0% found this document useful (0 votes)

3 views230 pages

Fe6 Wex

The document provides an overview of the FE-Fatigue software, detailing its purpose in estimating fatigue lives from finite element stress calculations. It includes copyright and warranty notices, usage guidelines, and a series of worked examples to illustrate the software's capabilities. The examples cover various fatigue analysis methods and demonstrate how to use the software effectively in design processes.

Uploaded by

Taah RIGO

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

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

3 views230 pages

Fe6 Wex

Uploaded by

Taah RIGO

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

Available Formats

Download as PDF, TXT or read online on Scribd

| Previous | Next | FeFat6 Menu | Main Menu |

FE-Fatigue
RELEASE 6.0

Copyright Notice
All of this documentation, and the software it describes, are copyrighted
with all rights reserved. Under copyright laws; Neither the documentation or
the software may be copied, photocopied, or reproduced in any way,
translated, or converted into any machine readable form or any electronic
medium, in whole or in part, without express written permission from nCode
International Ltd. (nCode). Failure to obtain such permission may result in
prosecution.

The software suite that comprises nSoft (the product) and FE-Fatigue and
its component programs and files, are the property of nCode. nSoft and FE-
Fatigue are protected under copyright law and are licensed for use only by
a user who has obtained the necessary license, or in the case of a multi
user licence may only be used by up to the maximum number of users
specified in the licence agreement. The sale, lease, hire rental, or any other
reassignment of the product to, or by, a third party without the prior written
consent of nCode is expressly forbidden.

All Worldwide Rights Reserved, 1999, 2000,2001,2002, 2003, 2004.

nSoft documents are viewed using Adobe Acrobat Reader.

“Acrobat® Reader Copyright © 1987-1996 Adobe Systems Incorporated.
All rights reserved. Adobe and Acrobat are trademarks of Adobe Systems
Incorporated which may be registered in certain jurisdictions.”

All other trademarks are the property of their respective owners.

Warranty Notice
Whilst nCode makes the product as reliable as is reasonably possible,
nCode does not warrant that the product will function properly under all
hardware platforms or software environments. Certain combinations of third
party software and/or manufacturers modifications to hardware and
software may impact upon the operation of nCode software.

nCode has tested the software and reviewed the documentation but nCode
MAKES NO WARRANTY, IMPLICIT OR EXPLICIT, WITH RESPECT TO
THE PRODUCT, TO ITS QUALITY, PERFORMANCE,
MERCHANTABILITY, OR FITNESS FOR A SPECIFIC USAGE. THIS
SOFTWARE AND DOCUMENTATION ARE LICENSED ‘AS IS’ AND YOU -
THE LICENSEE - ASSUME THE ENTIRE RISK AS TO THEIR QUALITY
AND PERFORMANCE, WHEN YOU USE THE PRODUCT.

Liability Notice
NCODE WILL NOT BE LIABLE FOR DAMAGES ARISING FROM THE
RESULTS, DIRECT OR INDIRECT, SPECIAL, INCIDENTAL, OR
CONSEQUENTIAL, OF THE LICENCEES USAGE OR MISUSAGE OF
THE PRODUCT, even if advised of the possibility of such damages. In
particular, and without prejudice to the foregoing, nCode has no liability for
any programs or data stored or used with nCode software, including the
costs of recovering such programs or data.

nCode International Ltd.

FE-Fatigue Worked Examples

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

About the worked examples. . . . . . . . . . . . . . . . . . . . . . 4

FE-Fatigue preferences. . . . . . . . . . . . . . . . . . . . . . . . . . 5

1. Simple S-N Example – Bracket analysis . . . . . . . . . . 8

2. Simple E-N example – Bracket analysis . . . . . . . . . . 21

3. Multiple Loads Example . . . . . . . . . . . . . . . . . . . . . . . 42

4. Multiaxial Assessment Example . . . . . . . . . . . . . . . . 63

5. Duty Cycle Analyzer Example . . . . . . . . . . . . . . . . . . 71

6. Transient Analysis and Modal Superposition. . . . . . 91

7. Multiaxial Strain-Life Example . . . . . . . . . . . . . . . . . . 110

8. Multiaxial Safety Factor Example. . . . . . . . . . . . . . . . 134

9. Spot Weld Analysis Example . . . . . . . . . . . . . . . . . . . 145

10. Multiple Curve S-N Example. . . . . . . . . . . . . . . . . . . 159

11. Seam Weld Analysis Example . . . . . . . . . . . . . . . . . 176

12. Temperature Corrected S-N Analysis Example . . . 194

13. Vibration Fatigue Example using NASTRAN . . . . . 212

Introduction
This document describes the use of the software product FE-Fatigue for estimating fatigue lives
from finite element stress calculations. For information about the theoretical basis of nCode’s
software, consult the nSoft online Fatigue Theory manual.
Text that is highlighted like this indicates the name of an input form, a file name, or an option
on a pick list, etc, that is to say something to which your attention should be drawn.

About FE-Fatigue
FE-Fatigue allows fatigue calculations to be performed in the early stages of the design process,
thereby allowing design analysts to assess the performance of designs with respect to durability.
Wherever structural stresses are calculated, fatigue calculations are usually also appropriate.
Where the type of component, material and manufacturing process are suitable, FE-Fatigue can
provide valuable additional design information which can reduce time to market, save money and
provide optimised designs earlier in the design process.
FE-Fatigue takes stress or strain data from a finite element (FE) calculation as input, and
combines this information together with details of the variation of load with time and cyclic
materials data. From this combination it estimates the fatigue damage for each supplied node or
element on the FE model. In order to support the many FE codes on the market and to provide a
consistent interface to the FE-Fatigue analyser, the FE data is translated from its original form
into a generic file format (the .FES file), whose details are supplied in the FES file format
document.
The capabilities of the software are dependent upon the ability of the translator to provide the
data in the correct form to the analyser. The different fatigue calculation methods require different
types of FE results, and the translator must be capable of providing the correct data in order to
complete the link and perform the analysis. Results are provided in a form suitable for loading
back into the appropriate FE post processor for contour plotting of life and damage.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 2

System Flowcharts
The basic system diagram is detailed in Figure 1 overleaf.
The important files are as follows:

• The Partial .FES file – This file contains the FE results and group information required for
the fatigue analysis.

• The Full .FES file – This includes the analysis parameters, and material property and
loading history reference information necessary to complete the fatigue analysis.

• Results files – Contain the results of the fatigue analysis in a variety of formats.
There are two basic ways of creating the partial .FES file; either using the FE-Fatigue results
translator FE2FES or by making use of an interface to FE-Fatigue in your FE pre/post-processor.
Third party interfaces may not support all the features of FE-Fatigue.
Fatigue analysis results are available in a variety of formats ([Link], Hypermesh, Medina,
SDRC UNV, ABAQUS .ODB) for convenient post-processing. In addition, results may be written
to the FE-Fatigue .FER file for postprocessing using the Studio FE-Display option (if licensed).
Studio FE-Display offers some unique features for postprocessing fatigue analysis results.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 3

Loads Materials

.FES
.FES
FE2FES .RPC
file .MDB
file
or
translator
.DAC

FE Partial FATFE Full

model/ .FES (FE-Fatigue .FES
results file main program) file

Pre/Post- .FES
.FES .FES
.FES
Fatigue Other
processor file
file
results
file
file
outputs

FE-Display Studio

Figure 1 Basic system flowchart

About the worked examples

The files for the following examples are supplied in the following directories (Windows and Unix):

• *.ncx files are in the system area directory: <installation dir>\nsdata

• *.ncl files are in the demonstrator files area: <installation dir>\demo

Note that the default installation directory is normally <drive>:\fe-fatigue6
For example: c:\fe-fatigue6\nsdata

Copy files macro

Each FE-Fatigue example has an associated macro (see Figure 2 below), which will copy all the
files required by the example into the current working directory. These can either be accessed
directly from the nsoft prompt, e.g. ncl @fe_bracket_copyfiles, or from the
Macros\System\ nCode\Copy File for FE-Fatigue Examples menu (see below).

Figure 2 Run an FE-Fatigue NCL macro from the prompt or via the Macros menu

Section Example Macro

1 Simple S-N Example fe_bracket_copyfiles.ncx
2 Simple E-N Example fe_bracket_copyfiles.ncx
3 Multiple Loads Example fe_shocktower_copyfiles.ncx
etc...
Detailed instructions for running each worked example are given where appropriate, starting at
example 1, a simple S-N example.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 5

FE-Fatigue preferences
NOTE: There are a number of preferences that can be set in FE-Fatigue which will affect the
behaviour and performance of the software, and in some cases will affect the answers. The
Preferences are accessible from the FE-Fatigue main menu (see below), which appears when
you open a full .FES file. The settings are stored in the home environment and may also be
modified using the environment manipulation module ENM. See the FATFE documentation for
details.

Figure 3 Preferences are accessed via the FATFE main menu

The preferences are divided into three categories:

• Analysis

• Monitor

Figure 4 Analysis preferences: default settings

Figure 5 Monitor preferences: default settings

Figure 6 Spotweld preferences: default settings

1. Simple S-N Example – Bracket analysis

In this example, you will use the Stress-Life (S-N) approach to calculate the fatigue life for the
constant amplitude loading of a bracket. The S-N approach typically relates nominal stresses to
total failure and is suited primarily to ‘long-life’ problems where there are a large number of cycles
to failure (> 103 or 104 cycles).
In this example, the input (partial .fes) file for FE-Fatigue is created using the FE2FES translator
to read the stress results from an ANSYS .rst file ([Link]).
Alternatively, the input file may be created within a FE pre/post-processor if it includes an
interface to FE-Fatigue.
Results may be post-processed either by creating a suitable results file for your post-processor in
FE-Fatigue, or by using the Studio FE-Display option.

1.1 Introduction
This example has a bracket bolted to a box section. A vertical load is applied to the end of the
bracket. This example will calculate how many times this load can be cycled (+/-) before failure
will occur. Exact fatigue material properties of the steel are unknown and will be estimated from
known data such as UTS (ultimate tensile strength).
The following files from the /demo directory are required for this example:
[Link]
[Link]

These files can be copied to your working directory using the Macros pulldown menu in the main
nSoft program: Macros\System\nCode\Copy Files for FE-Fatigue\Examples\Bracket
(fe_bracket_copyfiles.ncx)
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 9

1.2 Displaying the model in Studio

The model may be viewed either in your own pre-post processor (if it supports ANSYS .rst
files) or using the Studio FE-Display.
To use Studio FE-Display:

• Start the main FE-Fatigue interface, select Display from the Analysis Menu, and select
Studio Display/Reporting Tool (studio)
Studio will start and show an empty display area. To insert the model on the display page:

• Select Insert > FE Display

• In the file selection form, select [Link] from the working directory and click on
Open.
The model is shown on the current Studio page.

Figure 7 ANSYS model displayed using Studio

1.3 Translating ANSYS results using FE2ES

The next step is the read the stress data from the ANSYS results .rst file and convert this to a
FES file input deck for FE-Fatigue. We will use nCode’s FE translator program, fe2fes.

• Start the main nSoft interface, select FE-Fatigue from the nSoft Menu, and select
Generic FE to FES Translator (fe2fes)
fe2fes takes the user through a wizard-style interface to select which entities and which results
from an FE Analysis are to be used in a fatigue analysis.
On the first fe2fes form:

• Select Filename: [Link]

• Translator: Select translator from file extension (Or ANSYS rst)

Figure 8 FE2FES: File selection

1.4 Group selection

The first part of the form is the name and type of the FES file to be created.

• Output Filename: bracket_01

• File Type: Binary

This creates a binary FES file, bracket_01.fes.
On the Group Selection form, the part of the model for fatigue analysis is selected by either
Property or Material IDs.
The nodal data from shell elements are selected by:

• Solution Location: Element

• Group Type: Material

The Available Groups column lists all the property IDs found in the rst file. Select by clicking on
the bracket group:

• 1-MAT_1
and click the -> button to move it to the Selected Groups list.

Figure 9 FE2FES: Group selection

The “Combine selected groups into a single group in FES file” option can be checked if you wish
to make it easier to use the same fatigue properties across a whole model.
Click the ungreyed Next > button to move on to the next form. (You can also back up at any
stage using < Back.)

1.5 Results Selection

The Results Selection form lists all the results subcases found in the .rst file.

• Results Type: Stress

• Shell Layer: Bottom

In the group box for FER Options, check the boxes to select:

• Write model

• Write results
This creates a FER file based on the name of the FES file, in this case bracket_01.fer, which
contains the FE model data (nodes and elements, group information) and also the FE results as
used in the fatigue analysis. This data can be later viewed using Studio.
Select the available result set and click the -> button to move it to the Selected Results list.

Figure 10 FE2FES: Results selection

The final form is a summary of the translation.

Check the “Start FE-Fatigue with FES file” box and click on Finish and the fatfe FE-Fatigue
solver program is shown. (Alternatively, you can start fatfe separately and select
bracket_01.fes as the Input Fatigue Filename.)

1.6 FE-Fatigue analysis

In the FE-Fatigue form, enter the following:

• Analysis Type: S-N Analysis

• Stress Units: MPa

• Advanced Options: No

Figure 11 FATFE: Analysis options

It should be noted that FE-Fatigue leaves consistency of units largely up to the user. Care should
always be taken that the units of stresses and applied loads are consistent. For instance, a model
with dimensions in mm with loads in Newtons will by default produce stresses in N/mm2 or MPa.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 14

1.7 Fatigue Loading Setup

Next, you will be prompted to define the loading function. In this case, you are going to define a
simple constant amplitude, fully reversed load from +1 to –1.
Select t is 0.1000000E+01 from the list, and click on OK.
Select:

• Amplitude: Constant, Peak-Peak

Click on OK to fill the spreadsheet list.

Figure 12 FATFE: Loading input form

The list should now read: 1: t is 0.1000000E+01 Constant 1 -1

This definition of load will subject the FE model to a peak positive loading equivalent to the load
applied in the FE model followed by a peak negative loading equivalent to -1 times the load
applied in the FE model. This defines 1 cycle and the damage for this cycle is then used to
calculate the expected lifetime of the component in cycles.

Peak to peak

Peak to peak
1.5

0.5
Load scale factor

0
1 2 3 4

-0.5

-1

-1.5

Figure 13 Constant amplitude fully-reversed load

1.8 Material Information

The Generate option will be used to calculate generic S-N properties for this steel. Select MAT 1
in the list, and click on OK.

• Method: Generate

• Type: Steel

• UTS: 500

• E: 2.1E5

• Kf: 1

• Finish: No Finish

• Treatment: No Treatment

• Scale Factor: 1
and click on OK to fill in spreadsheet list.

Figure 14 FATFE: Material Input form

1.9 Running a Fatigue Analysis

The next question will ask if you want to go ahead and run the analysis – answer Yes. On the
Fatigue Analysis form, accept the defaults:

• Analysis Region: All Data in FES file

• Equivalent Units: 1 Repeats

• Safety Factor Analysis: None

Click on OK. At the “Results Filename Entry” screen select:

• Results Filename: bracket_01

• Output File Format: Choose an option suitable for your FE post-processor, or alternatively
choose the default and use use Studio FE-Display.

1.10 Reviewing Fatigue Results

During fatigue analysis, a window will pop-up showing the most damaged elements.

Figure 15 FATFE: Results Monitor

The most damaged element is 1755 with 97230 cycles to failure. Close the window and exit
FE-Fatigue.
NOTE: If your answers differ from those above, check your Preferences settings.
Analysis preferences should include the following:
Matrix size: 64
Fast analysis single load case: Yes
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 19

1.11 Displaying Results in Studio

Now you can post-process the results wither in your favoured FE results post-processor or using
Studio FE-Display. To use Studio FE-Display, the procedure is described below.
Restart Studio in the same directory as the bracket_01.fer file and all the supported file types
from the working directory will be made available.
NOTE: You can always recheck for files in the working directory, by selecting the double arrow
cursor above the available data window to load all files from current directory.
Alternatively, open the FER file in Studio by using File > Open Data Files and scan for
nCode S3 fatigue results files, and having found bracket_01.fer, move it to the right hand
selection list and choose ‘Add To File List’.

Figure 16 Fatigue results available in Studio

Note the first set of data in the FER file was written from fe2fes:
1 – Time 1.0000e+000: Stress:1
followed by a single set of fatigue data from fatfe because we analyzed the Bottom shell
surface:
1 – Stress-Life (S-N)
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 20

To show contour plots, first insert an FE Display as shown in section 1.2 above and read either
[Link] or bracket_01.fer.
To view the fatigue results, drag the 2- Stress-Life (S-N) result onto the model display. A
contour plot will automatically be displayed.
Automatically, the FE Display detects that it is a fatigue result. The Results/Legend property page
has a check box for Auto Format. With this checked, the contour type will default to an
appropriate contour plot:

Figure 17 Damage contour plot displayed using Studio

2. Simple E-N example – Bracket analysis

In this example, you will use the bracket model and perform a Strain-Life (E-N) analysis to
calculate life to initiate a crack with a variable amplitude loading function.
The Strain-Life method relates the local strain to life to initiate an engineering crack (1-2mm). The
S-N method uses the elastic stresses directly to calculate fatigue lives but the Strain-Life
approach in FE-Fatigue takes elastic stresses or strains from FE analysis and estimates the
actual elastic-plastic strain at each location.
In this example, the input (partial .fes) file for FE-Fatigue is created using the FE2FES translator
to read the stress results from an ANSYS .rst file ([Link]).
Alternatively, the input file may be created within a FE pre/post-processor if it includes an
interface to FE-Fatigue.
Results may be post-processed either by creating a suitable results file for your post-processor in
FE-Fatigue, or by using the Studio FE-Display option.

2.1 Introduction
This example problem has a bracket bolted to a box section. A vertical load is applied to the end
of the bracket. The example calculates how many times a variable amplitude load time history
([Link]) can be repeated before a crack will initiate. The example uses SAE standard
fatigue properties for 1008 steel.
The following files from the /demo directory are required for this example:
[Link]
[Link]

2.2 Displaying the model in Studio

The model may be viewed either in your own pre-post processor (if it supports ANSYS .rst
files) or using the Studio FE-Display.
To use Studio FE-Display:

• Select Insert > FE Display

• In the file selection form, select [Link] from the working directory and click on
Open.
The model will be shown on the current Studio page. (Other displays can be added to the same
page or a single view can be selected by using View > Interactive Mode.)
The model can be dynamically manipulated in the display by holding down either the Control or
Shift keys together with the mouse buttons. (Alternatively, function F2 and F3 keys can be used).
Put the cursor on the display and move mouse to manipulate the model. The controls are
summarized as:
Rotate – CTRL+ Left Mouse Button, or F2 + Left Mouse Button
Pan – CTRL + Right Mouse Button, or F2 + Right Mouse Button
Cursor Zoom – SHIFT + Left Mouse Button, or F3 + Left Mouse Button
Box Zoom – SHIFT + Middle Mouse Button, or F3 + Middle Mouse Button
Squiggle Zoom - SHIFT + Right Mouse Button, or F3 + Right Mouse Button
NOTE: The Controls Summary is also available by clicking the mouse icon on the toolbar.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 23

Figure 18 ANSYS model displayed using Studio

Take some time to investigate the various display options in Studio.

On the toolbar, there are icons for the following options (from left to right):

• Full Plot – scales the view to show the whole model

• Outline – shows edges of the model

• Mesh – shows an element mesh

• Fill – shows a fill/shaded view of the model (default).

• Contour – shows a contour plot. Requires a FER file results set to have already been
identified by dragging onto the model. If no data is present, it performs a Fill plot.

• View Options – set model views

X – View
Y – View
Z – View
Iso – View

Element by ID
Clear labels

• Model Summary – a pop-up list containing:

Model file name
Number of Nodes
Number of Elements
Number of Elements divided by type i.e., solids, shells, bars, others

• Controls Summary – lists the dynamic model controls available via mouse buttons

To further control the display, use the Property pages.

The Properties window may be in either “Dock” or “Pop” mode, depending on the setting under
View > Toolbars and Windows. If in “Pop” mode, the Property pages are not already visible on
screen and the Property pages are accessed by right-clicking on the FE Display.

Figure 19 Properties: Model Parameters

Under Model Parameters, there are various controls for the display of the model. By default,
when the model is rotated and dynamically manipulated, the model is shown as “Outline” plot
type.
To make the model rotate as the current display (i.e. Fill plot), change Move As: Outline to

• Move As: Plot Type

• Click Apply

Investigate the effect of some of the other parameters.

To show only the bracket part of the model, select the Property page for Groups. This page
enables different groups of elements to be added or removed from the display. It also enables the
colour and transparency of the groups to be controlled.

• Click on the check-boxes to uncheck the BEAM_5 and SHELL_4 groups

• Click Apply

Figure 20 Properties: Groups

Now, only the elements for property group SHELL_2 (main bracket) will be visible.

Figure 21 Displaying a Group

2.3 Translating ANSYS results using FE2FES

The next step is the read the stress data from the ANSYS results .rst file and convert this to a
FES file input deck for FE-Fatigue. We will use nCode’s FE translator program, fe2fes.

• Select Filename: [Link]

• Translator: Select translator from file extension (Or ANSYS rst)

Figure 22 FE2FES: File selection

Click Next > to move on to the next form.

2.4 Group Selection

The first part of the form is the name and type of the FES file to be created.

• Output Filename: bracket_01

• File Type: Binary

• Solution Location: Averaged node on element

• Group Type: Property

The Available Groups column lists all the property IDs found in the rst file. Select by clicking on
the bracket group:

• 2-SHELL_2
and click the -> button to move it to the Selected Groups list.

Figure 23 FE2FES: Group selection

2.5 Results Selection

The Results Selection form lists all the results subcases found in the rst file.

• Results Type: Stress

• Shell Layer: Top and Bottom

In the group box for FER Options, check the boxes to select:

• Write model

Figure 24 FE2FES: Results selection

The final form is a summary of the translation.

2.6 FE-Fatigue Analysis

In the first FE-Fatigue form, enter the following:

• Analysis Type: E-N Analysis

• Stress Units: MPa

• Advanced Options: No

Figure 25 FATFE: Analysis options

NOTE: Although stresses were selected from the ANSYS rst file, a Strain-Life (E-N) analysis can
be performed since FE-Fatigue converts the elastic stresses to strains. Elastic strains are
then converted to elastic-plastic strains using a Neuber correction.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 30

2.7 Fatigue Loading Setup

Select t is 0.1000000E+01 from the list, and click on OK.

Figure 26 FATFE: Loading input form

• Amplitude: Variable

• File Type: DAC

• Time History: [Link] (you can use the list button to select the [Link] file
from the current directory)

• Divide stress/strain by: 500

• Scale factor and offset: 1 and 0 (defaults)

Click on OK to fill in the list.
The Divide stress/strain by setting is used to normalize the stresses to stress/unit load.
Typically, this number would be the applied load in the FE model for this sub case.
In the Loading Input form, make sure Done is selected in the list and click on OK to move on to
the next form.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 31

2.8 Material Information

Select SHELL_2, and click on OK.

