Using ANSYS For Finite Element Analysis, Volume II Dynamic, Probabilistic Design and Heat Transfer Analysis PDF

EBOOKS Using ANSYS for Finite Element
ALTABEY • NOORI • WANG

FOR THE Analysis SUSTAINABLE STRUCTURAL
ENGINEERING
Dynamic, Probabilistic Design and Heat
SYSTEMS COLLECTION
LIBRARY Mohammad Noori, Editor
Transfer Analysis, Volume II
Create your own
Wael A. Altabey • Mohammad Noori • Libin Wang
Customized Content
Bundle—the more Over the past two decades, the use of finite element method as a
books you buy, design tool has grown rapidly. Easy to use commercial software, such as
Using ANSYS
the greater your ANSYS, have become common tools in the hands of students as well as
practicing engineers. The objective of this book is to demonstrate the
discount!
use of one of the most commonly used Finite Element Analysis software,
THE CONTENT
• Manufacturing
ANSYS, for linear static, dynamic, and thermal analysis through a series
of tutorials and examples. Some of the topics covered in these tutorials
include development of beam, frames, and Grid Equations; 2-D elastic-
for Finite
Engineering
• Mechanical
ity problems; dynamic analysis; composites, and heat transfer problems.
Element
Using ANSYS for Finite Element Analysis, Volume II

& Chemical
Analysis
These simple, yet, fundamental tutorials are expected to assist the users
Engineering with the better understanding of finite element modeling, how to control
• Materials Science modeling errors, and the use of the FEM in designing complex load
& Engineering bearing components and structures. These tutorials would supplement
• Civil & a course in basic finite element or can be used by practicing engineers
Environmental who may not have the advanced training in finite element analysis. Dynamic, Probabilistic
Engineering
• Advanced Energy Wael A. Altabey is an assistant professor in the department of Design and Heat
Technologies mechanical engineering, faculty of engineering, Alexandria University,
Alexandria, Egypt and has been a postdoctoral researcher at the Transfer Analysis
THE TERMS International Institute for Urban Systems Engineering, Southeast
• Perpetual access for University, Nanjing, China.
a one time fee
Mohammad Noori is a professor of mechanical engineering at California
Volume II
• No subscriptions or
Polytechnic State University in San Luis Obispo, California, USA, and a
access fees
fellow of the American Society of Mechanical Engineers. Dr Noori has
• Unlimited
over 34 years of experience as a scholar and educator. He has also been
concurrent usage
a distinguished visiting professor at the International Institute for Urban
• Downloadable PDFs
• Free MARC records
Systems Engineering, Southeast University, Nanjing, China. Wael A. Altabey
For further information,
Libin Wang is a professor and the dean of the school of civil engineer- Mohammad Noori
Libin Wang
ing at Nanjing Forestry University, in Nanjing, China. He has been an
a free trial, or to order,
contact: educator and scholar, for over 20 years, and has taught the subject of
sales@momentumpress.net finite element analysis both at the undergraduate and graduate level.
ISBN: 978-1-94708-322-6
Using ANSYS for
Finite Element
Analysis
Using ANSYS for
Finite Element
Analysis
Dynamic, Probabilistic Design and
Heat Transfer Analysis
Volume II
Wael A. Altabey, Mohammad Noori,

and Libin Wang
MOMENTUM PRESS, LLC, NEW YORK

Using ANSYS for Finite Element Analysis: Dynamic, Probabilistic
Design and Heat Transfer Analysis, Volume II
Copyright © Momentum Press®, LLC, 2018.
All rights reserved. No part of this publication may be reproduced, stored

in a retrieval system, or transmitted in any form or by any means—
electronic, mechanical, photocopy, recording, or any other—except for
brief quotations, not to exceed 400 words, without the prior permission
of the publisher.
First published by Momentum Press®, LLC

222 East 46th Street, New York, NY 10017
www.momentumpress.net
ISBN-13: 978-1-94708-322-6 (print)

ISBN-13: 978-1-94708-323-3 (e-book)
Momentum Press Sustainable Structural Systems Collection
Collection ISSN: 2376-5119 (print)

Collection ISSN: 2376-5127 (electronic)
Cover and interior design by Exeter Premedia Services Private Ltd.,

Chennai, India
10 9 8 7 6 5 4 3 2 1
Printed in the United States of America

Abstract
Finite Element Method (FEM) is a well-established technique for analyz-

ing the behavior and the response of structures or mechanical components
under static, dynamic, or thermal loads. Over the past two decades the use
of finite element analysis as a design tool has grown rapidly. Easy to use
commercial software have become common tools in the hands of students
as well as practicing engineers. The objective of this two volume book is
to demonstrate the use of one of the most commonly used Finite Element
Analysis software, ANSYS, for linear static, dynamic, and thermal anal-
ysis through a series of tutorials and examples. Some of the topics and
concepts covered in these tutorials include development of beam, frames,
and grid equations; 2-D elasticity problems; dynamic analysis; and heat
transfer problems. We are hoping these simple, yet, fundamental tutorials
will assist the users with the better understanding of finite element mod-
eling, how to control modeling errors, the safe use of the FEM in sup-
port of designing complex load bearing components and structures. There
are many good textbooks currently used for teaching the fundamentals of
finite element methods. There are also detailed users manuals available
for commercial software (ANSYS). However, those sources are useful for
advanced students and users. Therefore, there was a need to develop a
tutorial that would supplement a course in basic finite element or can be
used by practicing engineers who may not have the advanced training in
finite element analysis. That is the gap addressed by this book.
Keywords
ANSYS, composite materials, Dynamics, Failure analysis, Fatigue loads,

FEM, optimization, statistics
Contents
List of Figures ix
Preface xi
1 Dynamic Analysis 1
1.1 Tutorial 1: Harmonic Analysis of Structure 1
1.2 Tutorial 2: Modal Analysis of Structure 11
2 Composite Materials 19
2.1 Composites—A Basic Introduction 19
2.2 Modeling Composites Using ANSYS 26
2.3 Tutorial 3: Simply Supported Laminated Plate
Under Pressure 38
3 Probabilistic Design Analysis 69
3.1 Probabilistic Design 69
3.2 Probability Distributions 75
3.3 Choosing a Distribution for a Random Variable 91
3.4 Probabilistic Design Techniques 96
3.5 Postprocessing Probabilistic Analysis Results 98
3.6 Tutorial 4: Probabilistic Design Analysis of Circular
Plate Bending 107
4 APDL Programming 145
4.1 Create the Analysis File 145
4.2 Tutorial 5: Stress Analysis of Bicycle Wrench 148
4.3 Tutorial 6: Heat Loss from a Cylindrical Cooling Fin 155
viii • Contents
5 Design Optimization 167

5.1 Optimum Design 167
5.2 Design Optimization Using ANSYS 172
5.3 Tutorial 7: Design Optimization Tutorial 193
Bibliography 217
About the Authors 219
Index 221
List of Figures
Figure 2.1. Illustrating the combined effect on modulus of the

addition of fibers to a resin matrix. 21
Figure 2.2. Illustrates the tensile load applied to a composite
body.22
Figure 2.3. Illustrates the compression load applied to a
composite body. 22
Figure 2.4. Illustrates the shear load applied to a composite
body.23
Figure 2.5. Illustrates the loading due to flexure on a composite
body.23
Figure 2.6. Tensile strength of common structural materials. 24
Figure 2.7. Tensile modulus of common structural materials. 25
Figure 2.8. Specific tensile strength of common structural
materials.25
Figure 2.9. Specific tensile modulus of common structural
materials.26
Figure 2.10. Layered model showing dropped layer. 29
Figure 2.11. Sandwich construction. 32
Figure 2.12. Layered shell with nodes at midplane. 32
Figure 2.13. Layered shell with nodes at bottom surface. 33
Figure 2.14. Example of an element display. 36
Figure 2.15. Sample LAYPLOT display for [45/−45/−45/45]
sequence.37
Figure 3.1. The flow of information during a probabilistic design
analysis.75
Figure 3.2. The normal probability density function (pdf). 77
x • List of Figures
Figure 3.3. The normal cumulative distribution function (cdf). 78

