Table of Contents

About This Book................................................................................................... 4

Acknowledgement ............................................................................................... 8
1. Introduction to Fatigue ............................................................................... 10
1.1 Why Fatigue Analysis .................................................................................. 10
1.2 Static, Dynamic and Fatigue Analysis Comparison ..................................... 11
1.3 Definitions ................................................................................................... 11
1.4 Cyclic Loading Parameters .......................................................................... 14
1.5 Fatigue Fracture Growth Mechanism ......................................................... 15
1.6 Factors Influencing Fatigue Life .................................................................. 16
1.7 Various Approaches in Fatigue Analysis ..................................................... 19
2 Fatigue Concepts ........................................................................................ 21
2.1 Stress-Life Approach ................................................................................... 21
2.2 Strain-Life Approach ................................................................................... 23
2.3 Mean Stress Influence................................................................................. 27
2.4 Damage Model ............................................................................................ 29
2.5 Rainflow Cycle Counting ............................................................................. 30
2.6 Additional Correction Parameters .............................................................. 32
2.7 Fatigue Results ............................................................................................ 34
3 Uniaxial Fatigue Analysis ............................................................................ 36
3.1 Fatigue Loading ........................................................................................... 37
3.2 Fatigue Parameters (FATPARM).................................................................. 42
3.3 Fatigue Materials & Properties ................................................................... 45
3.4 Fatigue Load Step & Analysis Run ............................................................... 51
3.5 Fatigue Results ............................................................................................ 52
3.6 Tutorial: Uniaxial Stress-Life Fatigue Analysis of a Spring Link ................... 53
3.7 Tutorial: Uniaxial Strain-Life Fatigue Analysis of a Control Arm ................. 63
4 Multiaxial Fatigue Analysis.......................................................................... 72
4.1 Multiaxial Stress-Life Technical Background............................................... 73
4.1.1 Damage Models .................................................................................. 73
4.1.2 Minimum Circumscribed Ellipse ......................................................... 75

2
 4.1.3 Findley ................................................................................................. 76
4.1.4 Goodman ............................................................................................ 79
4.1.5 Dang Van ............................................................................................. 80
4.2 Damage Models in OptiStruct ..................................................................... 81
4.3 Analysis Procedure ...................................................................................... 82
4.4 Tutorial: Multiaxial Fatigue Analysis of a Knuckle ...................................... 84
5 Factor of Safety Analysis ............................................................................. 96
5.1 FOS Analysis Setup ...................................................................................... 97
5.2 Tutorial: FOS Analysis of a Spring Link ...................................................... 100
6 Seam Weld Fatigue Analysis ...................................................................... 105
6.1 Seam Weld Elements ................................................................................ 106
6.2 Seam Weld Properties Setup .................................................................... 111
6.3 Tutorial: Weld Fatigue Analysis of a Seam Welded Frame ....................... 118
7 Spot Weld Analysis ................................................................................... 127
7.1 Stress calculation ...................................................................................... 128
7.2 Spot Weld Fatigue Analysis Setup............................................................. 129
7.3 Tutorial: Spot Weld Fatigue Analysis of a Welded Beam.......................... 133
8 Vibration Fatigue ....................................................................................... 141
8.1 Sine Sweep Fatigue Analysis ..................................................................... 142
8.2 Tutorial: Fatigue Analysis under sine sweep loading ................................ 146
8.3 Random Response Fatigue Analysis.......................................................... 152
8.4 Tutorial: Fatigue Analysis of a cantilevered part under random loading . 155
9 HyperLife.................................................................................................. 163
9.1 Fatigue Capabilities in HyperLife............................................................... 163
9.2 Benefits of HyperLife................................................................................. 164
9.3 Fatigue methods in HyperLife ................................................................... 164
9.4 HyperLife Workflow .................................................................................. 165
10 eFatigue ................................................................................................... 167
10.1 Fatigue Calculator ..................................................................................... 168
10.1.1 Constant Amplitude Stress-Life Analysis .......................................... 169
Appendix A ...................................................................................................... 173
Appendix B ...................................................................................................... 174

3
About This Book
This study guide aims to provide a fundamental to advanced approach into the
exciting and challenging world of Structural Analysis. The focus will be on aspects of
Fatigue Analysis. As with our other eBooks we have deliberately kept the theoretical
aspects as short as possible.
The tool of choice used in this book is OptiStruct. Altair ® OptiStruct® is an industry
proven, modern structural analysis solver for linear and nonlinear structural problems
under static and dynamic loadings. OptiStruct is used by thousands of companies
worldwide to analyze and optimize structures for their strength, durability and NVH
(noise, vibration and harshness) characteristics.
In this eBook, we will describe in some detail, how to perform a Fatigue analysis with
different approaches and Optimization including:
• Uni-Axial and Multi-Axial Fatigue Analysis
• Factor of Safety Analysis (Dang Van)
• Weld Fatigue Analysis

Please note that a commercially released software is a living “thing” and so at every
release (major or point release) new methods, new functions are added along with
improvement to existing methods. This document is written using OptiStruct
2017.2.3. Any feedback helping to improve the quality of this book would be very
much appreciated.

Thank you very much.

Dr. Matthias Goelke
On behalf of the Altair University Team

4
Model Files
The models referenced in this book can be downloaded using the link provided in the
exercises, respectively.

Software
Obviously, to practice the “Fatigue Analysis Tutorials” you need to have access to
HyperWorks Student Edition 2019. As a student, you are eligible to download and
install the free Student Edition: https://altairuniversity.com/free-altair-student-
edition/

Support
In case you encounter issues (during installation
and also on how to utilize OptiStruct) post your
question in the moderated Support Forum
https://forum.altairhyperworks.com
It’s an active forum with several thousands of
posts – moderated by Altair experts!

5
Free eBooks
In case you are interested in more details about the “things” happening in the
background we recommend our free eBooks
https://altairuniversity.com/free-ebooks-2

6
Many different eLearning courses are available for free in the Learning & Certification
Program https://certification.altairuniversity.com/

For OptiStruct Fatigue Analysis, the prerequisite (or recommended) course is

Structural and Dynamic Analysis:
Learn Structural Analysis with OptiStruct
https://certification.altairuniversity.com/course/view.php?id=71

and

Learn Linear Dynamics with OptiStruct

https://certification.altairuniversity.com/course/view.php?id=100

7
Acknowledgement
A very special Thank You goes to all the many colleagues who contributed in different
ways:
The Editors, Gabriel Stankiewicz, Premanand Suryavanshi and Koushik
Chandrashekhar for creating and testing chapters contained in this eBook. Prakash
Pagadala and Obul Reddy Burra for helpful discussions and explanations. Rahul Rajan
for reviewing and compiling “Appendix” section into the book.
Rahul Ponginan for overall review of the book. For sure, your feedback and
suggestions had a significant impact on the “shape” and content of this book.
Junji Saiki, Warren Dias, Hariprasad Venkate Gowda and Ujwal Patnaik from
HyperWorks Core Development.

Mike Heskitt, Sean Putman, Nelson Dias, Pavan Kumar & Dev Anand for all the
support.
Nitin Gokhale from “Finite to Infinite” for all his passion about CAE, the inspirational
collaboration, and the friendship.
The entire OptiStruct Documentation team for putting together 1000’s of pages of
documentation and recently released OptiStruct verification and example section.
Lastly, the OptiStruct Development team deserves huge credit for their passion &
dedication! It is so exciting to see how OptiStruct has evolved throughout the last
couple of years.

Thank you very much.

Your Altair University Team

8
Disclaimer
Every effort has been made to keep the book free from technical as well as other mistakes.
However, publishers and authors will not be responsible for loss, damage in any form and
consequences arising directly or indirectly from the use of this book.
© 2020 Altair Engineering, Inc. All rights reserved. No part of this publication may be
reproduced, transmitted, transcribed, or translated to another language without the written
permission of Altair Engineering, Inc. To obtain this permission, write to the
attention Altair Engineering legal department at:
1820 E. Big Beaver, Troy, Michigan, USA, or call +1-248-614-2400.

9
1. Introduction to Fatigue
This chapter includes material from the book “Practical Finite Element Analysis
(From Finite to Infinite)”.

1.1 Why Fatigue Analysis

• Fatigue accounts for 90 % of service failures.
• Manufacturers give warranty on the components (in terms of kms or years).
Fatigue analysis helps in predicting life of the component in design phase itself.
Static or dynamic analysis can tell us about stress, displacement, acceleration etc.
but not how long the component will survive.
• Many a times static or dynamic analysis predicts location of failure not matching
with lab test or field failure and then analyst keep on thinking whether something
is wrong with boundary conditions or material properties or geometry of the
component. But when fatigue analysis is carried out using same static or dynamic
results, it reveals correct location of failure.
• Failures or crack usually initiates at surface. Life of the component depends on
surface condition (like grinding, induction hardening, shot preening etc.). Static or
dynamic analysis cannot take in to account these details while fatigue can.

What is fatigue?
Fatigue is the progressive and localized structural damage that occurs when a material
is subjected to cyclic loading. Fatigue is the most important failure mode for
mechanical components under cyclic loading. It is important to recognize that fatigue
failure is a probabilistic event, and that a good design against fatigue should involve
analyses, synthesis and testing.

10
1.2 Static, Dynamic and Fatigue Analysis
Comparison
Static Dynamic Fatigue
Force vary with respect to Force vary with respect to
Force is static (dead wt.)
time / frequency time / frequency
Calculations for life of the
Displacement is function of
Displacement is static structure when subjected to
time / frequency
repetitive load
Static & dynamic analysis
cannot tell how long the
component will survive for
given load. Also, there is no
No velocity & acceleration Velocity & acceleration
consideration for factors
due to constant or fixed develop due to variation of
like surface finish, heat
displacement displacement.
treatment, decarburizing,
alloying elements, realistic
representation of spot & arc
welds.
Damping force, inertia force
due to velocity &
Kx=F
acceleration
m x¨ + c x˙ + k x = F(t)
If force is doubled static &
dynamic stress will also be
Solution Time ↓ Solution Time ↑ doubled but fatigue life
would be reduced by a
10
factor of 2 .
Output – Stress,
Output – Stress, Displacement, Velocity, Output – Life of the
Displacement Acceleration with respect to component or safety factors
time/frequency

1.3 Definitions
What is Durability, Reliability and Fatigue?

Durability, Reliability and Fatigue are often loosely used for describing Fatigue related
analysis. There is slight difference in these three terms.

Durability describes overall life requirement, like to last for 2 years (or warranty
period).

Reliability includes a probability of failure, such as to have a 95 % chance of survival

(if we test 100 specimens, 95 will pass and 5 might fail).

11
Fatigue is the failure caused by application of repetitive load by the process of
initiation of cracks and growth.

What is Life?

Total Crack Life = Crack Initiation life + Crack growth Life

Crack Growth Life – Ductile Materials ↑

Brittle materials ↓

What is the criteria for transition from crack initiation to crack growth life?

Life till crack of the size 2 mm detected is crack initiation life and remaining life after
detection of crack is crack propagation or crack growth life.

12
S – N Curve :

- Life (abscissa) is always plotted on log scale while alternating stress on either linear
or log.

Low Cycle Fatigue (LCF): Life of component is less than 100000 cycles, applicable for
heavy duty application loading.

5
High Cycle Fatigue (HCF): Component subject to less sever loads and life > 10 cycles.

Infinite Life: Stress level below which material never fails is known as endurance limit
or fatigue limit. Never fails or infinite life is a relative term. For steel, test is stopped
6
after 2 * 10 cycles (in case if till then failure is not detected) and said to have infinite
life. This is the point where S-N curve slop changes and it becomes parallel to x-axis.

Unlike steel, non-ferrous alloys have no specific endurance limit (S-N curve never
become parallel to x-axis). Pseudo-endurance limit for these materials is stress value
8
corresponding to life = 5x10 cycles (something similar to proportionality limit for
brittle materials).

S-N curve shown above is based on constant amplitude rotating bending test (Shaft
subjected to pure alternating bending stress). Similar test could be conducted for
tension, compression, shear and torsional stress. Bending fatigue strength is higher
than tension / compression and torsional fatigue strength is the lowest.

13
Damage and Endurance Factor of Safety

• Damage is calculated for Low Cycle Fatigue (LCF) applications.

Damage = n/N = no. of cycles applied/ Total life

Damage < 1 ⇒ safe

Damage > 1 ⇒ fail

• Endurance Factor of safety is calculated for High Cycle Fatigue (HCF)

applications.
Endurance factor of safety = Endurance strength/ FE stress Amplitude

Endurance factor of safety < 1 ⇒ fail Endurance

factor of safety > 1 ⇒ safe

1.4 Cyclic Loading Parameters

Before we start with material fatigue theories, we need to understand the features of
cyclic loading, as they are the main part of equations presented later.

As visible on the picture above:

• One cycle of load is one segment of the cyclic load path, that when repeated
creates the complete load path, it is a piece of load path between two
maximum values or minimum values of load.

14
 • Stress range is the difference between maximum and minimum possible
stresses.
• Mean Stress is an average value of cyclic stress, calculated based on
maximum and minimum stress.
• Stress amplitude can be said to be a pulse or maximum deviation of stress
from a mean value (average) value – difference between maximum or
minimum value and mean stress.

NOTE: Stress amplitude is equal to half of stress range, therefore, cyclic load is
explicitly defined just by giving either stress range or amplitude.

1.5 Fatigue Fracture Growth Mechanism

During component’s life, the process of failure can last long for ductile materials and
can be sudden for brittle materials. Generally, failure due to fatigue is preceded by a
degradation procedure, which can be divided into four steps:
1) Crack nucleation
• Crack nucleation is a highly localized plastic deformation within the part
structure
• Crack nucleation is normally a surface phenomenon
• Crack nucleation is a stochastic process
2) Small crack growth in an elastic-plastic stress field
• Most of the fatigue life is consumed growing microcracks in the finite life
region
3) Macroscopic crack growth in a nominally elastic stress field
Crack growth takes place in one of three modes:
• Mode 1, tensile mode, opens the crack
• Mode 2, torsion mode, lengthens the crack by in-plane shear

15
 • Mode 3, also a torsion mode, lengthens the crack by out-of-plane shear

4) Final fracture

N = 1,000 N = 2,000 N = 10,000 N = 40,000

Image:
Ewing, J.A. and Humfrey, J.C. “The fracture of metals under repeated alterations of
stress”, Philosophical Transactions of the Royal Society, Vol. A200, 1903, 241-250

1.6 Factors Influencing Fatigue Life

Fatigue in a material will take place only if all of the following factors are present:
• Cyclic loading
• Tensile loading
• Plastic strain

Additionally, following factors are considered to influence fatigue strength (all these
factors can be set up in an OptiStruct fatigue analysis):

• Surface finish
Surface imperfections arising during manufacturing processes is a crucial factor
defining the fatigue life of components, this is because crack initiates at the surface.