Figure 27 FATFE: Material input form

Enter the following materials information:

• Material Name: SAE1008_91_HR

• Kf: 1

• Finish: No Finish

• Treatment: No Treatment

2.9 Running a Fatigue Analysis

Enter the following details:

• Analysis Region: All data in FES file

• Equivalent Units: 1 Repeats

• Safety Factor Analysis: None

Click on OK.
In the Results Filename screen select:

• Results Filename: bracket_01

• Shell calculation: Worst

• Output File Format: Default only (or select a format appropriate to your FE post-
processor)

• FER File Output: 2- Add to FER file (name as FES)

Figure 28 FATFE: Results Filename form

2.10 Reviewing Fatigue Results

The Results Monitor shows the results as the job is running.

Figure 29 FATFE: Results Monitor

The most damaged node should be 1985 with 6838 repeats of the time history to initiate a crack.
NOTE: If your answers differ from those above, check your Preferences settings.
Analysis preferences should include the following:
Matrix size: 64
Fast analysis single load case: Yes
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 34

The following warning message will appear. It is a result of the fact that we are using nodal
average stresses from thin shell elements. The individual elements stresses are calculated in
element co-ordinate systems (2D), but in order to get a valid average at the nodes, the stresses
must first be transformed to a common (global) co-ordinate system. The resulting stresses will
typically appear to be 3D stresses. If it is important to have stresses resolved to the plane of the
surface e.g. when carrying out a multiaxial assessment or analysis, either use shell element
centroid stresses or select the "surface nodes only" and "resolve data to local coordinate system"
options in FE2FES.

Figure 30 Results warning message

2.11 Displaying Results in Studio

Restart Studio in the same directory as the bracket_01.fer file and all the supported file types
from the working directory will be made available.
NOTE: You can always recheck for files in the working directory, by selecting the double arrow
cursor above the available data window to load all files from current directory.
Alternatively, open the FER file in Studio by using File > Open Data Files and scan for
nCode S3 fatigue results files, and having found bracket_01.fer, move it to the right hand
selection list and choose ‘Add To File List’.

Figure 31 Fatigue results displayed in Studio

Note the first set of data in the FER file was written from fe2fes:
1 – Time 1.0000e+000: Stress 1
followed by three sets of fatigue data from fatfe because we analyzed both shell surfaces:
2 – Strain-Life (E-N), Top Surface
3 – Strain-Life (E-N), Bottom Surface
4 – Strain-Life (E-N), Worst Case
If you left-click on the first set of data in the FER file and drag the entry onto a empty part of the
Studio page, you will get a table display of the stress components used in the fatigue analysis.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 36

Figure 32 Table display: Stress results

Similarly, you can drag one of the set of fatigue results onto an empty part of the Studio display.
For fatigue data, Studio will automatically sort the data to show maximum damage at the top of
the table.

Figure 33 Table display: Fatigue results

To contour the fatigue results, either read the appropriate results file into your favoured FE post-
processor, or alternatively use Studio FE-Display as shown below.
To show contour plots, first insert an FE Display as shown in section 2.2 above and read either
[Link] or bracket_01.fer. Drag the set of stresses onto the FE model (rather than on
an empty part of the page) and a contour plot will be displayed.
Use the FE Display properties to control the contour plot. The Properties window may be in either
“Dock” or “Pop” mode, depending on the setting under View > Toolbars and Windows. If in “Pop”
mode, the Property pages are not already visible on screen and the Property pages are accessed
by right-clicking on the FE Display.
Under Results/Legend, select:

• Results Type: Von Mises – Top Shell

Note that the contour plots can be rotated using the mouse controls described above. By default
when you rotate a contour plot it will switch to an outline plot. To be able to view the contour as
you rotate it, on the Model Parameters page change Move As: Outline to:

• Move As: Plot Type

• Select Apply

To view the worst case (shell) fatigue result, drag the 4- Strain-Life (E-N), Worst Case
result onto the model display.
Automatically, the FE Display detects that it is a fatigue result. The Results/Legend property page
has a check box for Auto Format. With this checked, the contour type will default to an
appropriate contour plot:
Damage:

• Spectrum: Blue to Red

• Intervals : Logarithmic
Life:

• Spectrum: Red to Blue

• Intervals : Logarithmic
Safety Factor:

• Spectrum: Red to Blue

• Intervals : Linear
All other plots:

• Spectrum: Blue to Red

• Intervals : Linear
Select the life result by selecting on the Results/Legend page:

• Results Type: Life Repeats

Figure 35

In this example, we have analyzed node locations. This means that a nodal or fringe type of
contour plot is produced. A further option on the Results/Legend page is provided for Smooth
Contours which interpolates the colours on the display for a blended contour. (If element results
are available, the default contour plot is a discrete elemental contour where each element is a
single colour.)
Some of the unique features of the display are:

• Specific legend labels for Static Failure and Beyond Cutoff

• Logarithmic scales with actual values (not Log life values)

Colours for these, for where no data is available, and for above and below a specified range can
be set on the Results/Legend page.
It is possible to adjust the contour display further by using the Model Parameters page to

• Turn off lighting

• Add mesh / outline and control its colour

Figure 36

To find most damaged locations, use the Find/Label toolbar option (the “binoculars” icon).
Select Node by ID and enter one or more IDs such as 1985, 617. The labels move with the
model and they can be obscured by fill or contour plots. So, mesh type plots often show the
labels most clearly.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 41

To easily create an image with a white background for a report, select the Print View option on
the Background page. This will automatically also turn the text to black. The FE display can then
be exported as an image file, as HTML or as a MS Word document.

3. Multiple Loads Example

This example introduces multiple simultaneous loads in a fatigue analysis. The principle of linear
superposition is used in the fatigue analysis to combine stresses from finite element analysis with
loading time histories. This approach is useful for simulating multiple loads on structures where
the dynamic (resonant) effects of the structure can be ignored. This is a widely used method in
the automotive industry. This example also includes re-analysis options and methods for
speeding up the analysis process.

3.1 Introduction
A front rail and shock tower section of a vehicle structure is subjected to 3 loads (x,y,z) using FE
analysis. The FE loads are applied as separate subcases to give the resulting stress due to each
load. Measured time history loads ([Link], [Link], [Link]) define the
loading and a strain-life analysis will be performed for the shock tower using material properties
defined by the user.
The following files are required for this example from the demo directory:
[Link]
shock.op2
[Link]
[Link]
[Link]
[Link]

These files can be copied to your working directory using the Macros pulldown menu in the main
nSoft/FE-Fatigue program: Macros\System\nCode\Copy Files for FE-Fatigue\Examples\
LinearStaticShockTower (fe_shocktower_copyfiles.ncx)
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 43

3.2 Displaying the model in Studio

If you wish to view the model before reading the results into FE-Fatigue, you can either use a
suitable FE pre/post-processor to read the NASTRAN input and/or results files (.dat and
.op2), or use Studio FE-Display as follows:

• Start the main FE-Fatigue interface, select Display from the Analysis Menu, and select
Studio Display/Reporting Tool (studio).
Studio will start and show an empty display area. To insert the model on the display page:

• Select Insert > FE Display

• In the file selection form, select [Link] from the working directory and click on Open.
The model is shown on the current Studio page.

Figure 38 Model displayed in Studio (Fill plot with Mesh)

3.3 Translating NASTRAN results using FE2FES

The next step is to read the stress data from the NASTRAN results op2 file and convert this to a
FES file input deck for FE-Fatigue. To achieve this, we will use nCode’s FE translator program,
fe2fes.
fe2fes takes the user through a wizard-style interface to select which entities and which results
from an FE Analysis are to be used in a fatigue analysis. NASTRAN op2 files which have been
created on either PC or on a UNIX workstation platform can be read by fe2fes.

Figure 39 FE2FES: Input file selection

On the first fe2fes form, enter the following:

• Filename: shock.op2

• Translator: Select translator from file extension (or NASTRAN op2)

3.4 Group Selection

The first part of the form is the name and type of the FES file to be created. In this case:

• Output Filename: shock_01

• File Type: Binary

This creates a binary FES file, shock_01.fes.
On the Group Selection form, the part of the model for fatigue analysis is selected by either
Property or Material IDs.
The nodal data from shell elements are selected by:

• Solution Location: Element

• Group Type: Material

The Available Groups column lists all the IDs found in the op2 file. Select by clicking on the
required group:

• 1-MAT_1
and click the -> button to move it to the Selected Groups list.

Figure 40 FE2FES: Group selection

3.5 Results Selection

The Results Selection form lists all the results subcases found in the op2 file.

• Results Type: Stress

• Shell Layer: Bottom(Z1)

In the group box for FER Options, check the boxes to select:

• Write model

• Write results
This creates a FER file based on the name of the FES file, in this case shock_01.fer, which
contains the FE model data (nodes and elements, group information) and also the FE results as
used in the fatigue analysis. This data can be later viewed using Studio.
Select the available results set and click the -> button to move them to the Selected Results list.

Figure 41 FE2FES: Results selection

The final form is a summary of the translation.

Check the “Start FE-Fatigue with FES file” box and click on Finish and the fatfe FE-Fatigue
solver program is shown. (Alternatively, you can start fatfe separately and select
shock_01.fes as the Input Fatigue Filename.)

3.6 FE-Fatigue Analysis

In the first FE-Fatigue form, enter the following:

• Analysis Type: E-N Analysis

• Stress Units: MPa

• Advanced Options: Yes

The Advanced Options screen is shown which allows certain fatigue parameters to be modified.

Figure 42 FATFE: Advanced Options form

The Advanced options include the ability to select Mean Stress Correction method (Smith-
Watson-Topper, Morrow or none for strain-life analysis) and also the Stress/Strain combination
method. The default is Absolute Maximum Principal which uses the largest principal
(positive or negative).
Accept the defaults on this form and click OK.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 48

3.7 Fatigue Loading Setup

Figure 43 FATFE: Fatigue loading setup form

Select the following:

• The first loadcase on the list and click OK

• Select time history using list button and choose [Link] (check that file type is set to
DAC)

• Divide Stress/Strain by: 100

Click OK to accept this data and fill in the table. Then select the following:

• The next loadcase on the list and click OK

• Select time history using list button and choose [Link]

• Divide Stress/Strain by: 100

Click OK to accept this data and fill in the table.
Then select the following:

• The next loadcase on the list and click OK

• Select time history using list button and choose [Link]

• Divide Stress/Strain by: 100

Click OK to accept this data and fill in the table.
Press F5 and view the options to display the time histories and Export/Import the ASCII version
of the table for external editing.
Selecting Display time histories will use MFD to display the multiple files. The time
histories each contain 6001 points.

Figure 44 Loading files displayed using MFD

The Export information to ASCII option writes the contents of the loading form to an
ASCII file named, in this case, shock_01.laf, shown below:
#
#LAF format - "Description","Type",filename,<channel>,divisor,
scalefactor,offset
#
"FZ Subcase-1",DAC,[Link],100,1,0
"FY Subcase-2",DAC,[Link],100,1,0
"FX Subcase-3",DAC,[Link],100,1,0

This enables the data in the loading form to be manipulated/ edited using a spreadsheet program
or text editor.
The Import information from ASCII options looks for a file named ‘jobname’.laf (in
this case shock_01.laf) and fills in the spreadsheet list from the file. This is very useful for
large numbers of load cases and avoids all the data being entered one field at a time.
Use the Return option to return to the loading form and when complete, ensure Done is selected
in the table and click OK.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 50

3.8 Material Information

The list of materials/groups (which appears in the first column) lists the material labels referenced
by the exported elements. (For a nodal analysis, the material label will be for the elements
associated with these nodes.) In this case, we have only exported one component and therefore
only one material is referenced (MAT_1). In general, this could be several materials in each
model perhaps for different strength steels, surface finishes etc. up to 100 different materials per
analysis.

• Select MAT_1 and click OK.

• Press F5 to view the options.

The import and export functions to ASCII are similar to the loading form (uses .maf files in this
case) for handling large numbers of materials.

Figure 45
With Materials manager selected, click OK. This starts mdm, the same materials manager used in
other nCode fatigue products such as FATIMAS EN/SN.
In this example, we will add a material to the database. (By default, this will add a material to the
[Link] file in the working directory. The default material database can be changed in mdm
using Preferences, Database select. For more information on mdm see the nSoft on-line
documentation.)
From the mdm main menu, select Create, data set 1
For Primary name type shock_steel (no spaces are allowed in this name).
Leave other descriptions blank and click OK.
The next form defines the static or monotonic data for the material.
Material Type (number or code):13
This is a plain carbon wrought steel with < 0.2% carbon, (pressing F1 gives a list of descriptions).
YS: Yield Strength (MPa): 253
UTS: Ultimate Tensile Strength (MPa): 363
E: Elastic modulus (MPa): 2.07E5
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 51

Leave the rest blank and click OK.

The next form is for the strain-life (E-N) fatigue data. Enter the following data:
Sf': Fatigue strength coefficient (MPa) 1297
b: Fatigue strength exponent -0.18
c: Fatigue ductility exponent -0.59
Ef': Fatigue ductility coefficient 0.93
n': Cyclic strain-hardening exponent 0.35
K': Cyclic strength coefficient (MPa) 1953
Nc: Cut-off (reversals) 2E8
SEe: Standard Error of Log(e) (Elastic) 0
SEp: Standard Error of Log(e) (Plastic) 0
SEc: Standard Error of Log(e) (Cyclic) 0
Click OK to accept this form; OK all subsequent forms to enter the material into the database.
To view the data graphically, select Graphical display > strain life plot

Figure 46 The graphical strain life plot

When complete, exit from the mdm main menu and you will return to the Material Input form.
With material1 selected, and method set to select, view the list of materials in the database;
shock_steel should be now on the list. Select shock_steel and accept the other defaults.
Click OK to accept this data and fill in the table.
When all of the above stages are complete, select Done in the list and click OK.

3.9 Running a Fatigue Analysis

Enter the following details:

• Analysis Region: All data in FES file

• Equivalent Units: 1 Repeats

• Safety Factor Analysis: None

Click on OK.
In the Results Filename screen select:

• Results Filename: shock_01

• Output File Format: Default only or select a results type suitable for your FE post-
processor.

• FER File Output: 2- Add to FER file (name as FES)

The analysis takes a few minutes and on completion the summary of results should show the
most damaged element to be 26283, with 1005 repeats.
NOTE: If your answers differ from those shown above, check the Analysis Preferences settings.
The Matrix Size should be set to 64.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 53

3.10 Displaying Results in Studio

The fatigue analysis results may be plotted by reading the appropriate results file back into your
FE post-processor. Alternatively, use the following approach to view the results in Studio FE-
Display.
Restart Studio in the same directory as the shock_01.fer file and all the supported file types
from the working directory will be made available.
NOTE: You can always recheck for files in the working directory, by selecting the double arrow
cursor above the available data window to load all files from current directory.
Alternatively, open the FER file in Studio by using File > Open Data Files and scan for
nCode S3 fatigue results files, and having found shock_01.fer, move it to the right hand
selection list and choose ‘Add To File List’.

Figure 47 Fatigue results files available in Studio

Note that three sets of data in the FER file were written from fe2fes, corresponding to the three
load cases:
1 – FZ=1000 N: Stress:1
2 – FZ=1000 N: Stress:2
3 – FZ=1000 N: Stress:3
followed by a single set of fatigue data from fatfe because we analyzed the Bottom shell
surface:
4 – Strain-Life (E-N)
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 54

To show contour plots, first insert an FE Display as shown in section 3.2 above and read either
shock.op2 or shock_01.fer. Drag the strain-life fatigue results onto the FE model (rather
than on an empty part of the page) and a contour plot will be displayed.
To zoom into the contour plot, shift + click on a point on the screen, then drag the mouse to zoom
in or out.

Figure 48 Results displayed using Studio

3.11 Re-analysis and extracting a strain time history

Obtaining fatigue results is not usually the end of the analysis task. If this was a real design
problem, some re-analysis may be required to check sensitivity to certain parameters such as
mean stress correction method, or magnitude of loads, or to meet design objectives.
Start FE-Fatigue again and select shock_01.fes.
Since the shock_01.fes file is a complete FES file which has already been solved, the main
FE-Fatigue menu will be automatically displayed.

Figure 49 The main FE-Fatigue menu form

Selecting Analyse displays the analysis form.
Select Analysis Region: by Nodes/Elements
This gives the ability to specify which nodes/elements to analyse. In this case, we have element
results and so if you give the worst case element id, 26283, you can perform a quick re-analysis.
If you do this you may wish to change the output name to avoid overwriting your previous results.
Selecting Edit job from the main FE-Fatigue menu, enables the job setup to be changed. (This
makes direct changes to the binary FES file so in practice it is recommended that you make a
copy of your original FES file and edit the copy). Review the job setup by exploring these forms.

Select Time History Extraction on the next form.

Figure 51 The node/element options form

This will create a linear elastic time history of the absolute maximum principal strain (or more
generally the component used in the analysis). The file is named shock_0126283.dac, i.e. the
fatigue job name followed by the entity id and is written to the working directory. A summary of
the max and min statistics of this signal is given, followed by the display of the time history using
QLD (Quick Look Display).

Figure 52 Display of a time series (.dac) file.

This time history can be used directly in other nSoft modules for further analysis. One application
would be to use the strain-life analysis module clf to perform what-if studies. clf enables back
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 57

calculations to be performed to find out which parameters would achieve a certain life. It also
creates damage matrices to identify how much damage is attributable to which size of cycles in
the time history.
Currently, care must be taken to redefine the fatigue parameters in clf to be the same as in FE-
Fatigue, and since FE-Fatigue calculates lives from a rainflow matrix rather than a time history,
the resulting lives can be slightly different. Results should however agree if a rainflow matrix is
exported from FE-Fatigue and used in clf.

3.12 Accelerating the Process

The two main contributors to fatigue analysis run-time are:

• Number of Nodes/Elements to analyse

• Number of points in the loading time histories (time steps)

In this shock tower example, the FE mesh is probably too coarse to achieve good stress results
and as meshes become finer so run times increase. Also time histories can be much longer than
the 6 seconds used in this case.
To shorten the time histories module pvxmul is provided. This performs peak valley slicing on
multiple time histories. It relies on the principle that fatigue is only dependant on cycles and
reduces each loading time history (channel) to a peak-valley sequence. The phase relationship is
maintained by retaining points in all the channels if a turning point (maxima or minima) occurs in
any of the channels.
Since this approach removes the time content (and hence frequency content) this is only valid for
quasi-static type analysis where there are no resonance effects.
To run PVXMUL, start the FE-Fatigue main interface (e.g. by typing fatigue6 at the unix prompt
or by clicking on the appropriate shortcut in Windows.
Within the FE-Fatigue interface, select Analysis Menu > FE-Fatigue > Multi-Channel Peak Valley
Slicing (pvxmul) or type pvxmul at the prompt.

• Input File Type: DAC

• Generic Input Filename: load (and press OK)

• Channels: ALL

• Output Filename: load

• Write time file: No

NOTE: The loading time histories need to use this naming convention of a filename stem (in this
case load) followed by a channel number (01,02,03 etc.).
Click OK and on the next form select:

• Gate Method: Cycles

percentage of the largest cycle in each channel.

• For Channel 1, select the cell in column G (% Gate) and enter 5.

• Repeat this action for other channels, or to speed things up press copy to fill down.

• Select File, OK to begin the peak valley slicing.

Three files will be created with a pvx extension ([Link], [Link], [Link]).
Display the new time histories using MFD (remembering to select files with pvx extension) and
the number of points has been reduced from 6001 points to 320 points.

Figure 53 The time histories displayed by MFD

Re-run the fatigue analysis using FE-Fatigue. Start FE-Fatigue from the nSoft menu, or by typing
fatfe at the prompt. Input shock_01.fes and select edit job > edit load input.

• Select the first load case and click OK.

• Change the time history by using the list function and select [Link].

Update all the loading files to pvx.

shock_01.laf and replace all dac by pvx. Use F5 on the loading form to import this ASCII file
to update the list.
Click OK to process the loading form to save the changes, and return to the main menu. Select
analyse and analyse all data in FES file. Change the results file name to shock_01_pvx and
run the job. Select FER File Output to be Add to FER File to add to the existing
shock_01.fer file and give the results case the description shock_01_pvx. The job should
run much quicker than before giving the most damaged element as 26283 with 1006 repeats
– virtually the same as before.
The other main influence on the length of the run time is the number of nodes or elements.
FE-Fatigue includes an auto-elimination option to pre-select nodes or elements on which to focus
the analysis.
This method exploits the fact that for most structures the high stresses are concentrated in local
areas and large proportions of the model experience little or no stress. Auto-elimination uses the
magnitude of the stress due to each loadcase and the magnitude of the load to calculate an
approximate worst case stress for each node or element. This value of stress is then used to filter
out the lower stressed nodes/elements.
From main menu of FE-Fatigue, select Analyse.

• Analysis Region: auto elimination (% model to retain)

• Auto Elimination Retention Factor: 25

• Auto Elimination Stress Threshold: 0 MPa

This will retain 25% of the shock cap for analysis with highest 'total' stresses. This should speed
up the analysis by almost 4 times.
Accept other defaults and click OK.
Change the results file name to shock_01_auto and run the job. Select FER File Output to be
Add to FER File to add to the existing shock_01.fer file and give the results case the
description shock_01_auto.
Note that although the analysis is much quicker, the results file and contour plots only contain
results for this subset of the model. Results will be identical to the shock_01_pvx results since
the loading functions used are the same. Use Studio to compare the results by dragging the
result sets from the FER file onto the page to create tables. Alternatively, use the results listing
module fatres (from the FE-Fatigue menu or Utilities > Results listing) to compare
shock_01_auto.fef and shock_01_pvx.fef. These are the ASCII files of the results of
each analysis in PATRAN format.
Additionally, a stress threshold can be set which can be used to remove further nodes or
elements from the analysis set. This allows an absolute 'fatigue limit', e.g 50 MPa, to be set such
that if the estimated worst case stress for a location is below this limit, then that location will not
be analysed.
This auto-elimination method is very useful but should be used with caution, especially where a
large proportion of the model is being discarded, and where other factors such as mean stresses
may increase damage.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 60

3.13 Using MTS RPC III time history files

A common time history file format in the automotive industry is MTS RPC III (Remote Parameter
Control as used with MTS testing equipment). The RPC III format allows multiple channels of
data in a single multiplexed binary file. FE-Fatigue enables this file format to be used directly, and
even in combination with nCode’s DAC file format. In this example, you will use an RPC III file of
the shock tower loads which have been “peak-valley” sliced to shorten them.
Re-start FE-Fatigue by any of the following methods:
On Unix:

• Start the FE-Fatigue interface by typing fatigue6 at the unix prompt, select FE-
Fatigue > FE-Fatigue Analysis (fatfe) from the Analysis Menu, and select
shock_01.fes; or

• Start FE-Fatigue directly by typing fatfe at the unix prompt and select shock_01.fes
Or, on NT:

• Start the nSoft interface from the Programs menu, select FE-Fatigue > FE-Fatigue
Analysis (fatfe) from the nSoft menu, and select shock_01.fes; or

• Start FE-Fatigue directly by typing fatfe at the prompt and select shock_01.fes
The main FE-Fatigue menu will appear which indicates a completely defined fatigue job has been
read in.

• Select Edit job from the FE-Fatigue main menu.

• Select edit Load input from the Edit FES file menu.