Figure 3.4. The uniform probability density function. 83
Figure 3.5. The uniform cumulative distribution function. 84
Figure 3.6. The lognormal probability density function for
four values of s.86
Figure 3.7. The lognormal cumulative distribution function for
four values of s.87
Figure 3.8. The Weibull probability density function for four
values of g.89
Figure 3.9. The Weibull cumulative distribution function for four
values of g.90
Figure 3.10. The graph of X1 and X2 illustrating bad sample
distribution.97
Figure 3.11. The graph of X1 and X2 illustrating good sample
distribution.98
Figure 3.12. The cumulative distribution function of the random
property X. 100
Figure 5.1. Optimization tree listing the optimization methods. 171
Figure 5.2. Optimization data flow. 175
Preface
Finite element method (FEM) is a well-established technique for analyz-

ing the behavior of mechanical and structural components of systems. In
recent years, the use of finite element analysis (FEA) as a design tool has
grown rapidly. Easy-to-use commercial software have become common
tools in the hands of students as well as practicing engineers.
In the first volume of this tutorial, we demonstrated the use of ANSYS
for Static Analysis of solid structures. In this volume we introduce the
following applications:
• Dynamic Analysis
• Composite Materials
• Probabilistic Design
• Heat Transfer
• Design Optimization Problems.
The main objective of this book is to serve as a practical tutorial to

help the readers gain insight into appropriate use of finite element mod-
eling, understand how to control modeling errors, benefit from hands-on
exercise at the computer workstation, and understand the safe use of the
FEM in support of designing complex load-bearing components and struc-
tures. There are many good textbooks already in existence that cover the
theory of FEMs. Similarly, there are detailed user’s manuals available for
commercial software (ANSYS). But, those are useful for advanced stu-
dents and users. Therefore, there was a need to develop a computer session
manual in line with the flow of a course and utilizing a software platform,
ANSYS, that is available in most engineering schools. Students will be
able to acquire the required level of understanding and skill in modeling,
analysis, validation, and report generation for various design problems.
This book could also be very helpful for the students of senior design
(Mechanical System Design) and FEA for Large Deformation Problems.
xii • Preface
In addition, it could be used for computer sessions of short courses on

stress analysis techniques and FEA offered by Mechanical Engineering
departments.
After giving a brief introduction to the finite element analysis and
modeling, various guided examples have been included in this book.
Several new tutorials have been developed and others adapted from

different sources including ANSYS manuals in 2-volumes, ANSYS
workshops, and Internet resources. Tutorials have been arranged in each
volume according to the flow of the course and covered topics, such as
solid modeling using 2D and 3D primitives available in ANSYS, dynamic
analysis (harmonic and modal analysis), and thermal analysis.
Chapter 1
Dynamic Analysis
Structural analysis is mainly concerned with finding out the behavior of a

physical structure when subjected to force. This action can be in the form
of load due to the weight of things such as people, furniture, wind, snow,
and so on, or some other kind of excitation such as an earthquake, shaking
of the ground due to a blast nearby, and so on. In essence all these loads are
dynamic, including the self-weight of the structure because at some point in
time these loads were not there. The distinction is made between the dynamic
and the static analysis on the basis of whether the applied action has enough
acceleration in comparison to the structure’s natural frequency. If a load is
applied sufficiently slowly, the inertia forces (Newton’s first law of motion)
can be ignored and the analysis can be simplified as static analysis. Structural
dynamics, therefore, is a type of structural analysis that covers the behavior
of structures subjected to dynamic (actions having high acceleration) load-
ing. Dynamic loads include people, wind, waves, traffic, earthquakes, and
blasts. Any structure can be subjected to dynamic loading. Dynamic analysis
can be used to find dynamic displacements, time history, and modal analysis.
A dynamic analysis is also related to the inertia forces developed by a
structure when it is excited by means of dynamic loads applied suddenly
(e.g., wind blasts, explosion, and earthquake).
Dynamic analysis for simple structures can be carried out manually,
but for complex structures finite element analysis can be used to calculate
the mode shapes and frequencies.
1.1 Tutorial 1: Harmonic Analysis

of Structure
In this tutorial, the harmonic analysis of a cantilever beam will be

addressed. Harmonic analysis is used to determine the response of a
2 • Using ANSYS for Finite Element Analysis
structure to harmonically time-varying loads. This tutorial was created

using ANSYS 7.0. The purpose of this tutorial is to explain the steps
required to perform harmonic analysis on the cantilever beam shown as
follows.
0.01 m
0.01 m
1.0 m
Modulus of elasticity (E) = 206800(106) N/m2
Density = 7830 kg/m3
We will now conduct a harmonic forced response test by applying a

cyclic load (harmonic) at the end of the beam. The frequency of the load
will be varied from 1– 100 Hz. The following figure depicts the beam
with the application of the load.
Cyclic load
Magnitude: 100 N
Frequency range: 1–100 Hz.
ANSYS provides three methods for conducting a harmonic analy-

sis. These three methods are the Full, Reduced, and Modal Superposition
methods. This example demonstrates the Full method because it is simple
and easy to use as compared to the other two methods. However, this
method makes use of the full stiffness and mass matrices and thus is the
slower and costlier option.
1.1.1 Step-by-Step ANSYS Solution
1.1.1.1 Preprocessing: Defining the Problem
Student should be able to make simple cantilever model himself or other-

wise use the following command list:
Dynamic Analysis • 3
1.1.1.2 The Command Log File
/TITLE, Dynamic Analysis

/FILNAME,Dynamic,0 ! This sets the jobname to “Dynamic”
/PREP7
K,1,0,0
K,2,1,0
L,1,2
ET,1,BEAM3
R,1,0.0001,8.33e-10,0.01
MP,EX,1,2.068e11
MP,PRXY,1,0.33
MP,DENS,1,7830
LESIZE,ALL,,,10
LMESH,1
FINISH
1.1.1.3 Solution: Assigning Loads and Solving
1. Define analysis type (Harmonic)

Solution > Analysis Type > New Analysis >
Harmonic ANTYPE,3
2. Set options for analysis type
Select Solution > Analysis Type > Analysis Options.
The following window will appear.
As shown, select the Full Solution method, the Real + imag-
inary DOF (degrees of freedom) printout format and do not use
lumped mass approx. Click “OK”
The following window will appear. Use the default settings

(shown as follows).
3. Apply constraints
Select Solution > Define Loads > Apply >
Structural > Displacement > On Nodes
The following window will appear once you select the node
at x=0 (Note small changes in the window compared to the static
examples):
Constrain all DOF as shown in the above window.

4. Apply Loads
Select Solution > Define Loads > Apply > Structural >
Force/Moment > On Nodes
Select the node at x=1 (far right). The following window will
appear. Fill it in as shown to apply a load with a real value of 100
and an imaginary value of 0 in the positive “y” direction.
Note: By specifying a real and imaginary value of the load we are
providing information on magnitude and phase of the load. In this
case the magnitude of the load is 100 N and its phase is 0. Phase
information is important when you have two or more cyclic loads
being applied to the structure as these loads could be in or out

of phase. For harmonic analysis, all loads applied to a structure
must have the SAME FREQUENCY.
5. Set the frequency range

Select Solution > Load Step Opts > Time/Frequency
> Freq and Substps...
As shown in the f o l l o w i n g window, specify a frequency
range of 0–100 Hz, 100 substeps, and stepped b.c. By doing this
we will be subjecting the beam to loads at 1 Hz, 2 Hz, 3 Hz, .....
100 Hz. We will specify a stepped boundary condition (KBC) as
this will ensure that the same amplitude (100 N) will be applied
for each of the frequencies. The ramped option, on the other hand,
would ramp up the amplitude where at 1 Hz the amplitude would
be 1 N and at 100 Hz the amplitude would be 100 N.
You should now have the following in the ANSYS Graphics

window
6. Solve the system

Solution > Solve > Current LS
SOLVE
1.1.1.4 Postprocessing: Viewing the Results
We want to observe the response at x=1 (where the load was applied)
as a function of frequency. We cannot do this with General PostProcess-
ing (POST 1); rather we must use TimeHist Post-Processing (POST26).
POST26 is used to observe certain variables as a function of either time
or frequency.
1. Open the TimeHist Processing (POST26) Menu