16
Any discontinuity on the surface is usually a weak spot, which presents lower fatigue
strength than the nominal value.

OptiStruct allows definition of surface treatment due to: polishing, grounding,

machining, hot rolling and forging. These processes of course decrease nominal
fatigue strength, which is initially assumed to be at ideal surface conditions. OptiStruct
then applies a proper correction factor. The below image shows a sample surface
finish parameter for steel

• Surface treatment
Surface of the material can be specially treated to increase fatigue resistance. In
OptiStruct, following surface treatment processes can be taken into account:
nitriding, shot peening and cold rolling. These processes generally increase surface
compressive residual stresses and consequently increase fatigue life of a component.
If one of these processes is chosen, OptiStruct will apply appropriate correction factor,
which will increase the material’s fatigue life.

• Notch presence
Notch is a typical example of crack initiation location, due to stress concentration.
Notches can be of very small dimensions, which in fact would not be in the discretized

17
model. User is therefore able to manually input a correction factor for notch
consideration.

• Residual stresses
Residual compressive stresses can significantly improve a material’s fatigue life.

NOTE: The correction factor mentioned in the notch presence section, is generally
meant to be used when any type of an influential factor for fatigue needs to be
considered. This factor can be used instead of defining the finish and treatment
processes.

18
 1.7 Various Approaches in Fatigue Analysis

Fatigue Analysis

Stress life Strain life Approach: Crack Propagation Vibration

Approach: Approach: Approach
• Low Cycle Frequency
• High Cycle Fatigue • Fracture Fatigue:
Fatigue • Crack Initiation Mechanics • Resonance
• Total Life Life LEFM, EPFM effect
• Rate of crack Input data for
• First Fatigue • Developed in
growth Life left fatigue
analysis 1960’s Elastic
Could be used calculations is
method to be and plastic
in combination dynamic
strains
developed with strain life analysis results
• Strain Vs
• Stress and approach to based on
Reversals Data predict total life
strain elastic Transient,
(ε-N curve) Frequency
Uses stress Vs
domain or
No. of cycles
Power Spectral
plot (S-N Density input
curve)

Types of Fatigue Analysis with OptiStruct:

• Uniaxial Fatigue Analysis
Uniaxial Fatigue Analysis, using S-N (stress-life) and E-N (strain-life) approaches for
predicting the life (number of loading cycles) of a structure under cyclical loading may
be performed by using OptiStruct.

• Multiaxial Fatigue Analysis

Multiaxial Fatigue Analysis, using S-N (stress-life), E-N (strain-life), and Dang Van
Criterion (Factor of Safety) approaches for predicting the life (number of loading
cycles) of a structure under cyclical loading may be performed by using OptiStruct.

• Spot Weld Fatigue Analysis (New Feature)

Allows for the study of fatigue performance of spot welds in structures (Based on a
paper by Rupp et al)

19
• Seam Weld Fatigue Analysis (New Feature)
The method implemented in OptiStruct is based on a research paper Fatigue Life
Prediction of MAG-Welded Thin-Sheet Structurespublished by M. Fermér, M
Andréasson, and B Frodin.

• Random Response Fatigue Analysis

The study of fatigue life of structures under Random Loading.

• Sine Sweep Fatigue Analysis

The Study of fatigue life of structures under Sinusoidal loading

20
2 Fatigue Concepts
Now that you are familiar with cyclic load, you will understand that figures describing
material behavior are expressed as a function of load cycles (N). It is not really time
that plays a role in case of fatigue, but the number of loading cycles.

Based on the initially observed state of the material, fatigue solutions are generally
characterized in one of two ways for a homogenous material:

• Safe Life: The material is free of initial visible defects: Stress-Life (S-N) or
Strain-Life (E-N)
• Damage Tolerant: The material contains an initial defect or crack: da/dN

In the absence of a pre-existing defect, most of the life of a part is spent in microcrack
growth within the plastic strain-dominated region. This is fundamental to the
development of both the stress-life and strain-life fatigue solutions.

2.1 Stress-Life Approach

Stress-Life approach is based on a chart, describing nominal stress as a function of the
number of load cycles. This chart is specific for a given material and it is called
Wöhler’s curve. The relationship between material strength and number of cycles is
the single most important factor for fatigue analysis.

21
Chart line composed on a basis of empirical data
N(cycle) axis expressed in log scale provides
possibility of linear assumption

The mathematical relationship between strength and number cycles can be modelled
based on a logarithmic chart:

𝑏1
𝑆 = 𝑆1(𝑁𝑓 )
Parameters needed (in OptiStruct):
S1 (or SR1): fatigue strength coefficient (mathematical point on the chart, not related
to real material property)

Nf: fatigue cycles to failure

b1: fatigue strength exponent

Stress-Life approach should be used when:

• High-cycle fatigue is considered (this is assumed to be above 10 000 cycles)

• Stresses oscillate within the elastic range and no plastic strains are seen
except at the tips of fatigue cracks.

Major Assumptions:
• Most of the life is consumed in nucleating cracks
• Elastic deformation
• Nominal stresses and material strength control the fatigue life

22
 • Accurate determination of Kf (Stress Concentration Factor) for each geometry
and material

Advantages:
• Changes in material and geometry can easily be evaluated
• Large empirical database for steel specimens with standard notch shapes

Limitations:
• Does not account for notch root plasticity
• Mean stress effects are often in error
• Requires empirical Kf for predictive results

2.2 Strain-Life Approach

Strain-Life is based on a graph, which presents a logarithmic relationship between
elastic and plastic strain components, together with total strain and reversals to
failure (1 load cycle consists of 2 reversals):

Where (parameters defined in OptiStruct, MATFAT extension):

 f
 a =  ea +  pa = (2 N f )b +  f  (2 N f )c
E
σf´: fatigue strength coefficient
b: is the fatigue strength exponent

23
εf´: fatigue ductility coefficient
c: is the fatigue ductility exponent

Strain-Life Approach should be used when:

• Cyclic loading subjects components to plastic strains (considering stresses in

the chart is no more useful)
• We are considering Low-Cycle Fatigue (number of cycles below 10 000)

Major Assumptions:
• Local stresses and strains control fatigue behavior
• Plasticity around stress concentrations
• Accurate determination of Kf

Advantages:
• Plasticity effects
• Mean stress effects

Limitations:
• Requires empirical Kf for good results
• Long life situations where surface finish and processing variables are
important

Since this is a nonlinear material problem, but the analysis type is not implicit nor
explicit to allow nonlinear solution, Neuber Correction is used to evaluate true stress
and strain in the model, which is modelled with elastic-plastic material. For more
information regarding how Neuber Correction is applied in Fatigue analysis, please
refer to an article by Axel Werkhausen using the below link. (FEMFAT software by
Magna Powertrain, Altair Partner Alliance):

24
http://innovationintelligence.com/the-neuber-correction-a-notch-above-the-rest-
for-linear-stress/

𝜎𝑒
𝜎

 =  e e

𝜀𝑒 𝜀

For the Strain-Life approach, further material properties need to be defined, which
are related to setting up a so called Cyclic Stress-Strain Curve.

Since in the plastic region stress-strain behavior is nonlinear, we need to consider

cyclic hardening of the material due to plastic strain. Please observe, the stress-time
relationship when subjected to constant cyclic strain.

25
From the above image, cyclic hardening and cyclic softening can be distinguished for
a material. Please note that if the material would be subjected to elastic deformation
only (Stress-Life approach) then the Stress-Strain chart would represent a line instead
of hysteresis loop.

By step-by-step construction of hysteresis loops, we can model a curve by connecting

stress-strain peaks from each loop – the created curve is then called Cyclic Stress-
Strain Curve.

26
Described by following equation:

Where (further parameters will be required for OptiStruct, MATFAT extension):

K': cyclic strength coefficient

n': strain cyclic hardening exponent

2.3 Mean Stress Influence

It should be mentioned, that afore presented graphs and equations are
experimentally evaluated relationships for fully reversed cyclic loading, which means
mean stress was equal to 0. However, in real conditions, cyclic loading may happen to
be of various kinds, where mean stress is far from the zero value, example of such
loading may be a pulsating load, where minimum stress is equal to zero, but maximum
stress is equal to stress range and two times greater than mean stress and stress

27
amplitude. Hence a few theories are available to introduce stress correction due to
mean stress.

For Stress-Life (S-N) approach:

Two correction graphs are developed for Haigh diagram (maximum stress range as a
function of mean stress):

• Goodman line
• Gerber curve

Notice that Goodman line shows linearly variable stress range for tensile mean stress
and ignores mean stresses in compression. The Goodman approach is recommended
for brittle materials.

Gerber curve treats both tensile and compressive mean stress influence onto stress
range. This correction model is recommended for ductile materials.

Please note that for UTS value of mean stress, the stress range for both corrections is
equal to zero. This in fact confirms that a constant stress exceeding UTS value leads
to material failure.

For Strain-Life (E-N) approach:

Two other correction methods are developed when Strain-Life approach is used:

28
 • Morrow
Morrow approach takes into account the fatigue stress coefficient (σf’) and subtracts
the value of mean stress in the equation according to the following formula:

Where:

σ0 – mean stress

This method is recommended when loading is mostly compressive.

• Smith, Watson and Topper (SWT)

This method introduces mean stress indirectly, i.e. the equation is modified by
amplitude to maximum stress fraction, which would be equal to 1 in case mean stress
is equal to 0 (stress amplitude = maximum stress) and would not influence the

equation:

σmax - maximum in 1 cycle

σa - amplitude in 1 cycle

This method should be used when loading is mostly tensile.

In OptiStruct, user may choose between these correction methods when setting up
analysis parameters (FATPARM card), no additional values are required.

2.4 Damage Model

Real fatigue load history usually does not provide constant parameters, instead it
consists of random peaks (random maxima and minima), which cannot be replaced
with any equivalent cyclic loading. Therefore, OptiStruct uses Palmgren-Miner’s linear

ni
D i =
Ni , f
 1.0

29
damage accumulation rule, to evaluate when damage in the material occurs. The
formula is as follows:

Where:

Ni,f – material fatigue life from S-N/E-N curve for the reference stress/strain of the
particular stress level i (mean correction should be applied)
ni – number of stress cycles at this load level i

To better explain that, let’s consider a few cases:

1) Constant stress of 250 MPa with material life of 100 000 cycles at this stress
(Wohler’s curve). In this case there would be one fragment in the equation,
where N = 100 000. The damage occurs when D exceeds 1, so it means there
need to be n = 100 000 to cause the damage. This sounds logical.
2) There are three different loading stages: 10 000 cycles with 190 MPa peak
stress, 50 000 cycles with 170 MPa peak stress and 40 000 cycles with 165
MPa peak stresses. Based on Wohler’s curve, the material life at 190 MPa
loading is 50 000 cycles, at 170 MPa 100 000 cycles and at 165 MPa 120 000
cycles. Now everything must be put into equation:

10 000 50 000 40 000

𝐷= + + = 0.2 + 0.5 + 0.3 = 1.0 ≥ 1.0
50 000 100 000 120 000

The value of D exceeds 1, therefore the damage will occur after this sequence
of loading.

2.5 Rainflow Cycle Counting

Rainflow Cycle counting is used in fatigue calculations to simplify the fatigue load
history, by omitting insignificant load peaks, which mostly do not influence the
damage process. The best understanding can be obtained by following an example:

30
 The following simple 9-point example is shown with its associated rainflow
algorithm:
LOAD HISTORY

5 D
4
F
H
A BC D E FG H I
3
2
-2 1 -3 5 -1 3 -4 4 -2
Force/Stress

1 B
0
-1 E
-2 A I
-3 C
-4 G
-5
Time

1. S=A; Y=|A-B|; X=|B-C|; X>Y. Y contains S, that is, point A.

S=B.
2. Y=|B-C|; X=|C-D|; X>Y. Y contains S, that is, point B.
S=C.
3. Y=|C-D|; X=|D-E|; X<Y.
4. Y=|D-E|; X=|E-F|; X<Y.
5. Y=|E-F|; X=|F-G|; X>Y.
Count |E-F| as one cycle and discard points E and F.
A cycle is formed by pairing range E-F and a portion of range F-G
6. Y=|C-D|; X=|D-G|; X>Y. Y contains S, that is, point C.
S=D.
7. Y=|D-G|; X=|G-H|; X<Y.
8. Y=|G-H|; X=|H-I|; X<Y. End of data.
9. Append uncounted points up to the starting point S to the end of the history

LOAD HISTORY LOAD HISTORY LOAD HISTORY

5 5 D D 5 D
4 4 H 4 H H
3 3 F 3 F F
2 2 2
Force/Stress

Force/Stress

Force/Stress

1 B 1 B 1 B
0 0 0
-1 -1 E E-1 E
-2 A -2 A -2 I A I I
-3 -3C C -3 C
-4 -4 G -4 G G
-5 -5 -5
Time Time Time

31
 D E F G H I A B C D
5 -1 3 -4 4 -2 -2 1 -3 5

10. S=D; Y=|H-I| ; X=|I-A|; X<Y. I = A eliminate point I.

11. Y=|H-A|; X=|A-B|; X<Y.
12. Y=|A-B|; X=|B-C|; X>Y.
Count |A-B| as one cycle and discard points A and B.

13. Y=|H-C|; X=|C-D|; X>Y.

Count |H-C| as one cycle and discard points H and C.

14. Y=|D-G|; X=|G-D|; X=Y. Y contains S, that is, point D.

Count |DG| as one cycle and stop

2.6 Additional Correction Parameters

Besides already mentioned correction parameters like Surface Finish or Treatment
(both defined in PFAT load collector), we can also define:

• Parameter for Rainflow counting: Gate in FATPARM card to define threshold

value for load peaks simplification:

32
 • For shell elements, which layer should be used to recover results – should it
be top or the worst from these two? (PFAT card load collector):

As you can see, PFAT load collector allows one to define layer for results, surface
properties and a general correction factor KF (user-defined real value).