Select:

• The first loadcase on the list and click OK

• File Type: RPC III

• Select time history using the list button and choose [Link]

• Channel number: 1

• Divide stress/strain by: 100

Click OK to accept this data and fill in the table.
Select:

• The next loadcase and click OK

• File Type: RPC III

• Select time history using the list button and choose [Link]

• Channel number: 2

• Divide stress/strain by: 100

Click OK to accept this data and fill in the table.
Select:

• The next loadcase and click OK

• File Type: RPC III

• Select time history using the list button and choose [Link]

• Channel number: 3

• Divide stress/strain by: 100

Click OK to accept this data and fill in the table.
NOTE: The nCode STUDIO program can be used to view RPC files.
When completed, ensure Done is selected in the table and click OK.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 62

To submit the fatigue analysis job, select Analyse from the FE-Fatigue main menu.
Select / enter the following parameters:

• Analysis region: All data in FES file

• Equivalent units: 1 Repeats

• Safety Factor Analysis : None

And click on OK.
Change the results name from shock_01 to shockrpc and click on OK.
The analysis should be relatively quick since the time histories have been edited but should be
equivalent to results from the previous exercises.

3.14 Further Comments on Using RPC III Files

In nSoft V5.3 and FE-Fatigue Release 4 or later versions, support for RPC files by default
assumes the standard MTS format. This means that on all platforms, nSoft is able to use RPC
files from MTS software without any further translation. In previous versions, RPC files needed to
be translated (using CONFIL) from VMS format to the local operating system format.
If users have existing RPC files in local operating system format, it will be necessary to convert
these RPC files to MTS format (VMS) using CONFIL. Alternatively, FE-Fatigue can use these
local format files if a home environment keyword is set (run the ENM module and set the keyword:
$RPCOPSY, value: LOCAL). To use MTS format files, either remove this keyword or set the
value to VMS.
There is also the option on the FE-Fatigue Preferences form, to set RPC III as the default time
history type rather than DAC; the loading form will then default to RPC III as the loading type.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 63

4. Multiaxial Assessment Example

4.1 Introduction
Multiaxial assessment can provide valuable insight into the behaviour of loading at a location of
interest. Understanding conditions that drive the fatigue process will aid selection of the most
appropriate analysis technique to be employed for predicting fatigue performance. The choice of
analysis technique often pivots on deciding how appropriate the use of uniaxial techniques will
be; the capability to make this decision is one that can allow the engineer to avoid unnecessary
use of multiaxial techniques that can be time consuming and possibly, significantly error prone.
In this example a cantilever beam is used. The beam is subjected to a combination of loads
acting at the free end of the beam - a vertical force acting perpendicular to the principle axis of
the beam and a moment acting about the principle axis of the beam.

4.2 Preparation of Example for Multiaxial Assessment

The files required for this example can be found in the nSoft installation directory, \ncode\demo.
These files can be used directly in the demo directory, or as recommended, copied into a new
directory of the user choice. Copying files to another directory can be performed with relative
ease by executing the macro Linear Static Cantilever; this macro can be found in
Macros > System > nCode > Copy Files for FE-Fatigue Examples.
The example described uses stresses recovered from the [Link] .op2 file using FE2FES,
but if the user wishes to use the interface within their own pre/post-processor to create the partial
.fes file, this is also possible, and files are provided to facilitate this.
The files required for this example are listed below (Note: [Link] and
[Link] for HyperMesh; [Link] for SDRC I-DEAS).
Cantilever.op2
[Link]
[Link]
[Link]
Cantilever_force.dac
Cantilever_moment1.dac
Cantilever_moment2.dac
Cantilever_moment3.dac

4.3 Reading Model and Stress Results into the Post Processor
The model and results can be read into the relevant post processor and viewed. Note the
existence of the two load cases. For this example:
The Finite Element results have been calculated for “unit” inputs for both Vertical force and
Moment.
The Cantilever model uses solid elements and has an additional “skin” of shell elements on the
surface.
If you want to view the model and stress analysis results in Studio FE-Display, first carry out the
translation using FE2FES, remembering to write the model and results to the .FER file.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 64

4.4 Writing the Partial FES File

Use FE2FES to extract the surface stresses from the skin of shell elements and write a partial
.fes file. Use the following selections on the Group Selection form (and accept the defaults for the
other fields).

• Output filename: [Link]

• Solution Location: Element

• Group: 1 - SHELL_2

Figure 55 FE2FES: Group Selection form

Click on Next > to proceed to the Results Selection form.

Here, select both loadcases, and if you are going to use Studio FE-Display to post-process the
results, check both the FER options.

Figure 56 FE2FES: Results Selection form

4.5 Fatigue Analyses

The main aim of this example is to demonstrate the use of the multiaxial assessment option in
FE-Fatigue. The process of assessing the behaviour of fatigue loading is performed post fatigue
analysis and usually focuses on the locations that have been shown to suffer the highest levels of
fatigue damage.
For this example, the vertical force time-history remains the same with different (moment) load-
cases used to demonstrate different types of fatigue loading behaviour. Fatigue analysis for this
example should be run in a similar way to previous examples. The following should be noted:

• Analysis type: E-N Analysis

• Stress Units: MPa

• Advanced Options: No

• Fatigue Loading Setup:

Subcase1 – cantilever_force.dac
Subcase2 – cantilever_moment1.dac
(Or cantilever_moment2.dac or cantilever_moment3.dac)

• Material Information: MANTEN

Following the analysis and results viewing, the following menu form is displayed:

Figure 57 The main FE-Fatigue menu form.

Selecting Utilities leads to the next form from which Node/Element options should be
selected; then select Worst case. The following form menu will appear:

Figure 58 Node/Element options form for selected element.

Select Multiaxial Assessment from the Node/Element options.

Give the generic output name: cantilever_1_ .
The following form will be shown.

Figure 59 FE Multiaxial Assessment form.

The operation of this Multiaxial assessment option will produce the following files:

.MAX Maximum principle stress vs.

time
.MIN Minimum principle stress vs.
time
.ABS Absolute maximum principle
stress vs. time
.VON Von Mises stress vs. time
.SHR Shear stress vs. time
.ANG Angle of max. principle vs.
time
.BAX Biaxiality ratio vs. time
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 68

The following Results form is produced:

Figure 60 Multiaxial assessment results form

NOTE: If your answers differ from those shown above, check the Analysis Preferences settings.
The Matrix Size should be set to 64.
The following menu form is then available:

Figure 61 Multiaxial assessment results.

From the multiaxial assessment results form a number of plot options are available. These plots
provide a visual method for assessing the loading behaviour at the element of interest. Each plot
type can be viewed by accepting the option.
For this example, as previously explained, there exist a number of different load-time histories in
order to create different types of loading behaviour. Each loading scenario will need to be run as
a separate fatigue analysis. This is most efficiently achieved by simply editing the load-input form
and re-running the fatigue job; post processing the multiaxial assessment as illustrated will
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 69

produce a results set.

The following combinations of load cases should give three different types of loading behaviour.
Cantilever_force.dac + Stress tensor stationary
cantilever_moment1.dac Biaxiality ratio = 0
Cantilever_force.dac + Stress tensor mobile
cantilever_moment2.dac Biaxiality ratio varies
Cantilever_force.dac + Stress tensor stationary
cantilever_moment3.dac Biaxiality ratio –ve

Cantilever_moment1.dac gives a unixial stress state at the critical location with a biaxiality ratio of
zero. This physically means there is a single dominant principal stress at this location. The angle
of the principal can also be seen to be constant. The fatigue analysis is valid for this condition.

Figure 62 Abs max princ stress vs Biaxiality ratio for cantilever_moment1.dac

Cantilever_moment2.dac gives a stress state that varies with time for the critical location. The
cross-plot of Abs. max principal stress vs Biaxiality ratio and Abs. max principal stress vs Angle
of principal shows the instantaneous values at each point in the loading time histories. The
mobility of the stress tensor indicates a non-proportional stress state that needs a more
advanced multiaxial fatigue analysis.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 70

Figure 63 Abs max princ stress vs Biaxiality ratio for cantilever_moment2.dac

Cantilever_moment3.dac gives a stress state with a constant ratio between the principal stress
and thus a fixed biaxiality ratio, in this case near –0.5.
The effect of this biaxial state of stress can be corrected for by selecting the Hoffman-Seeger
biaxiality correction in the FE-Fatigue Advanced Options form.

Figure 64 Abs max princ stress vs Biaxiality ratio for cantilever_moment3.dac

It is also possible to get an insight into the multiaxiality of the stresses by viewing the contour
plots created by FE-Fatigue in a pre-post processor. The Average Biaxiality plot will indicate a
zero value for a unixial condition but it is also necessary to view the standard deviation of
biaxiality to assess how much this value is varying with time. Since small values of stress tend to
cause scatter on these values, there is a preference in FE-Fatigue to exclude values associated
with small stresses from these calculations. By default stresses below 20% of the UTS are not
considered.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 71

5. Duty Cycle Analyzer Example

This example introduces multiple FE-Fatigue analyses to represent a Duty Cycle or Durability
Schedule. The principal of linear damage summation is used within the duty cycle analyzer to
combine multiple results.
It uses a multiple simultaneous load case FE-Fatigue analysis, and repeats this for multiple sets
of loading time histories. This is a widely used method within the automotive industry to compare
predictive fatigue results with a Proving Ground Durability Schedule.

5.1 Introduction
A front rail and shock tower section of a vehicle structure is subjected to 3 loads (x,y,z) using FE
analysis. The FE loads are applied as separate subcases to give the resulting stress due to each
load. Measured time history loads are available in RPCIII format that define the loading for each
event.
The duty cycle contains 18 different events. These represent 6 different road surfaces; surface
(1-6), and 3 different vehicle weight conditions; w1, w2 and w3. The total time for all 18 events is
2700 seconds (45 minutes) at 204.8 sample rate (~550,000 points per channel).
NOTE: For this example, weight conditions w2 and w3 have been generated by multiplying
weight condition w1 by 1.2 and 1.4 respectively. You can see this by using muledt to
overlay the RPC files: w1_surface1_.rsp, w2_surface1_.rsp and
w3_surface1_.rsp.
The files required for this example are:
[Link]
[Link]
[Link]
[Link]
[Link]
w1_surface1_.rsp
w2_surface1_.rsp
w3_surface1_.rsp
w1_surface2_.rsp
w2_surface2_.rsp
w3_surface2_.rsp
w1_surface3_.rsp
w2_surface3_.rsp
w3_surface3_.rsp
w1_surface4_.rsp
w2_surface4_.rsp
w3_surface4_.rsp
w1_surface5_.rsp
w2_surface5_.rsp
w3_surface5_.rsp
w1_surface6_.rsp
w2_surface6_.rsp
w3_surface6_.rsp
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 72

These files are stored in the installation directory, <installation dir>\demo.

Copy the required example files to your local directory either using system commands or using
the nCL Quickstart Example macros. If this macro is available it can be run by entering:
ncl @fe_dutycycle_copyfiles.ncx
at the nSoft5 prompt or accessed from the Macros > System > nCode > Copy Files for
FE-Fatigue Examples menu (see below).

Figure 65 Macros menu to Copy Files For FE-Fatigue Examples

5.2 Starting FE-Fatigue

From the nSoft5 interface, select the FE-Fatigue sub menu from the nSoft menu, and select
the module FE-Fatigue Analysis (fatfe). Or, enter fatfe at the command prompt.
Select [Link] as the partial FES input file and click on OK.
In the first FE-Fatigue form, select the following:

• Analysis Type: E-N Analysis

• Stress Units: MPa

• Advanced Options: Yes

The Advanced Options screen allows certain fatigue parameters to be modified.

Figure 66 The Advanced Options form.

Accept the defaults on this form and press OK.
NOTE: In the Preferences form, ensure “Rank without loading” is selected.

5.3 Fatigue Loading Setup

Figure 67 Fatigue Loading Setup

• The first loadcase on the list and click OK

• File Type: RPC III

• Time History: Use the list button to choose w1_surface1_.rsp

• Channel number: 1

• Divide Stress/Strain by: 100

• Scale factor, offset: 1, 0

Click OK to accept this data and fill in the table.
Then select:

• Next loadcase and click OK.

• File Type: remains RPC III and the time history file is still w1_surface1_.rsp, but the
channel number is automatically incremented to 2

• Divide Stress/Strain by: 100

• Scale factor, offset: 1, 0

Click OK to accept this data and fill in the table.
Then select:

• Next loadcase and click OK.

• File Type: remains RPC III and the time history file is still w1_surface1_.rsp, but the
channel number is automatically incremented to 3

• Divide Stress/Strain by: 100

• Scale factor, offset: 1, 0

Click OK to accept this data and fill in the table.
Ensure Done is selected in the table and click OK.

5.4 Material Information

The list of materials/groups (which appears in the first column) lists the material labels within this
partial FES file.
Select the following:

• The first material, mat2 on the list and click OK.

• Material: SAE1020_107_HR

• Surface finish: No finish

• Surface treatment: No treatment

Click OK to accept this data and fill in the table.
Ensure Done is selected in the table and click OK.

Figure 68 Figure Material Input

5.5 Submitting the Job

Do not submit the job yet.

Figure 69 Submit job form.

Click No on the “Analyse job now” form..

Figure 70 eXit the FE-Fatigue Analyser

Select eXit on the FE-Fatigue Analyser form, and click OK.

The shock FE-Fatigue job has been created, but has not been analyzed. The above steps have
taken a partial FES file, [Link], and created a full FES file, [Link] for input into the
Duty Cycle Analyser.

5.6 Starting Duty Cycle Analyser

From the nSoft5 interface, select the FE-Fatigue sub menu from the nSoft menu, and select
the module Duty Cycle Analyser (fatduty). Or, from the command prompt, enter
fatduty (alternatively, the duty cycle analyzer can be started from the Utilities menu in fatfe).
Select File > Open FES File to open the FES file. Select the file [Link] and click on
Open.

Figure 71 Open the file [Link]

Figure 72 First event in the Duty Cycle

Click Edit (Free) to edit the first event details.
Select the following:

• Change the event name to Weight 1 : Surface 1

• Set the number of repeats: 50

Click OK to accept this data and fill in the table.

Figure 73 Edit the event details

5.7 Adding Events to the Duty Cycle Analyser

Click Add to add an event to the list.
Select the following:

• RPC File

• Event Name: Weight 1 : Surface 2

• File Name: w1_surface2_

• File Extension: RSP

• Number of Repeats: 125

Click Accept to accept this data and enter the next event.
Select the following:

• RPC File

• Event Name: Weight 1 : Surface 3

• File Name: w1_surface3_

• File Extension: RSP

• Number of Repeats: 200

Click OK to accept this data and fill in the table. There should now be three events in the duty
cycle.

Figure 74 Three events in the Duty Cycle.

The full Duty Cycle Definition is shown in the following table.

Event name File name Repeats
Weight 1 : Surface 1 w1_surface1_ 50
Weight 1 : Surface 2 w1_surface2_ 125
Weight 1 : Surface 3 w1_surface3_ 200
Weight 1 : Surface 4 w1_surface4_ 25
Weight 1 : Surface 5 w1_surface5_ 100
Weight 1 : Surface 6 w1_surface6_ 150
Weight 2 : Surface 1 w2_surface1_ 40
Weight 2 : Surface 2 w2_surface2_ 100
Weight 2 : Surface 3 w2_surface3_ 160
Weight 2 : Surface 4 w2_surface4_ 20
Weight 2 : Surface 5 w2_surface5_ 80
Weight 2 : Surface 6 w2_surface6_ 120
Weight 3 : Surface 1 w3_surface1_ 20
Weight 3 : Surface 2 w3_surface2_ 45
Weight 3 : Surface 3 w3_surface3_ 80
Weight 3 : Surface 4 w3_surface4_ 10
Weight 3 : Surface 5 w3_surface5_ 40
Weight 3 : Surface 6 w3_surface6_ 60

Either use the Add button to add the remaining 15 events to the list, or load the existing Duty
Cycle File that contains the entire schedule of 18 events.
Select File > Load DCF File , respond No to saving the current Duty Cycle File file, and
load the file [Link].

Figure 75 Load an existing DCF file

The complete Duty Cycle is shown below.

Figure 76 The complete Duty Cycle

5.8 Define Job Options

From the main menu select Job Options > Analysis Options to define the analysis
options.

Figure 77 Select Analysis Options – 2% of model

Select:
Analysis Region: Auto elimination (% of model to retain)
Percentage of Model to Keep: 2
Stress threshold: 0
Equivalent units: 1 Schedules
For this model and with these time histories, this particular duty cycle analysis will take a couple
of minutes to complete.
From the main menu select Job Options > Results Options to define the results options.

Figure 78 Select Results Options

Select:

• Intermediate files: Keep All

• Final results file: Select Results Type from FES

• Results filename stem: Use FES stem

• Result:s filename event identifier: Use load case name

• Log file Default

This will output a results file for the duty cycle named [Link]. The output results file for each
event will have the naming convention shock_<RPC filename>.fef
The selection of Extra results file on the Results Option form will produce output results
files for third party post processors.
Select Job Options > Results Summary File Options to define the results summary
file options.

Figure 79 Select Results Summary Options

Select:

• Produce summary results file: Yes

• Number of calculation points: 10

• Column separator: Tab

• Results summary filename: Use FES stem

5.9 Calculate Duty Cycle

Click Calculate to start the duty cycle analysis.

Figure 80 Information display during Duty Cycle Analysis

During the analysis, the status for each event changes to show Results Available,
Running, No results, etc. The bottom left text area shows analysis progress through each
event and the bottom right text area indicates analysis progress through the whole duty cycle.
Do not do this now, but if you ever need to stop the job press the Abort Job button.
When the analysis is complete a summary of the most damaged nodes/elements is shown as a
Brief Results List.

Figure 81 The 10 Most Damaged Elements

These results show the combined damage for all events within the duty cycle. A warning
message is given to indicate that static failure has occurred in at least one element, in at least
one event, by exceeding the ultimate tensile strength of the material. Static failure in Strain-Life
analysis is indicated by a damage of exactly 2.0, and in Stress-Life by a damage of 1.0. It is
however possible for FE-Fatigue and the duty cycle analyzer to accumulate damage greater than
1 during an analysis thus predicting a fatigue failure not static failure.
Click OK to return to the duty cycle analyzer.

Figure 82 After analysis completion

NOTE: If your answers differ from those shown above, check the Analysis Preferences settings.
The Matrix Size should be set to 64.

5.10 Results Assessments

The Brief Results List shown at the end of the analysis can be redisplayed by selecting Results
> Brief Results List from the main menu. It is possible to cut and paste the values from
here into an editor.
Additionally a results summary file, [Link], is created and can be viewed by selecting
Results > View Results Summary File from the main menu.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 85

Figure 83 Header of Results Summary File

The header of the results summary file shows the number of elements in the summary, the FES
filename used in the fatigue analysis and all of the duty cycle events.

Figure 84 Results for element 26283

The body of the results file shows a summary of the results for each node/element. In this
example the results are tab separated which makes it easy for import into Excel, but causes
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 86

misalignment in a plain text view. The “Spaces” option in Results Summary File Options formats
the output for plain text view.
These results show the percentage damage contribution of each duty cycle event to the total
damage. For element 26283 event ’18 : Weight 3 : Surface 6’ the high damage value, 120,
indicates static failure where the load has exceeded the material UTS.
In practice if the stresses at this location are real there will be localized yielding and a
redistribution of the stresses to the area surrounding the plasticity. Alternatively, the high stresses
at this node/element causing the UTS to be exceeded may be because it is located at a loading
point, a constraint or other feature within the finite element model.
For this element the percentage damage contribution of each event is skewed because of the
very high damages associated with event 18.

Figure 85 Results for element 26469

The results for element 26469, further down the table, better illustrate the percentage damage
contribution from each event of the duty cycle.
By importing these results into a spreadsheet, it is possible to reorder the results by percentage
damage contribution per event. From this it can be seen that Surface 6, in all three weight
conditions, is the most damaging, and provides more than half of the total damage for this
element.
Surfaces 6, 2 and 3 provide more than 94% of the total damage for this element.
Surface 4 causes less than 1% of the total damage for this element.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 87

Element 26469
Event Description Damage %Contribution Cumulative
18 Weight 3 : Surface 6 0.09350 23.2 23.2
12 Weight 2 : Surface 6 0.09259 22.9 46.1
6 Weight 1 : Surface 6 0.04779 11.8 57.9
8 Weight 2 : Surface 2 0.04530 11.2 69.1
14 Weight 3 : Surface 2 0.04411 10.9 80.1
2 Weight 1 : Surface 2 0.02208 5.5 85.5
15 Weight 3 : Surface 3 0.01467 3.6 89.2
9 Weight 2 : Surface 3 0.01353 3.4 92.5
3 Weight 1 : Surface 3 0.00653 1.6 94.1
17 Weight 3 : Surface 5 0.00538 1.3 95.5
11 Weight 2 : Surface 5 0.00483 1.2 96.7
13 Weight 3 : Surface 1 0.00426 1.1 97.7
7 Weight 2 : Surface 1 0.00399 1.0 98.7
5 Weight 1 : Surface 5 0.00227 0.6 99.3
1 Weight 1 : Surface 1 0.00188 0.5 99.7
16 Weight 3 : Surface 4 0.00050 0.1 99.9
10 Weight 2 : Surface 4 0.00040 0.1 100.0
4 Weight 1 : Surface 4 0.00017 0.0 100.0
Total Damage 0.40378
Life 2.4766 Schedules

Figure 86 Duty cycle events ranked by percentage damage contribution

A results file [Link] will be created for the complete duty cycle schedule. In addition, each
event will create a results file shock_<event>_.fef. Processing these files with the utility
fef2fer will produce results files suitable for viewing in Studio FE-Display.
NOTE: Results for individual events are for 1 repeat of the load, and have not been scaled by the
number of repeats, they therefore do not reflect results for complete events. This enables
this scalar to be modified easily, and the life re-summed.

5.11 Accelerated Analyses

The following table gives an indication of the analysis time for various combinations of
accelerated analysis using Percentage of Model to Keep and
Percentage Peak-Valley Gate.
The actual acceleration factors that will be achieved are dependent upon your models and load
time histories. These are used to illustrate this example.
The total time for all 18 events is 2700 seconds (45 minutes) at 204.8 sample rate (circa 550,000
points per channel).
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 88

# Percentage Number of Percentage Number of Analysis Time

of Model to Elements Peak-Valley Points per
Keep Gate Channel
1 All 547 None ~550,000 3 hours 8 minutes
2 20 110 None ~550,000 38 minutes
3 10 55 None ~550,000 20 minutes
4 5 28 None ~550,000 11 minutes
5 2 11 None ~550,000 4 minutes
6 All 547 90 150* 4 minutes
7 All 547 80 246* 4 minutes
8 All 547 60 663* 4 minutes
9 All 547 40 1815* 4 minutes
10 All 547 20 6852** 7 minutes

* : 4608 points, each surface was 1 RPC3 frame of 256 points (18*256=4608)
** : 9216 points, combining surfaces with more than 1 RPC3 frame
Figure 87 Various Analysis Times for Accelerated Jobs
Analysis acceleration using both the Percentage of Model to Keep and Percentage
Peak-Valley Gate options are linear, and very effective.
The table also illustrates the effect of peak-valley gating using the RPC3 file format. For example,
#7 using an 80% cycles gate, reduces the circa 550,000 point original time histories down to 246
turning points, however the structure of the RPC3 file means that the actual number of points
used within the duty cycle is 4608. The event w1_surface1__pvx80_.rsp contains 21 turning
points, however to fill the RPC3 frame, these are padded with 235 values.
The RPC peak-valley gating has been applied using the macro [Link]. This macro
can be started using the command ncl@rpcpvxmul at the nSoft command line promt, or
Macros\System\nCode\FeFatigue\nCl macros menu. This macro pads the RPC3 frame
with the last value, which ensures no additional cycles are introduced.
This demonstrates that if RPC3 peak-valley files are used then their size is determined by the
RPC3 buffer size, and the number of frames per buffer. If nCode DAC files had been used here,
then the number of points in the file will be the actual number of turning points, leading to a faster
analysis.
Thus for RPC3 peak-valley files the optimum percentage gate is related to both the fatigue
damage predicted and using a full RPC3 buffer. There is no point in excessively gating the time
histories, only for the buffer to be padded.
The peak-valley gating for these events was able to make very large reductions in the number of
data points because there were only 3 channels within each RPC3 file. When working with a
more complex chassis component, or a full body structure there will be many more channels, i.e.
circa 60. The effect of peak-valley gating on these will be reduced.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 89

5.12 Comparing Fatigue Predictions for Accelerated Analsyes

Four tables are shown for elements; 26283, 26280, 26275 and 26276. These tables examine the
results for these elements in more detail, and compare them with different accelerated analyses.
The tables include the element ranking, the total duty cycle damage and life.
The tables show the expected reduction in fatigue damage as the peak-valley cycles gating
increases.