Select TimeHist Postpro from the ANSYS Main Menu.
2. Define variables
Here we have to define variables that we want to see plotted.
By default, Variable 1 is assigned either Time or Frequency. In our
case it is assigned Frequency. We want to see the displacement UY
at the node at x=1, which is node #2. (To get a list of nodes and
their attributes, select Utility Menu > List > nodes.)
Select Time Hist Postpro > Variable Viewer...
And the following window should pop up:
Select Add (the green “+” sign in the upper left corner) from this
window and the following window should appear:
We are interested in the Nodal Solution > DOF Solution >

Y-Component of displacement. Click OK.
Graphically select node 2 when prompted and click OK. The
“Time History Variables” window should now look as follows:
3. List Stored Variables

In the “Time History Variables” window click the “List” button,
three buttons to the left of “Add.” The following window will
appear listing the data:
4. Plot UY vs. frequency

In the “Time History Variables” window click the “Plot” button,
two buttons to the left of “Add.” The following graph should be
plotted in the main ANSYS window.
Note that we get peaks at frequencies of approximately 8.3

and 51 Hz. This corresponds with the predicted frequencies of
8.311 and 51.94 Hz.
To get a better view of the response, view the log scale of UY.
Select Utility Menu > PlotCtrls > Style >Graphs
> Modify Axis
The following window will appear:
As marked by an “A” in the above window, change the Y-axis

scale to “Logarithmic.”
Select Utility Menu > Plot > Replot
You should now see the following:
This is the response at node 2 for the cyclic load applied at this
node from 0–100 Hz.
For ANSYS version lower than 7.0, the “Variable Viewer”

window is not available. Use the “Define Variables” and “Store
Data” functions under TimeHist Postpro. See the help file for
instructions.
1.2 Tutorial 2: Modal Analysis of

Structure
In this tutorial, the modal analysis of a cantilever beam will be addressed.

Modal analysis—used to calculate the natural frequencies and mode
shapes of a structure. Different mode extraction methods are available.
This tutorial was created using ANSYS 7.0. The purpose of this tutorial
is to outline the steps required to do a simple modal analysis of the
cantilever beam shown as follows.
0.01 m
0.01 m
1.0 m
Modulus of elasticity (E) = 206800(106) N/m2
Density = 7830 kg/m3
1.2.1 Step-by-Step ANSYS Solution
1.2.1.1 Preprocessing: Defining the Problem
The simple cantilever beam is same as used in the previous tutorial.
1.2.1.2 Solution: Assigning Loads and Solving
1. Define analysis type

Solution >Analysis Type > NewAnalysis > Modal ANTYPE,2
2. Set options for analysis type:
Select Solution > Analysis Type > Analysis Options..
As shown, select the Subspace method and enter 5 in the “No.

of modes to extract.” Check the box beside “Expand mode shapes”
and enter 5 in the “No. of modes to expand.” Click “OK.”
Note that the default mode extraction method chosen is the
Reduced Method. This is the fastest method as it reduces the system
matrices to only consider the Master Degrees of Freedom (MDOFs)
(see the following figure). The Subspace Method extracts modes
for all DOFs. It is therefore more exact but it also takes longer to
compute (especially when the geometry is complex).
The following window will then appear:
For a better understanding of these options see the Commands

manual.
For this problem, we will use the default options so click on OK.
1. Apply constraints
Solution > Define Loads > Apply > Structural >
Displacement > On Keypoints Fix Keypoint 1
(i.e., all DOFs constrained).
2. Solve the system
Solution > Solve > Current LS
SOLVE
1.2.1.3 Postprocessing: Viewing the Results
1. Verify extracted modes against theoretical predictions

Select General Postproc > Results Summary...
The following table compares the mode frequencies (in Hz)

predicted by theory and ANSYS.
Mode Theory ANSYS Percent error

1 8.311 8.300 0.1
2 51.94 52.01 0.2
3 145.68 145.64 0.0
4 285.69 285.51 0.0
5 472.22 472.54 0.1
Note: To obtain accurate higher mode frequencies, this mesh
would have to be refined even more (i.e., instead of ten ele-
ments, we would have to model the cantilever using 15 or more
elements depending on the highest mode frequency of interest).
2. View mode shapes
Select General Postproc > Read Results > First Set
This selects the results for the first mode shape.
Select General Postproc > Plot Results > Deformed shape.
Select “Def + undef edge”
The first mode shape will now appear in the graphics window.
To view the next mode shape, select General Postproc > Read
Results > Next Set As above choose General Postproc > Plot
Results > Deformed shape. Select “Def + undef edge.”
The first four mode shapes should look like the following:
3. Animate mode shapes

Select Utility Menu (Menu at the top) > Plot Ctrls
> Animate > Mode Shape
Keep the default setting and click “OK.”

The animated mode shapes for Mode 3 is shown as follows. Mode 3
1.2.1.4 Using the Reduced Method for Modal Analysis
This method employs the use of MDOFs. These are DOFs that govern
the dynamic characteristics of a structure. For example, the MDOFs for
the bending modes of cantilever beam are:
Master degrees of freedom (MDOFs)
For this option, a detailed understanding of the dynamic behavior

of a structure is required. However, going this route means a smaller
(reduced) stiffness matrix, and thus faster calculations.
The steps for using this option are quite simple.
Instead of specifying the Subspace method, select the Reduced
method and specify 5 modes for extraction.
Complete the window as shown in the following:
Note: For this example both the number of modes and frequency range
was specified. ANSYS then extracts the minimum number of modes
between the two.
Select Solution > Master DOF > User Selected > Define
When prompted, select all nodes except the left-most node (fixed).
Select UY as the 1st degree of freedom (shown above). The same

constraints are used as above.
The following table compares the mode frequencies ( in Hz) pre-
dicted by theory and ANSYS (Reduced).
Mode Theory ANSYS Percent error

1 8.311 8.300 0.1
2 51.94 52.01 0.1
3 145.68 145.66 0.0
4 285.69 285.71 0.0
5 472.22 473.66 0.3
As you can see, the error does not change significantly. However,
for more complex structures, larger errors would be expected using the
reduced method.
Chapter 2
Composite Materials
2.1 Composites—A Basic Introduction
Composite materials have been used in structures for a long time. In recent
times, composite parts have been used extensively in aircraft structures,
automobiles, sporting goods, and many consumer products. Composite
materials are those containing more than one bonded material, each with
different structural properties. The main advantage of composite materials
is the potential for a high ratio of stiffness to weight. Composites used for
typical engineering applications are advanced fiber or laminated compos-
ites, such as fiberglass, glass epoxy, graphite epoxy, and boron epoxy. To
fully appreciate the role and application of composite materials to a struc-
ture, an understanding is required of the component materials themselves
and of the ways in which they can be processed; this article therefore looks
at basic composite theory and properties of materials used.
In its most basic form a composite material is one that is composed of at
least two elements working together to produce material properties that are
different to the properties of those elements on their own. In practice, most
composites consist of a bulk material (the “matrix”), and a reinforcement
of some kind, added primarily to increase the strength and stiffness of the
matrix. This reinforcement is usually in fiber form. Today, the most com-
mon man-made composites can be divided into three main groups:
• Polymer matrix composites (PMCs)—These are the most common

and will be discussed here. Also known as FRP—fiber-reinforced
polymers (or plastics)—these materials use a polymer-based resin
as the matrix, and a variety of fibers such as glass, carbon, and
aramid as the reinforcement.
• Metal matrix composites (MMCs)—Increasingly found in the
automotive industry, these materials use a metal such as aluminum
as the matrix, and reinforce it with fibers such as silicon carbide.
• Ceramic matrix composites (CMCs)—Used in very high tempera-

ture environments, these materials use a ceramic as the matrix and
reinforce it with short fibers, or whiskers such as those made from
silicon carbide and boron nitride.
2.1.1 Polymer Matrix Composites
Resin systems such as epoxies and polyesters have limited use for the
manufacture of structures on their own, since their mechanical properties
are not very high when compared to, for example, most metals. How-
ever, they have desirable properties, most notably their ability to be easily
formed into complex shapes. Materials such as glass, aramid, and boron
have extremely high tensile and compressive strength but in “solid form”
these properties are not readily apparent. This is due to the fact that when
stressed, random surface flaws will cause each material to crack and fail
well below its theoretical “breaking point.” To overcome this problem,
the material is produced in fiber form, so that, although the same num-
ber of random flaws will occur, they will be restricted to a small number
of fibers with the remainder exhibiting the material’s theoretical strength.
Therefore a bundle of fibers will reflect more accurately the optimum per-
formance of the material. However, fibers alone can only exhibit tensile
properties along the fiber’s length, in the same way as fibers in a rope.
It is when the resin systems are combined with reinforcing fibers such
as glass, carbon, and aramid that exceptional properties can be obtained.
The resin matrix spreads the load applied to the composite between each
of the individual fibers and also protects the fibers from damage caused by
abrasion and impact. High strengths and stiffnesses, ease of molding com-
plex shapes, high environmental resistance all coupled with low densities,
make the resultant composite superior to metals for many applications.
Since PMCs combine a resin system and reinforcing fibers, the properties
of the resulting composite material will combine something of the proper-
ties of the resin on its own with that of the fibers on their own, as surmised
in Figure 2.1.
Overall, the properties of the composite are determined by:
• The properties of the fiber.