• Certainty of survival (by default 0.5 – 50%) (SURVCERT in FATPARM). Since

Wöhler’s curve is a statistical approximation of a number of experimental
data, it makes half of the experimental data points lay under the Wöhler’s
curve and half above, this is a result of a stochastic nature of crack nucleation
and growth. This

simply says that if a component reaches maximum number of cycles defined

by Wöhler’s curve, there is actually a 50% chance that the real fatigue
strength at that load value is actually more than the statistically approximated
value by Wöhler’s curve.

33
 By defining a statistical parameter called Standard Error of Log10N cycles (SE
in MATFAT card), we specify the scatter of real data. Having defined the
scatter of real data, we can specify what percentage of certainty we need to
have so that our model actually survives the loading when reaching the
Wöhler’s curve point. Certainty of Survival then offsets the curve, so that if
we, for instance, defined 0.9 (90%) certainty of survival (SURVCERT), 90% of
real data (distribution specified by Standard Error) lays above the offset curve.

2.7 Fatigue Results

What type of stress (max. principal or vonMises or max. shear stress) is used for
fatigue calculation?
Fatigue calculations are based on absolute max. principal stress or signed vonMises
or signed max. shear stress.

Signed principal stress or absolute principal stress:

This term is commonly used in fatigue analysis. Fatigue calculations are based on
amplitude and mean stress. It has been observed that if the calculations are just based
on only maximum principal stress or only minimum principal stress then stress range
(σ– σ) is less and leads to higher fatigue life. Remedy is to find max. value out of the
two at a point over given period of time and then find the stress range or amplitude
and mean stress based on this data(say max. principal stress at a node at time 1 sec.
is + 250 and min. principal is – 400 then absolute principal stress = -400, collect the
data over a period of say 10 sec and then find stress range based on max. and min
absolute stress values out of the 10)

Signed vonMises stress:

vonMises and max. shear stress values are always positive. If these values are used
for fatigue calculations then stress range would be reduced to half resulting in higher
fatigue life. Remedy is to find sign of absolute principal stress at the point at a given
time instance and assign it to corresponding value of vonMises or max. shear stress.
Say absolute max. principal stress is -300 at time 2 sec and corresponding vonMises
stress at time = 2 sec is 315 then signed vonMises stress would be -315.

34
In general max. principal stress is recommended over vonMises and max. shear.
vonMises and max. shear stress are not directional i.e. direction of crack propagation
could be better answered by using max. principal stress.

For fatigue calculations what kind of stress is recommended; average stress or

unaverage, element (centroidal) or nodal stresses?
Nodal stresses are recommended over elemental, similarly unaverage over average.
In some situations, this might lead to lesser fatigue life than actual but from design
point of view it is always safer and recommended. In OptiStruct element stresses are
used for fatigue

35
3 Uniaxial Fatigue Analysis
Setup of a Fatigue analysis involves creating a specific card structure, which is
conveniently presented on the figure below:

SUBCASE

FATDEF FATPARM FATSEQ

PFAT FATEVNT

MATFAT FATLOAD

TABFAT

The Fatigue subcase definition consists of three main parts:

• FATDEF – data containing fatigue material parameters and component

properties, which influence fatigue life.
• FATSEQ – one master load collector with card image FATSEQ gathers the
whole of fatigue loading.
• FATPARM – defines all the remaining parameters except material and
components properties and loading. It defines the kind of fatigue analysis
(uniaxial or multiaxial) and concept that is used (S-N or E-N), correction
method for mean stress, Neuber Correction, Rainflow cycle counting
parameters, certainty of survival and many more for each kind of fatigue
analysis.

36
3.1 Fatigue Loading
Fatigue loading setup allows a very comprehensive approach to setup a loading
history. In Uniaxial Fatigue analysis, multiple load histories are merged by means of
superposition. Generally, there are four different card images which are used to
model a loading structure:

SUBCASE

FATDEF FATPARM FATSEQ

PFAT FATEVNT

MATFAT FATLOAD

TABFAT

1) Fatigue Table (TABFAT)

At the bottom of the hierarchy is a TABFAT load collector, this is a one column
tabular data containing load peaks scaling factors. This is a single load history
sequence. This can be created manually, by choosing TABFAT card image for
a load collector. TABLEFAT_NUM is used to define the number of load peaks.
By clicking on Data: y, user can enter all the load peaks scaling factors

The other option is to import a ready tabular data of .xlsx or .csv extension.
User needs to open a TABLE Create utility available in Utility browser. Then
importing options are available, TABFAT card image must be chosen:

37
2) Fatigue Load (FATLOAD)
FATLOAD collector is used to associate TABFAT load peaks to a static
loadstep, which means this card will contain following information:

SUBCASE

FATDEF FATPARM FATSEQ

PFAT FATEVNT

MATFAT FATLOAD

TABFAT

LOAD (location
and direction)
Static Subcase
SPC
TABFAT

Magnitude of
FATLOAD
LOAD

Scale

Offset

38
 As visible on the chart, FATLOAD defines one single fatigue load with final
magnitudes. TABFAT load peaks are multiplied by Magnitude of LOAD
entered manually by user, which location and direction is referred by LOAD
in static subcase. SCALE is used to apply further scaling for the load
magnitudes and OFFSET is defined for offsetting the load history in time.
Example:

3) Fatigue Event (FATEVNT)

Next comes the Fatigue Event setup (FATEVNT card image load collector). This
is basically a definition of a full event, that gathers all FATLOADs taking place
at one time. If more than one FATLOAD is used, OptiStruct uses superposition
to create an uniaxial loading based on these FATLOADs.

SUBCASE

FATDEF FATPARM FATSEQ FATLOAD1

PFAT FATEVNT FATEVNT ...

FATLOADn
MATFAT FATLOAD

TABFAT

39
4) Fatigue Sequence (FATSEQ)
On the top of the ladder is the Fatigue Sequence (FATSEQ). As the name says,
it defines a sequence of loading. For that, either FATEVNT or other FATSEQ
cards can be used. Each referred FATEVNT or FATSEQ (it is then called sub-
sequence) is defined together with a number instances, i.e. the number of
times they will be repeated. FATSEQ card allows a setup of very complex
loading, such as is visible on the example below:

SUBCASE

FATDEF FATPARM FATSEQ

PFAT FATEVNT

MATFAT FATLOAD

TABFAT

40
An example of FATSEQ definition is shown below. Please note that initially N
number of instances cannot be entered through Entity Editor. User needs to
right click on FATSEQ load collector and choose Card Edit to enable N setup
(image shows N values entered through Card Edit and viewed in Entity Editor,
with already activated N setup)
Here: FATEVNT_1 will be first repeated 100 000 times, next when the first
event is done FATEVNT_2 will be repeated 100 000 times and finally
FATEVNT_3 is going to be conducted for 100 000 times.

41
3.2 Fatigue Parameters (FATPARM)
One of the three subcase entries in Fatigue load step is dedicated for FATPARM load
collector, within this card, all relevant analysis parameters are defined. Therefore,
below are presented parameters that are applicable for Uniaxial Fatigue Analysis:

SUBCASE

FATDEF FATPARM FATSEQ

PFAT FATEVNT

MATFAT FATLOAD

TABFAT

• Type of analysis (TYPE)

Type of fatigue analysis that is defined.

SN = Stress Life
TYPE
EN = Strain Life
FOS = Factor of Safety Analysis

Within this option we can select the approach that will be used for analysis. As
mentioned before, SN approach should be used when fatigue loading is limited to
elastic range and we are considering high-cycle fatigue (above 10 000 cycles). EN
approach on the other hand is useful for low-cycle fatigue (below 10 000 cycles) and
when the cyclic loading causes plastic deformation.

• Stress usage parameters (STRESS field)

This field is specially dedicated for choosing the proper fatigue theories regarding
stress calculation. Different options are available for SN and EN type, therefore user
needs to be aware of the fatigue theories, which were presented earlier on in this
book.

42
 Stress-Life (TYPE = SN) Strain-Life (TYPE = EN)

Default = ABSMAXPR
(ABSMAXPR = Abs Max Principal MAXPRINC = Max Principal
MINPRINC = Min Principal VONMISES = von Mises
Stress
SGVON = Signed von Mises TRESCA = Tresca
COMBINE combination
SGMAXSHR = Signed Max
method.
SGTRESCA = Signed Tresca Shear
XNORMAL = X Normal YNORMAL = Y Normal
ZNORMAL = Z Normal XYSHEAR = X-Y Shear
YZSHEAR = Y-Z Shear ZXSHEAR = Z-X Shear)

GOODMAN = Default SWT – Smith-Watson-Topper = Default

Mean stress GERBER - Gerber mean stress correction MORROW – Morrow correction
CORRECT correction GERBER2 – ignores effect of negative mean MORROW2 – ignores effect of negative
method. stress mean stress
SODERBERG NONE

FE analysis The Unit is necessary because the S-N/E-N curve (MATFAT card) might be defined in
STRESSU Stress Tensor different unit, and FEA stress needs to be converted before looking up the fatigue life
Unit. for a given stress level on the S-N curve.

Neuber NEUBER = Default

correction NONE
PLASTIC Not applicable
for EN
analysis

• Rainflow cycle counting parameters (RAINFLOW field)

During analysis, OptiStruct converts loading data using Rainflow counting method and
user may want to check and update following parameters:

Rainflow data type. LOAD = Load-time history = Default

RTYPE
STRESS = Stress-time history

Relative fraction of The reference value is the maximum range

GATEREL maximum gate range. multiplied by GATEREL, and used for gating out
small disturbances or "noise" in the time series.

43
A few comments of Rainflow data type:

RTYPE = LOAD is valid when there is only one static load case defined in an event. The
cycle counting results (load Ranges and Means) will be scaled by combined FEA stress.
Doing rainflow counting on load-time is much faster than doing it on stress-time
(RTYPE=STRESS), especially when the load-time history is complex and contains a large
number of time points, but it is less accurate.

If the event contains multiple static load cases, RTYPE will automatically be set to
STRESS because there will be stress super-positioning among different load cases;
doing rainflow counting on load-time history cannot deal with it. The stress-time
history has the same length as load-time, while each point of the stress time is the
combined stress value where the stress tensor is FEA stress scaled by y point value of
the corresponding load-time history.

• Certainty of Survival (CERTNTY)

As mentioned before in this book, this parameter is based on Standard Error, provided
in MATFAT card to define the scatter of real data used for creating Wöhler’s curve.
The value that is entered in SURVCERT field depicts percentage of component’s
survival chances. This parameter offsets the Wöhler’s curve from initial position
(averaged curve, which when reached, gives 50% chances of survival as half of the real
data points are below and half above the curve).

Certainty of survival based on the scatter of the S-N curve.

SURVCERT
Default = 0.5 (0.0 < Real < 1.0).

44
3.3 Fatigue Materials & Properties
The MATFAT bulk data card is used in addition to an existing material card to include
material fatigue properties. Its continuation lines depend on whether SN or EN
analysis is to be run.

SUBCASE

FATDEF FATPARM FATSEQ

PFAT FATEVNT

MATFAT FATLOAD

TABFAT

MATFAT is activated by checking the box next to MATFAT in MATi material card. First
there are some general parameters that are defined regardless of Fatigue analysis
type:

Defines the units of stress values specified on the YS, UTS, SRI1, FL, Sf, and Kp
UNIT
fields

STATIC Indicates that static material properties are defined in the following fields.

YS Yield strength.

UTS Ultimate tensile strength.

UTS or YS is used in mean stress correction (SN) and surface finish correction (SN and
EN). If both UTS and YS are defined, UTS will be used both UTS and YS cannot be blank.

45
 • Material data for S-N curve

If the analysis type is chosen to be SN. Material data for SN section need to be defined
to model the S-N curve. Following parameters are available:

SR1 Fatigue Strength Coefficient.

B1 First Fatigue Strength Exponent.

In one-segment SN curve, this is the cycle limit of endurance.

NC1
In two-segment SN curve, this is the transition point

B2 Second Fatigue Strength Exponent.

FL Fatigue Limit; No damage occurs if the stress range is less than FL.

SE Standard Error of Log(N)

These parameters then model one of following curve:

Fig 1a Fig 1b

46
 Fig 2

Figure 1a: One-segment S-N curve in log-log scale (b2=0) (Nc1 is not defined or less
conservative than FL)

Figure 1b: One-segment S-N curve in log-log scale (b2=0) (FL is not defined or less
conservative than Nc1)

Figure 2: Two-segment S-N curve in log-log scale

NOTE: Stress values are defined for Stress Range, not Stress Amplitude.

• Material data for E-N curve

This section should be filled when the analysis type EN is chosen.

SF Fatigue strength coefficient.

B Fatigue strength exponent.

C Fatigue ductility exponent.

EF Fatigue ductility coefficient.

NP Cyclic strain-hardening exponent.

KP Cyclic strength coefficient.

NC Reversal limit of endurance.

47
 SEE Standard Error of (Log(N) from elastic strain.

SEP Standard Error of (Log(N) from plastic strain.

First four parameters are used to model E-N curve:

NP and KP are used for defining Cyclic Stress-Strain Curve. For more information,
please refer to previous chapter about E-N approach.

• Component Fatigue Features (PFAT)

All the component features, connected with manufacturing processes or geometrical
shapes, which are used for correction of component’s fatigue limit are defined within
PFAT load collector. It must be noted that PFAT is not a property card image, i.e. it is
not an actual property definition which links components and materials, rather a load
collector card image.

48
 SUBCASE

FATDEF FATPARM FATSEQ

PFAT FATEVNT

MATFAT FATLOAD

TABFAT

If shell elements are used, it is necessary to specify the

Region Layer for appropriate layer or specify surface of results to use Top
Layer
shell elements. or Bottom. Worst case is then the worst result of Top and
Bottom (the one with larger damage).

Default = NONE (NONE, POLISH, GROUND, MACHINE,

Material Surface
HOTROLL, FORGE or any float value between 0.0 and 1.0)
Finish, a result of
Finish When it is a float value, it will be used to modify the fatigue
manufacturing
limit by multiplying it with the original fatigue limit.
process.

Default = NONE (NONE, NITRIDED, SHOTPEEN, COLDROLL

Material Surface
or any float value greater than 0.0)
Treatment for
Treatment When it is a float value, it will be used to modify the fatigue
Material S-N
limit by multiplied with the original fatigue limit.
Curve.

Fatigue strength reduction factor takes into account the

effect of notch effects, size effects, and loading type
Fatigue strength influence,
Kf
reduction factor. Kf = Cnotch Csize Cloading
where, Cnotch, Csize, and Cloading are correction factors for
notch effect, size effect, and loading type influence.