Percentage Percentage Ranking # Duty Cycle Duty Cycle

of Model to Peak-Valley Damage Life (Sched-
Keep Gate ules)
20 None 1 127.154* 0.00786445
10 None 1 127.154* 0.00786445
5 None 1 127.154* 0.00786445
All 90 1 122.4 0.00816993
All 60 1 125.6 0.00796178
All 20 1 126.9 0.00788022

* : Static failure, stress exceeded material UTS, occurred in one or more events
Figure 88 Comparison for Element 26283

Percentage Percentage Ranking # Duty Cycle Duty Cycle

of Model to Peak-Valley Damage Life (Sched-
Keep Gate ules)
20 None 2 1.45641* 0.686618
10 None 2 1.45641* 0.686618
5 None 2 1.45641* 0.686618
All 90 2 0.52 1.9
All 60 2 1.119 0.893655
All 20 2 1.384 0.722543

* : Static failure, stress exceeded material UTS, occurred in one or more events
Figure 89 Comparison for Element 26280

Percentage Percentage Ranking # Duty Cycle Duty Cycle

of Model to Peak-Valley Damage Life (Sched-
Keep Gate ules)
20 None 5 0.751304 1.33102
10 None 5 0.751304 1.33102
5 None 5 0.751304 1.33102
All 90 5 0.294 3.40136
All 60 4 0.667 1.49925
All 20 5 0.74 1.35135

Figure 90 Comparison for Element 26275

Percentage Percentage Ranking # Duty Cycle Duty Cycle

of Model to Peak-Valley Damage Life (Sched-
Keep Gate ules)
20 None 10 0.516636 1.9356
10 None 10 0.516636 1.9356
5 None 10 0.516636 1.9356
All 90 9 0.207 4.8309
All 60 9 0.449 2.2271
All 20 10 0.504 1.9841

Figure 91 Comparison for Element 26276

6. Transient Analysis and Modal Superposition

In structures where the loading can significantly excite resonant modes, linear static
superposition may not accurately calculate the resulting stress time histories. When resonances
are excited, displacements in certain locations will be greater than if static loads are applied, and
therefore stresses in the model also increased.
A transient finite element analysis will account for these dynamic effects by representing the
mass, stiffness and damping in the structure and also by including dynamic loading time histories
in the analysis. NASTRAN can perform a Direct Transient analysis (sol 109) and in this case
each time step is a separate problem to be solved. In this way, the CPU time required for large
numbers of time steps can be prohibitive as model size increases. Another major issue is disk
space required to store all the results for a large number of time steps.
An alternative is Modal Transient analysis (sol 112) where the dynamic properties of the FE
model are approximated by a modal representation. Once the required modes of the structure
are calculated, the calculation of the transient response is relatively rapid and in a majority of
cases there will be significant CPU time savings by using a modal approach. However, the modal
approach raises the question of the required number of modes to sufficiently represent the
structure, and the issue of disk space for results is the same as the direct method.
There is an alternative method of handling the data from modal transient analysis which can be
more efficient for fatigue analysis. In this approach the required modal superposition is performed
inside FE-Fatigue rather than in NASTRAN, and this is explored in the following example.

6.1 Introduction
In this example, a 250mm cantilever beam is excited by two simultaneous loads at its free end,
1. A torsion load, constant amplitude 200 [Link] sine wave at 80 Hz
2. A variable amplitude vertical load (b_vert.dac). A frequency analysis of the time history has
shown two significant frequencies, 65 Hz and 82 Hz.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 92

Figure 92 b_vert. dac loading time history

Figure 93 frequency content of b_vert.dac (power spectral density)

A NASTRAN modal analysis (sol 103) has been performed and calculated the first 10 modes of
the beam, the first 5 modes being:
1. 10.04 Hz 1st bending mode
2. 31.10 Hz 1st lateral mode
3. 62.91 Hz 2nd order bending mode
4. 77.63 Hz Torsion mode
5. 176.08 Hz 3rd order bending mode
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 93

Figure 94 2nd order bending mode – 62.91 Hz (Von Mises stress)

Figure 95 Torsion mode – 77.63 Hz (Von Mises stress)

The torsion load is close in frequency to the torsional mode and the 65 Hz content in the variable
amplitude load will tend to excite the 2nd order bending mode (62.91 Hz).
This example will calculate how many times this loading can be repeated before a crack will
initiate, using HYPRESS20 steel material properties. The following files are required for this
example from the demo directory:
b_modal.dat
b_modal.op2
b_modal.res
b_tran.dat
b_tran.op2
b_tran.res
b_transp.dat
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 94

b_transp.pch
b_vert.dac

These files can be copied to your working directory using the Macros pull down menu in the main
nSoft program: Macros\System\nCode\Copy Files for FE-Fatigue
Examples\ModalSuperpositionCantilever (fe_cantilever2_copyfiles.ncx)

6.2 Process the Modal Results using FE2FES

The modal results are in the NASTRAN output2 file b_modal.op2.
FE2FES is used to convert the stress from this file into input files for FE-Fatigue and studio
FE-Display.

• Start FE2FES by selecting FE-Fatigue > Generic FE to FES Translator from the pull down
Analysis Menu.

• Select Filename, and pick the NASTRAN results file b_modal.op2 as input.

• Select Next>.
The cantilever is primarily modeled using solid brick elements (CHEXA8), but in this example, very
thin shell elements (1E-4 mm thick) have been used to recover the surface stresses. This has the
advantages of:
a) Defining a reduced element set of interest (surface only)
b) Providing surface resolved stresses (which aids multiaxial assessment)
c) Post processing can be clearer on an element basis rather than averaged nodal contours.
On the Group Selection form of FE2FES:

• Set the Output Filename to b_modal_01.

• Set the File Type to Binary.

• Set the Solution Location to Element.

• Set the Group Type to Property.

Select 1-SHELL_39005 in the Available Groups window and transfer it to the Selected Groups
window using the -> button.
Select the Next> button to complete the group selection.

Figure 96 Group selection in FE2FES

On the Results Selection form of FE2FES:

• Set the Shell Layer to Top (Z2).

• Set the Load Case Description to Description 1.

• Select all the load cases in the Available Results window by picking the Select All
button beneath the window. Transfer these load cases to the Selected Results window
using the -> button.

• In FES Options, ensure that Memory based translation is checked.

• In FER Options, check the Write results option.

• Select the Next> button to complete the results selection.

Figure 97 Results selection in FE2FES

Two files have been created, b_modal_01.fes is the input file for FE-Fatigue, and
b_modal_01.fer is a results file to allow viewing of the modal stresses in Studio FE-Display.
The actual values of stress for each mode are not in themselves directly significant since the
eigenvectors (i.e. the displacements) are normalized with respect to each other. In this example,
the modes are ‘mass’ normalized which means the eigenvectors are scaled by forcing the mass
of each mode to be unity (1.0). Alternatively, modes could be normalized by setting the maximum
deflection in each mode to be unity. This is set on the NASTRAN EIGRL card.

6.3 Process the Transient Results using FE2FES

A modal transient analysis (sol 112) has been defined in b_tran.dat and this NASTRAN
deck can be viewed using a text editor. The modal definition uses the same EIGRL card as the
above modal analysis. (Alternatively, a restart on the modal analysis could have been used.
Restarting as a transient analysis would have prevented the modes having to be recalculated.)
The torsion load is defined using a TLOAD2 card to define a constant amplitude 200 [Link] sine
wave at 80 Hz
The variable amplitude load of b_vert.dac is defined using a TLOAD1 card with reference to a
TABLED1 card of paired X-Y time history data.
This TABLED1 card was created using an nCL script, [Link] which is supplied with FE-
FATIGUE. This script converts multiple DAC files into TABLED1 cards.
Damping of 5% of critical is defined on the TABDMP1 card.
The TSTEP cards defines the required number of output steps and in this case it follows the
b_vert.dac time history:

• 1 second of data

• 750 time steps

• 1.333E-3 seconds between each step

These results are in the NASTRAN output2 file b_tran.op2.
FE2FES is used to convert the stress from this file into an input file for FE-Fatigue.

• Start FE2FES by selecting FE-Fatigue > Generic FE to FES Translator from the pull down
Analysis Menu.

• Select filename, and pick the NASTRAN results file b_tran.op2 as input.

• Select Next>.
On the Group Selection form of FE2FES:

• Set the Output Filename to b_tran_01.

• Set the File Type to Binary.

• Set the Solution Location to Element.

• Set the Group Type to Property.

Select 1-SHELL_39005 in the Available Groups window and transfer it to the Selected Groups
window with the -> button.
Select the Next> button to complete the group selection.
On the Results Selection form of FE2FES:

• Set the Shell Layer to Top (Z2).

• Set the Load Case Description to Description 1.

Select all the load cases in the Available Results window by picking the Select All button
beneath the window. Transfer these load cases to the Selected Results window using the ->
button.
In the FES Options, check both Produce Transient FES File and Memory based
translation.
In the FER Options, ensure that Write model and Write results are deselected.

Figure 98 Transient results selection in FE2FES

Select the Next> button to complete the results selection.

The translation of the transient dynamic stresses is now complete, and FE2FES may be exited
by selecting the Finish button at the bottom of the summary panel.
The file b_tran_01.fes has been created, and may be used as an input file for FE-Fatigue.

6.4 Running FE-Fatigue

Start FE-Fatigue by selecting FE-Fatigue > FE-Fatigue Analysis from the pull down Analysis
Menu.

• As Input Fatigue Filename, select b_tran_01.fes.

• Select OK.
In the first FE-Fatigue form, select the following:

• Analysis Type: E-N Analysis

• Stress Units: MPa

• Advanced Options: No
The loading form is not required for a transient data in a FES file since the time histories of stress
are already defined.
Complete the Material form with the following information, then click OK:

• Select pshell_39005, and click on OK

• Method: Select

• Material Name: hypress20

• Kf: 1

• Finish: No Finish

• Treatment: No Treatment

To exit the Material form, ensure Done is selected and click on OK..

Figure 99

• Analysis Region: All data in FES file

• Equivalent Units: 0.01666 minutes ( = 1 second)

• Safety Factor Analysis: None

Click on OK.

• Set the Output File Format to Default only.

• Set FER File Output to 2-Add to FER file (name as FES).

• Enter a Results Description of dynamic transient results.

• Set Create FPP file to Yes.

6.5 Reviewing Fatigue Results

During analysis, the 10 Most Damaged Elements summary screen (see below) shows that the
shortest life occurs at element 656 with a life of 158 minutes (9481 repeats).

Figure 100 Summary of results

The results can be viewed using Studio FE-Display (see below)

Figure 101 Life in minutes from transient analysis

6.6 Modal Superposition using FE-Fatigue

An alternative method of performing a fatigue analysis from dynamic stress data is using modal
superposition. In this section, you will repeat the above analysis using modal superposition and
then compare the results.
Modal superposition uses FE-Fatigue to perform linear superposition of modal stresses with time
histories of modal responses. This superposition is the same calculation that is done within
NASTRAN when stresses are requested in a modal transient analysis.
The NASTRAN deck, b_transp.dat, shows the required output request to recover the modal
responses from a modal transient analysis. The only change from the previous analysis is the
output request should only be:
SDISP(PUNCH)=ALL
This creates an ASCII punch file of the time history response of each mode. In comparison with
the 120MB punch file of transient stress results, this punch file of modal responses,
b_transp.pch, is only 1.2 MB.
To convert pch file data to DAC files of modal responses:
Start nSoft5 and set directory to where the example files are.
Run nCL script [Link] by starting nSoft5 interface and type at the nSoft prompt:
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 103

• ncl @modalsup (or use the pulldown menu Macros\System\nCode\FE-Fatigue\

Modal_Superposition [Link])

• Select b_transp.pch

• Define output LAF filename as b_modal_01

The LAF file is a list of the created time histories and this automates the association of time
history with modal stresses. The time histories created are named:
b_transp_m01.dac – for response of mode 1
b_transp_m02.dac – for response of mode 2
etc..

Figure 102

Figure 103 [Link]

Start FE-Fatigue by selecting FE-Fatigue > FE-Fatigue Analysis from the pull down Analysis
Menu.

• As Input Fatigue Filename, select b_modal_01.fes.

• Click OK.
In the first FE-Fatigue form, select the following:

• Analysis Type: E-N Analysis

• Stress Units: MPa

• Advanced Options: No
FE-Fatigue should then detect that a LAF file exists (b_modal_01.laf) with the same name as
the current FES file and suggest that it be used. Answer Yes to using the LAF file. This will
automatically complete the loading form and will show the association between the modal
stresses and the modal response time histories.

Figure 104 Loading form

At this point, it is perhaps worth taking a look at the DAC files of modal response to get an insight
into the response of the structure.

• On the Loading form, select F5 for options and select Display time histories
The time histories show the response in generalized coordinates which is in effect the amount
each mode contributes to the total solution at each instant in time.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 105

The magnitude of the response therefore gives the contribution:

• Mode 1, which is vertical bending 10.04 Hz, is being excited by the vertical load but, since
the frequency of the vertical load is much higher, the response dies down after the initial
transient response. It is possible to see the 10 Hz frequency of the mode in the time
history by counting the 10 ‘sine waves’ of reducing magnitude.

• Mode 2, which is the lateral mode 31.1 Hz, is not being excited by the vertical and torsion
loads and consequently has a response close to zero (1e-14).

• Mode 3, which is the second order bending, has the largest response as expected from
the frequency content of the vertical loading.

• Mode 4, which the torsion mode, is being excited by a constant amplitude sinusoidal load
and the response can be seen to build up from zero to a constant amplitude response as
limited by the level of damping (in this case 5% of critical).

Figure 105 Modal response time histories

Exit mfd to return to the Loading form. To exit the Loading form, ensure Done is selected and
click on OK.
The Material form should be completed as the above transient analysis with the following:

• Select material1, and click on OK

• Method: Select

• Material Name: hypress20

• Kf: 1

• Treatment: No Treatment

• Scale factor: 1
Click on OK.
On the Analysis form:

• Analysis Region: All data in FES file

• Equivalent Units: 0.01666 minutes

• Safety Factor Analysis: None

Click on OK.

• Set the Output File Format to Default only.

• Set FER File Output to 2-Add to FER file (name as FES).

• Enter Results Description dynamic transient results.

• Set Create FPP file to Yes.

Click on OK.
On completion, the summary of most damaged elements shows the same fatigue lives as for the
full transient FES file – shortest life is at element 656 with a life of 158 minutes (9481 repeats).
NOTE: If your answers differ from those shown above, check the Analysis Preferences settings.
The Matrix Size should be set to 64.

6.7 Comparison of Results

To compare the results from the full transient and modal superposition:

• Use Studio FE-Display to compare the transient dynamic and modal superposition results.

• Use results listing in FE-Fatigue (fatres) to compare b_tran.fef and b_transp.fef

To further compare results, export time histories of strain for element 656 by running FE-
Fatigue with b_tran_01.fes and b_modal_01.fes and selecting:

• Utilities: Node/element options, Worst case, Time history

extraction
Overlay and cross-plot the resulting DAC files to show the results are identical using mfd (multi-
file display), which can be started from the main nSoft menu.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 107

It is also worth checking the multiaxial assessment, especially since in this case we are applying
a torsion and bending load simultaneously. The cross-plots of biaxiality ratio and angle with
absolute maximum principal show that both are generally fixed and the fatigue analysis is valid
for this location.

Figure 106 Biaxiality ratio vs abs max principal

Figure 107 Angle of principal vs abs max principal

6.9. Summary and Further Considerations

Summary of Modal Superposition process:
1) Run NASTRAN modal analysis (SOL = 103) and request stress (STRESS = ALL). Use
EIGRL card to select frequency range of interest and/or the number of modes. Create a FES file
of modal stresses using FE2FES, ensuring that the Produce Transient FES File option is
not checked on the Results Selection form.
2) Run NASTRAN modal transient (SOL = 112) with same EIGRL card. Define the time history
loading in the NASTRAN deck using TABLED1 cards. (Use [Link] to convert DAC to
TABLED1 cards). Request the output to be SDISPLACEMENTS (PUNCH) = ALL. This gives the
modal response time histories. If 6 modes are solved for, this will give an ASCII punch file with 6
time histories.
3) Run [Link] to convert punch file to DAC files of modal response.
4) Run FE-Fatigue to set up the fatigue analysis. The LAF file created by [Link]
associates modal stresses with the appropriate modal response time history.
5) Run fatigue analysis and post process as usual.
By using the process of modal superposition we have shown that the same fatigue lives can be
calculated, but that the volume of data required is considerably reduced.

Disk space required for Full Transient:

Transient punch file 120 MB
FES file of transient stress (b_tran.fes) 6.3 MB

Disk space required for Modal Superposition:

Modal response punch file (b_transp.pch) 1.2 MB
DAC files of modal response 0.04 MB
FES file of modal stress (b_modal_01.fes) 0.7 MB
The reason for this large saving in disk space is that FE-Fatigue performs the linear
superposition and rainflow cycle counts for each location, and only needs to store the rainflow
matrix of stress. FE-Fatigue does not need to store the time history of stress and therefore saves
a large amount of space.
The time for the transient NASTRAN analysis is essentially the same for both methods, although
more time is required to write the large results files and the FES file for the transient analysis.
More time is required by FE-Fatigue for the modal superposition approach since it also needs to
perform the linear superposition, but this is not usually prohibitive.
It should be noted that another advantage of using modal superposition is that the fatigue
analysis itself can be significantly accelerated by:

• auto-elimination to remove low stress elements

• peak valley slicing to shorten the modal responses. This is valid since the modal
response already include the dynamic transient effects and the calculation is directly
analogous to linear static superposition. In other words, once the modal responses are
calculated, the fatigue analysis only requires cycles of stress (or strain) to be identified.
The modal response time histories also provide valuable engineering insight into the contribution
of each mode to the total solution, as illustrated in this example. The effect of “removing” a mode
can also be investigated by applying a 0.0 scale factor to the modal response time history for that
mode.
For application to other engineering problems, further considerations may be required. In the
NASTRAN analysis, there is an issue of the number of modes required to accurately model the
dynamics. The accuracy of local responses of a structure can be improved by the use of residual
flexiblity in NASTRAN (using PARAM,RESVEC,YES). This allows static loads to be used as
additional ‘modes’, which enable the deflection and stresses near a load application point to be
captured without having to use an extremely large number of additional, and higher frequency,
modes. It may also be necessary to account for the effect of gravity on a structure, which may
increase stresses or change the mean stress level.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 110

7. Multiaxial Strain-Life Example

7.1 Introduction
The Fatigue Design and Evaluation (FDE) committee of the Society of Automotive Engineers
(SAE) conducted an extensive round-robin testing program on a notched shaft [1]. The program
involved combined bending and torsion loading of the shaft until fatigue failure. Over 100 notched
specimens were tested, and input loads and observed lives to failure recorded. The stress-strain
histories produced at the notch root were multiaxial, either proportional or non-proportional.
Geometry of the notched shaft is given in Figure 108. The shaft was designed such that fatigue
failures occur at the notch with 5 mm root radius.
Material of the shaft was SAE1045 steel in hot-rolled and normalized condition. All specimens
were produced from a single heat. Table 1 summarizes monotonic and cyclic parameters of this
material, obtained from separate sets of tests.
In this example, fatigue life predictions based on FEA results from a refined model of the shaft
are compared with experimentally observed lives to failure.

Figure 108 SAE notched shaft geometry.

The following files are required for this example:

[Link] shaft3.op2 [Link]
bend_op.dac [Link] torq_op.dac

Monotonic properties
Yield strength (0.2%) 380 MPa
UTS 621 MPa
νel 0.28
νupl 0.5
Axial cyclic Torsional cyclic
K' 1258 MPa Ko' 614 MPa
n' 0.208 no' 0.217
σf' 948 MPa τf' 505 MPa
b -0.092 bo -0.097
εf' 0.26 γf' 0.413
c -0.445 co -0.445
E 202000 MPa G 79000 MPa

Table 1: SAE1045HR material parameters

7.2 FEA mesh description

The geometry of the SAE notched shaft was modeled in HyperMesh [2] and then meshed using
hexagonal and pentagonal solid elements, Figure 110. The area near the notch root was carefully
refined, Figure 112. Since solid elements cannot correctly model surface stresses, which are
needed for fatigue life computations, the shaft model was wrapped with thin shell elements
(10-4 mm thickness) with the same material properties as the solid elements. That way, the
surface stresses are correctly recovered. This is the recommended procedure if the solid
elements are included in a FE model. Although the shell elements do increase computational
burden somewhat, they substantially increase the accuracy of the stress-strain predictions.
Shells should be chosen thin enough such that they do not influence the stress-strain response
of the solid elements. There were total of 22140 elements, out of which there are 15192 solid
and 3780 shell elements in the model.
Boundary and loading conditions are displayed below. The loads were applied at 150 mm
distance from the notch root.

7.3 Displaying the model in Studio

To view the model prior to starting the analysis, we will use the nSoft Studio module.
Alternatively, if you do not have access to the Studio FE-Display plug-in (for example, if you do
not have a license), the model may be viewed using a suitable FE pre/post-processor to read the
NASTRAN input file.

• Start the main nSoft interface, select Display from the nSoft Menu, and select Studio
Display/Reporting Tool (studio)
Studio will start up and show an empty display area. To insert the model on the display page:

• Select Insert > FE Display

• In the file selection form, select shaft3.op2 from the working directory and click on
Open.
The model is shown on the current Studio page.

Figure 110 FEA model of notched shaft, displayed using Studio

NOTE: The “mottled” display is due to the fact that the display shows shells on the surface of
solids. To turn off solids and display shells only, do the following:

• Right-click on the FE Display and choose Properties > Groups.

Figure 111 Groups property page

• Select the Property Group Type and then deselect everything except the shell layer.

• Click Apply to show changes.

Shift + left mouse click activates the Zoom feature, allowing you to zoom into the notch region
(see below).
A mesh display can be selected by right-clicking on the FE Display and choosing the relevant
option in the Properties > Model Parameters property page.