• The properties of the resin.
• The ratio of fiber to resin in the composite (fiber volume fraction
(FVF)).
• The geometry and orientation of the fibers in the composite.
Composite Materials • 21
Fiber
Tensile stress
FRP composite
Resin
Strain
Figure 2.1. Illustrating the combined effect on
modulus of the addition of fibers to a resin matrix.
The ratio of the fiber to resin derives largely from the manufacturing
process used to combine resin with fiber. However, it is also influenced by
the type of resin system used, and the form in which the fibers are incor-
porated. In general, since the mechanical properties of fibers are much
higher than those of resins, the higher the fiber volume fraction the higher
will be the mechanical properties of the resultant composite. In practice
there are limits to this, since the fibers need to be fully coated in resin to
be effective, and there will be an optimum packing of the generally cir-
cular cross-section fibers. In addition, the manufacturing process used to
combine fiber with resin leads to varying amounts of imperfections and
air inclusions. Typically, with a common hand lay-up process as widely
used in the boat-building industry, a limit for FVF is approximately 30–40
percent. With the higher quality, more sophisticated and precise processes
used in the aerospace industry, FVFs approaching 70 percent can be suc-
cessfully obtained.
The geometry of the fibers in a composite is also important since fibers
have their highest mechanical properties along their lengths, rather than
across their widths. This leads to the highly anisotropic properties of com-
posites, where, unlike metals, the mechanical properties of the compos-
ite are likely to be very different when tested in different directions. This
means that it is very important when considering the use of composites to
understand at the design stage, both the magnitude and the direction of the
applied loads. When correctly accounted for, these anisotropic properties
can be very advantageous since it is only necessary to put material where
loads will be applied, and thus redundant material is avoided.
It is also important to note that with metals the material supplier
largely determines the properties of the materials, and the person who
fabricates the materials into a finished structure can do almost nothing

to change those “in-built” properties. However, a composite material is
formed at the same time, as the structure is itself being fabricated. This
means that the person who is making the structure is creating the prop-
erties of the resultant composite material, and so the manufacturing pro-
cesses they use have an unusually critical part to play in determining the
performance of the resultant structure.
2.1.1.1 Loading
There are four main direct loads that any material in a structure has to
withstand: tension, compression, shear, and flexure.
Tension: Figure 2.2 shows a tensile load applied to a composite. The

response of a composite to tensile loads is very dependent on the
tensile stiffness and strength properties of the reinforcement fibers,
since these are far higher than the resin system on its own.
Compression: Figure 2.3 shows a composite under a compressive
load. Here, the adhesive and stiffness properties of the resin system
are crucial, as it is the role of the resin to maintain the fibers as
straight columns and to prevent them from buckling.
Shear: Figure 2.4 shows a composite experiencing a shear load. This
load is trying to slide adjacent layers of fibers over each other.
Under shear loads the resin plays the major role, transferring
the stresses across the composite. For the composite to perform
well under shear loads the resin element must not only exhibit
good mechanical properties but must also have high adhesion to
the reinforcement fiber. The interlaminar shear strength (ILSS) of
Figure 2.2. Illustrates the tensile load applied to a composite body.
Figure 2.3. Illustrates the compression load applied to a composite body.

Figure 2.4. Illustrates the shear load applied to a composite body.
Figure 2.5. Illustrates the loading due to flexure on a composite body.
a composite is often used to indicate this property in a multi-layer

composite (“laminate”).
Flexure: Flexural loads are really a combination of tensile, com-
pression, and shear loads. When loaded as shown in Figure 2 . 5 ,
the upper face is put into compression, the lower face into tension,
and the central portion of the laminate experiences shear.
2.1.1.2 Comparison With Other Structural Materials
Owing to the factors described previously, there is a very large range

of mechanical properties that can be achieved with composite materi-
als. Even when considering one fiber type on its own, the composite
properties can vary by a factor of 10 with the range of fiber contents
and orientations that are commonly achieved. The comparisons that fol-
low therefore show a range of mechanical properties for the composite
materials. The lowest properties for each material are associated with
simple manufacturing processes and material forms (e.g., spray lay-up
glass fiber), and the higher properties are associated with higher technol-
ogy manufacture (e.g., autoclave molding of unidirectional glass fiber
prepreg), such as would be found in the aerospace industry. For the other
materials shown, a range of strength and stiffness (modulus) figures are
also given to indicate the spread of properties associated with different

alloys, for example.
Figures 2.6 and 2.7 clearly show the range of properties that different
composite materials can display. These properties can best be summed up
as high strengths and stiffnesses combined with low densities. It is these
properties that give rise to the characteristic high strength and stiffness to
weight ratios that make composite structures ideal for so many applica-
tions. This is particularly true of applications, which involve movement,
such as cars, trains, and aircraft, since lighter structures in such applica-
tions play a significant part in making these applications more efficient.
The strength and stiffness to weight ratio of composite materials can best
be illustrated by the following graphs that plot “specific” properties. These
are simply the result of dividing the mechanical properties of a material
by its density. Generally, the properties at the higher end of the ranges
illustrated in the previous graphs (Figures 2.6 and 2.7) are produced from
the highest density variant of the material. The spread of specific proper-
ties shown in the following graphs (Figures 2.8 and 2.9) takes this into
account.
2,800
2,400
2,000
Tensile modulus (MPa)
1,600
1,200
800
400
0
Aramid composites
Woods
Al.Alloys
Titanium
Steels
E-Glass composites
S-Glass composites
HS carbon composites
IM carbon composites
Figure 2.6. Tensile strength of common structural materials.

Specific tensile strength Tensile modulus (GPa)
0
30
60
90
120
150
180
210
0
200
400
600
800
1,000
1,200
1,400
1,600
1,800
2,000
Woods Woods
Al.Alloys Al.Alloys
Titanium Titanium
Steels Steels
E-Glass composites E-Glass composites
S-Glass composites S-Glass composites
Aramid composites Aramid composites
HS carbon composites Figure 2.7. Tensile modulus of common structural materials. HS carbon composites
Figure 2.8. Specific tensile strength of common structural materials.

120
110
100
90
Specific tensile modulus
80
70
60
50
40
30
20
10
0
Aramid composites
Al.Alloys
Titanium
Steels
E-Glass composites
S-Glass composites
HS carbon composites
Woods
Figure 2.9. Specific tensile modulus of common structural materials.
2.2 Modeling Composites Using ANSYS
ANSYS allows you to model composite materials with specialized ele-

ments called layered elements. Once you build your model using these
elements, you can do any structural analysis (including nonlinearities such
as large deflection and stress stiffening).
Composites are somewhat more difficult to model than an isotropic
material such as iron or steel. You need to take special care in defining the
properties and orientations of the various layers since each layer may have
different orthotropic material properties. In this section, we will concen-
trate on the following aspects of building a composite model:
• Choosing the proper element type.