49
As you can see, all of these parameters are aimed to apply correction value for Fatigue
Limit due to geometrical features of component. When a specified manufacturing
process is used, OptiStruct applies appropriate value of correction.

• Regions for Fatigue Analysis (FATDEF)

FATDEF card (also load collector card image) is more similar to classical property
setup, since this aims to define the regions that are included in fatigue analysis and
associate PFAT component features to them. For Uniaxial Analysis, it can be either
referring to an element set (ELSET) or actual properties (PTYPE). FATDEF card allows
multiple associations, so that each property in the model can have a different PFAT
card assignment:

SUBCASE

FATDEF FATPARM FATSEQ

PFAT FATEVNT

MATFAT FATLOAD

TABFAT

ELSET/Property1
PFAT1 ID
ID
NUMIDS = x
FATDEF ... ...

ELSET/Propertyx
PFATx ID
ID

50
3.4 Fatigue Load Step & Analysis Run
The last step is the creation of Fatigue load step. As visible on the fatigue analysis
setup image, shown a few times before, three load collectors are referred in Fatigue
load step, that is: FATDEF, FATPARM and FATSEQ. Fatigue load step definition looks
as follows:

Before you run the analysis, make sure that the analysis type is Uniaxial:

Activation:

• Uniaxial Fatigue Analysis is run by default in any fatigue analysis

• Make certain that the MAXLFAT field on the FATPARM Bulk Data Entry is set
to NO or blank.
• Also make certain that the Type field on the FATPARM Bulk Data Entry is not
set to FOS.

51
Input and Control:

• The Type field on the FATPARM entry can be used to select the Stress-Life or
Strain-Life approach.
• Fields are available on the SN and EN continuation lines on the MATFAT bulk
data entry to provide additional information.

3.5 Fatigue Results

Fatigue analysis output request is required to obtain life, damage and Rain Flow
results. They can be requested either by:
• GLOBAL OUTPUT REQUEST card.
• Fatigue subcase, under OUTPUT option.

DAMAGE:

Where TYPE should be chosen to be one of the following:

• SUB – total damage from the fatigue subcase.

• EVENT – damage contribution for all FATEVNTs and total damage will be
available.
• ALL – total damage from fatigue analysis.

Reminder: Damage is calculated from Palmgren-Miner’s damage accumulation

model:
ni
D i =
Ni , f
 1.0

As can be seen, the equation consists of a sum of damages coming from one certain
load i, therefore when SUB is selected damage fraction from each FATEVNT can be
presented.

52
LIFE:

Life tells us the number of cycles each part (element) of the component is able to
withstand with current fatigue loading.

Rain Flow:

This entry can be used to output the Rainflow Cycle count history for a particular
fatigue subcase. The Cycle number, Stress Range, Mean Stress, and Damage, for each
fatigue event for each element are output to .rnf file.

3.6 Tutorial: Uniaxial Stress-Life Fatigue Analysis

of a Spring Link
This exercise introduces user to set up Uniaxial Fatigue Analysis with Stress-Life
approach. The Spring link will be subjected to cyclic loading in 3 different directions in

53
a sequence: cyclic bending in two directions and simple tension-compression. Each
event will consist of 100 000 cycles, which in total gives 300 000 cycles of loading.

Open the model file: altair_spring_link.hm

1. Review the model, notice that three static subcases: Vertical, Longitudinal and
Radial are already created, they are necessary for FATLOADs setup.

2. Create one TABFAT load collector to define simple reversal cycle:

• Create new load collector named TABFAT and set card image to TABFAT
• Define the number of table lines: TABLEFAT_NUM = 2
• Click on Data: y and enter following values: -1, 1

3. Define 3 FATLOAD load collectors for each static subcase and assign it with
created TABFAT
• Create new load collector named FATLOAD_vertical
• Set card image to FATLOAD
• Under TID: choose created TABFAT one cycle table
• For LCID choose Vertical load step.
• Set load magnitude of 1.0 in LDM field
• Set load scale of 1.0 in SCALE field
• Define the two remaining FATLOADs as on the images below:

54
*Note:
LDM is used to scale the units if the unit of loads in FEA and test are different.
It is best advised to maintain consistency of units between FEA and TEST and
thereby keeping the value of LDM as 1.

For ex:

Units of Load in FE = Load History Unit in the TestData, then LDM would be 1.

4. Create three Fatigue Events - FATEVNT, one for each FATLOAD.

• Create new load collector and name it FATEVNT_1
• Set card image to FATEVNT
• Set FATEVNT_NUM_FLOAD equal to 1

55
 • Click on table icon next to Data: FLOAD
• Choose FATLOAD_vertical load collector in the pop-up window.

• Repeat the steps to define two remaining Events, choose appropriate

FATLOADs for each. See images below.

5. Create Fatigue Sequence, which will order the before created Events in
following order: FATEVNT_1, FATEVNT_2, FATEVNT_3.
• Create new load collector named FATSEQ
• Set card image to FATSEQ
• In FATSEQ_NUM = field, enter 3, to define 3 entries for a sequence.

56
• Click on Data: FID,… table icon
• Choose Fatigue Events as visible on the image below.

Notice that it is not possible to enter the number of instances N for each Event
here. This must be done through Card Edit panel.

• Right Click on FATSEQ load collector and choose: Card Edit

• For each Event, enter 100 000 instances under N(i) fields. See image below.
• Click return.

• Click again on table icon in Entity Editor for FATSEQ load collector to verify if
N numbers are already present.

57
 • Close the pop-up window. Fatigue loading is now fully defined.

6. Now define Components Fatigue Features with PFAT load collector

• Create new load collector named PFAT
• Set card image to PFAT
• Define parameters like on the image below:

7. Assign Fatigue Features load collector PFAT to solid property with FATDEF load
collector
• Create new load collector named FATDEF
• Set card image to FATDEF
• Check the box next to PTYPE and then check the box next to PSOLID
• Click on table icon next to Data: PID, … to associate PFAT load collector with
spring_link solid property.

58
8. Define Fatigue Material properties within MATFAT extension to existing material
MAT1
• Review if the structural material properties are as on the image below:

• Check the box next to MATFAT extension

• Check the box next to SN, to define that Stress-Life curve will be provided
• One can specify if the curve is Amplitude based on Range based by selecting
A/R on MATFAT. By default it is Range based
• Define Wöhler’s curve parameters as on the image below:

9. Define Fatigue Analysis Parameters within FATPARM load collector

• Create new load collector with name FATPARM
• Set card image to FATPARM

59
• Set TYPE of the analysis to SN
• Set stress combination method COMBINE to Absolute Max Principal
ABSMAXPR

• Set mean stress correction method to GOODMAN

• Stress units STRESSU should be MPA
• Rainflow RTYPE set to STRESS, since we have multiple events in the analysis

60
 • Set the Certainty of Survival SURVCERT to 0.92 (this can be defined because
Standard Error in MATFAT card has be defined)

10. Create and set up a Fatigue load step.

• Create new load step named Fatigue
• Set Analysis Type to Fatigue
• Refer FATDEF load collector in FATDEF subcase entry
• Refer FATPARM load collector in FATPARM subcase entry

61
 • Refer FATSEQ load collector in FATSEQ subcase entry

11. Add output requests for Damage and Life results

• Remain in Entity Editor of Fatigue load step
• Check the box next to OUTPUT in SUBCASE OPTIONS
• In OUTPUT options, check the box next to DAMAGE, set FORMAT to H3D,
TYPE to ALL, OPTION to ALL
• Check the box next to LIFE, set FORMAT to H3D and OPTION to ALL.

12. Run the analysis and postprocess the results

• In OptiStruct panel, set export options to custom, run options to analysis and
click OptiStruct.
• When analysis is done, load results to Hyperview
• Switch Subcases to Subcase 4 (Fatigue)
• View contours for Damage (s)

62
As you can see, the maximum Damage is 0.05102, which tells us that the component
will fail when such loading would be applied approximately another 19 times. You
can also view the results for each Fatigue Event.

3.7 Tutorial: Uniaxial Strain-Life Fatigue Analysis

of a Control Arm
This exercise introduces the user to setting up strain-life (E-N) analysis on a control
arm model shown below using a manual setup method. This model is primarily 3-
dimensional elements overlaid with a skin of 2D shells. The available loads include
braking and vertical loading conditions and both will be used in this fatigue analysis.

63
Problem Setup

You should copy the file: en_carm.hm, load1.csv, load2.csv

Step 1: Open the model en_carm.hm in HyperMesh Desktop

Step 2: Create TABFAT TIMEHIST1 for the time history curve load1.csv and TIMEHIST2
for load2.csv

Step 3: Create a new FATLOAD card FATVERT to associate the TIMEHIST1 time history
curve with the Vertical load case, an LDM of 1, and a Scale of 5. Similarly,
create another FATLOAD FATBRAKE with TID TIMEHIST2 and LCID Brake
using LDM 1 and Scale 5.

64
Step 4: Create a new FATEVNT with 2 FATEVNT_NUM_FLOAD entries using FATVERT
as the first FLOAD in the sequence and FATBRAKE as the second FLOAD in the
sequence

Step 5: Define the fatigue sequence of a single cycle (N=1) of the previously created
FATEVNT in a new FATSEQ card

65
Step 6: Create a FATPARM card with analysis type set to EN. Use the signed von Mises
SGVON stress combination, SWT stress correction method, MPA stress units,
NEUBER plasticity correction with STRESS rainflow type and survival certainty
of 0.5.

66
Step 7: Activate the MATFAT section of the material properties for Aluminum to
include the following fatigue parameters for strain life analysis:

67
Step 8: Create a new PFAT load collector and set the following properties

Step 9: Create a new FATDEF fatigue definition load collector that links the skin shell
elements to the fatigue element property PFAT

68
Step 10: Create a new load step for fatigue analysis named Fatigue

Step 11: Request output using Control Cards

Step 12: Run the analysis

1. In the OptiStruct panel under the Analysis page, ensure that the export options: is
set to all, the run options: is set to analysis.

2. Click OptiStruct to run the fatigue analysis, exporting the *.fem file with the default
naming convention.

Step 13: Review the results in HyperView

69
1. When the run has completed, from the OptiStruct panel, click on the HyperView
button to launch the post-processing report en_carm.mvw which was
automatically created from the analysis.

2. Using the subcase selector in the Results Browser, change the active subcase to
Subcase 3 (Fatigue).

3. Expand the Components list in the Results Browser and click the mesh display icon
next to PSOLID_2 and PSOLID_5 to turn off the display of those components.

4. Click the Shaded Mesh Display button to show the displayed mesh with mesh
lines.

5. Enter the Contour panel and set the Result type to Damage (s). Click Apply to
contour the displayed mesh with damage results.

6. Click on the Legend tab in the Contour panel and click Edit Legend… to bring up
the Edit Legend checkbox. Enter the values as shown below and click Apply to
apply the new contour maximum value and OK to exit the dialog box. (for
Logarithmic Display)

70
EXERCISE RESULTS: en_carm.h3d

71
4 Multiaxial Fatigue Analysis
The main difference between Uniaxial and Multiaxial Fatigue Analysis is the stress
conversion:

• In Uniaxial Fatigue Analysis, OptiStruct converts the stress tensor to a scalar

value using user-defined combined stress method (von Mises, Maximum
Principal Stress, and so on).
• In Multiaxial Fatigue Analysis, OptiStruct uses the stress tensor directly to
calculate damage. Multiaxial Fatigue Analysis discussed in the following
sections are based on the assumption that stress is in the plane-stress state.

Only free surfaces of structures are of interest in multiaxial fatigue analysis in

OptiStruct.

• For solid elements, a shell (membrane) skin is automatically generated by

OptiStruct, shell elements are used as-is.

If the loading is uniaxial, that means there is one principal stress produced, uniaxial
analysis is used.

If the loading is proportional biaxial, which means that different loading contributors
are acting cyclically in the same phase, resulting in one directional resultant stress,
then multiaxial analysis is used.

If the loading is non-proportional multiaxial, that means different load components

are in different cyclic phases, then multiaxial analysis is used.

72
4.1 Multiaxial Stress-Life Technical Background
(Ref:https://www.efatigue.com/multiaxial/background/stresslife.html)

Fatigue occurs on the surface where one of the principal stresses is usually zero. As a
result, multiaxial fatigue problems are usually biaxial in nature. Here we have
assumed that the stress normal to the free surface is zero in all the calculations. Any
stress concentration effects must be included in the input loading data.

Multiaxial Fatigue Calculators are used for constant amplitude loading where the
largest cycle is assumed to do all of the fatigue damage. The Multiaxial Fatigue
Analyzer counts cycles and sums fatigue damage for all cycles. More damage models
are available with the Multiaxial Fatigue Analysis.

4.1.1 Damage Models

There are many damage models used for multiaxial fatigue analysis. We have
implemented some of the most common and useful methods including Findley, Sines,
Goodman and Dang Van.

A common feature of many stress based multiaxial fatigue criteria is that they are
expressed as a general form that includes both shear stress amplitude τa and normal
stress σn during a loading cycle. Multiaxial fatigue models differ in the interpretation
of how the shear stress amplitude and normal stress are defined. One of the
fundamental problems is to evaluate the effective amplitude and mean values of the
shear/normal stresses on a material plane.

73
Fatigue cracks begin on or near the surface. On the surface the stress state is one of
plane stress so that there are only three stress components σx, σy and τxy. All other
stresses are zero. For these stress states, the accumulation of fatigue damage that will
eventually lead to fatigue cracks will occur on planes that are oriented either 45° or
90° to the free surface.

The normal stress is easy to evaluate on any plane. Its direction is fixed, and the
magnitude varies. But the shear stress can vary in both magnitude and direction.
There are two shear stresses acting on the 45° plane. An in-plane stress, τA and an out-
of-plane shear stress τB. How should these two shear stress be combined? One
possibility is to consider both shear stresses separately because only one of the two
slip systems will be activated in an individual grain of the material.

• OptiStruct searches for the most damaging plane by 10 degrees of θ.

• On each plane, OptiStruct assesses damage using both tensile crack damage
model and shear crack damage model.
• Tensile Crack Growth takes place along planes perpendicular to the free
surface.
• Shear Crack Growth takes place along planes oriented at 45° to the free
surface.
• The shear stress A is an in-plane shear stress and causes microcracks to grow
along the surface.
• B is a maximum out-of-plane shear, microcracks growth into the material
• At the end of a search, OptiStruct reports damage at the most damaging
plane which is a critical plane.

