define heat fluxes between A and B, respectively, and the
contacted body.

MAR101, Appendix A, November 2015

A-9

RIGID 2D HEAT TRANSFER ANALYSIS

For plane stress and plane strain elements:

For axisymmetric elements:

Notice that C, D and E have the same temperature.

For rigid bodies allowing heat transfer, the procedure is similar to
deformable contact with heat transfer.

MAR101, Appendix A, November 2015

A - 10

RIGID 3D STRESS ANALYSIS

Transformation to local system (x, y,z)
x and y are tangential and z is normal to contacted segment or
contacted surface
Define prescribed displacement component

Dwrigidbody = displacement component in local z direction

following rigid body movement

MAR101, Appendix A, November 2015

A - 11

RIGID 3D HEAT TRANSFER ANALYSIS

Define heat fluxes between A, B, C and D, respectively, and the
contacted body.

MAR101, Appendix A, November 2015

A - 12

RIGID 3D HEAT TRANSFER ANALYSIS

Continue:
Where
h = film coefficient
T = temperature

And

Notice that E, F, G, H, I, J, K, L and M have the same temperature.

For rigid bodies allowing heat transfer, the procedure is similar to
deformable contact with heat transfer.
MAR101, Appendix A, November 2015
Copyright 2015 MSC.Software Corporation