• Defining the layered configuration.
• Specifying failure criteria.
• Following modeling and postprocessing guidelines.
2.2.1 Choosing the Proper Element Type
The following element types are available to model layered compos-

ite materials: SHELL99, SHELL91, SHELL181, SOLID46, and
SOLID191. Which element you choose depends on the application, the

type of results that need to be calculated, and so on. Check the individ-
ual element descriptions to determine if a specific element can be used
in your ANSYS product. All layered elements allow failure criterion
calculations.
2.2.1.1 SHELL99—Linear Layered Structural Shell Element
SHELL99 is an 8-node, 3-D shell element with six degrees of freedom

at each node. It is designed to model thin to moderately thick plate and
shell structures with a side-to-thickness ratio of roughly 10 or greater.
For structures with smaller ratios, you may consider using SOLID46.
The SHELL99 element allows a total of 250 uniform-thickness layers.
Alternately, the element allows 125 layers with thicknesses that may vary
bilinearly over the area of the layer. If more than 250 layers are required,
you can input your own material matrix. It also has an option to offset the
nodes to the top or bottom surface.
2.2.1.2 SHELL91—Nonlinear Layered Structural Shell Element
SHELL91 is similar to SHELL99 except that it allows only up to 100

layers and does not allow you to input a material property matrix. How-
ever, SHELL91 supports plasticity, large-strain behavior, and a special
sandwich option, whereas SHELL99 does not. SHELL91 is also more
robust for large deflection behavior.
2.2.1.3 SHELL181—Finite Strain Shell
SHELL181 is a 4-node 3-D shell element with six degrees of freedom

at each node. The element has full nonlinear capabilities including large
strain and allows 255 layers. The layer information is input using the
section commands rather than real constants. Failure criteria is available
using the FC commands.
2.2.1.4 SOLID46—3-D Layered Structural Solid Element
SOLID46 is a layered version of the 8-node, 3-D solid element,

SOLID45, with three degrees of freedom per node (UX, UY, UZ). It is
designed to model thick-layered shells or layered solids and allows up to
250 uniform-thickness layers per element. Alternately, the element allows

125 layers with thicknesses that may vary bilinearly over the area of the
layer. An advantage with this element type is that you can stack several
elements to model more than 250 layers to allow through-the-thickness
deformation slope discontinuities. The user-input constitutive matrix
option is also available. SOLID46 adjusts the material properties in the
transverse direction permitting constant stresses in the transverse direc-
tion. In comparison to the 8-node shells, SOLID46 is a lower order ele-
ment and finer meshes may be required for shell applications to provide
the same accuracy as SHELL91 or SHELL99.
2.2.1.5 SOLID191—Layered Structural Solid Element
SOLID191 is a layered version of the 20-node 3-D solid element

SOLID95, with three degrees of freedom per node (UX, UY, UZ). It is
designed to model thick-layered shells or layered solids and allows up to
100 layers per element. As with SOLID46, SOLID191 can be stacked to
model through-the-thickness discontinuities. SOLID191 has an option to
adjust the material properties in the transverse direction permitting con-
stant stresses in the transverse direction. In spite of its name, the element
does not support nonlinear materials or large deflections.
In addition to the layered elements mentioned earlier, other composite
element capabilities exist in ANSYS, but will not be considered further in
the chapter:
• SOLID95, the 20-node structural solid element, with KEYOPT

(1) = 1 functions similarly to a single-layered SOLID191 including
the use of an orientation angle and failure criterion. It allows non-
linear materials and large deflections.
• SHELL63, the 4-node shell element, can be used for rough,
approximate studies of sandwich shell models. A typical applica-
tion would be a polymer between two metal plates, where the bend-
ing stiffness of the polymer would be small relative to the bending
stiffness of the metal plates. The bending stiffness can be adjusted
by the real constant RMI to represent the bending stiffness due to
the metal plates, and distances from the middle surface to extreme
fibers (real constants CTOP, CBOT) can be used to obtain output
stress estimates on the outer surfaces of the sandwich shell. It is not
used as frequently as SHELL91, SHELL99, or SHELL181, and
will not be considered for anything other than sandwich structures
in this section.
• SOLID65, the 3-D reinforced concrete solid element, models

an isotropic medium with optional reinforcing in three different
user-defined orientations.
• BEAM188 and BEAM189, the 3-D finite strain beam elements,
can have their sections built up with multiple materials.
2.2.2 Defining the Layered Configuration
The most important characteristic of a composite material is its layered

configuration. Each layer may be made of a different orthotropic material
and may have its principal directions oriented differently. For laminated
composites, the fiber directions determine layer orientation. Two methods
are available to define the layered configuration:
• By specifying individual layer properties.

• By defining constitutive matrices that relate generalized forces and
moments to generalized strains and curvatures (available only for
SOLID46 and SHELL99).
2.2.2.1 Specifying Individual Layer Properties
With this method, the layer configuration is defined layer-by-layer from

bottom to top. The bottom layer is designated as layer 1, and additional
layers are stacked from bottom to top in the positive Z (normal) direction
of the element coordinate system. You need to define only half of the lay-
ers if stacking symmetry exists.
At times, a physical layer will extend over only part of the model. In
order to model continuous layers, these dropped layers may be modeled
with zero thickness. Figure 2.10 shows a model with four layers, the sec-
ond of which is dropped over part of the model.
For each layer, the following properties are specified in the element
real constant table
[R, RMORE, RMODIF]
Main Menu> Preprocessor> Real Constants
Accessed with REAL attributes.
Layer 2
4
3 4 is dropped
2 3
1 1
Figure 2.10. Layered model showing dropped layer.

• Material properties (via a material reference number MAT).

• Layer orientation angle commands (THETA).
• Layer thickness (TK).
Layered sections may also be defined through the Section Tool.

Prep>Sections>Shell—Add/edit
For each layer, the following are specified in the section definition
through the section commands; or through the Section Tool (SECTYPE,
SECDATA) (accessed with the SECNUM attributes).
• Material properties (via a material reference number MAT)

• Layer orientation angle commands (THETA)
• Layer thickness (TK)
• Number of integration points per layer (NUMPT).
Material Properties as with any other element, the MP command:

Main Menu> Preprocessor> Material Props>
Material Models> Structural> Linear> Elastic>
Isotropic or Orthotropic
Is used to define the linear material properties, and the TB command is
used to define the nonlinear material data tables (plasticity is only available
for elements SOLID191 and SHELL91). The only difference is that the
material attribute number for each layer of an element is specified in the
element’s real constant table. For the layered elements, the MAT command:
Main Menu> Preprocessor> Meshing> Mesh Attributes>
Default Attribs
Attribute is only used for the DAMP and REFT arguments of the MP
command. The linear material properties for each layer may be either isotro-
pic or orthotropic (see Linear Material Properties in the ANSYS Elements
Reference). Typical fiber-reinforced composites contain orthotropic materi-
als and these properties are most often supplied in the major Poisson’s ratio
form (see the ANSYS, Inc. Theory Reference). Material property directions
are parallel to the layer coordinate system, which is defined by the element
coordinate system and the layer orientation angle (described as follows).
Layer orientation angle—This defines the orientation of the layer
coordinate system with respect to the element coordinate system. It is the
angle (in degrees) between X-axes of the two systems. By default, the
layer coordinate system is parallel to the element coordinate system. All
elements have a default coordinate system that you can change using the
ESYS element attribute [ESYS]
Main Menu> Preprocessor> Meshing> Mesh Attributes>
Default Attribs
You may also write your own subroutines to define the element and
layer coordinate systems (USERAN and USANLY); see the Guide to
ANSYS User Programmable Features for details.
Layer thickness—If the layer thickness is constant, you only need
to specify TK(I), the thickness at node I. Otherwise, the thicknesses at the
four corner nodes must be input. Dropped layers may be represented with
zero thickness.
Number of integration points per layer—This allows you to deter-
mine in how much detail the program should compute the results. For very
thin layers, when used with many other layers, one point would be appropri-
ate. But for laminates with few layers, more would be needed. The default
is three points. This feature applies only to sections defined through the sec-
tion commands. Currently, the graphical user interface (GUI) only allows
layer real constant input of up to 100 layers. If more layers are needed for
SHELL99 or SOLID46, the R and RMORE commands must be used.
2.2.2.2 Defining the Constitutive Matrices
This is an alternative to specifying the individual layer properties and is avail-

able as an option KEYOPT (2) for SOLID46 and SHELL99. The matrices,
which represent the force–moment and strain–curvature relationships for the
element, must be calculated outside the ANSYS program as outlined in the
ANSYS, Inc. Theory Reference. They can be included as part of the solution
printout with KEYOPT(10). The main advantages of the matrix approach are:
• It allows you to incorporate an aggregate composite material

behavior.
• A thermal load vector may be supplied.
• The matrices may represent an unlimited number of layers.
The terms of the matrices are defined as real constants. Mass

effects are incorporated by specifying an average density (real constants
AVDENS) for the element. If the matrix approach is used, detailed results
in each layer cannot be obtained since individual layer information is not
input.
2.2.2.3 Sandwich and Multiple-Layered Structures
Sandwich structures have two thin faceplates and a thick, but relatively
weak, core. Figure 2.11 illustrates sandwich construction.
Faceplate
Core (at least 1/2 of local thickness)