7.4 Linear static FEA solution

In this case, the bending force and torque were applied as two separate static sub-cases to the
FEA model, and solved using NASTRAN [3]. The magnitudes of the bending force and torque
were 1000 N and 1000 Nmm, respectively. The static sub-cases will be combined within
FE-Fatigue with actual bending and torque loading histories using the principle of superposition
to produce stress histories for all or selected nodes of the model.
When using multiaxial EN analysis with linear static stress superposition, FE-Fatigue corrects
linear elastic stress-strain results to elastic-plastic results using multiaxial Neuber correction with
Mroz-Garud kinematic plasticity model.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 116

7.5 Translating NASTRAN results using FE2ES

The next step is the read the stress data from the NASTRAN results .op2 file and convert this to
a FES file input deck for FE-Fatigue. We will use nCode’s FE translator program, fe2fes.

• From the main nSoft/FE-Fatigue interface, select FE-Fatigue from the Analysis Menu, and
select Generic FE to FES Translator (fe2fes)
fe2fes takes the user through a wizard-style interface to select which entities and which results
from an FE Analysis are to be used in a fatigue analysis.
On the first fe2fes form:

• Select Filename: shaft3.op2

• Translator: Select translator from file extension (Or NASTRAN op2)

Figure 113 FE2FES: File Selection

7.6 Group selection

The first part of the form is the name and type of the FES file to be created.

• Output Filename: shaft3_01

• File Type: Binary

This creates a binary FES file, shaft3_01.fes.
On the Group Selection form, the part of the model for fatigue analysis is selected by either
Property or Material IDs.
The nodal data from shell elements are selected by:

• Solution Location: Element

• Group Type: Property

The Available Groups column lists all the property IDs found in the op2 file. Select by clicking on
the bracket group:

• 1-SHELL_23
and click the -> button to move it to the Selected Groups list.

Figure 114 FATFE: Group Selection

7.7 Results Selection

The Results Selection form lists all the results subcases found in the op2 file.

• Results Type: Stress

• Shell Layer: Top(Z2)

In the group box for FER Options, check the boxes to select:

• Write model

• Write results
This creates a FER file based on the name of the FES file, in this case shaft3_01.fer, which
contains the FE model data (nodes and elements, group information) and also the FE results as
used in the fatigue analysis. This data can be later viewed using Studio.
Select the available result sets and click the -> button to move them to the Selected Results list.

Figure 115 FE2FES: Results Selection

The final form is a summary of the translation.

Check the “Start FE-Fatigue with FES file” box and click on Finish and the fatfe FE-Fatigue
solver program is shown. (Alternatively, you can start fatfe separately and select
shaft3_01.fes as the Input Fatigue Filename.)
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 119

7.8 Using FE-Fatigue

When FE-Fatigue is started, the prompt for the partial (or full) FES file appears:

Figure 116 FE-Fatigue: Job form.

If this was the first time that this fatigue analysis was run, then a prompt for partial to full FES file
completion appears, Figure 117. The partial FES file contains in this case only static sub-case
results per node or element, whereas the full FES file also contains particular fatigue analysis
related parameters.

Figure 117 FE-Fatigue: Partial to full FES file completion form.

Select Fatemi-Socie as the damage model and click OK.

Figure 118 FE-Fatigue: Advanced Options form.

Next, the Loading form appears, Figure 119. Select each sub-case in turn and populate the fields
as shown in Figure 120 and Figure 121. The stress results from the static load sub-cases are
divided by the values of static loads applied to the FEA model (1000 N for the bending load and
1000 Nm for the torque).
Note that [Link] and [Link] are normalized in-phase loading histories (Figure 122),
while BEND_OP.DAC and TORQ_OP.DAC are normalized 90° out-of-phase loading histories,
Figure 123. They are scaled during stress superposition within FE-Fatigue according to the
applied scale factor. In this case, the scale factor for bending was 12333 N, which corresponds to
the bending moment of 1850 Nm (the bending force was applied at 150 mm distance from the
notch root, thus 1850/0.15 = 12333 N). The scale factor for the torque was 2,550,000 Nmm,
since the dimensions in the FEA model are given in mm. Zero offset on these loading histories
was applied.
Select Done when all the fields are populated.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 121

Figure 119 FE-Fatigue: Loading form

Figure 120 FE-Fatigue Loading form: define bending sub-case

Figure 121 FE-Fatigue loading form: define torque sub-case

Figure 122 Normalized in-phase loading histories.

Figure 123 Normalized 90° out-of-phase loading histories.

Next the Material form appears, Figure 124. Select material SAE1045_shaft from the given list,
and click Done when completed.

Figure 124 FE-Fatigue: Material Input form.

Select Yes when asked if you want to Analyze job now, and from the form in Figure 125
select Analysis Region by Material/Group (group SHELL_23 in this case is a group of
surface shell elements).
Or, use auto-elimination (%model to retain) and 10% retention to speed up the analysis.

Figure 125 FE-Fatigue: Analysis form

On the Results Filename Entry form:

• Set the Output File Format to suit your FE post-processor. If you are using Studio
FE-Display to postprocess the results, you may wish to leave this option set to Default
only.

• Set the FER File Output to 2-Add to FER file (name as FES). This will generate a
FER file so that results may be displayed in Studio.

Figure 126 FE-Fatigue: Results Filename form

During the analysis, the Results Monitor displays the most damaged elements (see below).
Note that completing the analysis may take several minutes.

Figure 127 FE-Fatigue: Results Monitor

7.7 Displaying results

It should now be possible to post-process the fatigue analysis results using your FE post-
processor by importing the appropriate results file created in the previous step. Alternatively, the
fatigue results written to the FER file may be postprocessed using Studio FE-Display. The
procedure for doing this is as follows:
Restart Studio in the same directory as the shaft3_01.fer file and all the supported file types
from the working directory will be made available.
NOTE: You can always recheck for files in the working directory, by selecting the double arrow
cursor above the available data window to load all files from current directory.
Alternatively, open the FER file in Studio by using File > Open Data Files and scan for
nCode S3 fatigue results files, and having found shaft3_01.fer, move it to the right hand
selection list and choose ‘Add To File List’.

Figure 128 Fatigue results available in Studio

Note that two sets of data in the FER file were written from fe2fes, corresponding to the two
load cases:
1 – VERTICAL Z LOAD: Stress
2 – X MOMENT: Stress: 2
followed by a single set of fatigue data from fatfe because we analyzed the Top(Z2) shell
surface:
3 – Multiaxial E-N
The fatigue results (shaft3_01.fer) may be loaded into a Studio FE Display and displayed as
Damage contours (see below).
First, insert an FE Display as shown in section 5.3 above and read in shaft3.op2 or
shaft3_01.fer. Then, drag a set of stresses onto the FE model (rather than on an empty part
of the page) and a contour plot will be displayed.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 128

To control the contour plot, right click on the FE Display and select Properties…
Under Results/Legend, select:
Results Type: Damage
To display only shell elements, right click on the FE Display and select Properties…
Under Groups, select:
Display: Shell_23 (deselect all other Groups)
Click OK to update the display.

Figure 129 SAE notched shaft: Damage contours

To zoom into the display, shift + left-click on the relevant area of the display and then drag to
zoom in
.

Figure 130 Zoom shows damage contours at the notch root

7.8 Comparison with experimental data

Now that the critical location is known, a variety of different fatigue analyses can be performed on
it, either by keeping the same loads and changing the damage model (Edit Job/Edit
Analysis options), or keeping the same damage model and changing the loads by applying
different scaling factors to them (Edit Job/Edit Load Input). Two damage models were
employed in this example: Fatemi-Socie and Wang-Brown [4].
Table 2 lists experimental results and predictions for a variety of in-phase loadings, while table 3
lists the same for a variety of 90° out-of-phase loadings. Note that the loading functions for the
out-of-phase loading were changed from [Link] and [Link] to BEND_OP.DAC and
TORQ_OP.DAC, respectively.
Experimentation of this type can be performed on a reduced set of nodes by selection of the
analysis region to be specified nodes rather than by Group. Either a single node ID, e.g. 53119,
can be entered directly or the group of nodes specified in the file [Link] from the previous
run can be used by copying [Link] to another name, e.g. [Link] and specifying
@[Link] in the node/element input field. This will greatly reduce calculation time.
It should be emphasized that the critical location and the associated lives to failure change
slightly with the changing levels of non-proportionality. Also, the predicted lives to failure change
with both increasing number of points in the loading histories and with changing the waveforms of
the histories. Several different cases have been tested, with number of points in the histories
ranging from 5 to 500, and with waveforms of the histories ranging from triangular, to trapezoidal,
to sine waves. Overall, the differences in life predictions between the tested cases were rather
small (within 15% on life for either the Fatemi-Socie or the Wang-Brown method). The analyses
here are presented for one critical element only (element 53119), with triangular loading histories,
and with only peaks and valleys representing the time histories (thus only 4 to 5 points in the
histories).
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 131

Specimen Rb Rt Mb Mt Nf [cycles] Nf Nf
ID [Nm] [Nm] Experimental Fatemi-Socie Wang-Brown
IL-XR3-1 -1 -1 1850 2550 5,113 10804 11564
JD-BR3-1 -1 - 2800 - 8,262 7356 13400
JD-XR3-1 -1 -1 2000 2100 12,050 14117 15691
JD-ZR3-1 -1 -1 1250 2700 10,420 17915 17965
IL-ZR3-1 -1 -1 1150 2700 12,700 20091 19750
IL-YR3-1 -1 -1 2325 1350 11,380 13168 19041
IL-YR3-2 -1 -1 2325 1350 12,090 13168 19041
AOS-BR3-1 -1 - 2600 - 13,060 10448 19418
IL-BR3-2 -1 - 2600 - 13,760 10448 19418
JD-BR3-2 -1 - 2600 - 18,310 10448 19418
AOS-BR3-1 -1 - 2586 - 17,450 10721 19956
JD-TR3-1 - -1 - 3000 9,528 31032 23237
IL-TR3-1 - -1 - 3000 14,720 31032 23237
IL-XRN- I -1 - 1355 2550 11,630 19364 19970
IL-XR3-2 -1 - 1850 2100 11,500 17855 19339
RN-XR3-1 -1 - 1850 2100 11,630 17855 19339
IL-ZR3-1 -1 -1 851 2700 17,730 27385 25690
IL-ZR3-2 -1 -1 840 2700 24,540 27658 25921
BC-TR3-1 -1 - 80 2534 33,330 70913 56535
IL-YR3-1 -1 -1 1720 1350 51,780 46686 62297
IL-YR3-2 -1 -1 1720 1350 65,800 46686 62297
BC-TR2-1 - -1 - 2400 101,100 101200 78247
IL-TR2-1 - -1 - 2400 164,070 101200 78247
IL-ZR2-1 -1 -1 780 2180 142,700 77060 73766
IL-ZR2-2 -1 -1 780 2180 169,500 77060 73766
IL-ZR2-3 -1 -1 570 2180 177,800 98250 93174
IL-ZR2-4 -1 -1 570 2180 186,700 98250 93174
IL-XR2-1 -1 -1 1220 1710 163,700 91804 99621
BC-YR2-1 -1 -1 1680 960 65,049 71629 113600
JD-XR2-1 -1 -1 1220 1710 158,100 91804 99621
RN-XR2-1 -1 -1 1220 1700 124,500 92911 101600
JD-BR2-1 -1 - 1875 - 106,700 55050 115800
BC-BR2-1 -1 - 1875 - 117,700 55050 115800
JD-YR I - I -1 -1 1680 900 153,800 74053 122200
IL-YR2-2 -1 -1 1550 1090 159,900 93349 137100
IL-XR2-2 -1 -1 1550 1090 220,500 93349 137100
JD-YRI-I -1 -1 1300 1400 226,000 139300 155100
RN-TRI-1 - -1 - 2000 1,293,000 292400 233200
IL-TR1 - 1 - -1 - 2000 2,238,000 292400 233200
GKN-TRI-I - -1 - 2000 2,000,000 292400 233200
JD-ZR2-1 -1 -1 845 1800 396,800 156800 164200
RN-BR2-1 -1 - 1730 - 83,600 85902 188500
IL-BR2-1 -1 - 1730 - 184,300 85902 188500
IL-BR2-2 -1 - 1730 - 132,300 85902 188500
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 132

AOS-BR2-1 -1 - 1730 - 228,300 85902 188500

AOS-BR2-2 -1 - 1708 - 249,900 92381 204200
RN-XRI-l -1 -1 990 1390 587,000 320600 356800
IL-XRI-l -1 -1 990 1390 1,194,000 320600 356800
AOS-BR1-1 -1 - 1475 - 403,800 219100 532300
JD-BR1-1 -1 - 1475 - 709,000 219100 532300
IL-YR1-1 -1 -1 1250 880 722,500 343100 539100
RN-YR1-1 -1 -1 1250 880 747,000 343100 539100
AOS-BR1-2 -1 - 1460 - 764,000 233300 570500

Table 2: SAE notched shaft; in-phase, constant amplitude tests; linear elastic static FEA solutions and
multiaxial Neuber analysis; element 53119.

Specimen Rb Rt Fi Mb Mt Nf Nf Nf
ID [Nm] [Nm] [cycles] FS WB
JD-ZR03-1 -1 -1 90 1150 2700 13,110 24620 34655
JD-XR03-1 -1 -1 90 1850 2100 27,470 57610 112400
JD-XR03-2 -1 -1 90 1800 2100 24,620 59487 131100
BC-XR03-1 -1 -1 90 1698 2242 10,840 38432 170700
JD-YR03-1 -1 -1 90 2300 1325 23,980 19640 37467
JD-ZR03-2 -1 -1 90 770 2180 157,100 72304 113000
BC-XR03-1 -1 -1 90 1295 1710 45,580 224300 113700
JD-XR02-1 -1 -1 90 1220 1710 213,800 233300 427400
JD-XR02-2 -1 -1 90 1220 1710 266,200 233300 427400

Table 3: SAE notched shaft; 90° out-of-phase, constant amplitude tests; linear elastic static FEA solutions
and multiaxial Neuber analysis; element 53119.

The plot in Figure 131 summarizes the results from the two tables. Almost all of the predictions
fall within a factor of 3 with respect to the experimental results, and about 80% of them fall within
a factor of 2. This represents very good correlation, especially considering that simple and
inexpensive linear elastic static FE analyses were employed and their results corrected to elastic-
plastic states using the multiaxial Neuber correction.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 133

Figure 131 Comparison of experimental vs. predicted fatigue lives; linear static superposition with multiaxial
Neuber correction; node 53119.

The following abbreviations apply to Figure 131:

FS-IP Fatemi-Socie in-phase
FS-OP Fatemi-Socie 90 degrees out of phase
WB-IP Wang-Brown in-phase
WB-OP Wang-Brown 90 degrees out of phase

7.9 References
1. Multiaxial Fatigue: Analysis and Experiments, G. Leese, D. Socie, eds., SAE AE-14, SAE International,
1989.
2. Altair Engineering, HyperMesh v.4.0 Users Manual, 2000.
3. MSC/NASTRAN Users manual, MacNeal-Schwindler Corporation, 2000.
4. FE-Fatigue Theory Manual, nCode International, 2001.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 134

8. Multiaxial Safety Factor Example

Some structures experience multiaxial loadings in critical locations, that is, the stress state in the
critical locations is characterised as having more than one significant principal stress, and/or the
principal stress directions change with time. These issues are described and illustrated in worked
examples 4 and 7, which describe methods for understanding stress states and making finite life
fatigue calculations. However, some components are expected to survive very many critical
loading cycles during their lifetime and for these components it is not realistic to design them for
a finite fatigue life. These comments apply to many engine components such as connecting rods
and crankshafts. For such components it is more usual to design them using a safety factor
approach which allows the critical loading cycles to be compared to a fatigue or endurance limit
criterion. The simple uniaxial safety factor approach is fine for many cases, but where the loading
is multiaxial, especially when it is non-proportional, something more sophisticated is required,
such as the Dang Van or McDiarmid methods which are included in FE-Fatigue.

8.1. Introduction
This example describes the application of the Dang Van multiaxial safety factor method to the
fatigue analysis of a shaft which is subjected simultaneously to bending and torsional loads. The
shaft used is the SAE notched shaft used in a SAE round-robin test program (see below).

Figure 132 SAE notched shaft showing applied bending load.

Stress Analysis was carried out on the shaft using NASTRAN with two load-cases, representing a
bending load of 1000N (applied at the location shown above) and a torsional load of 1000 Nmm
applied at the same location. In both cases the shaft was constrained around the parallel section
as shown above. The corresponding bending and torsional loading histories are illustrated below.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 135

Figure 133 Bending (shaft01) and torsion (shaft02) loading histories to be applied to the notched shaft.
This example will carry out multiaxial safety factor analysis on this component to determine
whether or not it is fit for service, which in this case requires that it be able to survive an infinite
number of repeats of the loading history, with a load safety factor of 1.1.
The following files are required for this example from the demo directory
[Link]
shaft3.op2
[Link]
[Link]
[Link]
These files can be copied to your working directory using the Macros pull down menu in the main
FE-Fatigue program. Select the example
MultiaxialSafetyFactor(fe_safetyfactor_copyfiles.ncx)

8.2 Review model and stress analysis results

The stress analysis was done in NASTRAN, with two static load cases, representing torsional
and bending loads as described above. The model is composed of linear solid elements, with a
thin skin of shell elements. Some rigid elements are used to distribute the applied load.
The practice of applying shell elements to the surface of a solid model has these advantages:

• It gives a reduced set of elements for which the stress analysis results may be recovered
and post-processed – most fatigue analyses need only consider the surface stresses.

• Fatigue analysis normally requires the surface stresses, and the stresses recovered from
the shell elements will be of better quality than the nodal stresses, extrapolated from the
solid element gauss points, that would otherwise be used.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 136

• The shell element stresses are normally recovered in element coordinate systems which
have the advantage of being in the plane of the surface. This makes multiaxial
assessment easier.
The model and results may be reviewed using any suitable FE post-processor. NASTRAN .op2
files and Hypermesh database and results files are available. Alternatively, the model may be
viewed using Studio FE-Display.
In order to view the stress analysis results in Studio FE-Display, it is necessary first to write them
to the .FER file. This is achieved by checking the appropriate FER option when running the
FE2FES translator (see the next section for details).
Because the model has a thin skin of shells over a solid model, the best results post-processing
is achieved by displaying only the shell elements before making a contour or element fill plot.
Note the different stress distribution and stress states caused by the two subcases.
8.3 Write FES file using Generic FE to FES translator
In this example, the FES file is generated directly from the FE stress analysis results file (in this
case a NASTRAN .op2 file) using the Generic FE to FES translator. The process is as follows:

• Start the main nSoft interface in your working directory and from the nSoft menu, select
FE-Fatigue and then Generic FE to FES Translator (fe2fes).The interface
should appear as shown below.

Figure 134 FE2FES user interface

• Enter (or select) shaft3.op2 in the Filename entry field. By default, fe2fes will
recognise the op2 file by its filename extension. It can also read ABAQUS .fil files,
ANSYS .rst files and SDRC .unv files, all recognised by their file extension. If any
results file selected has a different extension you will need to select the correct translation
using the Translator field. Click Next to continue.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 137

• Now the Group Selection form appears. You must use this form to select the elements
and results location to be used in the fatigue analysis. Set the Output Filename to shaft
and leave the file type as Binary. Set the Solution Location to Element. This means we
will be making fatigue calculations using the element centroid stresses. We will be
selecting elements for analysis by element property set, so leave the Group Type as
Property. We would like to use the shell elements only, and all the required shells are in
a property set called SHELL_23. Click on this item to highlight it and then transfer it to the
Selected Groups box by clicking on the right arrow. The form should now appear as
shown below. Click Next to continue.

The next form is the Results Selection form. We only have stress results here so the Results
Type field is greyed out. Because the surface shells have negligible bending, it doesn’t matter
which layer we recover the stresses for, so we leave the Shell Layer as Top(Z2). Leave the load
case description field alone, and then select the two results cases 1-VERTICAL Z LOAD and
2-X MOMENT and transfer to the Selected Results box using the right arrow. The form
should now appear as shown in Figure 136.
Check both FER options to write the FE model and stress analysis results to the FER file.
Click on Next and then Finish to complete the translation.

Figure 136

8.4 Run Multiaxial Safety Factor Analysis

The next stage is to carry out the multiaxial safety factor analysis using the fatigue analyser
FATFE. First here is a review of the loading and material data that will be used in the analysis:

• The loading histories [Link] and [Link] may be viewed using the multi-
file-display module MFD. MFD may be started by selecting nSoft Menu/Display/
Multi-File Display (mfd). Type shaft01,shaft02 in the Input Filenames field
and click OK to view. The file shaft01 represents the bending loading (in N) and
shaft02 is the torque (in Nmm).

• Review the material data by starting the Material Databank Manager MDM – select nSoft
Menu/FE-Fatigue/Material Databank Manager (mdm). At the main form select
Load/data set 1 and select the material PSCI_DV_Test. This is a spheroidal
graphite cast iron. This material has only the minimum data necessary for a multiaxial
safety factor calculation. Use Tabulate/Data values to view the data.
Now we will set up and run the fatigue analysis:

• From the main nSoft interface, start the module FATFE either by typing fatfe at the
prompt, or selecting nSoft Menu/FE-Fatigue/FE-Fatigue Analyser (fatfe).
On the Fatigue Jobname Entry form, enter or select the Input Fatigue Filename
shaft. This will select the file [Link]. Click OK to continue.

• On the next form set the Analysis Type to multiaxial safety Factor and set
Advanced Option to Yes. Click OK to continue.

• On the Advanced Options form, there are two choices to be made – the damage
model (Dang Van or MacDiarmid) and the hardening Parameter. We will use the Dang
Van method. The Hardening Parameter is used in Dang Van to control the rate of growth
(isotropic hardening) of the yield surface. The Dang Van uses an iterative method, and
this parameter controls the rate of convergence – larger values will give faster
convergence at the expense of some accuracy. A value of 0.05 is a reasonable
compromise for most loadings, though for very short loading sequences a smaller value
may be advisable. Click OK to continue.

• On the Loading Input form we must associate the static loading cases with their
corresponding loading histories and appropriate scaling factors. First select load case
1 (VERTICAL Z LOAD). Type or select [Link] in the Time History field and
enter 1000 in the field named Divide stress/strain by. Repeat for load case
2 (X MOMENT) and loading history [Link]. The factor of 1000 is to account for
the fact that the load defined in the FE for both loadcases was 1000. The form should
appear as in Figure 137. When finished, select Done and click on OK to continue.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 140

Figure 137

• On the Material Input form, select the group SHELL_23 and apply the material name
PSCI_DB_Test. Click Done and OK to continue, then YES to continue with the analysis.

• The Dang Van method uses an iterative approach which can be time consuming. We can
use auto-elimination to reduce the analysis time by concentrating on the more highly
stressed parts of the structure. On the Analysis Form, change the Analysis Region from
All data in FES file (the default) to auto elimination (% of max Stress). Set
the Auto Elimination Retention Factor (%) to 50. This means that the analysis
will proceed for each element only if the calculated auto-elimination stress exceeds 50%
of the maximum value in the model. This means that the calculation is only carried out for
the critical regions around the notch. Click OK to continue.

• On the Results Filename Entry form, select the appropriate Output File
Format for your post-processor. In this example we have described the use of
Hypermesh, but you could equally well use SDRC (Universal File), ANSYS or Default only
(PATRAN).

• For post-processing fatigue analysis results in Studio FE-Display, set FER File Output to
2 - Add to FER file (name as FES).

• During the analysis, a Results Summary screen appears as shown below..