74
4.1.2 Minimum Circumscribed Ellipse
(de Freitas M., Li B, Santos JLT. Multiaxial Fatigue and Deformation: Testing and
Prediction, 2000, ASTM STP 1387)

An example of the MCE is shown below.

First, the shear stress loading history is determined for a plane. For proportional
loading this will always be a straight line. Nonproportional loading histories will have
some complex shape. In this case, the the major axis of the ellipse is defined as the
smallest circle that will contain all of the shear stresses. This is denoted RA in the

75
example. Using the major axis as a reference, the minor axis, RB is defined as the
maximum distance of any stress point in the loading history from the major axis. An
equivalent shear stress is then used in the fatigue calculations. This shear stress may
be used with any damage model.

𝜏𝑎 = √𝑅𝐴 2 + 𝑅𝐵 2

4.1.3 Findley
(Findley, W. N., "A Theory for the Effect of Mean Stress on Fatigue of Metals Under
Combined Torsion and Axial Load or Bending," Journal of Engineering for Industry,
Nov. 1959, 301-306)

Findley reviewed much the same experimental data as Sines but came to a slightly
different model. He suggested that the normal stress σn, acting on a shear plane might
have a different linear influence on the allowable alternating shear stress, Δτ/2.

∆𝜏
( + 𝑘𝜎𝑛 )max= 𝑓
2

Parabolic forms were investigated but Findley concluded that the linear form was
sufficient to describe the experimental data. Here k is a material constant that is
related to the materials' sensitivity to normal stresses and f is directly related to the
materials' fatigue strength.

Mohr's circle for both tension and torsion are show below. Both loading conditions
have the same shear stress. In an R= -1 fatigue test, both tension and torsion loading
will have the same shear stress range acting on a small microcrack located on
maximum shear stress plane. But Mohr's circle shows that tension loading will also
have a normal stress cycle with the same range as the shear stress acting on this shear
plane. The result is that for the same shear stress range, tension loading will always
be more damaging than torsion.

76
Findley's model differs from Sines' or other yield criteria models in that it identifies
the stress acting on a specific plane within the material. These are termed critical
planes and can be defined as one or more planes within a material subject to a
maximum value of a damage criterion. Fatigue life is controlled by the combination of
stresses and strains acting on a critical plane. Findley identifies a critical plane for
fatigue crack initiation and growth that is dependent on both alternating shear stress
and maximum normal stress. The combined action of shear and normal stresses is
responsible for fatigue damage and the maximum value of the quantity in
parentheses is used rather than the maximum value of shear stress. This concept is
easily illustrated for torsion loading. The stresses σθ and τθ on a plane oriented at an
angle θ can be computed from the applied shear stress τxy as

𝜎𝜃 = 𝜏𝑥𝑦 sin(2𝜃)
𝜏𝜃 = 𝜏𝑥𝑦 cos(2𝜃)

Values of τθ + kσθ for a value of 0.3 are plotted below and failure is expected to occur
on the plane that has the largest combination of Δτ/2 + kσn.

77
Findley criterion for torsion

The maximum value of the parameter occurs at angles of about 82° and -8°. In the
case of pure torsion, it can be shown as:

∆𝜏
√1 + 𝑘 2 =𝑓
2

and for pure bending or axial loading with a stress amplitude σa and maximum stress
σmax

√(𝜎𝑎 )2 + 𝑘 2 (𝜎𝑚𝑎𝑥 )2 + 𝑘𝜎𝑚𝑎𝑥 = 2𝑓

The constant k is determined experimentally by performing fatigue tests involving two

or more stress states. Examples of solutions for k based on the ratio of fatigue strength
for two stress states are given below.

78
For ductile materials, k typically varies between 0.2 and 0.3.

The Findley criterion is often applied for the case of finite long-life fatigue. An
equation of the form

is used where tf* is computed from the torsional fatigue strength coefficient, tf', using

The correction factor typically has a value of about 1.04.

4.1.4 Goodman
The traditional Goodman diagram can be used with the principal stresses to evaluate
fatigue under multiaxial loading. Care must be taken when using principal stresses for
loading which involves both tension and compression. The principal stresses and their
directions are shown below for both tension and compression loading.

79
By convention, the principal stresses are always ordered so that σ1 >= σ2 >= σ3. Note
that during a simple tension compression loading the principal stress direction rotates
90°. To avoid these problems, the normal stresses are computed for each plane and
the range of normal stress acting on a plane is used to compute fatigue lives. The
Goodman correction is used to account for mean stresses on each plane. The plane
with the lowest fatigue life is considered the critical plane.

4.1.5 Dang Van

(Dang-Van, K., "Macro-Micro Approach in High-Cycle Multiaxial Fatigue," in Advances
in Multiaxial Fatigue, ASTM STP 1191, D.L. McDowell and R. Ellis, Eds., American
Society for Testing and Materials, Philadelphia, 1993, 120-130.)

Dang Van has proposed an endurance limit criterion based on the concept of
microstresses within a critical volume of the material. This model arises from the
observation that fatigue crack nucleation is a local process and begins in grains that
have undergone plastic deformation and form characteristic slip bands. It is
hypothesized that because cracks usually nucleate in intragranular slip bands, the
microscopic shear stress on a grain must be an important parameter. In the same way,
it is reasoned that the microscopic shear stress on a grain must be an important
parameter. In the same way, it is reasoned that the microscopic hydrostatic stress will
influence the opening of these cracks or slip bands. The simplest failure criterion
involving these two variables is a linear combination

80
𝜏(𝑡) + 𝑎𝜎ℎ = 𝑏

where τ(t) and σh are instantaneous microscopic shear stress and hydrostatic stress
and a and b are constants. The constant b is the fatigue strength determined from a
torsion test and a is related to the sensitivity of the material to hydrostatic stress.

4.2 Damage Models in OptiStruct

Since there are two different crack growth mechanisms in multiaxial fatigue, two
damage models are used simultaneously by OptiStruct, one for tensile crack growth
search and second for shear crack growth search. Following damage models are used
by default:
Stress-Life (S-N) Approach:

• Goodman for Tensile Crack Growth

• Findley for Shear Crack Growth
Strain Life (E-N) Approach:

• Smith-Watson-Topper (SWT) for Tensile Crack Growth

• Fatemi-Socie and Brown-Miller for Shear Crack Growth

Damage models are selected by default, if user wants to select them manually
anyway, the option MDMGMDL allows choice of multiple damage models for
multiaxial analysis.
It is available in FATPARM load collector.

81
4.3 Analysis Procedure
OptiStruct determines the damage assessment search according to the following
methods:
SN Flowchart:

Elastic Stress

Proportional Yes
Loading?

No
Assess damage. No Plane
search is required.

Assess damage. Plane

search is required

End

EN Flowchart:

Two models are available for multiaxial strain life analysis.

For non-proportional multiaxial cyclic loading:

• OptiStruct runs Jiang-Sehitoglu plasticity model to calculate the total strain and
elasto-plastic stress.
• Note that in OptiStruct, the isotropic hardening part is removed from Jiang-
Sehitoglu's original model.
• The Jiang-Sehitoglu model is the OptiStruct default for this loading type.

For proportional biaxial loading:

• OptiStruct runs (by default), the Hoffman-Seeger method to calculate stress and
strain from pseudo stress.

82
• Proportional Biaxial Loading is typically a result of a single static load referenced
in FATEVNT, however if the CHK field on the FATPARM Bulk Data Entry is set to
NO, then the Jiang-Sehitoglu plasticity model is used regardless of loading.

*Jiang-Sehitoglu plasticity
Nominal Stress model
**HS Biaxial model

Proportional Yes Simple Plasticity

Loading? correction**

No

Full Plasticity Model*

Assess damage. No Plane
search is required.

Assess damage. Plane

search is required

End

Setup tips for multiaxial analysis:

• In order to activate multiaxial fatigue analysis in a model, set MAXLFAT field on

the FATPARM Bulk Data Entry to YES.
• Multiple Damage models are allowed in E-N Multiaxial Fatigue Analysis. The
MDMGMDL continuation line on the FATPARM Bulk entry can be used to define
damage models. By default, the SWT and FS damage models are active.
• In E-N Multiaxial Fatigue Analysis, the CHK field on the FATPARM Bulk Data entry
can be used to choose the treatment method for Proportional Loading.
• The MXLMSTRN field on the MATFAT bulk data entry can be used to set the
maximum strain value for E-N Multiaxial Fatigue Analysis.

83
• Fields are available on the SN and EN continuation lines on the MATFAT bulk data
entry to provide material and damage model control information.
• The automatic skin generation for multiaxial analysis is done within by default
activated SURFSTS, YES in FATPARM card.

4.4 Tutorial: Multiaxial Fatigue Analysis of a

Knuckle
This exercise introduces the user to setting up multiaxial fatigue analysis on a knuckle
model.

Problem Setup

You should copy the file: altair_knuckle.hm, load1.csv, load2.csv

Step 1: Open the model altair_knuckle.hm in HyperMesh Desktop

Step 2: Create TABFAT cards to represent the time history curves

Tip: Import load1.csv into TABFAT load1 and import load2.csv into TABFAT load2.

84
Step 3: Associate the time history curves & load cases using the Matrix Browser

1. Create a new FATLOAD named FATLOAD_FX_900457 referencing LOAD1, LCID (1)

IR-FX_900457. Set LDM to 1 and Scale to 1000.

85
2. Duplicate this loadcase 9 times to produce a total of 10 FATLOAD cards with
identical information.

Tip: Modifying the individual FATLOAD cards names and information can be
completed through the Entity Editor or, alternatively, through the Matrix Browser as
detailed below.

3. Open the Matrix Browser by selecting Tools > Matrix Browser from the pull-down
menu.

4. When the new worksheet comes up, go to the selectors at the bottom of the screen.
Select DataSource HMdata, Entities loadcols and click Query.

5. The HyperMesh Desktop panel area shows a loadcols entity selector. Click on the
entity selector and select the FATLOAD_FX_900457 load collector and its 9

86
 duplicates and click proceed to close the entity selector panel and return to the
Matrix Browser.

*Tip: The ids of the load collectors selected should be populated in the first column
of the Matrix Browser.

6. Click to highlight the first column in the worksheet of the Matrix Browser. When
the first Datanames entries are populated, select the options name, FATLOAD,
FATLOAD_LCID, FATLOAD_LDM, FATLOAD_OFFSET, FATLOAD_SCALE, and
FATLOAD_TID.

7. Click Query to load these entries into the Matrix Browser alongside their
corresponding loadcol id.

8. Change the data within the Matrix Browser as shown below:

87
*Tip: Note that when entries in a card are changed within the Matrix Browser, the
default settings of the Matrix Browser update the corresponding database entries
within the HyperMesh Desktop database as well. Please be aware of the values and
data entries that you are editing.

*Tip: Note that in the Matrix Browser, certain integer data types such as ID numbers
may be shown as a floating point value with separator.

*Tip: Note that all of the Z-loadcase IDs utilize FATLOAD_TID 13 and the X-loadcase
IDs use FATLOAD_TID 12.

9. Switch to the Model Browser tab to review the new load collectors and verify the
information. Once the information is correct, close the Matrix Browser.

*Tip: There is no need to save the existing worksheet data from the Matrix Browser.

Step 4: Create a new fatigue event card that references all ten of the multiaxial
fatigue loads

88
Step 5: Define the fatigue sequence using a FATSEQ card with N=100

*Tip: Right-click on FATSEQ in the Model Browser and Card Edit to enable and update
N.

89
Step 6: Define the FATPARM card for multiaxial stress with Stress rainflow counting
type and activate the option for free surface damage membranes. more
than one damage models can be assigned by the user, if the user just
specifies one, the tool calculates only for that Damage model, user can
specify both shear and tension damage models and the tool gives out the
worst result. (Refer FATPARM card)

90
Step 7: Extend the existing material properties to include fatigue parameters

91
Step 8: Create a new PFAT load collector and set the following properties:

Step 9: Create a new FATDEF fatigue definition

Tip: The knuckle component’s solid elements utilize a PSOLID property

92
Step 10: Create a new load step for fatigue analysis named Fatigue

93
Step 11: Request Damage and Life outputs in H3D format

*Tip: New output request cards can be created by accessing the menu selection Tools
> Create Cards > <first letter of card name>. Output request cards are marked with
[OR] after their name. Example: DAMAGE [OR] is an output request for damage
results.

Step 12: Run the analysis in OptiStruct

Step 13: Review the results in HyperView with a contour plot of the knuckle
component, showing the damage

*Tip: Edit legend and input values as shown below

94
EXERCISE RESULTS: altair_knuckle.h3d

95
5 Factor of Safety Analysis
Uniaxial and Multiaxial type of Fatigue analysis is conducted in order to evaluate an
overall damage that is summed up throughout the loading history. Regardless if
certain load peaks exceed the Ultimate Tensile Stress or not, each load peak has its
contribution to the overall damage sum. This kind of approach is only useful when we
are sure that the entire load history will remain below the strength limit of the
material.

What if the load history is clearly irregular and we are not sure whether a single load
peak can cause a failure?

Factor of Safety Analysis (Dang Van) investigates if the component will fail during its
entire loading history. The calculation of maximum stresses for each element is done
and factor of safety is calculated with respect to stress limit defined within shear
stress and hydrostatic pressure chart:

The calculation of Factor of Safety is done within a following formula:

Where:

τ(t) - is the Shear Stress based on the deviatoric part of microscopic stress
p(t) - is the hydrostatic part of macroscopic Stress
a and b - are material constants (resp. TFL and HSS in MATFAT)

96
The approach of Factor of Safety analysis to investigate whether the component will
fail during the entire load history is also referred to as Dang Van Criterion.

Examples of parts which may be analyzed using Dang Van approach include:

• Automobile components which undergo multiaxial cyclic loading

• Propeller shafts
• Jet turbine blades

Fatigue crack initiation usually occurs at zones of stress concentration such as

geometric discontinuities, fillets, notches and so on. This phenomenon takes place in
the microscopic level and is localized to certain regions like grains which have
undergone local plastic deformation in characteristic intra-crystalline bands.

The FOS calculation involves evaluating the Microscopic Deviatoric Shear Stress and
Macroscopic Hydrostatic Stress as functions of time for all time steps over the entire
loading cycle. The worst FOS value (time domain) is selected and output for each
element in the fatigue space. If FOS is less than 1, the component cannot experience
infinite life. The FOS, Hydrostatic Stress, and Microscopic Shear output are controlled
by the I/O Options.