Faceplate
Figure 2.11. Sandwich construction.
You can model sandwich structures with SHELL63, SHELL91, or

SHELL181. SHELL63 has one layer but permits sandwich modeling
through the use of real constants. You can modify the effective bending
moment of inertia and the distance from the middle surface to the extreme
fibers to account for the weak core. KEYOPT(9) = 1 of SHELL91 speci-
fies the sandwich option. The core is assumed to carry all of the transverse
shear; the faceplates carry none. Conversely, the faceplates are assumed
to carry all (or almost all) of the bending load. Only SHELL91 has this
sandwich option.
SHELL181 models the transverse shear deflection using as
energy equivalence method that makes the need for a sandwich option
unnecessary.
2.2.2.4 Node Offset
For SHELL181 using sections defined through the section commands,

nodes can be offset during the definition of the section, using the
SECOFFSET command. For SHELL91, and SHELL99 the node off-
set option (KEYOPT(11)) locates the element nodes at the bottom, mid-
dle, or top surface of the shell. The following figures illustrate how you
can conveniently model ply drop off in shell elements that are adjacent
to each other. In Figure 2.12, the nodes are located at the middle surfaces
(KEYOPT(11) = 0) and these surfaces are aligned. In Figure 2.13, the
Nodes located at the midplane with KEYOPT(11) = 0
Figure 2.12. Layered shell with nodes at midplane.

Ply 3
Ply 2
Ply 1
Nodes located on bottom surface with KEYOPT(11) = 1
Figure 2.13. Layered shell with nodes at bottom surface.
nodes are located at the bottom surfaces (KEYOPT(11) = 1) and these

surfaces are aligned.
2.2.3 Specifying Failure Criteria
Failure criteria are used to learn if a layer has failed due to the applied loads.
You can choose from three predefined failure criteria or specify up to six fail-
ure criteria of your own (user-written criteria). The three predefined criteria
are:
• Maximum Strain Failure Criterion, which allows nine failure

strains.
• Maximum Stress Failure Criterion, which allows nine failure
stresses.
• Tsai-Wu Failure Criterion, which allows nine failure stresses and
three additional coupling coefficients. You have a choice of two
methods of calculating this criterion. The methods are defined in
the ANSYS, Inc. Theory Reference.
The failure strains, stresses, and coupling coefficients may be tem-

perature-dependent. See the ANSYS Elements Reference for details
about the data required for each criterion. To specify a failure criterion,
use either the family of TB commands or the family of FC commands. The
TB commands are TB, TBTEMP, and TBDATA.
Main Menu> Preprocessor> Material Props> Failure Criteria
A typical sequence of commands to specify a failure criterion using
these commands is shown as follows.
TB,FAIL,1,2 ! Data table for failure criterion, material 1,

! no. of temperatures = 2
TBTEMP,,CRIT ! Failure criterion key

TBDATA,2,1 ! Maximum Stress Failure Criterion (Const. 2 =
1) TBTEMP,100! Temperature for subsequent
failure properties TBDATA,10,1500,,400,,10000
! X, Y, and Z failure tensile stresses (Z value!
set to a large number)
TBDATA,16,200,10000,10000 ! XY, YZ, and XZ failure shear
stresses
TBLIST
TBTEMP,200! Second temperature
TBDATA,...
The FC → commands → are FC, FCDELE, and FCLIST commands:

Material Models> Structural> Nonlinear> Inelastic>
Non-Metal Plasticity> Failure Criteria
Main Menu> General Postproc> Failure Criteria
A typical sequence of commands to specify a failure criterion using
these commands is shown as follows.
FC,1,TEMP,, 100, 200 ! Temperatures

FC,1,S,XTEN, 1500, 1200 ! Maximum stress components
FC,1,S,YTEN, 400, 500
FC,1,S,ZTEN,10000, 8000
FC,1,S,XY , 200, 200
FC,1,S,YZ ,10000, 8000
FC,1,S,XZ ,10000, 8000
FCLIST, ,100 ! List status of Failure Criteria at 100.0 degrees
PRNSOL,S,FAIL ! Use Failure Criteria
The TB commands (TB, TBTEMP, and TBDATA) can be used only

for SHELL91, SHELL99, SOLID46, or SOLID191, but the FC and
FCLIST commands can be used for any 2-D or 3-D structural solid element
or any 3-D structural shell element. See the ANSYS Commands Reference
for a discussion of the TB, TBTEMP, TBDATA, TBLIST, FC, FCDELE,
and FCLIST commands. Some notes about specifying failure criteria:
• The criteria are orthotropic, so you must input the failure stress or
failure strain values for all directions. (The exception is that com-
pressive values default to tensile values.)
• If you don’t want the failure stress or strain to be checked in a par-

ticular direction, specify a large number in that direction (as shown
in the previous example).
User-written failure criteria may be specified via user subroutines

USRFC1 through USRFC6. These subroutines should be linked with the
ANSYS program beforehand; see the ANSYS Advanced Analysis Tech-
niques Guide for a brief description of user-programmable features.
2.2.4 A
dditional Modeling and Postprocessing
Guidelines
Some additional guidelines for modeling and postprocessing of composite

elements are presented in the following list.
1. Composites exhibit several types of coupling effects, such as cou-

pling between bending and twisting, coupling between extension
and bending, and so on. This is due to stacking of layers of differing
material properties. As a result, if the layer stacking sequence is not
symmetric, you may not be able to use model symmetry even if the
geometry and loading are symmetric, because the displacements
and stresses may not be symmetric.
2. Interlaminar shear stresses are usually important at the free edges of
a model. For relatively accurate interlaminar shear stresses at these
locations, the element size at the boundaries of the model should
be approximately equal to the total laminate thickness. For shells,
increasing the number of layers per actual material layer does not
necessarily improve the accuracy of interlaminar shear stresses.
With elements SOLID46, SOLID95, and SOLID191, however,
stacking elements in the thickness direction should result in more
accurate interlaminar stresses through the thickness. Interlaminar
transverse shear stresses in shell elements are based on the assump-
tion that no shear is carried at the top and bottom surfaces of the
element. These interlaminar shear stresses are only computed in the
interior and are not valid along the shell element boundaries. Use of
shell-to-solid submodeling is recommended to accurately compute
all of the free-edge interlaminar stresses.
3. Since a large amount of input data is required for composites, you
should verify the data before proceeding with the solution. Several
commands are available for this purpose:
• ELIST (Utility Menu> List> Elements) lists the nodes and
attributes of all selected elements.
• EPLOT (Utility Menu> Plot> Elements) displays all selected

elements. Using the /ESHAPE,1 command (Utility Menu>
PlotCtrls> Style> Size and Shape) before EPLOT causes
shell elements to be displayed as solids with the layer thick-
nesses obtained from real constants or section definition (see
Figure 2.14. This example uses element SHELL99 with /ESH-
APE turned on). It also causes SOLID46 elements to be dis-
played with layers.
• /PSYMB, LAYR,n (Utility Menu> PlotCrls> Symbols) fol-
lowed by EPLOT displays layer number n for all selected
layered elements. This can be used to display and verify each
individual layer across the entire model.
• /PSYMB, ESYS,1 followed by EPLOT displays the element
coordinate system triad for those elements whose default coor-
dinate system has been changed.
• LAYLIST (Utility Menu> List> Elements> Layered Ele-
ments) lists the layer stacking sequence from real constants
and any two material properties for SHELL99, SHELL91,
SOLID46, and SOLID191 elements. You can specify a range
of layer numbers for the listing.
LIST LAYERS 1 TO 4 IN REAL SET 1 FOR ELEMENT
TYPE 1
TOTAL LAYERS = 4 LSYM = 1 LP1 = 0 LP2 = 0 EFS =
.000E+00
NO. ANGLE THICKNESS MAT
Figure 2.14. Example of an element display.