Figure 138

Click OK to continue. You should now see the main FATFE form shown below.

8.5. Post-process the Results

The most obvious way to post-process the results is to make a contour plot of the calculated
safety factors. However it is also possible to visualise in detail the results at individual elements
(or nodes) by making a Dang Van plot.

• From the FATFE main menu, select Utilities/Node/Element Options. Enter

[Link] in the Result Filename field, and click on OK twice to select the Worst Case
Element. Accept the default Scatter Plot on the next form to view the Dang Van plot
for the worst element in the model. The resulting plot is illustrated below. It shows how the
stress path, in terms of microscopic shear stress and hydrostatic stress, compares to the
fatigue limit criterion (the sloping line). If the two do not cross, no fatigue damage is
predicted to occur.

Figure 140
Safety factors of less than 1 indicate that fatigue failure may occur. In this case we have a safety
factor everywhere greater than 1 (see Figure 138), and the worst safety factor is also larger than
the design requirement of 1.1.
It is now also possible to post-process the results in your FE-postprocessor, or alternatively in
Studio FE-Display. To use Studio FE-Display:
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 143

• From the main FE-Fatigue (or nSoft) interface, start Studio by typing studio at the
prompt.

• Studio will start and show an empty display area. To insert the model on the display page:

• Select Insert > FE Display.

• Select one of the three files offered - [Link], [Link] and shaft.op2. All
contain the model geometry and can be displayed. Click on Open.
The model will be shown on the current Studio page. (Other displays can be added to the same
page or a single view can be selected using View > Interactive Mode).
The model can be dynamically manipulated in the display by holding down either the Control or
Shift keys together with the mouse buttons. (Alternatively, function F2 and F3 keys can be used).
Put the cursor on the display and move the mouse to manipulate the model. The controls are
summarized as:
Rotate - CTRL+ Left Mouse Button, or F2 + Left Mouse Button
Pan - CTRL + Right Mouse Button, or F2 + Right Mouse Button
Cursor Zoom - SHIFT + Left Mouse Button, or F3 + Left Mouse Button
Box Zoom - SHIFT + Middle Mouse Button, or F3 + Middle Mouse Button
Squiggle Zoom - SHIFT + Right Mouse Button, or F3 + Right Mouse Button
NOTE: The Controls Summary is also available by clicking the mouse icon on the toolbar.

Use the Properties dialog box (if not already visible, this can be displayed by right clicking on the
model and selecting "Properties") to display only the Group containing the surface skin of shell
elements SHELL_23.
An element fill plot of the multiaxial safety factors may now be made as follows:
In the data tree in the Available Data window, expand the branch "nCode S3 fatigue results.
Click and drag 3-Multiaxial Safety Factor on to the model.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 144

You should now see something similar to the following figure, highlighting the critical area and the
worst safety factors.

Figure 141 Safety factor results plot

9. Spot Weld Analysis Example

This example performs a Stress-Life fatigue analysis of the spot welds in the shock tower model
used in previous examples. The spot welds are represented by bar elements.

Figure 142
The method uses the bar element cross sectional forces and moments to calculate the "structural
stresses" in each spot weld nugget and the adjacent sheets according to the methodology
described by Rupp, Störzel and Grubisic [1]. These stresses are used to make fatigue life
predictions at intervals around the perimeter of the nugget of the spot weld using the S-N
method.
Details of the spot weld method used in FE-Fatigue and alternative modeling strategies are given
in the FE-Fatigue theory manual.
References: [1] Rupp, A., Störzel, K. and Grubisic, V. (1995) "Computer Aided Dimensioning of Spot-
Welded Automotive Structures". SAE Technical Paper 950711.

9.1 Introduction
A front rail and shock tower section of a vehicle structure is subjected to 3 loads (x,y,z) using FE
analysis. A NASTRAN model of the structure uses CBAR elements to represent the spot welds.
Element forces are recovered using the FORCE = ALL command in NASTRAN. The FE loads are
applied as separate subcases to give the resulting forces and moments in the welds due to each
load. Measured time history loads are defined in RPC III format (force_spot.rsp in this case),
and a spot weld fatigue analysis will be performed using the generic spot weld SN data in the
materials database.
This example uses nCode’s Studio FE Display to show the model and results and uses the
FE2FES translator to read the stresses from a NASTRAN file.
The following files are required for this example:
[Link]
shock_spot.op2
force_spot.rsp
[Link]
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 146

These files can be copied to your working directory using the Macros pulldown in the main FE-
Fatigue / nSoft program:
Macros\System\nCode\Copy Files for FE-Fatigue Examples\
fe_spotweld_copyfiles.ncx

9.2 Displaying the model in Studio

If you wish to view the model before reading the results, use a pre-post processor (e.g.
[Link], HyperMesh etc.) to import the NASTRAN bulk data file [Link], or alternatively
use Studio FE-Display as shown below.

• Start the main nSoft interface, select Display from the nSoft Menu, and select Studio
Display/Reporting Tool (studio)
Studio will start and show an empty display area. To insert the model on the display page:

• Select Insert > FE Display

• In the file selection form, select [Link] from the working directory and click on Open.
The model is shown on the current Studio page.

Figure 143 Model displayed in Studio (Fill plot with Mesh)

9.3 Translating NASTRAN results using FE2FES

The first step is to read the spot weld forces and moments from the NASTRAN results op2 file
and convert this to a FES file input deck for FE-Fatigue. To achieve this, we will use nCode’s FE
translator program, fe2fes.
From the nSoft/FE-Fatigue main interface, select FE-Fatigue from the Analysis menu, then
select Generic FE to FES Translator (fe2fes).
fe2fes takes the user through a wizard-style interface to select which entities and which results
from an FE Analysis are to be used in a fatigue analysis. NASTRAN op2 files which have been
created on either PC or on a UNIX workstation platform can be read by fe2fes.

Figure 144
On the first fe2fes form, enter the following:

• Filename: shock_spot.op2

• Translator: Select translator from file extension (or NASTRAN op2)

9.4 Group Selection

The first part of the form is the name and type of the FES file to be created. In this case:

• Output Filename: shock_spot_01

• File Type: Binary

This creates a binary FES file, shock_spot_01.fes.

Figure 145 FE2FES: Group Selection

On this Group Selection form, the part of the model for fatigue analysis is selected. Nodes or
elements can be selected by either Property or Material IDs. In this case, all the beams
have the same PBAR card (PBAR 1000).
NOTE: The Solution Location drop down menu is greyed out in this case since the op2 only
contains element forces. In general, this selects whether a fatigue analysis is to be
performed at node or element locations.
The bar elements are selected by Group Type: Property
The Available Groups column lists all the Property IDs found in the op2 file. Select the group
by clicking on the first group with elements referring to PBAR card 1000; 1- BAR_1000 and click
the -> button to move to the Selected Groups list.
Click the ungreyed Next > to move on to the next form. (You can also back up at any stage
using < Back.)
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 149

9.5 Results Selection

The Results Selection form lists all the results subcases found in the op2 file.
The Results Type drop down menu is greyed out in this case since the op2 only contains
element forces and therefore defaults to Spot Weld. In general the user can select Stress,
Strain or Spot Weld data.

Figure 146 FE2FES: Results Selection

The Load case description field enables you to toggle between available Title
descriptions in the op2 file to identify loadcases in FE-Fatigue.

• Select all the three available result sets and click the -> (move to right) button to move
to the Selected Results list.
In the group box for FER Options, check the boxes to select:

• Write model

• Write results
This creates a FER file based on the name of the FES file, in this case shock_spot_01.fer,
which contains the FE model data (nodes and elements, group information) and also the FE
results as used in the fatigue analysis. This data can be later viewed using Studio.

• Click Next > to move on to the next form.

The final fe2fes form (shown below) gives a summary of the translation.

Figure 147 FE2FES: Summary screen

• Check the Start FE-Fatigue with FES file box and then click on Finish. The
fatfe FE-Fatigue solver will appear. Alternatively, you can start fatfe separately and
select shock_spot_01.fes as the Input Fatigue Filename.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 151

9.6 Starting Fatigue Analysis

Since the FES file is already defined as a spot weld analysis, the analysis type in the first
FE-Fatigue form defaults to Spot Weld Analysis. The FES file contains force/moment data
and the sheet thicknesses; the next stage is to complete the FES file by defining the loading and
material information for the fatigue analysis. Accept the Analysis form defaults and click OK.

Figure 148 FATFE: Analysis options

9.7 Loading Information
For the loading information, select:

• The first loadcase (VERTICAL) on the list and click OK

• File Type: RPC III

• Time History: use the list button and choose force_spot.rsp

• Channel number: 1

• Divide stress/strain by: 1000

(The calculated NASTRAN forces/moments were due to a 1000 N applied load and these are
normalized by dividing by 1000.)
Click OK to accept this data and fill in the table. Then select:

• The next loadcase (LATERAL) and click OK

• File Type: RPC III

• Time History: use the list button and choose force_spot.rsp

• Divide stress/strain by: 1000

Click OK to accept this data and fill in the table by selecting:

• The next loadcase (HORIZONTAL) and click OK

• File Type: RPC III

• Time History: use the list button and choose force_spot.rsp

• Channel number: 3

• Divide stress/strain by: 1000

Click OK to accept this data and fill in the table
You can directly view/edit RPC III files using the muledt display module. Pressing F5 on the
FE-Fatigue loading form will give the option to view time histories. The ASCII export/ import
options also directly support RPC files and channel numbers.
When completed, ensure Done is selected in the table and click OK.

Figure 149 FATFE: Loading Input

9.8 Material Information

The Material Information form defines the fatigue properties to be input for the analysis. In this
case, there is only one group BAR_1000 since all the bar elements in this example referred to the
same PBAR card (PBAR id 1000). In general from NASTRAN op2 files, multiple groups would
be possible if bar elements exist with different property or material cards.
Select the following for BAR_1000:

• Sheet material: spot_sheet_generic

• Nugget material: spot_nugget_generic

• Nugget diameter: AUTO

• Scale factor: 1
Click on OK to enter this data for group BAR_1000.

Figure 150 FATFE: Material Input

Separate SN curves are defined for the sheet steel and for the spot weld nugget.
The Nugget diameter can be defined per group of bar elements (for instance a spot weld
diameter of 5.0 mm could be entered here). However, spot weld diameters tend to be defined by
the thickness of the sheets being welded and so FE-Fatigue includes a feature to automate the
definition of weld nugget size. If AUTO (or –1) is entered, a lookup table [Link] is used
to define the nugget diameter based on the thickness of the thinnest sheet at each weld.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 154

An example [Link] file is:

SPOT WELD DEFINITION FILE
NUGGET_DIAMETER=BY_THICKNESS
0.3,3.5
0.8,4.0
1.2,5.0
2.0,5.5
3.0,6.0

where the first column is the minimum sheet thickness, and the second column is the
corresponding nugget diameter. This example file is found in the nssys directory in the nSoft
installation. Alternatively you can put a [Link] file in the nSoft home or in the working
directory where it will take precedence over the central [Link] file in the nssys directory.
FE-Fatigue does not interpolate between the data points in the [Link] file. Rather, it
works as follows:
The thickness of the inner sheet from each spotweld is compared to the table and the value of
thickness in the table which is nearest to but less than or equal to the thickness of the sheet is
identified. The corresponding diameter from the table is assigned to that spotweld.
Select Done and then click on OK to exit the form. Then select Yes to analyse the job now.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 155

9.9 Analysis Form

Figure 151 FATFE: Analysis form

On the Analysis form, select:

• Analysis Region: All weld elements

• Equivalent Units: 1 Repeats

• Analyse Nugget: No

• Three Sheet Correction: No

• Number of Calculation Points: 18

The number of calculation points defines the number of locations around each weld to be
analysed. A larger number of calculation points gives a more detailed and probably more
accurate answer. However, the analysis will take longer. A choice of 18 calculation points means
that calculations will be made at 20 degree intervals around the perimeter of each spotweld.

• Click on OK

• For results filename, accept the default shock_spot_01 and select the following.

• Output File Format: Select an output format suitable for your post-processor. Note that not
all post-processors have convenient tools for viewing spot weld results.

• FER File Output: 2- Add to FER file (name as FES) (to allow viewing of the
results in Studio FE-Display).

• Click on OK to start the analysis.

9.10 Reviewing Fatigue Results

During the analysis, the most damaged elements are listed. The most damaged weld is element
32597 with 50797 repeats to failure. The location of most damage is on sheet 2, where sheet
2 is defined by the connectivity of the bar in NASTRAN.
Sheet 1 is the sheet connected to the first node on the CBAR card and Sheet 2 is the sheet
connected to the second node.

Figure 152 FATFE: Results Monitor

NOTE: If your answers differ from those shown above, check the Analysis Preferences settings.
The Matrix Size should be set to 64.
Click OK to exit this list.
The analysis will have created several files, namely:
shock_spot_01.fef – ASCII results file of worst damage result at each weld
shock_spot_01.spt – ASCII results file of all damage results at each weld
shock_spot_01.fer – nCode fatigue results file
shock_spot_01.ent – ASCII list of most damaged elements
In addition, the following files are optionally produced:
shock_spot_01.hyp – HyperMesh results file (results as .fef)
shock_spot_01.unv – SDRC universal file
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 157

shock_spot_01.ans – file suitable for importing into ANSYS using the nCode-ANSYS macro
shock_spot_01.bof – Medina results file
shock_spot_01.frs – FEDEM results file
shock_spot_01.tml – PATRAN 2.5 format results template file (enables .fef file to be easily
read back into PATRAN)
For further information on the results, the Utilities menu provides several options including:

• Results Listing, which gives a spreadsheet summary of the shock_spot_01.fef

file using fatres.

• Node/Element options which enable:

– Time history extraction of stress at the most damaged locations around the weld.
– Full results listing which gives results on a per angle basis for a weld (.spt file).
– Polar plot to graphically view the results on an angle basis for a weld.

9.11 Displaying Weld Results using Studio

It should now be possible to post-process the results in your FE-postprocessor. FE post-
processors differ in their abiltity to post-process spotweld data effectively. At the time of writing
[Link] has good facilities, and Hypermesh may also be used effectively. Alternatively, use
Studio FE-Display as follows.
Start Studio in the same directory as the shock_spot_01.fer file and all the supported file
types from the working directory will be made available.
NOTE: You can always recheck for files in the working directory, by selecting the double arrow
cursor above the available data window to load all files from current directory.
Alternatively, open the FER file in Studio by using File > Open Data Files and scan for
nCode S3 fatigue results files, and having found shock_spot_01.fer, move it to the right
hand selection list and choose ‘Add To File List’.
To show contour plots, first insert an FE Display and read either [Link] or
shock_spot_01.fer.
Drag the set of spot weld fatigue data onto the FE model (rather than on an empty part of the
page) and a marker plot will be displayed.

Figure 153 Spot weld fatigue results displayed using Studio

Use the available settings in the Properties > Model Parameters dialog if you need to change the
look of the plot.
Press Shift + left click and drag to zoom in/out of the display.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 159

10. Multiple Curve S-N Example

This example uses multiple S-N curves to account for mean stress. This method is an alternative
to the standard mean stress correction methods such as Goodman or Gerber. It allows the user
to define and use multiple mean stress curves for a given material, each curve representing a
different level of mean stress or load ratio. Instead of correcting for mean stress, FE-Fatigue
interpolates between the curves according to the mean stress or load ratio of each cycle.

10.1 Introduction
This example uses a Z-shaped specimen (illustrated below) which is used to reproduce complex
loadings seen at a cut-out in a rotor head of a helicopter under service loading. The applied
loading on the specimen is in the Y-direction. The loading is represented by a modified version of
the standard loading spectrum ROTRIX8, and the example uses FE-Fatigue to make a S-N
calculation, using multiple S-N curves.

Figure 154 Z-shaped specimen

The following files are required for this example from the demo/ directory:
z_spec.fil
z_spec.inp
z_spec.hm
z_spec.odb
[Link]

10.2 Writing a Partial FES file using the nCode translator (FE2FES)
The first task is to import the stress data from the static stress results file (z_spec.fil) to a partial
input file for FE-Fatigue.
From the Analysis menu pull-down, select FE-Fatigue, and then the generic FE to FES
Translator (FE2FES). Alternatively, type FE2FES in the command prompt line to start the
module.
In the first FE2FES form, select the following:

• Filename: z_spec.fil

• Translator: ABAQUS fil

• Next >

In the Group Selection form, enter the FES filename to be created: z_spec_01, keep the file
type: binary, and the group type: set. From the available groups, select only the shell elements
for the analysis by clicking on the SHELLS group and using the across arrow. The form should
look as below:

Figure 156

To move to the next form, click Next >.

In the Results Selection form, choose both Top and Bottom layer of the shell elements, select
the only load-case available: Z specimen with 2nd order elements. Keep all the
remaining options as their default settings.
If you wish to view the results using Studio FE-Display, tick the FER Options check boxes.

Figure 157

Click Next > to start the file translation process.

The translation should only take a few seconds before the Summary form below appears, tick the
box for starting FE-Fatigue, and click on Finish.

10.3 Starting FE-Fatigue

FE-Fatigue will start directly in the Analysis form. Complete it as below, choosing the Analysis
Type as Multiple curve SN. Choose Yes for the Advanced Options.

Figure 159 Analysis form

In the Advanced Options form, ensure the Mean Stress Correction is Interpolate, and keep
the rest of the answers as defaults.

Figure 160 Advanced Options

At the Loading form, apply the load history [Link] to the load-case Z Specimen
with… as follows:

Figure 161 Loading form

Do not forget to set the Scale Factor to 300.

At the Material Input form, choose the nest of curves called Z-Specimen from the default
[Link] database which is automatically copied to the working directory.

To view the database, press F5 and select Materials edit. The materials data ASCII file
[Link] should contain the following section:
-------------------------------------------------------
MATERIAL NAME = Z-Specimen
DATA TYPE = STRESS
MATERIAL TYPE = 300
MEAN TYPE = R-ratio
UTS = 1.03e+003
E = 1.17e+005
R-ratio = -0.5
STRESS VALUES = 800 731 696 669 643 620 598
578 561 545 530 516 504 493 483 474
465 458 451 445
LIFE VALUES = 1000 2510 3980 6310 1E4 1.58E4
2.51E4 3.98E4 6.31E4 1E5 1.58E5 2.51E5 3.98E5 6.31E5
1E6 1.58E6 2.51E6 3.98E6 6.31E6 1E7
R-ratio = 0.1
STRESS VALUES = 557 507 485 465 447 431 416
403 390 379 369 359 351 343 336 330
324 319 314 310
LIFE VALUES = 1000 2510 3980 6310 1E4 1.58E4
2.51E4 3.98E4 6.31E4 1E5 1.58E5 2.51E5 3.98E5 6.31E5
1E6 1.58E6 2.51E6 3.98E6 6.31E6 1E7
R-ratio = 0.5
STRESS VALUES = 368 334 320 307 295 284 274
265 257 250 243 236 231 226 221 217
213 210 207 204
LIFE VALUES = 1000 2510 3980 6310 1E4 1.58E4
2.51E4 3.98E4 6.31E4 1E5 1.58E5 2.51E5 3.98E5 6.31E5
1E6 1.58E6 2.51E6 3.98E6 6.31E6 1E7
R-ratio = 0.9
STRESS VALUES = 117 107 102 98 94 90 88
85 82 79 77 76 74 72 70 69
68 67 66 65
LIFE VALUES = 1000 2510 3980 6310 1E4 1.58E4
2.51E4 3.98E4 6.31E4 1E5 1.58E5 2.51E5 3.98E5 6.31E5
1E6 1.58E6 2.51E6 3.98E6 6.31E6 1E7
-------------------------------------------------------

This shows an example of how the data is stored as a general look up table of Stress Range
versus Life for different R-ratios. An R-ratio is defined as the (minimum stress/maximum stress).
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 168

Consider the following example.

Figure 163

If maximum stress for a cycle is 200 MPa and minimum is -100 MPa then
Mean stress = (200 + (-100)) /2 = 50 MPa
and
R-ratio = -100/200 = -0.5
It is also possible to define the curves by specifying a constant mean stress level directly rather
than a constant R-ratio.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 169

In order to view these curves graphically, an nCL macro has been provided ([Link]). This can
be run by typing ncl @dsn at the prompt of the main nSoft window.
Select the materials database [Link] and choose the material Z-Specimen.
This nCL uses a display program RQP to show the nest of curves to be used in the analysis.
Select File > Exit in RQP when done viewing the files.

Figure 164 Plot of nest of S-N curves for Z-Specimen

10.4 Submitting Fatigue Analysis

The next on-screen question will ask if you want to run the analysis. Answer No, so that the
correct analysis preferences are set. If preferences are not set to the same as those indicated
below, answers may differ.
Select Preferences from the Post-Processing menu and Analysis from the sub-menu to
display the Analysis Preferences form.

Figure 165
The program will ask if you wish to save these preferences. If you answer No, the preferences
will apply until you exit the solver. If Yes, they will be stored in your home directory and will apply
to future runs of FATFE within your user account.
Now select Analyse.
On the Fatigue Analysis form, accept the defaults:

• Analysis Region: All Data in FES file

• Equivalent Units: 1 Repeats

• Safety Factor Analysis: None

On the Results Filename Entry form select the following:

• Enter Results Filename: z_spec_01

• Shell Calculation: Worst

• Solid Calculation: None

• Output File format: Default only

• FER file output: 2 - Add to FER file (name as FES)

• Results description: Fatigue life from application of ROTRIX8

Click OK.

Figure 166

NOTE: If you wish to use another display program, for example Hypermesh, you can select it as
an alternative Output File format. The FEF and FER files will still be created.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 172

10.5 Reviewing Fatigue Results

During the fatigue analysis, a window pops up showing the most damaged elements.

Figure 167 Fatigue results summary

The results of all the elements in the fatigue results file z_spec_01.fef may also be viewed
using the nSoft module fatres.
Ignore any warning messages about 3D elements – this is related to the calculation of biaxiality
ratio.
NOTE: If your answers differ from those shown above, check the Analysis Preferences settings.
The Matrix Size should be set to 64.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 173

10.6 Making Contour Plots

To show the results in Studio, insert an FE-Display and select either the original FIL file, or the
FER file created by FE2FES, as the source of the model. Go to 'Properties' and in 'Groups'
deselect the solids.
Drag on your fatigue results (top surface, bottom surface and worst case should all be available)
from the Available data window and use the properties to select the results to view. Damage and
Life results are automatically shown in logarithmic scaling, for example:

Figure 168 Contour plot in Studio

Numerical results can also be shown in Studio by dragging a fatigue results set onto a blank
page. They are automatically sorted, with the elements with shortest lives at the top of the list.