5.1 FOS Analysis Setup

1) The FOS analysis is activated in FATPARM card by choosing TYPE as FOS. No other
parameters are required in FATPARM card, DANGVAN criterion is active by
default

97
2) MAXLFAT must be set to NO or left blank
3) In MATFAT extension card, FOS field should be edited to define the shear stress –
hydrostatic pressure relationship (a and b constants):
a. HSS is “a” constant
b. TFL is “b” constant

The Tfl field on the MATFAT Bulk Data Entry can be used to define either a value
or an integer (table – multiple slopes) to specify the Failure zone for Factor of
Safety Analysis.

98
 The integer value references the identification number of a TABLES1 Bulk Data
Entry that defines the intersection points. The X-values represent Hydrostatic
Pressure, and the Y-values represent Shear.

The STHETA and SSHEAR fields on the MATFAT Bulk Data Entry can be used to
define the Safe Zones for Factor of Safety Analysis.

4) I/O Option: FOS (H3D, SUB) = ALL

99
5.2 Tutorial: FOS Analysis of a Spring Link
This exercise introduces the user to setting up factor of safety analysis using the Dang
Van method on a spring link model. Unlike previous models which were set up from
the beginning of the fatigue process, this model will be brought into HyperMesh
Desktop already set up for fatigue analysis. The model will then be modified to provide
factor of safety as a result and analyzed.

Problem Setup

You should copy the file: altair_spring_link_FOS.hm

Step 1: Open altair_spring_link_FOS.hm in HyperMesh Desktop

Step 2: Modify the MATFAT of the material card for MAT1_FATIGUE, deactivating
the

100
SN section and activating the FOS section, setting TFLmax to 102 and HSS to 0.424

101
Step 3: Update the FATPARM card for factor of safety analysis

Step 4: Update the output requests using the GLOBAL_OUTPUT_REQUESTS of the

control cards to request FOS output in H3D format

102
Step 5: Run the analysis in OptiStruct

Step 6: Review the results in HyperView, contouring the FOS(FOS) result type in
HyperView

Tip: Use the Edit Legend dialog box to import the FOS_legend.tcl file and create the
legend as shown in the results image.

103
EXERCISE RESULTS: spring_link_FOS.h3d

104
6 Seam Weld Fatigue Analysis
This type of analysis is used to investigate seam welds exclusively. The method
implemented in OptiStruct is based on a research paper Fatigue Life Prediction of
MAG-Welded Thin-Sheet Structures published by M. Fermér, M Andréasson, and B
Frodin. The method is a hot spot stress approach applicable to thin metal sheets. Hot
spot stress is calculated from grid point forces at weld line. The method showed a
good agreement with laboratory test results for sheet thickness between 1.0 mm and
3.0 mm. The method typically requires two S-N curves. One is a bending S-N curve
which is dominated by bending stress, and the other is a membrane S-N curve which
is dominated by membrane stress.

Note on stress resultants

When considering an element with constant cross section, let’s say cylindrical,
membrane stress results purely from compressive or tensile forces, it causes pure
tensile or compressive part of a stress that is constant throughout the thickness.

Bending stress however is resultant from bending moments, it is a pure bending part
of the stresses, characteristic stress field is provided which linearly changes from
tensile to compressive part through neutral axis.

105
A typical seam weld presents a few characteristic features that results from the
welding process, the features are named as follows:

6.1 Seam Weld Elements

In HyperWorks, seam welds are modelled as a row of CQUAD4 elements, which are
usually connected to joined components in toe locations – face elements. It can either
be a row of face elements, or a combination of face elements, root elements or in
other than T-Junction welds, also a row of throat elements. All of these possible

combinations are presented later in this chapter. Generally, in order to define that a
certain row of elements represents a seam weld, FATSEAM load collector is used to
assign weld features to existing elements. An example of T-Junction seam weld:

For FE modelling of seam welds, a few rules must be stated:

106
 • Weld elements should be CQUAD4 elements, CTRIA3 elements are allowed
for corner weld
• The connection line should be weld elements and adjacent elements define
the toe location
• There should be one or two rows of CQUAD4 weld elements, 3 rows are
allowed in case weld fillets in both toe locations are necessary
• The thickness of the weld element is the same as the effective throat.
• Elements are defined as weld elements by referring them in FATSEAM load
collector
• Element normal direction must point outward of the weld, that is opposite to
root of the weld

In the discretized FE model, seam welds can be modelled as shown on the pictures
below. Since in the FATSEAM card user must choose between four types of welds:

• FILLET
• OVLAP (Overlap)
• LOVLAP (Laser Overlap)
• LEOVLAP (Laser Edge Overlap)

They were also grouped with respect to these options below:

Fillet

• T-Junction
A choice between these combinations is dependent on the depth of penetration of
seam weld (see the light blue cross sections of welds on the pictures below). As visible
on the pictures, deep seam welds are modelled together with root elements, whereas
seam welds, which are meant not to penetrate the components much are
represented only with face elements.

107
• Cross-joint

108
The same situation and rules as with T-Joint, when it comes to choice of weld
modelling.

Overlap

Laser Edge Overlap

Laser Overlap

The failure that is expected in seam welds is actually predicted to be in certain

locations, these locations are always assumed by OptiStruct. For Fillet and Overlap
welds they are located in the elements adjacent to weld elements, near toe of root
locations:

109
Fatigue Bulk Data Cards: FATSEAM

The FATSEAM bulk data card is used to define parameters for seam weld fatigue
analysis

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

FATSEAM ID WTYPE

PSHELL PID1 PID2 ... ... ... ... PID7

Where:

Each FATSEAM card must have a unique ID. The FATPARM Subcase
ID
Information entry may reference this identifier.

WTYPE Weld type.

PSHELL Indicates that the following property fields reference PSHELL entries.

PIDi Property identifier of weld elements

FATSEAM associates weld type WTYPE to PSHELL property, which in turn contains
assignment of material to a group of CQUAD4 elements.

110
6.2 Seam Weld Properties Setup
The structure for seam weld analysis is similar to the “normal” fatigue analysis,
loading

SUBCASE

FATDEF FATPARM FATSEQ

FATSEAM FATEVNT

PFATSMW FATLOAD

MATFAT TABFAT

setup is exactly the same except the difference in FATPARM card and FATDEF
structure:

Top Stress Fraction

(TOPSTR)
PSHELL
FATSEAM
FATDEF
Seam weld features

BRATIO, TREF,
PFATSMW
TREF_N

PFATSMW
Besides the additional FATSEAM card, described in the chapter before, PFATSMW
property is used instead of PFAT load collector. Please notice that this time it is the
card image of a property.

Following parameters are defined within PFATSMW property:

111
BRATIO Critical bending ratio

TREF Reference thickness for thickness effect consideration

TREF_N Exponent for the thickness effect consideration

• BRATIO
The upper and lower curves on the picture below are referred to as Bending SN curve
and Membrane SN curve, respectively. It is recommended that membrane SN curve
should be used when membrane stress dominates in an element and bending SN
curve should be used when bending stress dominates. Interpolation between the two
curves may be carried out depending on the degree of bending dominance.

Generally, bending ratio is defined as a contribution of maximum stress coming from

bending moments in overall stress in weld damage locations (toe, root, throat)

σB – is maximum bending stress

σM – is maximum membrane stress.

112
For all shell elements, considered in weld fatigue, an average bending ratio is
evaluated and compared to critical bending ratio BRATIO. If average bending ratio
(rBAVG) is less than or equal to the critical bending ratio (rBCRIT), then the Membrane SN
curve is used to assess damage. If average bending ratio is greater than the critical
bending ratio, then an SN curve that is interpolated between membrane SN curve and
the bending SN curve is used.

• TREF and TREF_N

The thickness correction process is for size effect correction. SN curves are based on
test results from a particular size of the specimen. In reality, the stress vs life curve
may vary depending on specimen size. Therefore, thickness correction parameters
can be used to correct for this effect. It may be applied based on the thickness T of
each shell element under consideration for Fatigue calculation (that is, toe, root, or
throat element). The calculations are:

If T ≤ TREF, then there is no Thickness Correction.

If T > TREF, OptiStruct increases stress based on the equation below:

113
The default values are 1.0 and 0.2 respectively. The defaults are in inches (English
units), if the metric system is used, then the values should be modified accordingly.

Thickness Correction can be turned on or off using the corresponding THCKCORR field
on the FATPARM bulk data entry for Seam Weld Fatigue Analysis.

FATDEF

FATDEF card, as in previous analyses, is used to assign properties to analyzed regions,

usually PFAT property was assigned to PSOLID or PSHELL property. This time
PFATSMW is assigned to FATSEAM load collector, as FATSEAM already contains
assignment of weld features to PSHELL property.

Top Stress Fraction

(TOPSTR)
PSHELL
FATSEAM
FATDEF
Seam weld features

BRATIO, TREF,
PFATSMW
TREF_N

FATDEF allows multiple pairs of FATSEAM and PFATSMW. Top Stress Fraction
(TOPSTR) is used to define that only the highest stress results are written to output,
the stresses out of this fraction will result in no stress in the output.

114
FATPARM

The Fatigue parameters are defined as usual, except SMWLD section must be checked
and METHOD of weld fatigue analysis must be chosen (only VOLVO is available).

(“Volvo Method” or “Hot Spot” Stress approach is used to calculate the hot spot stress
from the nodal force at weld line)

115
When additional SMWLD is activated, FATPARM card can look as follows:

Of course, all the setup possibilities for fatigue calculation are available as for a
uniaxial analysis. The SMWLD section allows following parameters to be defined:

Flag indicating that the following parameters are used for seam weld
SMWLD
fatigue analysis.

Seam weld fatigue analysis method.

METHOD
Default = VOLVO (VOLVO or blank)

Mean stress correction indicator.

CORRECT
Default = NONE (FKM or NONE)

Certainty of survival.
SURVCERT
Default = SURVCERT value on CERTNTY continuation line (0.0 < Real < 1.0)

Thickness correction flag.

THCKCORR
Default = YES (YES or NO)

116
MATFAT

The MATFAT card parameters for Seam Weld Fatigue Analysis are filled in SMWLD
section, it requires a setup of S-N or E-N diagrams, the parameters are as follows:

SMWLD Indicates that seam weld fatigue analysis fatigue material properties follow.

Mean Stress Sensitivity parameter for mean stress correction based on FKM
FKMMSS_SM
guidelines for seam weld fatigue analysis

A/R Defines the interpretation of the defined SN curve.

Fatigue strength coefficient. It is the stress range intercept of SN curve at 1

SR1_SMi
cycle in log-log scale.

The first fatigue strength exponent. It is the slope of the first segment of SN
B1_SMi
curve in log-log scale.

NC1_SMi In one-segment SN curve, this is the cycle limit of endurance

The second fatigue strength exponent. It is the slope of the second segment
B2_SMi
of SN curve in log-log scale

FL_SMi Fatigue Limit

SE_SMi Standard Error of Log(N)

Notice that the parameters used for building a S-N curve are exactly the same as for
“normal” fatigue analysis, but each of them has a suffix _SMi. Here i = 1 for bending
SN and i = 2 for membrane SN.
For more information, please refer to OptiStruct Help documentation.

117
6.3 Tutorial: Weld Fatigue Analysis of a Seam
Welded Frame
This exercise illustrates seam weld fatigue through the use of a frame structure. Users
will review the welds and the partial fatigue setup available and will continue the
modeling by creating the necessary cards to complete the fatigue setup.

Problem Setup

You should copy the file: Frame_SeamWeld.hm

Step 1: Open the model Frame_SeamWeld.hm in HyperMesh Desktop

118
Step 2: Review the components of the model, paying particular attention to the seam
weld elements and locations. Also review the existing fatigue load collectors to
determine the relationship between the load steps and fatigue loading curves.

119
Step 3: Create a new FATPARM card and activate the SMWLD section and set the
METHOD to VOLVO

120
Step 4: Set up a new FATSEAM card for FILLET weld types that uses the PSHELL from
the seam weld property

Step 5: Create a new property card of type PFATSMW which sets up necessary
properties for analyzing seam welds in fatigue analysis

121
Step 6: Create a new FATDEF card referencing the FATSEAM id as the fatigue analysis
entity type

122
Step 7: Extend the existing material properties to include fatigue parameters for
seam welding.

123
Step 8: Create a new load step for seam weld fatigue analysis

Step 9: Run the analysis in OptiStruct

Step 10: Review the results in HyperView

*Tip: For seam weld fatigue analysis, results are available for the elements directly
connected to the weld locations

124
EXERCISE RESULTS: Frame_SeamWeld.h3d

*Edit legend and input values as shown below for logarithmic display of result.

Click Apply > ok

125
126
7 Spot Weld Analysis
Before a process of spot welds analysis in OptiStruct will be presented, a few words
about what actually a spot weld is – A Spot Weld is created between two adjacent
sheet elements as a result of current transfer through thicknesses of these sheets.
The electric current that is transferred from one electrode to another causes heating
of the metal

parts due to electric resistance of the metals. Highly concentrated current

(transferred through a small area) brings the metal parts to molten state and due to
pressure applied by electrodes a joining of melted sheet parts occurs. The area at the
joint is called a nugget.

Spot weld fatigue analysis allows for the study of fatigue performance of spot welds
in structures. Currently, only Stress-Life (SN) based spot weld fatigue analysis is
supported. spot weld fatigue can only be applied to spot welds between two shells.
The spot weld location is defined by three attributes, sheet 1, sheet 2, and the nugget.
The sheets are defined by shell elements, and the nugget is defined by CWELD, CBAR,
CBEAM, or CHEXA elements. The nugget can be directly connected to the shells or
RBE2/3 elements can be used to connect the nugget to the shells.

The nugget needs to be modelled between two nodes of a shell parts. It seems that
when the actual thicknesses of the shells are considered, the nugget should be a
length of zero, since the shell parts are in contact. However, the shell meshes are

127
represented usually with zero thickness and so the nodes are placed in the middle of
the theoretical thicknesses – in that case the length of the nugget is equal to:

𝑡1 +𝑡2
𝐿=
2

Where t1 and t2 are real thicknesses of the shells.

t1

Cross section of
joined shells
Nugget element

t2

7.1 Stress calculation

Fatigue analysis for spot welds involves examining the weld at three distinct locations,
the center planes of the two attachment sheets at the points of attachment and at
the center of the nugget, based on a paper by Rupp et al. The cross-sectional forces
and moments at each of the three locations is determined and used to calculate

corresponding stresses. These stresses are then used to calculate Fatigue Damage
using Rainflow counting and the SN approach.