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

1 45.0 0.250 1
2 -45.0 0.250 2
3 -45.0 0.250 2
4 45.0 0.250 1
------------------------
SUM OF THK 1.00
• LAYPLOT (Utility Menu> Plot> Layered Elements) dis-
plays the layer stacking sequence from real constants in the
form of a sheared deck of cards (see Figure 2.15). The layers
are crosshatched and color coded for clarity. The hatch lines
indicate the layer angle (real constant THETA) and the color
indicates layer material number (MAT). You can specify a range
of layer numbers for the display.
• SECPLOT (Main Menu> Preprocessor> Sections> Shell>
Plot Section) displays the section stacking sequence from sec-
tions in the form of a sheared deck of cards (see Figure 2.15). The
sections are crosshatched and color coded for clarity. The hatch
lines indicate the layer angle (THETA) and the color indicates
layer material number (MAT) defined by the SECDATA com-
mand. You can specify a range of layer numbers for the display.
4. By default, only data for the bottom of the first (bottom) layer, top
of the last (top) layer, and the layer with the maximum failure crite-
rion value are written to the results file. If you are interested in data
for all layers, set KEYOPT(8) = 1. Be aware, though, that this may
result in a large results file.
Layer# Material#
1
1
Theta
2
2
45 2
3
–45
4 1
–45
45
Figure 2.15. Sample LAYPLOT display for [45/−45/−45/45] sequence.

5. Use the ESEL S,LAYER command to select elements that have

a certain layer number. If an element has a zero thickness for the
requested layer, the element is not selected. For energy output, the
results are applicable only to the entire element; you cannot get
output results for individual layers.
6. Use the LAYER command (Main Menu> General Postproc>
Options for Outp) in POST1 (or LAYERP26 (Main Menu>
TimeHist Postpro> Define Variables) in POST26) to specify the
layer number for which results are to be processed. The SHELL
command (Main Menu> General Postproc> Options for Outp or
Main Menu> TimeHist Prostpro> Define Variables) specifies a
TOP, MID, or BOT location within the layer. The default in POST1
is to store results for the bottom of the bottom layer, and the top
of the top layer, and the layer with the maximum failure criterion
value. In POST26, the default is layer 1. If KEYOPT(8) = 1 (i.e.,
data stored for all layers), the LAYER and LAYERP26 commands
store the TOP and BOT results for the specified layer number. MID
values are then calculated by average TOP and BOT values. If
KEYOPT (8) = 2 is set for SHELL181 during SOLUTION, then
LAYER and LAYERP26 commands store the TOP, BOTTOM,
and MID results for the specified layer number. In this case, MID
values are directly retrieved from the results file. For transverse
shear stresses with KEYOPT(8) = 0, therefore, POST1 can only
show a linear variation, whereas the element solution printout or
KEYOPT(8) = 2 can show a parabolic variation.
7. By default, POST1 displays all results in the global Cartesian coor-
dinate system. Use the RSYS command (Main Menu> General
Postproc> Options for Outp) to transform the results to a different
coordinate system. In particular, RSYS,SOLU allows you to dis-
play results in the layer coordinate system if LAYER is issued with
a nonzero layer number.
2.3 Tutorial 3: Simply Supported

Laminated Plate Under Pressure
In this tutorial, a simply supported square cross-ply laminated plate is

subjected to a uniform pressure po. The stacking sequence of the plies
is symmetric about the middle plane. Determine the center deflection δ
(Z-direction) of the plate due to the pressure load and the von Mises stress
distribution.
Y Z
Po
uz = 0; ux = 0;
0° t
90° t
h X
90°
0°
a X
uz = 0;
uy = 0
uz = 0;
uy = 0
uz = 0; ux = 0
a
Geometric
properties Material properties Loading
Ex = 25 × 106 N/m2
a = 10 m Ey = 1 × 106 N/m2
h = 0.1 m υxy = 0.25 (Major Poisson’s ratio) po = 1.0 N/m2
t = 0.025 m Gxy = G = 0.5 × 106 N/m2
Gyz = 0.2 × 106 N/m2
2.3.1 Approach and Assumptions
A quarter of the plate is modeled due to symmetry in geometry, material

orientation, loading, and boundary conditions. One model using Linear
Layered Structural Shell Elements is analyzed. 3-D 8-node Layered Struc-
tural Solid Elements SOLID46 elements can also be used.
Note that PRXY is used to directly input the major Poisson’s ratio. EZ
(explicitly input) is assumed to be equal to EY.
2.3.2 Summary of Steps
1. Set preferences.
2. Define element types and options.
3. Define real constants.
4. Define material properties.
5. Build geometry.
6. Generate mesh.
7. Verification of data.
8. Apply displacement constraints.
UZ = 0; UX = 0
Symmetric boundary
condition (ux=0)
a X
UZ = 0;
UY = 0
UZ = 0; Symmetric boundary
condition (uy=0)
UY = 0
UZ = 0; UX = 0
9. Apply pressure load.

10. Obtain solution.
11. Review results.
12. Exit the ANSYS program.
2.3.3 Step-By-Step Analysis
Utility Menu >File>Change title

Enter Simply Supported Laminated Plate (SHELL99) then OK.
1. Set preferences
In preparation for defining materials, you will set preferences so
that only materials and elements that pertain to a structural analysis
are available for you to choose. To set preferences:
Main Menu> Preferences
Turn on structural filtering.

OK to apply filtering and close the dialog box.
Before going to the next step, enter preprocessor by selecting

Preprocessor from the ANSYS Main Menu.
2. Define element types and options

In any analysis, you need to select from a library of element types
and define the appropriate ones for your analysis. For this analysis,
you will first use element type, SHELL99. SHELL99 may be used
for layered applications of a structural shell model. It allows up to
250 layers. If more than 250 layers are required, a user-input con-
stitutive matrix is available.
Main Menu> Preprocessor Element Type> Add/Edit/Delete
Add an element type.

Structural shell family of elements.

Choose the 8-noded layered shell element (SHELL99).
OK to apply the element type and close the dialog box.
Close the element type dialog box.
3. Define real constants

For this analysis, since the assumption is plane stress with thick-
ness, you will enter the thickness as a real constant for PLANE82.
To find out more information about PLANE82, you will use the
ANSYS Help System in this step by clicking on a Help button from
within a dialog box.
Main Menu> Preprocessor> Real Constants>
Add/Edit/Delete
Add a real constant set.

OK for SHELL99.
OK to real constant set number 1.

Enter 4 for number of layers and 1 for symmetric.
Enter material number, angle, and thickness for each layer. Because
of symmetry (0/90/90/0), enter data only for the first two layers.
Layer 1: 1, 0, 0.025
Layer 2: 1, 90, 0.025
OK to define the real constant and close the dialog box.
Close the real constant dialog box.
4. Define material properties

Material Models
Double-click on Structural, Linear, Elastic, Orthotropic.

Enter 25.0E6 for EX.
Enter 1.0E6 for EY and EZ.
Enter 0.25 for PRXY and PRXZ.
Enter 0.01 for PRYZ.
Enter 0.5E6 for GXY and GXZ.
Enter 0.2E6 for GYZ.
OK to define material property set and close the dialog box.
Material> Exit
5. Build geometry
There are several ways to create the model geometry within
ANSYS, some more convenient than others. Decide where the ori-
gin will be located and then define the rectangle primitive relative
to that origin. The location of the origin is arbitrary. Here, use the
center of the plate. ANSYS does not need to know where the origin
is. Simply begin by defining a rectangle relative to that location. In
ANSYS, this origin is called the global origin.
Main Menu> Preprocessor> Modeling> Create>

Areas> Rectangle> By Dimensions
Enter the following: X1=0, X2=5, Y1=0, Y2=5

OK to close the dialog box.
6. Generate mesh
One nice feature of the ANSYS program is that you can automat-
ically mesh the model without specifying any mesh size controls.
This is using what is called a default mesh. If you’re not sure how
to determine the mesh density, let ANSYS try it first. Instead you
will specify a global element size to control overall mesh density.
Main Menu> Preprocessor> Meshing> Mesh Tool
Set Global Size control.
Type in 0.5.
OK
Choose Area Meshing. Click on Mesh.