Figure 169 Table display of fatigue results

A report page with multiple graphics and numerical results can be generated, for example:

Figure 170 Report generated by Studio

11. Seam Weld Analysis Example

The method used in this example is based on an approach originally proposed by Volvo Car
Corporation and Chalmers University of Technology in Gothenburg and developed in cooperation
with nCode. This method has been developed specifically for automotive components welded
from thin sheets (1-3 mm in thickness). Full details of the method are included in the FE-Fatigue
theory documentation but it can be summarised briefly as follows:
1. The structure should be meshed predominantly with quadrilateral elements (NASTRAN
CQUAD4) representing the mid-planes of the metal sheets, and the weld bead represented
by a single row of shells with thickness corresponding to the thickness of the weld throat.
The nodes of these elements should lie at the weld toe. The mesh around the weld should
be regular with elements of around 5 mm in size, and triangular elements should be avoided.
See Figure 171. The row of shell elements representing the weld bead must have their
surface normals (element Z-axes) pointing outwards – i.e., towards the welder.
2. When running the NASTRAN analysis, the parameter SNORM must be set e.g.
PARAM,SNORM,55.0 and stresses recovered to the output2 file using the CUBIC option.
See Figure 172 for an example file header.
3. Unaveraged node-at-element stresses are recovered for the nodes at the weld toes, based
on the elements adjacent to the weld (not the weld elements themselves) and used to make
the fatigue calculations. The worst result for each node is reported to the results file.
4. The method determines the contribution of bending to the total stress and from this
determines whether the weld is essentially “stiff” or “flexible”. Each of these cases requires a
different S-N curve.

Figure 171 Seam weld modelling of a simple T-joint

SOL 101
CEND
$
TITLE = static analysis
SPC=1
SUBCASE 1
SUBTITLE = Simple loadcase 100 N
LOAD = 2
STRESS(CUBIC,PLOT,PRINT)=all
DISPLACEMENT(PLOT)=all
BEGIN BULK
PARAM,POST,-1
PARAM,AUTOSPC,YES
PARAM,EPZERO,1.E-5
PARAM,PRGPST,NO
PARAM,K6ROT,1.
PARAM,SNORM,55.
$
$ GRID Data
$
GRID 1 0.0 0.0 0.0

11.1 Introduction
The example considered here is a rear trailing suspension arm which is constrained at 3
locations – corresponding to the swing-arm mount, and the upper and lower lateral arm mounts.
The arm is subjected to 6 simultaneous loading histories representing 3 orthogonal forces and 3
moments applied at the wheel centre. The model is illustrated below.

Figure 173 Model of trailing arm showing three of the static load-cases

Analysis was carried out in NASTRAN using 6 loading cases:

• 1000 N in the x-direction

• 1000 N in the y-direction

• 1000 N in the z-direction

• 1000 Nmm in the x-direction

• 1000 Nmm in the y-direction

• 1000 Nmm in the z-direction

The component is relatively stiff, so the fatigue analysis is to be carried out using the linear static
superposition approach. To speed up the analysis, the loading histories have been peak-valley
sliced using PVXMUL with a cycle gate of 10% of the total range.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 179

The filtered loading histories are shown below.

X-Force
Test wheel_force, Chan 1 N
2134

-2134
0 2000 4000 6000 8000 1E4
Y-Force
Test wheel_force, Chan 2 N
2142

-2142
0 2000 4000 6000 8000 1E4
Z-Force
Test wheel_force, Chan 3 N
6781

-6781
0 2000 4000 6000 8000 1E4
X-Moment
Test wheel_force, Chan 4 Nmm
7.266E4

-7.266E4
0 2000 4000 6000 8000 1E4
Y-Moment
Test wheel_force, Chan 5 Nmm
8.529E4

-8.529E4
0 2000 4000 6000 8000 1E4
Z-moment
Test wheel_force, Chan 6 Nmm
1.8E4

-1.8E4
0 2000 4000 6000 8000 1E4

Time (Seconds)
Time Sel=Pick Cursor=Signal Auto Rnd=Off Mark Snap=Off Chan Sel=List

Figure 174 Peak-valley sliced loading files.

In this exercise you will carry out a fatigue life prediction using the “Volvo” approach to determine
whether or not this component is fit for service. The component is required to survive 100 repeats
of the loading history. The required files are:
susp_arm.dat
susp_arm.op2
susp_arm.hm
susp_arm.res
wheel_force.rsp

These files are located in the demo directory. They should be copied to your working directory
before starting the exercise by using the Macros pull-down in the main FE-Fatigue/nSoft
program.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 180

11.2 Review model and stress analysis results

The model used in this analysis has a row of shell elements representing a weld, as described
above. Unaveraged nodal stresses are recovered to the OP2 file using the STRESS(CUBIC)
option in NASTRAN. The stresses used in the analysis are the unaveraged stresses from the
weld toe elements at the weld toe nodes. Fatigue analyses are carried out at each node based
on the unaveraged contributions to stress from each of the weld toe elements and the fatigue
results are reported for the worst case. The choice of elements and surfaces depends on the
selection made when running the FE2FES translator. See the FE-Fatigue Theory manual for
more details.
As described above, results exist for 6 loadcases. If you wish, you may now review the model
and stress analysis results using a suitable FE post-processor.
Alternatively, the model may be viewed using the Studio FE-Display.
You will see that the weld we are going to consider is represented by elements with the property
set PSHELL22553 and where it joins PSHELL1001 and PSHELL7.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 181

11.3 Generate FES file using FE2FES

In this example the FES file is generated directly from the FE stress analysis results file (a
NASTRAN op2 file) using the Generic FE to FES translator FE2FES. Run the translation as
follows:

• Start the main nSoft interface in your working directory in the usual manner and from the
nSoft menu, select FE-Fatigue and then Generic FE to FES Translator
(fe2fes). The interface should appear as illustrated below.

Figure 175 FE2FES main form

• After the file has been scanned, the Group Selection form appears. By default, the output
filename will be called susp_arm_01.fes. Set the Solution Location to Seam
Weld. Then select the group (property set) containing the weld elements. For this
analysis, we need to select SHELL_22553 – highlight this item and transfer it to the
Selected Groups box by clicking on the right arrow.
.

Figure 176 Group Selection form

• The form should now appear as above. Click on Next to continue.

The next form is the Weld Toe Element and Surface Selection form.

• Choose the Intelligent Selection option. This looks at the geometry of the model and,
based on the selected weld element groups, identifies the correct weld toe elements and
surfaces to use in the fatigue analysis. For more details of this and the other options, see
the FE-Fatigue Theory manual.

• Checking the ASCII output option and specifying a filename will result in a comma
separated list of the selected weld toe elements being written to the output file. This may
be read into your FE post-processor as a way of checking that the correct elements are
identified. The Intelligent Selection option is quite reliable, so long as the meshing rules
are followed, but occasionally additional spurious elements may be included.

Figure 177 Intelligent Selection of weld toe elements

• Click on Next to continue.

• The next form is the Results Selection form. Click on Select All to select all the
available loadcases and then click on the right arrow to transfer these to the Selected
Results box. The screen should now look as below. Click on Next to continue.

Figure 178 Results Selection form

• Finally, the Summary screen is shown.

Figure 179 Translation summary form

11.4 Fatigue analysis using FATFE

Firstly, open the FES file and add Advanced analysis options, as shown below:

• Start the FE-Fatigue Analysis (FATFE) either by using the nSoft menu system or by typing
fatfe at the nSoft command prompt. On the Fatigue Jobname Entry form, either input or
select the file susp_arm_01.fes.

• This file is the partial FES file created by the translator. It is now necessary to complete
the form by adding in analysis, loading and material information. On the form illustrated
below, select Advanced Options and click on OK.

• On the Advanced Options form, set the Mean Stress Correction to Yes.

• Set the Miner’s Sum to 0.3. Experience indicates that a Miner’s Sum setting of 0.3-0.5
is suitable for variable amplitude loadings.

• Note the other available options. Then click on OK.

Figure 180 Partial to Full Job Completion form

Loading information form

The Loading Information form is used to establish the correspondence between the applied static
loading cases and the loading histories.
Fill in the Loading Information form as follows:

• Select the first loadcase on the list (1000N IN X-DIRECTION) and click on OK.

• Set the File Type to RPCIII.

• Select or enter the Time History wheel_force.rsp.

• For Channel Number, enter 1.

• In the box labelled Divide stress/strain by enter 1000.

• Click on OK.
Continue and enter the information for the other 5 loadcases. When you have finished, the
Loading Information form should appear as shown below.

Figure 181 Completed Loading Information form

If you want to view the loading histories you can create a multi-file display by pressing F5 to
access the options.
When you have finished, select Done and click on OK to proceed to the next form.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 188

Material Input form

On the Material Input form, S-N curves are associated with different parts of the model. First we
will use the Material Manager to view the S-N curves from the database (note that this step is not
essential for running an analysis).

• Press F5 and select Materials Manager to start the Material Database Manager.

• Select Load, data set 1 and select seam_steel_stiff from the list of available
datasets.

• Select Load, data set 2 and select seam_steel_flex.

• Select Graphical display to compare the S-N curves.

In the fatigue analysis, FATFE will use one or other of these curves depending on whether the
stress at the weld toe is predominantly due to bending (flexible) or normal (stiff) loading. The
curves in the database are generic curves which should give reasonable results for welds in
typical automotive steels.

• Now exit from the graphics and the Material Database Manager to return to the Material
Input form

• Select the group SHELL_22553 from the form and click OK.

• For the Flexible S-N curve select seam_steel_flex and for the Stiff S-N curve select
seam_steel_stiff.

• The seam weld approach includes a mean stress correction option. This uses two mean
stress sensitivity values M1 and M2. M1 applies to load ratio R < 0 and M2 applies to load
ratio R > 0. Only M1 is stored in the material database, so it is necessary to supply the
ratio of M1/M2 to be used here. The normal value is 2.5

• One of the main features of the seam weld approach is that it distinguishes between
cases where the loading on the weld toe is predominantly due to bending and when it is
predominantly due to normal load. A stress ratio r is determined by looking at the top and
bottom surface stresses at the weld. When the loading is pure bending the value of r is 1
and when it is pure normal load the value is 0. The Transition ratio defines the value of r
at which the software switches from one S-N curve to the other. The decision about which
curve to use at an individual node is based on a weighted average of the bending ratio.
The average is weighted in proportion to the square of the stress. If the weighted average
value of r is less than the Transition ratio, the "stiff" curve will be used. Otherwise, the
"flexible" curve is used.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 189

• The form should now look as below. Click OK to continue, and reply Yes to the ensuing
question.

Figure 182 Completed Material Input form

Run the analysis

The FATFE Analysis form now appears

• Click OK to accept the defaults on this form.

• On the Results Filename Entry form, make sure you select the Output File Format
appropriate for your post-processor.

• Alternatively, if you wish to review the results in Studio FE-Display, select the option
2 -Add to FER file (name as FES) from the FER File Output options.

• Click OK to initiate the analysis.

The results for the “worst” elements are listed by the Fatigue Monitor as the analysis proceeds
(see below).
The most damaged node should be 34358 with 568 repeats.

Figure 183 The Fatigue Monitor

When the analysis is complete, acknowledge any warning messages and then click on OK to exit
the Fatigue Monitor.
NOTE: If your answers differ from those shown above, check the Analysis Preferences settings.
The Matrix Size should be set to 64.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 191

11.5 Review results

FE-Fatigue provides a number of ways of viewing the results of an analysis. We will now look at
some of these methods.

• From the FATFE main menu, select Utilities > Results Listing. Select the
name of the job you have just run (should be susp_arm_01) and accept the defaults on
the subsequent forms to view the results for the most damaged nodes (see below).

Figure 184 Results listing in FATRES

• From the FATFE main menu, select Utilities > Node/Element options. Accept
the defaults on the next two forms to generate detailed time history information for the
worst node. The results are viewed by default in Studio. For each element contributing
to the fatigue result at the worst node, five time histories are generated, being the bending
ratio, bottom surface stress, top surface stress, bending stress and direct stress. Exit from
Studio and return to FATFE.
The fatigue analysis results may also be displayed by reading the appropriate results file into
your FE post-processor.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 192

Alternatively, the results may be displayed using Studio FE-Display.

• From the main FE-Fatigue (or nSoft) interface, start Studio by typing studio at the
prompt.
Studio will start and show an empty display area. To insert the model on the display page:

• Select Insert > FE Display

• Select one of the three files offered – susp_arm.dat, susp_arm.op2 and

susp_arm_01.fer. All contain the model geometry and can be displayed. Click on
Open.
The model will be shown on the current Studio page. (Other displays can be added to the same
page or a single view can be selected using View > Interactive Mode)
The model can be dynamically manipulated in the display by holding down either the Control or
Shift keys together with the mouse buttons. (Alternatively, function F2 and F3 keys can be used).
Put the cursor on the display and move mouse to manipulate the model. The controls are
summarized as:
Rotate – CTRL+ Left Mouse Button, or F2 + Left Mouse Button
Pan – CTRL + Right Mouse Button, or F2 + Right Mouse Button
Cursor Zoom – SHIFT + Left Mouse Button, or F3 + Left Mouse Button
Box Zoom – SHIFT + Middle Mouse Button, or F3 + Middle Mouse Button
Squiggle Zoom – SHIFT + Right Mouse Button, or F3 + Right Mouse Button
NOTE: NOTE: The Controls Summary is also available by clicking the mouse icon on the toolbar.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 193

A fringe plot of the seam weld fatigue results may now be made as follows:

• In the data tree in the Available Data window, expand the branch nCode S3 fatigue
results.

• Click and drag the seam weld results on to the model

By adjusting the position of the model and modifying the properties (right click over the model if
these are not visible) you should be able to produce something similar to the following figure.

Figure 185 Contour plot in Studio

12. Temperature Corrected S-N Analysis Example

This example performs an S-N analysis on a piston model. Due to the steady state temperature
distribution within the piston, the fatigue properties of the material vary throughout the model. A
set of S-N data curves for different temperatures is interpolated to determine the fatigue
properties at each location. (Similar corrections can also be performed for Strain-Life analysis)

12.1 Introduction
A piston model is loaded via a distributed pressure load across the crown, and restrained from
translating at the wrist pin. This data is provided in ABAQUS results file [Link]. The model
contains 2nd order tetrahedral elements but the surface has been ‘coated’ with thin shell elements
to recover the surface stresses and reduce the data to be analysed.

Figure 186 Stress distribution

The piston also experiences a steady state temperature distribution due to combustion with a
maximum temperature in excess of 300°C. This temperature data has been calculated at the
nodes of the model and is provided in an ASCII file, piston_temp.tds.

Figure 187 Temperature distribution

The materials data of the aluminium alloy is provided in the [Link] materials file. The
material is named AL_PISTON and fatigue data is provided for 20, 200, 345°C.

Figure 188 Materials data

Given the above material data and temperature distribution, the task is to determine how many
times the pressure load can be applied (zero to maximum load) before failure occurs.
The following files are required for this example from the demo directory:

• [Link]

• piston_temp.tds
These files can be copied to your working directory using the Macros pulldown in the main
FE-Fatigue/nSoft program:
Macros\System\nCode\Copy Files for FE-Fatigue Examples\AbaqusPiston.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 197

12.2 Reading FE Results File

If you wish to view the model before reading the results, use a suitable pre-post processor to
import the ABAQUS input file [Link].
Alternatively, the model may be viewed using Studio FE-Display.

Figure 189 ABAQUS input file

The first task is to import the stress data from the static stress results file ([Link]) to a
partial input (.FES) file for FE-Fatigue.
From the Analysis menu pull-downs, select FE-Fatigue, and then the Generic FE to FES
Translator (FE2FES). Alternatively, type FE2FES in the command prompt line to start the
module.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 198

At the first FE2FES form, select the following:

Figure 190 FE2FES: selecting the input file

• Filename: [Link]

• Translator: Select translator from file extension (or select ABAQUS fil)

In the Group Selection form, enter the FES filename to be created: piston_01, keep the file
type: Binary and Solution Location: Averaged node on element. ”Averaged node on
element” means that nodal results from adjacent elements are averaged together to achieve an
averaged stress state per node. (Nodal stress values are used since the temperatures have been
calculated at the nodes and the location of stress and temperature must be consistent i.e. either
both at elements or both at nodes.)
With the Group Type: Set, select only the shell elements for the analysis by clicking on the
STRESS_COAT group and move to Selected Groups list the using the right arrow.
The form should look as below:

Figure 191 FE2FES: Group Selection

To move to the next form, press Next>.

In the Results Selection form, choose the Top layer of the shell elements, select the only
loadcase available: 1-Pressure Load and move it to the Selected Results list.
Make sure you check the Load Temperature Data box, before clicking on Next>.

Figure 192 FE2FES: Results Selection

Having selected Load Temperature Data, FE2FES provides the ability to also translate
temperature data into the FES file. This data may be input from an FE results file such as an
ABAQUS fil file. In this case, we will read the data from an ASCII file. Select:

• Load Data From ASCII : ASCII File (.tds)

• Filename: piston_temp.tds

Figure 193 FE2FES: loading temperature data from ASCII file

This ASCII file has a simple format of node number and temperature, ie:

1 180.95
2 193.98
3 180.95
4 193.80
5 181.54
6 195.01
7 181.28
8 196.16
9 218.44
10 204.84
11 213.14

It also possible to add temperatures to a pre-existing FES file by using the LOADTEMP module.
More information on this (and keyword information which can added to temperature file) can be
found in the [Link] documentation.
Having clicked Next>, FE2FES will scan the file for data and provide a list of available
temperature data.

Figure 194 FE2FES: Temperature Data Set Selection

Enter a Default Temperature: 20 then select Next>.

The translation should only take a few seconds before the summary form appears. Select the box
for starting FE-Fatigue, and click on Finish.

Figure 195 FE2FES: Summary screen

12.3 Starting FE-Fatigue

FE-Fatigue will start directly in the Analysis form. Complete it as below, choosing the Analysis
Type as S-N Analysis. Accept the defaults for the other input fields.

Figure 196 FATFE: Selecting Analysis Type

On the Loading Input form, click OK to edit the Pressure Load loadcase. Select Amplitude:
Constant and Zero-Peak. This cycles the stress from zero to the stress due the pressure load.
When completed, ensure Done is selected in the table and click OK.

Figure 197 FATFE: Loading Input form

On the Material Input form, click OK to enter the material data for the STRESS_COAT group.
Select the material AL_PISTON.

Figure 198 FATFE: Material Input form

Because the FES file already contains temperatures, the method for defining material data
defaults to be Thermal. This means the list of materials comes from the ASCII “high-
temperature” materials file [Link] in the working directory. This file can contain data for both
S-N and E-N. For S-N, this is essentially a look-up of Stress Range versus Life for different
temperatures. For E-N, strain-life parameters are defined for different temperatures.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 206

The ASCII file can be viewed from the Material form by pressing F5 and selecting High
Temperature Database.
A typical file is shown below:
NCODE MATERIAL TEMPERATURE DATABASE #1.0
TEMPERATURE UNITS = CELSIUS
STRESS UNITS = MPa
-------------------------------------------------------
MATERIAL NAME =AL_PISTON
DATA TYPE = STRESS
MATERIAL TYPE =101
MAXIMUM TEMPERATURE = 400
TEMPERATURE =20
UTS =390
E =70000
STRESS VALUES = 900 180 60 30
LIFE VALUES =1000 1E8 1E12 1E15
TEMPERATURE =200
UTS =350
E =70000
STRESS VALUES = 720 144 50 25
LIFE VALUES =1000 1E8 1E12 1E15
TEMPERATURE =345
UTS =310
E =70000
STRESS VALUES = 530 90 35 18
LIFE VALUES =1000 1E8 1E12 1E15

Because the FES file already contains temperatures at each node, the Temperature field is
defaulted to BY_LOCATION which means these temperatures will be used. If a FES file does not
contain temperatures or you wish to override the values it is possible to enter a temperature
value for a whole group.
Fatigue properties for the temperature at each node are calculated by interpolating between the
curves in the material file. If the temperature of the node is above the highest temperature curve
in the material file, FE-Fatigue will extrapolate the fatigue properties but only up to a maximum
temperature defined in the material dataset (in this case 400°C). If this temperature is exceeded,
the analysis will stop with an error. For temperatures below the lowest temperature curve (in this
case 20°C), the properties of the lowest curve are used.
(It is also possible to select materials from the standard materials database ([Link]) by
changing the Method of definition from Thermal to Select, which will ignore temperature
effects.)
When completed, ensure Done is selected in the table and click OK. On the next form, choose to
carry out the analysis.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 207

12.4 Submitting Fatigue Analysis

Select\enter the following parameters then click on OK:

• Analysis Region: All data in FES file

• Equivalent Units: 1 Repeats

• Safety Factor Analysis: None

The Results Filename Entry screen is shown.

Figure 199 FATFE: Results filename entry screen

On the Results Filename Entry screen, select:

• Enter Results Filename: piston_01

• Output File Format: Select here a suitable format for your FE post-processor.

• FER File Output : 2- Add to FER file (This writes the data to the FER file for
postprocessing in Studio)

• Results description: <leave blank>

The Results Monitor will appear and show the damage for nodes as they are calculated.

Figure 200 FATFE: Results Monitor

The shortest life should be at node 7131 with a life of 9.5522E+10 repeats.
NOTE: If your answers differ from those shown above, check the Analysis Preferences settings.
The Matrix Size should be set to 64.
If you wish to assess the effect of temperature in this case, you can choose Edit Job >
Material Input and change the temperature for the group to be a constant 20°C and then
re-run the job. (Ensure you change the output results filename to avoid overwriting the original
results). The life will increase by more than a factor of 10.
Results can be listed by using the results listing module fatres (accessed from the FE-Fatigue
menu or Utilities > Results listing). This will read the ASCII .fef results file.
Alternatively, the graphics and reporting module Studio is able to list the binary .fer files (i.e.
piston_01.fer) as a grid display.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 209

12.5 Making Contour Plots

The fatigue analysis may now be displayed by reading the appropriate results file into your FE
post-processor.
Alternatively, the results may be displayed using Studio FE-Display. To do this:

• From the main FE-Fatigue (or nSoft) interface, start Studio by typing studio at the
prompt.
Studio will start and show an empty display area. To insert the model on the display page:

• Select Insert > FE Display

• Select one of the three files offered – [Link], [Link] and piston_01.fer.
All contain the model geometry and can be displayed. Click on Open.
The model will be shown on the current Studio page. (Other displays can be added to the same
page or a single view can be selected using View > Interactive Mode)
The model can be dynamically manipulated in the display by holding down either the Control or
Shift keys together with the mouse buttons. (Alternatively, function F2 and F3 keys can be used).
Put the cursor on the display and move mouse to manipulate the model. The controls are
summarized as:
Rotate – CTRL+ Left Mouse Button, or F2 + Left Mouse Button
Pan – CTRL + Right Mouse Button, or F2 + Right Mouse Button
Cursor Zoom – SHIFT + Left Mouse Button, or F3 + Left Mouse Button
Box Zoom – SHIFT + Middle Mouse Button, or F3 + Middle Mouse Button
Squiggle Zoom – SHIFT + Right Mouse Button, or F3 + Right Mouse Button
NOTE: NOTE: The Controls Summary is also available by clicking the mouse icon on the toolbar.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 210

A contour plot of the fatigue analysis results may now be made as follows:

• In the data tree in the Available Data window, expand the branch nCode S3 fatigue
results.

• Click and drag the Stress-Life (S-N) results on to the model.

• The default results shown are fatigue damage.

The Properties window can be used to modify what is seen on the screen by changing
background colour, model appearance, groups and results (e.g. damage, life, temperature). If the
Properties window is not visible, click the right mouse button over the model and select
Properties. It should be possible to produce something similar to the following figure.