The failure may occur due to two possible stress conditions:

128
 • Failure at sheet due to radial tensile stress
• Failure at nugget due to bending and shear stress

7.2 Spot Weld Fatigue Analysis Setup

The structure of Spot Weld Fatigue analysis setup is very similar to Seam Weld Fatigue,
the difference is in the FATDEF entries – FATSEAM card is not defined in this case and
FATDEF card associates the PFATSPW properties directly to nugget property (PBARL,
PBAR, PBEAML, PBEAM, PWELD, PSOLID)

SUBCASE

FATDEF FATPARM FATSEQ

PTYPE FATEVNT

PFATSPW FATLOAD

MATFAT TABFAT

129
PFATSPW

Like in the seam weld fatigue PFATSMW is used, in spot weld fatigue the fatigue
properties are defined with a PFATSPW property card image.

SPTFAIL Damage assessment option type.

The value of α used to determine the AUTO option on the SPTFAIL field.
ALPHA
Default = 3.5.

Effective diameter of the spot weld that is generated using Hexahedron

HEXA_D
(CHEXA) elements

TREF Reference thickness for thickness effect consideration

TREF_N Exponent for the thickness effect consideration

SF Stress scale factor

The most important parameter is to define SPTFAIL, to choose whether the damage
will be assessed at nugget or sheet location, or both.

Following options are available:

Default = SHEET (SHEET, NUGGET, ALL, or AUTO)

SHEET – Only assess sheet damage at the spot weld location.
NUGGET – Only assess nugget damage at the spot weld location.
ALL – Assess both sheet and nugget damage at the spot weld location.
AUTO – Automatically assess damage based on the following condition, if:
𝐷 < 𝛼√𝑇 assess nugget damage at the spot weld location
𝐷 ≥ 𝛼√𝑇 assess sheet damage at the spot weld location
Where, D is the diameter of the weld connector, 𝛼 is set using the ALPHA field, and T
is the thickness of the shell.

TREF and TREF_N parameters are used in the same manner as in Seam Weld Fatigue
(see the chapter 6 for more information)

130
Beside the TREF, TREF_N fields, the stresses can be scaled directly within SF entry. By
default, it is 1.0.

MATFAT
Material Fatigue properties for S-N approach for Spot Weld Fatigue Analysis are
defined in the SPWLD field in MATFAT entry.

Indicates that the fatigue material properties for spot weld fatigue analysis are
SPWLD
to follow.

Mean Stress Sensitivity parameter for mean stress correction based on FKM
FKMMSS_SP
guidelines for spot weld fatigue analysis

A/R Defines the interpretation of the defined SN curve

Fatigue strength coefficient. It is the stress range intercept of SN curve at 1

SR1_Spi
cycle in log-log scale

The first fatigue strength exponent. It is the slope of the first segment of SN
B1_Spi
curve in log-log scale

In one-segment SN curve, this is the cycle limit of endurance

NC1_Spi
In two-segment SN curve, this is the transition point

The second fatigue strength exponent. It is the slope of the second segment
B2_Spi
of SN curve in log-log scale

FL_Spi Fatigue Limit.

SE_SPi Standard Error of Log(N)

Notice that again the parameters are the same as for Seam Weld and for the Wöhler’s
curve parameters Spi suffix is used, where i stays for:

i = 1 for the sheet 1 properties

i = 2 for the sheet 2 properties
i = 3 for the nugget properties

131
FATPARM
The FATPARM card has a SPWLD section for Spot Weld Fatigue parameters. They are
the same as for Seam Weld section, except the NANGLE parameter, which is available
for the spot welds to define how many angles are to be examined for stress calculation
(see Help Documentation for more information) default is 18 degrees.

Flag indicating that the following parameters are used for spot weld fatigue
SPWLD
analysis.

METHOD Spot weld fatigue analysis method

CORRECT Mean stress correction indicator

SURVCERT Certainty of survival.

THCKCORR Thickness correction flag

NANGLE Number of angles to be examined on the sheet and nugget

FATDEF
As mentioned before, FATDEF card in case of Spot Weld Analysis assigns PFATSPW
properties directly to the property of nugget. The Top Stress Fraction can also be
defined.

Top Stress Fraction

(TOPSTR)

Nugget elements,
FATDEF PTYPE
Material

PFATSPW SPTFAIL.... etc.

132
7.3 Tutorial: Spot Weld Fatigue Analysis of a
Welded Beam

• Load existing file Fatigue_Spotweld_cbar_ini.hm

• Load in User Profiles, OptiStruct
• Change representation to ”1D Detailed Element Representation”

• Edit the existing Material “Steel” (MAT1) with the values as presented below

133
 • Define Spotweld properties (σN Curves) by activating the MATFAT box

HINT: Here 1, 2, 3 represent sheet1, sheet2, and nugget

• Create Property PFATSPW; Card Image PFATSPW

134
 • set SPTFAIL to SHEET

• Create Load Collectors FATDEF

➢ Card Image FATDEF

➢ set PTYPE to PBARL

135
• Assign Fatigue Property to SpotWeld with FATDEF

• Create Load Collector FATPARM

• Set the card image to FATPARM
• STRESS -> COMBINE - > SGVN
• CERTNTY, Survcert set to 0,5
• SPWLD checked
• Set CORRECT to FKM
• Enable thickness correction by setting THCKCORR to YES

136
• Create Load Step and refer the FATDEF, FATPARM and FATSEQ subcase
entries by appropriate load collectors

137
• Start OptiStruct ANALYSIS

• Post-Processing in HyperView, Options, Visualisation, BAR, “Cylinder “

• Make 2 Windows to display

• displacement for static load on left side
• damage for fatigue analysis on rigid side

138
• Edit legend as per the values displayed

139
140
8 Vibration Fatigue
In this section, we talk about Fatigue in frequency domain.

1. Sine Sweep Fatigue

2. Random Response Fatigue
Sine sweep fatigue

• Cyclic load

• Stress amplitudes and their no. of cycles are determined by sweeping rate
and FRF analysis results.

• Eventually, it’s fatigue in time domain.

Random response fatigue

• Not cyclic. But we all know the random signal causes damage.

• Stochastic approach.

• Probability of no. of cycles of certain stress amplitudes are suggested by

✓ Dirlik
✓ Lalanne
✓ Steinberg’s 3 bands
✓ Narrow band

141
E-N in vibration fatigue

Support of vibration fatigue based on Strain-Life in OptiStruct

• Random vibration fatigue

• Sine sweep fatigue
No additional input is required to activate eN vibration fatigue. Just specify EN in
FATPARM. Optimization is supported in eN random vibration fatigue.

8.1 Sine Sweep Fatigue Analysis

Sine Sweep Fatigue Analysis: Sine sweep is a well-known method of vibration testing
which is conducted to evaluate the vibration behavior of a specimen. It is the study of
fatigue life of structures under Sinusoidal Loading.

• The setup is very similar to a Frequency Response Analysis setup, with an

additional Fatigue subcase, and additional required Sine Sweep parameters.

• The LCID field on the FATLOAD entry references the subcase ID of the
Frequency Response Analysis subcase.
• Frequency Response Stress results from the Frequency Response Analysis are
used to calculate the frequency-dependent stress range, which are
subsequently used to calculate Fatigue damage based on sweep rate.
• Both Stress-Life and Strain-Life approach is available for Sine Sweep Fatigue
analysis.

Explanation on how to perform Frequency Response Analysis is not covered in this

book. To know more about it, kindly go through our eBook titled “Learn Dynamic
Analysis with Altair OptiStruct”.

142
Acceleration Profile and Sweep rate

Steps for sine sweep fatigue

• Perform FRF analysis with given freq. range

• Get stress amplitude at the i-th freq.
• Figure out no. of cycles at the i-th freq.
• Calculate damage at the i-th feq.
• Add up damages through all the freq.
• Multiply the damage and no. of sweeps to get final damage

No. of cycles (n) at a freq, f

• If Hz/sec : f/v → freq. dependent

• If Oct/min : 1/(v *ln2) → feq. independent

FATPARM Bulk Cards

• The TYPE field on the FATPARM Bulk Data Entry can be used to identify stress-
life (SN) or strain-life (EN) sine sweep fatigue analysis.

143
 • The SWEEP continuation line on the FATPARM Bulk Data Entry can be used to
define the frequencies (NF or DF fields) at which the sine sweep fatigue
calculations are performed. Additionally, the STSUBID field can be used to
identify a static subcase for mean stress correction.

FATLOAD Bulk Cards

• The LCID field on FATLOAD Bulk Data Entry references a Frequency Response
Analysis subcase.
• The SWEEP continuation line should be specified on FATLOAD Bulk Data Entry,
and corresponding sine sweep parameters, via SR (sweep rate) and SRUNIT
(sweep rate unit) fields are required.

FATSEQ Bulk Cards

• The N# fields on the FATSEQ Bulk Data Entry identifies the number of sweeps
of the corresponding FID# reference and the T# fields define the time in sec.

Sine sweep Fatigue: Input Deck

144
145
8.2 Tutorial: Fatigue Analysis under sine sweep
loading
This exercise introduces the user to setting up stress-life (S-N) analysis on a spring
link arm subjected to frequency response (FRF) loading. The coupling of a FRF
subcase to a fatigue simulation is similar to setting up fatigue analysis a linear static
loadcase with the addition of a few important parameters.

Problem Setup

You should copy the file: 7a_sine_seep_fatigue_spring_link.

Step 1: Open the model 7a_sine_seep_fatigue_spring_link.hm in

HyperMesh Desktop

146
Step 2: Create a new FATLOAD card FATLOAD to reference the frf load case,
SWEEP <checked>, SR 0.1, SRUNIT HZPS.

Step 3: Create a new FATEVNTcontaining the single FATLOAD

147
Step 4: Define the fatigue sequence of a single cycle of the previously created
FATEVNT in a new FATSEQ card

148
Step 5: Create a FATPARM card with analysis type set to SN. Use the VONMISES
stress combination, MPA stress units, surface stress YES and survival certainty of
0.92. Check the Rainflow option RTYPE to STRESS and SWEEP <checked>.

Note: Make sure the MATFAT section is activated and set for
MAT1_FATIGUE material: UNIT: MPA; UTS: 450; SN: <checked>;
MEAN_STRESS_SENSITIVITY: 1slope; SRI1: 2401; B1: -0.135135; NC1: 2e6;
FL: 338; SE: 0.3099.

Step 6: Create a new PFAT load collector and set the following properties

Step 7: Create a new FATDEF fatigue definition load collector that links the plate
shell elements to the fatigue element property PFAT

149
150
 Step 8: Create a new load step for fatigue analysis named Fatigue

Step 9: Run the analysis in OptiStruct

Step 10: Review the results in HyperView as a contour plot of damage

151
8.3 Random Response Fatigue Analysis

Random Response Fatigue analysis is available to study the fatigue life of structures
under Random Loading. The setup is similar to a Random Response Analysis setup,
with an additional Fatigue subcase. The LCID field on the FATLOAD entry references
the subcase ID of the Random Response Analysis subcase. Power Spectral Density
(PSD) results from the Random Response Analysis are used to calculate Moments (mn)
that are used to generate the probability density function for the number of cycles vs
the stress range. The PSD Moments are calculated based on the Stress PSD generated
from the Random Response Analysis.

Explanation on how to perform Random Response Analysis is not covered in this book.
To know more about it, kindly go through our eBook titled “Learn Dynamic Analysis
with Altair OptiStruct”.

OptiStruct supports Analysis only for the following modules/approaches:

• Uniaxial SN with Von-Mises stress and approximated principal stress

• Spot weld (CBAR, CWELD, CHEXA)

• Seam weld

Currently following options are available to define the damage model:

• Dirlik

• Lalanne

• Narrow band

• Steinberg's 3 bands
Note that multiple models can be specified. OS reports the worst damage. Mean
stress correction requires static load subcase to apply static stress offset.

152
 Outputs include:

Damage/life of shell, solid, spot weld elements (CWELD, CBAR, CHEXA) and seam weld
element (CQUAD4)

Fatigue Bulk Data Cards: FATPARM extensions for random fatigue

SUBCASE

FATDEF FATPARM FATSEQ

PFAT FATEVNT

MATFAT FATLOAD

TABFAT

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

FATPARM ID TYPE MAXLFAT

STRESS COMBINE CORRECT STRESSU PLASTIC SURFSTS

… … … … … …

FOS FOSTYPE

RDMGMDL DM1 DM2 DM3

RANDOM FACSREND SREND NBIN DS TEXP STSUBID

153
Where:

Indicates Random Response Damage Model information is to

RDMGMDL
follow.

Damage models to be used in Random Response fatigue analysis.

DMi
Multiple damage models can be specified in the three DMi fields.

Indicates that parameters for Random Response Fatigue are to

RANDOM follow; only when the LCID field references a Random Response
Analysis Subcase.

FACSREND Calculates the upper limit of the stress range (SREND)

Can be used to directly specify the upper limit of the stress

SREND range.

Default = SREND based on FACSREND

Calculates the width of the range of stress ranges for which the
NBIN
probability is calculated

DS Can be used to directly define the width of the stress ranges

Total Exposure time of the initial source time domain Random

TEXP
Loading Sample

References the subcase ID of a Static Subcase to account for

STSUBID mean stress correction with any loading that leads to a mean
stress different from zero

154
Random Response Fatigue: Input Deck

8.4 Tutorial: Fatigue Analysis of a cantilevered

part under random loading
This exercise introduces the user to setting up stress-life (S-N) analysis on a
cantilevered shaped plate subjected to random loading. The coupling of a random
subcase to a fatigue simulation is similar to setting up fatigue analysis a linear static
loadcase with the addition of a few important parameters.

155
The random loading setup is complete, containing two PSDF of individual subcases
and one coupled PSDF relating those two subcases. Outputs are carefully selected to
limit the amount of output from the random response analysis, limiting the results to
what is necessary to produce the fatigue analysis.

Problem Setup

You should copy the file: RandomFatigue.hm

Step 1: Open the model RandomFatigue.hm in HyperMesh Desktop

156
Step 2: Create a new FATLOAD card FATLOADto reference the randomload case with
an LDM of 1,and a Scale of 0.1.

Step 3: Create a new FATEVNTcontaining the single FATLOAD

157
Step 4: Define the fatigue sequence of a single cycle of the previously created
FATEVNT in a new FATSEQ card

Step 5: Create a FATPARM card with analysis type set to SN. Use the VONMISES stress
combination, MPA stress units, and survival certainty of 1e-3. Check the RDMGMDL
box to enable damage modeling for random fatigue and select the DIRLIK model as
DM1. Check the RANDOM box and set a FACSREND of 6.0.