A - 13

MAR101, Appendix A, November 2015

A - 14

A User-Material Subroutine Incorporating Single Crystal Plasticity in The Abaqus Finite Element Program
No ratings yet
A User-Material Subroutine Incorporating Single Crystal Plasticity in The Abaqus Finite Element Program
47 pages
Total NAS102
No ratings yet
Total NAS102
363 pages
Multiscale Modeling of Composite Materials and Structures With ANSYS
100% (1)
Multiscale Modeling of Composite Materials and Structures With ANSYS
12 pages
MSC Software - Presentation
No ratings yet
MSC Software - Presentation
22 pages
Accuracy of Finite Element Model-Based Multi-Organ Deformable Image Registration
No ratings yet
Accuracy of Finite Element Model-Based Multi-Organ Deformable Image Registration
14 pages
MSC Marc Course
50% (2)
MSC Marc Course
368 pages
Mechanical Behavior of Materials 01 PDF
100% (2)
Mechanical Behavior of Materials 01 PDF
54 pages
INT AC19 Training Series Vol 1
No ratings yet
INT AC19 Training Series Vol 1
43 pages
Developing Wmi Solutions A Guide To Windows Management Instrumentation.9780201616132.25091
100% (2)
Developing Wmi Solutions A Guide To Windows Management Instrumentation.9780201616132.25091
652 pages
Benchmarks PDF
100% (1)
Benchmarks PDF
1,761 pages
ANSYS Workbench 2021 R1: A Tutorial Approach, 4th Edition
From Everand
ANSYS Workbench 2021 R1: A Tutorial Approach, 4th Edition
Prof. Sham Tickoo
No ratings yet
Modal Transient Response
No ratings yet
Modal Transient Response
24 pages
AWS90 Ch05 Modal
No ratings yet
AWS90 Ch05 Modal
24 pages
MSC Patran Gap Elements
100% (1)
MSC Patran Gap Elements
22 pages
1972-Large Deformation Isotropic Elasticity - On The Correlation of Theory and Experiment - Ogden
100% (1)
1972-Large Deformation Isotropic Elasticity - On The Correlation of Theory and Experiment - Ogden
21 pages
Free Vibration Analysis of Composite Laminated Beams
No ratings yet
Free Vibration Analysis of Composite Laminated Beams
18 pages
5 Mscfatigue Tutor
No ratings yet
5 Mscfatigue Tutor
24 pages
Nas101 S+N+B Workbook PDF
100% (1)
Nas101 S+N+B Workbook PDF
372 pages
Visco Plasticity
No ratings yet
Visco Plasticity
18 pages
Models - Sme.cohesive Zone Debonding
No ratings yet
Models - Sme.cohesive Zone Debonding
18 pages
COMSOL Cohesive Zone Debonding
No ratings yet
COMSOL Cohesive Zone Debonding
22 pages
Czm-On The Practical Application of The Cohesive Model
No ratings yet
Czm-On The Practical Application of The Cohesive Model
25 pages
T1793 101
No ratings yet
T1793 101
10 pages
Aws90 Ws Fatigue A12
No ratings yet
Aws90 Ws Fatigue A12
34 pages
An Integrated Approach For Transient Rolling of Tires: Abaqus Technology Brief
No ratings yet
An Integrated Approach For Transient Rolling of Tires: Abaqus Technology Brief
6 pages
19 COMNE MSC Abaqus Marc PerformanceComparison
No ratings yet
19 COMNE MSC Abaqus Marc PerformanceComparison
8 pages
General Preprocessing: Chapter Three
No ratings yet
General Preprocessing: Chapter Three
72 pages
Calculation of Contact Stress
100% (1)
Calculation of Contact Stress
23 pages
Understanding Frequency Domain Viscoelasticity in Abaqus
No ratings yet
Understanding Frequency Domain Viscoelasticity in Abaqus
13 pages
NCodeDL Intro 17.0 L01 Lecture Slides Fatigue Introduction
No ratings yet
NCodeDL Intro 17.0 L01 Lecture Slides Fatigue Introduction
28 pages
Analysis of Wear Using LS - DYNA
No ratings yet
Analysis of Wear Using LS - DYNA
10 pages
Workshop 4 Minimizing The Weight of A Composite Wing
No ratings yet
Workshop 4 Minimizing The Weight of A Composite Wing
20 pages
Modeling Surface-Bonded Structures With ABAQUS Cohesive Elements: Beam-Type Solutions
No ratings yet
Modeling Surface-Bonded Structures With ABAQUS Cohesive Elements: Beam-Type Solutions
27 pages
Tutorial 12 VCCT
No ratings yet
Tutorial 12 VCCT
15 pages
Workshop 7 Thermal Contact Resistance: WS7-1 NAS104, Workshop 7, March 2004 2004 MSC - Software Corporation
No ratings yet
Workshop 7 Thermal Contact Resistance: WS7-1 NAS104, Workshop 7, March 2004 2004 MSC - Software Corporation
24 pages
3398-Lingap Oct2001 Update
No ratings yet
3398-Lingap Oct2001 Update
21 pages
Bracket Model - Random Response
No ratings yet
Bracket Model - Random Response
16 pages
Application of User Defined Subroutine U PDF
100% (2)
Application of User Defined Subroutine U PDF
58 pages
Vibration Fatigue Analysis in MSC - Nastran: JULY 2011
No ratings yet
Vibration Fatigue Analysis in MSC - Nastran: JULY 2011
23 pages
Modal FRA With Nastran
No ratings yet
Modal FRA With Nastran
13 pages
Pyton Ls Dyna
No ratings yet
Pyton Ls Dyna
9 pages
Abaqus Interaction Module
No ratings yet
Abaqus Interaction Module
10 pages
Lec12 PDF
No ratings yet
Lec12 PDF
25 pages
Tutorial I2 - Beam
No ratings yet
Tutorial I2 - Beam
14 pages
General Abaqus Checklist
0% (1)
General Abaqus Checklist
26 pages
Evaluation of Damping in Dynamic Analysis of Structures
No ratings yet
Evaluation of Damping in Dynamic Analysis of Structures
4 pages
AM2007 ChabocheABST
No ratings yet
AM2007 ChabocheABST
2 pages
FEA Benchmark For Dynamic Analysis of Perforated Plates
No ratings yet