Figure 201 Studio contour plot

12.6 Further considerations

This is a straightforward example of using temperature corrections and in practice there may be
other considerations:
Mechanical Properties in Stress Analysis
It is worth noting that if the influence of temperature on material mechanical properties, such as
Young's modulus and Poisson's ratio, is significant, the steady state temperature should be taken
into account in the FE analysis when calculating the stresses due to a mechanical load.
Thermal Stresses
This example only considers mechanical stresses due to the applied load. A temperature
distribution will cause expansion and therefore thermal strains. Thermal stresses are caused by
constrained thermal strains. In the steady state, thermal stresses effectively shift the mechanical
mean stresses and consequently affect the fatigue life. The effects of these thermal stresses can
be simulated in FE-Fatigue by creating a FES file of the mechanical stresses and including the
thermal stresses as an additional loadcase. The thermal stresses can be added as a constant
stress offset by using with a constant time history of unit magnitude.
Variable Temperatures and Thermal Cycling
The life prediction described in this example is limited to a constant steady-state temperature. In
practice, however, temperatures may vary. A reasonable approach can be to use the maximum
temperature as an equivalent constant temperature. However, variable temperatures have an
additional effect of causing cycles of thermal stress which themselves can cause fatigue damage.
These can be analysed with FE-Fatigue by creating a FES file of thermal stresses (typically a
transient FES file with the stresses at one temperature to the stresses at a different temperature).
Again, an assumption would be required for the effective temperature distribution for the fatigue
properties. It should be noted, however, that in such cases there may be a benefit to using
thermo-mechanical fatigue models that use time-step temperature and stress results rather than
steady state results. This is particularly the case at elevated temperatures (above 50% of melting
temperature) where there are creep-fatigue interactions. These effects are not accounted for in
this release of FE-Fatigue but are currently the subject of an extensive research collaboration
between nCode and several major corporations.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 212

13. Vibration Fatigue Example using NASTRAN

This example performs a vibration fatigue S-N analysis using Power Spectral Density stress
results from NASTRAN v2001. FE-Fatigue calculates a PSD of maximum principal stress then
calculates the expected magnitude of cycles of physical stress (using the Dirlik method). Stress-
Life analysis is used to compute the resulting damage per second of vibration.

13.1 Introduction

Enforced
vertical
acceleration
PSD

Figure 202 Cantilever beam sample

In this example, a 250 mm cantilever beam is excited by an enforced vertical acceleration at one
end. This input is in the form of a PSD of acceleration of g^2/HZ and is defined by a simple
spectrum:
0 - 40 Hz 0.01 g^2/Hz
50 - 150 Hz 5.0 g^2/Hz
160 - 240 Hz 0.1 g^2/Hz
250 - 300 Hz 0.01 g^2/Hz

Figure 203 Input acceleration PSD

A modal analysis of the beam has shown the first 5 modes of vibration to be:
1 0.04 Hz 1st bending mode
2 31.10 Hz 1st lateral mode
3 62.91 Hz 2nd order bending mode
4 77.63 Hz Torsion mode
5 176.08 Hz 3rd order bending mode
Therefore the loading spectrum will primarily excite the 2nd order vertical bending mode at
62.91 Hz.

Figure 204 2nd order bending mode – 62.91 Hz (Von Mises stress)

This example calculates how many seconds the beam can withstand this loading before failure
occurs, using HYPRESS20 steel material properties.
The following files are required for this example from the demo directory:
b_vib.dat
b_vib.op2
These files can be copied to your working directory using the Macros pulldown in the main
FE-Fatigue/nSoft program:
Macros\System\nCode\Copy Files for FE-Fatigue Examples\NastranVibration
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 214

13.2 NASTRAN input file

The NASTRAN input file b_vib.dat shows the required job setup and can be reviewed using a
text editor. The job is defined similarly to a standard modal frequency response analysis
(SOL 111), with the first step calculating the frequency stress response to a unit input with
frequency. In this case, an enforced acceleration of 1 g (9810 mm/s^2) with frequency is
achieved using the large mass method. Modes up to 900 Hz are used in the calculation. An
additional case control card, RANDOM, signals that a further step is performed for random
vibration analysis. The input is defined using RANDPS and TABRND cards to define the loading
PSD spectra:
Case control cards :
RANDOM = 100
Bulk data cards:
RANDPS,100,1,1,1.,0.,333
TABRND1,333,LINEAR,LOG
,0.0000,1.000E-2,10.0000,1.000E-2,20.0000,1.000E-2,30.0000,1.000E-2,
,40.0000,1.000E-2,50.0000,5.0000,60.0000,5.0000,70.0000,5.0000,
,80.0000,5.0000,90.0000,5.0000,100.0000,5.0000,110.0000,5.0000,
,120.0000,5.0000,130.0000,5.0000,140.0000,5.0000,150.0000,5.0000,
,160.0000,0.1000,170.0000,0.1000,180.0000,0.1000,190.0000,0.1000,
,200.0000,0.1000,210.0000,0.1000,220.0000,0.1000,230.0000,0.1000,
,240.0000,0.1000,250.0000,1.000E-2,260.0000,1.000E-2,270.0000,1.000E-2,
,280.0000,1.000E-2,290.0000,1.000E-2,300.0000,1.000E-2,ENDT

(Only a single load is used in this case, however multiple RANDPS cards can refer to the same
RANDOM id to enable multiple simultaneous PSDs and Cross Spectra loading )
5% critical damping is defined on the TABDMP cards.
The PSD stresses are requested using STRESS(PLOT,PSDF)=10 and recovered at frequencies
determined by the FREQ3 card. This distributes more frequencies near to resonant frequencies of
the model.
NOTE: NASTRAN v2001 or later is required for the PSD results to be written to the op2 file.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 215

13.3 Reading FE Results File

If you wish to view the model before reading the results, use your FE pre/post-processor to
import the NASTRAN input file b_vib.dat. Alternatively, the model may be viewed using Studio
FE-Display.
This first task is to import the stress data from the static stress results file (b_vib.op2) to a
partial input file for FE-Fatigue.
From the nSoft Menu pull-downs, select FE-Fatigue, and then the Generic FE to FES
Translator (FE2FES). Alternatively, type FE2FES in the command prompt line to start the
module.

Figure 205 FE2FES: Input Parameters form

In the first FE2FES form, select the following:

• Filename: b_vib.op2

• Translator: Select translator from file extension (or select NASTRAN op2)

In the Group Selection form, enter the FES filename to be created: b_vib_01, keep the file type:
Binary and select Solution Location: Element.
With the Group Type: Property, select only the shell elements on the surface of the beam for
the analysis by clicking on the SHELL_39005 group and move to the Selected Groups list by
using the right arrow.
The form should then look as below:

Figure 206 FE2FES: Group Selection form

To move to the next form, click Next>.

In the Results Selection form:

• Results Type: Random Vibration Stress

• Shell Layer: Top (Z2)

• Memory Based Translation: Yes

• Write Model: Yes

Select all 64 frequencies available and move them to the Selected Results list then click Next>.
The translation combines the component stress data from NASTRAN to create a FES file of PSD
of maximum principal stress for each element.

Figure 207 FE2FES: Results Selection form

13.4 Starting FE-Fatigue

FE-Fatigue will start directly in the Analysis form, and the analysis type will be automatically set
to Vibration Fatigue S-N. The units are MPa and the user may wish to review the
Advanced Options form where default the cycle counting method is set to Dirlik.

Figure 208 FATFE: Analysis form

Since the loading is fully defined in the FE analysis, when you OK the Analysis form the next
form is the Material definition form.

13.5 Materials information

Figure 209 FATFE: Material Input form

Enter the following information at the Material Input form:

• Material Name: HYPRESS20

• Strength Reduction (Kf): 1

• Finish: No Finish

• Treatment: No Treatment

• Scale Factor: 1

• Offset: 0

• Click on OK and choose to Analyse Job Now.

13.6 Submitting Fatigue Analysis

Select the following parameter then click on OK:

• Analysis Region: All data in FES file

On the Results Filename Entry form, select

• Enter Results Filename: b_vib_01

• Output File Format: Select a file format suitable for your FE post-processor.

• Output Life Units : Seconds

• FER File Output : 2- Add to FER file (name as FES). This allows post-
processing of the fatigue analysis results in Studio.

Figure 210 FATFE: Results Filename Entry form

NOTE: The Results Description field is optional.

Click on OK to start the analysis.

Figure 211 FATFE: Results Monitor

The Results Monitor will appear and show the fatigue lives for elements as they are calculated.
Note that there is not currently a method implemented for ranking a vibration fatigue analysis and
therefore the user should wait until the job is complete before reviewing the results.
The shortest life should be at element 408 with a life of 9601 seconds to failure. Note that results
are directly reported in units of time, not in terms of 'repeats' as with time history simulations.
NOTE: If your answers differ from those shown above, check the Analysis Preferences settings.
The Matrix Size should be set to 64.

13.7 Reviewing results

To review the results in more detail, they can be listed by using the results listing module fatres
(available from the FE-Fatigue menu or Utilities > Results listing). This will read the
ASCII .fef results file.

Figure 212 FATRES: selecting the .fef file

Select the Life Units as Seconds and view the top 20 damaged elements.

Figure 213 FATRES: Preferences screen

Further results available include the Irregularity factor which is a statistical measure of
the time history which could be inferred from the PSD. It is defined as the ratio of number of zero
crossings in the time history to the number of peaks. An irregularity factor of 1.0 is a pure sine
wave of single frequency and irregularity factor of 0.0 indicates white noise. In this case, an
irregularity factor of over 0.9 shows the response is being primarily dominated by a single
frequency. Root M0 is defined as the square root of the area under the Stress PSD for each
element and is effectively an RMS value of the stress response. Note the Angle Range is not
currently used.

Figure 214 FATRES: Results listing

Further post-processing can enable the stress PSD for each element to be viewed or the
histogram of expected cycles per second to be extracted.
From the FATFE menu select:

• Utilities > Node/Element Options

Use the default results file b_vib_01.fef and use the worst case element 408.
Select:

• PSD Extraction
The PSD of stress^2/Hz for element 408 (b_vib_01408.mdf) will be displayed using [Link]
plot shows that the frequencies used in the analysis are unequally spaced (due to FREQ3
NASTRAN card). Use the Log Y Axes setting to see more detail in the response.

Figure 215 Stress PSD viewed using TPD module

It should be noted that although the FREQ3 card has distributed frequencies near resonances,
accuracy would be improved if there were more points near the significant frequency of the 2nd
order bending mode near 63 Hz. Insufficient detail in the frequency response can either miss
peaks which will reduce predicted damage, or insufficient points either side of a resonance will
cause too wide a peak which will be overly damaging. The selection of damping and response
frequencies can have a significant effect on results from a vibration fatigue analysis.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 223

13.8 Making Contour Plots

The fatigue analysis results may now be displayed by reading the appropriate file into your FE
post-processor.
Alternatively, the results may be displayed using Studio FE-Display. To do this:

• From the main FE-Fatigue (or nSoft) interface, start Studio by typing studio at the
prompt.
Studio will start and show an empty display area. To insert the model on the display page:

• Select Insert > FE Display

• Select one of the three files offered – b_vib.dat, b_vib.op2 or b_vib_01.fer. All
contain the model geometry and can be displayed. Click on Open.
The model will be shown on the current Studio page. (Other displays can be added to the same
page or a single view can be selected using View > Interactive Mode).
The model can be dynamically manipulated in the display by holding down either the Control or
Shift keys together with the mouse buttons. (Alternatively, function F2 and F3 keys can be used).
Put the cursor on the display and move mouse to manipulate the model. The controls are
summarized as:
Rotate – CTRL+ Left Mouse Button, or F2 + Left Mouse Button
Pan – CTRL + Right Mouse Button, or F2 + Right Mouse Button
Cursor Zoom – SHIFT + Left Mouse Button, or F3 + Left Mouse Button
Box Zoom – SHIFT + Middle Mouse Button, or F3 + Middle Mouse Button
Squiggle Zoom – SHIFT + Right Mouse Button, or F3 + Right Mouse Button
NOTE: The Controls Summary is also available by clicking the mouse icon on the toolbar.
| Previous | Next | FeFat6 Menu | Main Menu |
FE-Fatigue Release 6
Worked Examples Page 224

A contour plot of the fatigue analysis results may now be made as follows:

• In the data tree in the Available Data window, expand the branch nCode S3 fatigue
results.

• Click and drag the Vibration Fatigue S-N results on to the model.

• The default results shown are fatigue damage.

Figure 216 Fatigue results displayed in Studio

Note that the distribution of fatigue damage also confirms that the random vibration analysis is
primarily exciting the 2nd order bending mode.

Fefategs
No ratings yet
Fefategs
229 pages
Fe Safe
No ratings yet
Fe Safe
13 pages
nCode Design Life Fatigue Analysis Guide
100% (2)
nCode Design Life Fatigue Analysis Guide
34 pages
Fatigue Evaluation with nCode DesignLife
No ratings yet
Fatigue Evaluation with nCode DesignLife
34 pages
DesignLife Fatigue Analysis Software - 477
No ratings yet
DesignLife Fatigue Analysis Software - 477
8 pages
Dynamic Fatigue Analysis Techniques
No ratings yet
Dynamic Fatigue Analysis Techniques
60 pages
Fatigue 2013 Doc Quickstart
No ratings yet
Fatigue 2013 Doc Quickstart
932 pages
FATLAB: Open-Source Fatigue Analysis Tool
No ratings yet
FATLAB: Open-Source Fatigue Analysis Tool
12 pages
CAE Associates - Fatigue in Ansys
No ratings yet
CAE Associates - Fatigue in Ansys
23 pages
Simulia Fe Safe
No ratings yet
Simulia Fe Safe
6 pages
MSC Nastran Fatigue Analysis Setup
No ratings yet
MSC Nastran Fatigue Analysis Setup
4 pages
FEMFAT Flyer e Web
No ratings yet
FEMFAT Flyer e Web
8 pages
Evaluating Fatigue
No ratings yet
Evaluating Fatigue
14 pages
Introduction To Fe Safe
No ratings yet
Introduction To Fe Safe
22 pages
Fe Safe User
No ratings yet
Fe Safe User
395 pages
Track 1 1440 Fatigue Modelling
No ratings yet
Track 1 1440 Fatigue Modelling
35 pages
fe-safe 2024 User Guide for Fatigue Analysis
No ratings yet
fe-safe 2024 User Guide for Fatigue Analysis
393 pages
Fatigue Failure Analysis Techniques
No ratings yet
Fatigue Failure Analysis Techniques
17 pages
Fesafe User
No ratings yet
Fesafe User
395 pages
nCode DesignLife: Fatigue Analysis Software
No ratings yet
nCode DesignLife: Fatigue Analysis Software
8 pages
Fe-Safe User Guide (Dassault Systemes Simulia Corp.) (2022) (114514731) Libgen - Li
No ratings yet
Fe-Safe User Guide (Dassault Systemes Simulia Corp.) (2022) (114514731) Libgen - Li
393 pages
DOC11531 - MSCNastran2018 - EFatigue
No ratings yet
DOC11531 - MSCNastran2018 - EFatigue
449 pages
CAEfatigue VIBRATION Installation Guide
No ratings yet
CAEfatigue VIBRATION Installation Guide
13 pages
Fatigue Analysis in Ansys Workbench
No ratings yet
Fatigue Analysis in Ansys Workbench
31 pages
Femfat Modules
0% (1)
Femfat Modules
4 pages
LS-DYNA Fatigue Analysis Techniques
No ratings yet
LS-DYNA Fatigue Analysis Techniques
11 pages
Principles of Fatigue Analysis - MSC Software
No ratings yet
Principles of Fatigue Analysis - MSC Software
2 pages
CAEfatigue 2021.4 Technical FAQ
No ratings yet
CAEfatigue 2021.4 Technical FAQ
13 pages
Aerospace Spectrum File Support Guide
No ratings yet
Aerospace Spectrum File Support Guide
8 pages
nCode DesignLife: Fatigue Analysis Tool
No ratings yet
nCode DesignLife: Fatigue Analysis Tool
8 pages
Using FEA Results For Fatigue Analysis
No ratings yet
Using FEA Results For Fatigue Analysis
7 pages
SolidWorks Simulation Help - Fatigue Life Analysis
No ratings yet
SolidWorks Simulation Help - Fatigue Life Analysis
31 pages
Caefatigue Software 2022.2 (HTML) : © 2026 Hexagon Ab And/Or Its Subsidiaries
No ratings yet
Caefatigue Software 2022.2 (HTML) : © 2026 Hexagon Ab And/Or Its Subsidiaries
6 pages
FEMFAT 4.8 Report File Overview
No ratings yet
FEMFAT 4.8 Report File Overview
4 pages
COSMOSM Fatigue Analysis Modules
No ratings yet
COSMOSM Fatigue Analysis Modules
51 pages
Hexagon Documentation Site Export
No ratings yet
Hexagon Documentation Site Export
14 pages
Chapter 14 Fatigue
No ratings yet
Chapter 14 Fatigue
12 pages
Fatigue Analysis for Finite Element Models
No ratings yet
Fatigue Analysis for Finite Element Models
6 pages
Introduction to Fatigue Analysis
No ratings yet
Introduction to Fatigue Analysis
59 pages
2019 Inn Conf Day2 Failure Crack Initiation Propagation
No ratings yet
2019 Inn Conf Day2 Failure Crack Initiation Propagation
65 pages
Understanding Fatigue Analysis in Materials
No ratings yet
Understanding Fatigue Analysis in Materials
23 pages
Fatigue Analysis of Piping with Caesar II
No ratings yet
Fatigue Analysis of Piping with Caesar II
4 pages
Fatigue Analysis in ANSYS Mechanical
No ratings yet
Fatigue Analysis in ANSYS Mechanical
12 pages
Fatigue Failure in Mechanical Components
No ratings yet
Fatigue Failure in Mechanical Components
10 pages
Fatigue Analysis Using S-N Method
No ratings yet
Fatigue Analysis Using S-N Method
8 pages
7 Christian Gaier Multi Axial Fatigue Analysis With The Fe Post Processor Femfat Utmis 2010
100% (2)
7 Christian Gaier Multi Axial Fatigue Analysis With The Fe Post Processor Femfat Utmis 2010
50 pages
Fatlab User Manual: Fatigue Analysis Guide
No ratings yet
Fatlab User Manual: Fatigue Analysis Guide
59 pages
Durability and Fatigue Life Analysis Using MSC Fatigue
No ratings yet
Durability and Fatigue Life Analysis Using MSC Fatigue
2 pages
ASME Fatigue App Documentation
No ratings yet
ASME Fatigue App Documentation
21 pages
MSC Fatigue 2016: Theory Guide
No ratings yet
MSC Fatigue 2016: Theory Guide
236 pages
Fatigue Analysis in CAESAR II
100% (1)
Fatigue Analysis in CAESAR II
14 pages
Strain-Life Fatigue Life Calculation Program
No ratings yet
Strain-Life Fatigue Life Calculation Program
91 pages
Fatigue Analysis in ANSYS: Overview and Methods
No ratings yet
Fatigue Analysis in ANSYS: Overview and Methods
70 pages
Preload and Operating Load in Fatigue Analysis
No ratings yet
Preload and Operating Load in Fatigue Analysis
5 pages
CAE Fatigue Analysis Workshop Overview
No ratings yet
CAE Fatigue Analysis Workshop Overview
47 pages
Walker Alternating Stress Analysis
No ratings yet
Walker Alternating Stress Analysis
22 pages
ANSYS NCode DesignLife - Mentley
No ratings yet
ANSYS NCode DesignLife - Mentley
20 pages
Fatigue Analysis for Plant Engineering
100% (1)
Fatigue Analysis for Plant Engineering
25 pages
Understanding Fatigue in Materials
No ratings yet
Understanding Fatigue in Materials
5 pages
Nonlinear Vertical Vibration of Tension Leg Platforms
No ratings yet
Nonlinear Vertical Vibration of Tension Leg Platforms
14 pages
MSC Thesis Marina Cerd Alonso
No ratings yet
MSC Thesis Marina Cerd Alonso
89 pages
Change Management in Organizations
No ratings yet
Change Management in Organizations
1 page
Coupled vs Uncoupled TLP Models Analysis
No ratings yet
Coupled vs Uncoupled TLP Models Analysis
17 pages
Luxottica's Vertical Integration Strategy
No ratings yet
Luxottica's Vertical Integration Strategy
8 pages
Luxottica's Competitive Edge in Eyewear
No ratings yet
Luxottica's Competitive Edge in Eyewear
14 pages
Entrepreneurial Marketing & SME Sustainability
No ratings yet
Entrepreneurial Marketing & SME Sustainability
10 pages
Understanding Intellectual Property Types
No ratings yet
Understanding Intellectual Property Types
5 pages
Karas Lawsuit
No ratings yet
Karas Lawsuit
39 pages
DGX Pod Reference Design Whitepaper
No ratings yet
DGX Pod Reference Design Whitepaper
15 pages
Global Market Entry Strategies
100% (1)
Global Market Entry Strategies
7 pages
Effect, The
No ratings yet
Effect, The
75 pages
Legal Issues in Oil and Gas Industry
No ratings yet
Legal Issues in Oil and Gas Industry
40 pages
Event Planning Checklist Template
No ratings yet
Event Planning Checklist Template
4 pages
ANSYS Mechanical APDL Modeling and Meshing Guide
No ratings yet
ANSYS Mechanical APDL Modeling and Meshing Guide
230 pages
Trend Micro Business License Agreement
No ratings yet
Trend Micro Business License Agreement
12 pages
Primesim Hspice Si
No ratings yet
Primesim Hspice Si
351 pages
2022 Administrative Law Pre-bar Notes
No ratings yet
2022 Administrative Law Pre-bar Notes
12 pages
Pharmacy Opening Requirements and Tips
No ratings yet
Pharmacy Opening Requirements and Tips
4 pages
Gains from Trade and Interdependence
No ratings yet
Gains from Trade and Interdependence
30 pages
LICENSE
No ratings yet
LICENSE
3 pages
Delhi High Court Ruling on Ghazipur Slaughterhouse Lease
No ratings yet
Delhi High Court Ruling on Ghazipur Slaughterhouse Lease
9 pages
Cybersecurity Management Concepts Overview
No ratings yet
Cybersecurity Management Concepts Overview
19 pages
ArcReach SuitCase 12 Parts List
No ratings yet
ArcReach SuitCase 12 Parts List
12 pages
Understanding the TRIPS Agreement
No ratings yet
Understanding the TRIPS Agreement
37 pages
Meta Research Participation Agreement
No ratings yet
Meta Research Participation Agreement
9 pages
Power BI Client API Lab: Filtering & Menu
No ratings yet
Power BI Client API Lab: Filtering & Menu
11 pages
Sesa Elektronik Software License Confirmation
No ratings yet
Sesa Elektronik Software License Confirmation
4 pages
NVIDIA App License Agreement Summary
No ratings yet
NVIDIA App License Agreement Summary
7 pages
Necromancer's Apprentice Adventure Guide
No ratings yet
Necromancer's Apprentice Adventure Guide
7 pages
MTH800 User Guide: Key Features
No ratings yet
MTH800 User Guide: Key Features
344 pages
TI Slap161 PDF
No ratings yet
TI Slap161 PDF
32 pages
ICB Industry Classification Overview
No ratings yet
ICB Industry Classification Overview
9 pages
Stabling Charges in Railway Operations
No ratings yet
Stabling Charges in Railway Operations
66 pages
The Donkeys
No ratings yet
The Donkeys
486 pages
2020 Kilbaha Exam 1 Solutions
No ratings yet
2020 Kilbaha Exam 1 Solutions
13 pages
Slua 565
No ratings yet
Slua 565
6 pages