Tip: FACSREND is used with the RMS stress from the random response analysis to
calculate the upper stress range for fatigue analysis. For more information, see the
HyperWorks Help Reference Guide for the FATPARM entry.

158
159
Step 6: Activate the MATFAT section of the material properties for Aluminum to
include the following fatigue parameters for stress life analysis:

160
Step 7: Create a new PFATload collector and set the following properties

Step 8: Create a new FATDEF fatigue definition load collector that links the plate
shell elements to the fatigue element property PFAT

161
Step 9: Create a new load step for fatigue analysis named Fatigue

Step 10: Run the analysis in OptiStruct

Step 11: Review the results in HyperView as a contour plot of damage

162
9 HyperLife
HyperLife is an easy-to-learn fatigue analysis software developed on a solver-neutral
framework. It provides a comprehensive toolset for durability analysis, directly
interfacing with all the major finite element analysis (FEA) results files. With an
embedded material library, HyperLife enables prediction of fatigue life under
repetitive loadings experienced in a wide range of industrial applications.

9.1 Fatigue Capabilities in HyperLife

1. Stress Life (SN) and Strain Life (EN)
Uniaxial and Multiaxial assessment options with multiple Mean stress correction
theories. Various stress combination methods are available for Uniaxial assessment.
Critical plane implementation for Multiaxial assessment.

For Stain-Life (EN), Neuber plasticity option available for Uniaxial and Jiang-Sehitoglu
plasticity model (default) for Multiaxial assessment.

2. Factor of Safety
Factor of Safety calculations based on Dang Van fatigue limit criterion which is used
to predict if a component will fail in its entire load history.

3. Weld Fatigue
Weld Fatigue assessment type is available for Spot Welds and Seam Welds.

Spot Weld: Structural stress method implementation when spot weld nuggets are
idealized with CHEXA, CWELD and CBAR/CBEAM element representation.

Seam Weld: Structural stress method implementation for Seam welds idealized with
plate or shell elements. The approach is based on VOLVO method. Supported weld
type is FILLET weld and the weld lines (root and toe) are automatically identified seam
weld fatigue analysis based on VOLVO method.

163
9.2 Benefits of HyperLife
1. Easy to Learn
The easy to use and easy to learn GUI will bring efficiency to simulation teams. The
intuitive user experience enables novice to expert users to perform Fatigue life
predictions at ease.

2. Solver Neutral
The commonly used FEA result data from leading solvers can be interfaced as input
file.

3. Accelerated Decision Making

Perform multiple variations of analysis settings without starting a new session there
by accelerating decision making and saving cost.

9.3 Fatigue methods in HyperLife

HyperLife supports following methods in different categories.

164
9.4 HyperLife Workflow
The following illustrated workflows show the steps involved in performing a fatigue
analysis in HyperLife.

Fatigue Module Fatigue Material Load Spectrum

Fatigue Calculation

Estimated life

165
 Fatigue Calculation

Load
Fatigue Fatigue Spectrum Estimated Life
Module Material

1 2 3 4

166
10 eFatigue
In the eFatigue website (https://www.efatigue.com/), you will find everything you
need to make a fatigue assessment of a structure or component.

The site is divided into several Fatigue Technologies listed in the left sidebar. These
are grouped by types of problems being analyzed. Start by selecting a Fatigue
Technology from the left sidebar. Constant Amplitude is a good place to start.

Each technology is organized into major sections: Fatigue Calculators, Fatigue

Analyzers, Finders and Technical Background. Fatigue Calculators and Fatigue
Analyzers do the calculations. Finders help you get the necessary input data, including
stress concentration factors and a database of standard material properties. Technical
Background describes all the steps and equations used in the calculations.

Every Fatigue Calculator and Fatigue Analyzer has an example problem that will guide
you through a complete analysis. After selecting an analysis type, click on the Learn
By Example button for a step-by-step guide through all of the input data and results.

eFatigue also includes several documented benchmark problems containing all the
information needed for fatigue analysis. Here you can use the Fatigue Calculators or
Fatigue Analyzers to compute fatigue lives and then compare the results of your
analysis with the experiments. These benchmark problems will give you experience in
using eFatigue and increase your confidence in your ability to solve other more
complex fatigue problems.

Fatigue Analyzers are used for more computationally intensive problems such as
processing finite element models and variable amplitude analysis from large data
files. Simply upload your file and process it directly. An eFatigue subscription is
required to access Fatigue Analyzers.

167
10.1 Fatigue Calculator
Fatigue Calculator is a portion of eFatigue website (https://www.efatigue.com/)
which contains technologies and tools that are needed for accurate fatigue
assessments with an interface that is easy for the non-expert to navigate.

One can do the following things quickly with a Fatigue Calculator:

• Conduct a simple fatigue analysis

• Compare results from different methods

• Find a stress concentration factor

• Find material properties

Each Fatigue Calculator includes a "Learn by Example" button that will take you
through its input parameters and methods. Databases for material properties, stress
concentration factors and stress intensity factors are included. Using a Fatigue
Calculator, any engineer can quickly and easily conduct a fatigue or a durability
analysis.

Fatigue analysis methods are based on:

• Stress-Life
• Strain-Life
• Crack Growth

Fatigue Calculators are available in the following sections of eFatigue.

• Constant Amplitude
• Multiaxial
• Probabilistic

In the next section, Constant Amplitude Stress-Life fatigue calculator is explained

with example values. Similarly, one can do any combination from the earlier stated
fatigue analysis methods and fatigue sections of eFatigue website.

168
10.1.1 Constant Amplitude Stress-Life Analysis
All structures and mechanical components that are cyclically loaded can fail by
fatigue. With limited input data, constant amplitude fatigue analysis is used to make
a simple and quick estimate of the likely fatigue performance or durability.

Constant Amplitude Analysis provides analysis for simple constant amplitude loading
for all of the methods. It is typically found in power transmission applications such as
shafts, gears etc. It is frequently used in the early stages of design to set the overall
stress levels and to select appropriate materials. Many design and testing
specifications are written in terms of constant amplitude loading.

The stress-life method is typically used for long life situations (millions of cycles)
where the stresses are elastic. This method is often referred to as infinite life design.

One can make use of the following button: Experienced user on/off. Turn it off if you
are an experienced user and don’t need explanation of each fatigue step. Turn it on
for assistance in performing each fatigue step.

Make use of the button: Learn by example which performs a fatigue analysis using
example values.

The steps involved in performing a constant amplitude stress life analysis with
example values are as follows:

Loading
Loads can be entered as either the maximum and minimum values or as
the alternating stress and mean stress. Note that, Alternating stress is one half of
the stress range.

Specified values:

Stress Amplitude, Sa or ea = 50 MPa

Mean Stress, Sm or em = 10 MPa

169
Material
You may load a material from the database by selecting it and clicking on “Load
Material” or browse the database by clicking the “Material Property Finder” button
or specify individual properties directly. Clicking “Material Property Estimator” will
show the default properties that are computed from the input values.

We have selected 1020 cold rolled steel from the Material Property Finder. Default
values will be used for any properties left blank.

For finite life design, the entire stress-life curve must be known. Stress life curves are
characterized by a slope and an intercept.

Specified values:

Ultimate Strength, Su = 455 MPa

Elastic Modulus, E = 206800 MPa

Intercept, Sf′ = 814 MPa

Slope, b = -.118

Default values:

Fatigue Limit, SFL = 159 MPa

Fatigue Limit Cycles, NFL = 1.00E+06

170
Modifying Factors
Some of the important modifying factors are: Surface finish factor (kSF), the loading
factor (kL ) and the size factor (ksize).

Fatigue usually starts at the surface so that the quality of the surface finish is very
important. The surface finish becomes even more important as the strength of the
material increases.

Either specify the modifying factor directly or choose a finish from the drop-down box.
If you don't know, select None and a default value of 1 will be used.

Specified values:

Surface Finish Type = ground

Loading Factor Type = axial

Default values:

kSF = 0.939

kL = 0.923

ksize = 1.000

Stress Concentration Factor

All mechanical components contain some form of stress concentrators which can
cause cracks to form. A well-designed structure will have stress concentrations
ranging between 2 and 4. Here, we select a stress concentration for a circular hole
where Kt = 3. You could obtain stress concentration factors for other geometries from
the Stress Concentration Finder. If you leave Kt blank, a default value of Kt = 1 will be
used.

Small stress concentrations are less effective in fatigue than predicted by Kt. A fatigue
notch factor (effective stress concentration in fatigue) is used to account for this
effect. It is related to the size of the local stress gradient and material strength.

Either specify Kf directly or enter Kt and the radius.

171
Specified values:

Kt = 3

Use Fatigue Notch Factor = No

Radius, r = 1 mm

Safety Factor
The Safety Factor represents how much you have underestimated the strength of the
material in order to ensure a safe design with a life equal to the fatigue limit.

If you wish to calculate a safety factor, leave this section blank and click the Calculate
Life / Safety Factor button below. If you wish to calculate the stresses, you must
specify the desired safety factor and provide some additional information about the
mean stresses.

In this example, we want to calculate Life / Safety Factor. So, we will leave this ‘Safety
Factor’ section blank.

If we choose to calculate stresses, all the stress information in the loading section will
be ignored and only safety factor will be considered.

Results
Click on the button ‘calculate Life / Safety Factor’. Any parameters determined during
the analysis are given in the Computed section. The equivalent stress Seq is computed
from the Goodman diagram. The various modifying factors change the slope of the
material SN curve in the analysis. The value of the adjusted slope is given.

Results obtained are as follows:

Nf = 486000

Smax or emax = 60 MPa

Smin or emin = -40 MPa

Seq = 54 MPa

beq = -0.208

172
Appendix A
An extended version of Fatigue analysis subcase tree:

SUBCASE

FATDEF FATPARM FATSEQ

Fatigue analysis parameters:
Regions for fatigue analysis. A sequence of FATEVNT and
-type of analysis (SN,EN,FOS),
This card is used to FATSEQ collectors. There is
-Uniaxial/Multiaxial
associate PFAT features with one major FATSEQ, which is
-stress calculation
element sets/actual referenced in subcase entry.
-rainflow
properties.
-multiaxial damage
-survival certainty FATEVNT
PFAT -Seamweld controls
Combination of
Defines component -Spotweld controls
FATLOADs taking place
fatigue properties: -Random response damage
at the same time.
shell element layer, model
finish, treatment, -random response
fatigue strength parameters FATLOAD
Association of TABFAT
correction Kf…
load peaks to a static
loadstep. Additional
MATFAT parameters: Load
Material properties for magnitude, scale,
the following fatigue offset.
models: SN, EN, FOS,
Spot Weld, Seam Weld. TABFAT
Tabular loading time
history: one column of
user-defined number
of load scaling factors

173
Appendix B
This section discusses new features regarding OptiStruct Fatigue analysis.

Fatigue Analysis based on a Local Submodel

Feature supported version: OptiStruct 14.0.220 or above

In Global-local analysis, a full model is solved using two or more submodels. One
submodel represents the full structure, and parts of the structure with small details,
which require relatively higher accuracy, can be modeled as local submodels with a
fine mesh. Displacements from the coarser global model are interpolated and applied
to the finer mesh of the local model at the transfer zone. This allows for
the local model to be driven by the results of the global model.

• This feature may help improve results in models with local stress
concentrations. It allows faster solution time as only parts of the structure is
being resolved with a fine mesh.

• Solid-to-Solid, Shell-to-shell, and shell-to-solid zooming are supported.

• It is an approximate solution, under the assumption that the displacements

are correct in the coarser global model.

Until Version 14.0.220 only linear static analysis was supported

In V14.0.220 Fatigue analysis is also supported

Setup is similar to setting up any fatigue analysis

174
Can pick either global subcase or a local subcase

Can pick either global element set, or local element set

Example

• Symmetric Boundary Conditions

• Pressure Load applied on the plate

• Fatigue Analysis of both global and local submodel

175
Results

Amplitude Based MATFAT

Feature supported version: OptiStruct 2017.2.1 or above

The SN curve considered for fatigue calculation could be either Stress-range V/s No.
of Cycles or Stress-Amplitude V/S No. of cycles. Until Version 2017.2, only Stress-range
based curves were supported. In Version 2017.2.1, we support both. The default is
still stress-range based curve.

176
How to setup in OptiStruct

• The A/R field on the MATFAT card can be used to switch between stress-
range and stress-amplitude based curves.

• A/R defines the interpretation of the defined SN curve.

➢ A: The SN curve is defined based on Amplitude.

➢ R: The SN curve is defined based on Range.

Static Loading Applied Sequentially in Fatigue

Analysis
Feature supported version: OptiStruct 2017.2.1 or above

Static loads can now be applied sequentially in fatigue analysis from Version 2017.2

Until Version 2017.2, static loads could only be super-imposed in fatigue analysis

The SQNTL flag on the FATEVNT entry can be used to switch the applied Static
Fatigue loads to sequential loading instead of super-positional loading.

177
TID field on FATLOAD card needs to be blank

Format

Random Response Fatigue Analysis with Strain-Life

Approach
Feature supported version: OptiStruct 2017.2.2 or above

Static loads can now be applied sequentially in fatigue analysis from Version 2017.2
Strain-Life approach is supported for Random Fatigue Analysis.

Stress life approach is already available.

Only Uniaxial Analysis is supported with combined stress as von-Mises stress.

The TYPE field on the FATPARM Bulk Data Entry can be used to identify stress-life (SN)
or strain-life (EN).

178
Multiple Mean Stress SN curve Definition
Feature supported version: OptiStruct 2017.2.2 or above

Multiple SN data can be defined for different mean stress through the FATMCRV entry
which can be referenced on the SNCM continuation line of the MATFAT entry. The
input on FATMCRV would be the mean stress and TABLEXN IDs. TABLEXN defines
stress vs life. Alternatively, TABLEXN can directly be referenced in MATFAT and in that
case, there is single SN curve in analysis.

• Stress type - A(Amplitude,) R(Range), or MAX (Maximum stress) is supported.

• Available analysis type is static, transient, and random response analysis. No

support for seam weld / spot weld.

• The intermediate curves are interpolated for the damage evaluation.

• Mean stress correction is redundant when Multiple mean stress SN curves are
defined.

179
Multiple Mean Stress SN curve Definition: Bulk Cards

180