FEA Benchmark For Dynamic Analysis of Perforated Plates
18 pages
Ball To Plate Impact With Element Tutorial 9: Ball To Plate Impact With Element Deletion Ball To Plate Impact With Element
No ratings yet
Ball To Plate Impact With Element Tutorial 9: Ball To Plate Impact With Element Deletion Ball To Plate Impact With Element
8 pages
2010 Elementbibliothek LSDyna Cadfem PDF
No ratings yet
2010 Elementbibliothek LSDyna Cadfem PDF
78 pages
Stress Strain Curve True Multilinear Formula
No ratings yet
Stress Strain Curve True Multilinear Formula
58 pages
Altair - RADIOSS - Moscow - 160311 - 001a PDF
No ratings yet
Altair - RADIOSS - Moscow - 160311 - 001a PDF
49 pages
ANSYS 19.2 Worksheet
No ratings yet
ANSYS 19.2 Worksheet
40 pages
Composites Summary
No ratings yet
Composites Summary
30 pages
Finite Element Method
From Everand
Finite Element Method
Gouri Dhatt
1/5 (1)
Damage Mechanics in Metal Forming: Advanced Modeling and Numerical Simulation
From Everand
Damage Mechanics in Metal Forming: Advanced Modeling and Numerical Simulation
Khemais Saanouni
4/5 (1)
Sec00 Mar101 Toc
No ratings yet
Sec00 Mar101 Toc
6 pages
Marc 2008 r1 User's Guide
100% (1)
Marc 2008 r1 User's Guide
1,606 pages
Docslide - Us Mar 102
100% (1)
Docslide - Us Mar 102
460 pages
Marc 2008 r1 Volume A: Theory and User Information
No ratings yet
Marc 2008 r1 Volume A: Theory and User Information
794 pages
Mar 104
No ratings yet
Mar 104
1,845 pages
Advanced Modeling To Simulate Real-World Behavior: Datasheet
No ratings yet
Advanced Modeling To Simulate Real-World Behavior: Datasheet
2 pages
MD R2 Nastran Release Guide
No ratings yet
MD R2 Nastran Release Guide
276 pages
Laser Welding On GTD-111
No ratings yet
Laser Welding On GTD-111
5 pages
Thermocouple Type K (Chromel-Alumel) Sensitivity Temperature Range
No ratings yet
Thermocouple Type K (Chromel-Alumel) Sensitivity Temperature Range
1 page
Introduction To Magnesium PDF
No ratings yet
Introduction To Magnesium PDF
12 pages
Seminar AIU59 - 8-7-2016
No ratings yet
Seminar AIU59 - 8-7-2016
28 pages
Hot Deformation of Microalloyed Steel
100% (1)
Hot Deformation of Microalloyed Steel
56 pages
Surface Treatment For Erosion Resistance: 28 July 2015
No ratings yet
Surface Treatment For Erosion Resistance: 28 July 2015
59 pages
CNS MCQ Quiz
No ratings yet
CNS MCQ Quiz
2 pages
04
No ratings yet
04
7 pages
Anti Homomorphism in Fuzzy Subgroups: Dr. S. Subramanian, (M.Sc.,M.Phil., B.Ed., PGDCA., PH.D.)
No ratings yet
Anti Homomorphism in Fuzzy Subgroups: Dr. S. Subramanian, (M.Sc.,M.Phil., B.Ed., PGDCA., PH.D.)
7 pages
How To Setup The Sony SR PC4 For Ethernet Transfer 12-02-01
No ratings yet
How To Setup The Sony SR PC4 For Ethernet Transfer 12-02-01
4 pages
OS Lecture 05 Synchronization
No ratings yet
OS Lecture 05 Synchronization
28 pages
SRM University Applied Mathematics MA513 Syllabus
No ratings yet
SRM University Applied Mathematics MA513 Syllabus
1 page
Radwag Ps r2
No ratings yet
Radwag Ps r2
3 pages
Internet of Things: The Next Big Thing: Country Adoption
No ratings yet
Internet of Things: The Next Big Thing: Country Adoption
4 pages
Demo 50 Disha SSC Mathematics Guide 2022 Edition English Medium
No ratings yet
Demo 50 Disha SSC Mathematics Guide 2022 Edition English Medium
50 pages
Allama Iqbal Open University, Islamabad (Department of Computer Science) Warning
No ratings yet
Allama Iqbal Open University, Islamabad (Department of Computer Science) Warning
2 pages
Chapter 4_ Operators in Java _ Solutions for Class 10 ICSE Logix Kips Computer Applications with BlueJ Java _ KnowledgeBoat
No ratings yet
Chapter 4_ Operators in Java _ Solutions for Class 10 ICSE Logix Kips Computer Applications with BlueJ Java _ KnowledgeBoat
17 pages
1 PB PDF
No ratings yet
1 PB PDF
15 pages
Menu - 634641194757402500 - CS6106 Compiler Design Lab Assignments
No ratings yet
Menu - 634641194757402500 - CS6106 Compiler Design Lab Assignments
3 pages
Dorathi Paranjothi Resume
No ratings yet
Dorathi Paranjothi Resume
6 pages
Pat Ds350c
No ratings yet
Pat Ds350c
68 pages
User Manual - PTRC - View Historic Returns - From 1st April-2016
No ratings yet
User Manual - PTRC - View Historic Returns - From 1st April-2016
8 pages
SPCC Cheat
No ratings yet
SPCC Cheat
1 page
SecondHand Bike Preference - Marketing
No ratings yet
SecondHand Bike Preference - Marketing
49 pages
List Members List Members Catch Automatic Code Generation
No ratings yet
List Members List Members Catch Automatic Code Generation
87 pages
Switching Interview Questions
No ratings yet
Switching Interview Questions
3 pages
See What We Can Do For
No ratings yet
See What We Can Do For
17 pages
YouTube Video Lectures
No ratings yet
YouTube Video Lectures
2 pages
CNS (KCS074) CT1 2022-23
No ratings yet
CNS (KCS074) CT1 2022-23
2 pages
Opc Ua
No ratings yet
Opc Ua
6 pages
Ipocc Taskflow Editor en
No ratings yet
Ipocc Taskflow Editor en
126 pages
Excel2007 VBA
No ratings yet
Excel2007 VBA
107 pages
Elementary Graph Theory: Robin Truax March 2020
No ratings yet
Elementary Graph Theory: Robin Truax March 2020
15 pages
Dali Wallstation Specsheet FWB
No ratings yet
Dali Wallstation Specsheet FWB
3 pages