Pick All for the area to be meshed (in picking menu). Close any
warning messages that appear.
Close the Mesh Tool.
7. Verification of data
Since a large amount of input data is required for composites, you
should verify the data before proceeding with the solution. Several
commands are available for this purpose. Some of them are demon-
strated as follows:
EPLOT (Utility Menu> Plot> Elements) displays all
selected elements. Using the /ESHAPE,1 command (Utility
Menu> PlotCtrls> Style> Size and Shape) before EPLOT causes
shell elements to be displayed as solids with the layer thicknesses
obtained from real constants or section definition.
Utility Menu> PlotCtrls>Style>Size and Shape
Turn on display of element key then Ok
Utility Menu> Plot Elements

Select isometric view
Zoom in to clearly see the four layers along the thickness direction.
Pick the zoom tablet.
Select the region Enlarged view of the corner showing layers.
LAYPLOT (Utility Menu> Plot> Layered Elements)

This displays the layer stacking sequence from real constants
in the form of a sheared deck of cards. The layers are crosshatched
and color coded for clarity. The hatch lines indicate the layer
angle (real constant THETA) and the color indicates layer material
number (MAT). You can specify a range of layer numbers for the
display.
Utility Menu> Plot> Layered Elements
Select the element to be displayed (only one element can be

selected at one time).
Enter the range of layers to be displayed for the selected element.

OK.
Figure shows the configuration 0/90/90/0 for the layered element.
2.3.4 Apply Loads
The beginning of the solution phase. A new, static analysis is the default,
so you will not need to specify analysis type for this problem. Also, there
are no analysis options for this problem.
Utility Menu > Plotctrls > Numbering
Select LINE numbers
Utility Menu > Plot > Lines
8. Apply displacement constraints

You can apply symmetric displacement constraints directly to two
lines.
Main Menu> Solution> Define Loads> Apply> Structural>
Displacement> Symmetry B.C.>On Lines
Pick the two lines of symmetry (Line numbers 1 and 4).

OK (in picking menu).
Now, apply displacement on the two simply supported plate edges.

Main Menu> Solution> Define Loads> Apply>
Structural> Displacement>On Lines
Pick the line at the top (Line numbers 3).

Click on UX.
Enter 0 for zero displacement.
APPLY to apply constraints and return to dialog box.
Pick the line at the top (Line numbers 3).

Click on UZ.
Pick the line at the right (Line numbers 2).

Click on UY.
Pick the line at the right (Line numbers 2).

Click on UZ.
OK to apply constraints and return to dialog box.
Utility Menu> Plot Lines
9. Apply pressure load

Now apply the uniform pressure load on the top surface. The
ANSYS convention for pressure loading is that a positive load
value represents pressure into the surface (compressive).
Main Menu> Solution> Define Loads> Apply>
Structural> Pressure> On Areas
Pick the area.

OK.
Enter 100 for VALUE.

Pressure distribution is shown by face-out lines. It can be

changed to arrows demonstrated as follows.
Utility Menu > Plotctrls > Symbols

Select Arrows for Pressure.
OK.
10. Obtain solution

Main Menu> Solution> Solve> Current LS
Review the information in the status window, then choose

File> Close (Windows), OK to begin the solution. Choose Yes to
any Verify messages that appear.
Then Close the information window when solution is done.
ANSYS stores the results of this one load step problem in the
database and in the results file, Jobname.RST (or Jobname.RTH
for thermal, Jobname.RMG for magnetic, and Jobname.RFL for
fluid analyses). The database can actually contain only one set of
results at any given time, so in a multiple load step or multiple sub-
step analysis, ANSYS stores only the final solution in the database.
ANSYS stores all solutions in the results file.
11. Review results

The beginning of the postprocessing phase.
Main Menu> General Postproc> Plot Results>
Contour Plot> Nodal Solu
Choose DOF Solution item to be contoured.
SCROLL DOWN AND CHOOSE Z- COMPONENT
OF DISPLACEMENT (UZ).
Main Menu> General Postproc> Plot Results>

Contour Plot> Nodal Solu
Choose Stress item to be contoured.
Scroll down and choose von Mises (SEQV).
To see the stress distribution along the thickness,

First select bottom view, then
Rotate the view once about positive x-axis.
Zoom the region of interest.

List reaction solution:

Main Menu> General Postproc> List Results>
Reaction Solu
OK to list all items and close the dialog box.
Scroll down and find the total vertical force, FZ.
Total FZ = Pressure × Area = 100 × (5 × 5) = 2,500 N
File> Close (Windows).
PRINT FZ REACTION SOLUTIONS

PER NODE
NODE FZ
2 -43.064
22 85.605
23 -171.83
.
.
58 -34.133
59 -69.352
60 -34.873
61 -70.354
TOTAL VALUES
VALUE -2500.0
12. Exit the ANSYS program

When exiting the ANSYS program, you can save the geometry and
loads portions of the database (default), save geometry, loads, and
solution data (one set of results only), save geometry, loads, solution
data, and postprocessing data (i.e., save everything), or save noth-
ing. You can save nothing here, but you should be sure to use one
of the other save options if you want to keep the ANSYS data files.
1. Toolbar: Quit.
2. Choose Quit—No Save!
3. OK.
Congratulations! You have completed this tutorial.

Chapter 3
Probabilistic Design
Analysis
3.1 Probabilistic Design
Probabilistic design is an analysis technique for assessing the effect of

uncertain input parameters and assumptions on your model. A probabilis-
tic analysis allows you to determine the extent to which uncertainties in
the model affect the results of a finite element analysis. An uncertainty (or
random quantity) is a parameter whose value is impossible to determine
at a given point in time (if it is time-dependent) or at a given location (if
it is location-dependent). An example is ambient temperature; you can-
not know precisely what the temperature will be one week from now in
a given city. In a probabilistic analysis, statistical distribution functions
(such as the Gaussian or normal distribution, the uniform distribution,
etc.) describe uncertain parameters.
Computer models are expressed and described with specific numer-
ical and deterministic values; material properties are entered using cer-
tain values, the geometry of the component is assigned a certain length or
width, and so on. An analysis based on a given set of specific numbers and
values is called a deterministic analysis. Naturally, the results of a deter-
ministic analysis are only as good as the assumptions and input values
used for the analysis. The validity of those results depends on how correct
the values were for the component under real-life conditions.
In reality, every aspect of an analysis model is subjected to scatter
(in other words, is uncertain in some way). Material property values are
different if one specimen is compared to the next. This kind of scatter
is inherent for materials and varies among different material types and
material properties. For example, the scatter of the Young’s modulus for
many materials can often be described as a Gaussian distribution with
standard deviation of ±3–5 percent. Likewise, the geometric properties

of components can only be reproduced within certain manufacturing tol-
erances. The same variation holds true for the loads that are applied to a
finite element model. However, in this case the uncertainty is often due to
a lack of engineering knowledge. For example, at elevated temperatures
the heat transfer coefficients are very important in a thermal analysis, yet it
is almost impossible to measure the heat transfer coefficients. This means
that almost all input parameters used in a finite element analysis are inex-
act, each associated with some degree of uncertainty.
It is neither physically possible nor financially feasible to eliminate
the scatter of input parameters completely. The reduction of scatter is typ-
ically associated with higher costs either through better and more precise
manufacturing methods and processes or increased efforts in quality con-
trol; hence, accepting the existence of scatter and dealing with it rather
than trying to eliminate it makes products more affordable and production
of those products more cost-effective.
To deal with uncertainties and scatter, you can use the ANSYS Proba-
bilistic Design System (PDS) to answer the following questions:
• If the input variables of a finite element model are subjected to scat-

ter, how large is the scatter of the output parameters? How robust
are the output parameters? Here, output parameters can be any
parameter that ANSYS can calculate. Examples are the tempera-
ture, stress, strain, or deflection at a node, the maximum tempera-
ture, stress, strain, or deflection of the model, and so on.
• If the output is subjected to scatter due to the variation of the input
variables, then what is the probability that a design criterion given
for the output parameters is no longer met? How large is the prob-
ability that an unexpected and unwanted event takes place (what is
the failure probability)?
• Which input variables contribute the most to the scatter of an output
parameter and to the failure probability? What are the sensitivities
of the output parameter with respect to the input variables?
Probabilistic design can be used to determine the effect of one or

more variables on the outcome of the analysis. In addition to the prob-
abilistic design techniques available, the ANSYS program offers a set of
strategic tools that can be used to enhance the efficiency of the probabi-
listic design process. For example, you can graph the effects of one input
variable versus an output parameter, and you can easily add more samples
and additional analysis loops to refine your analysis.
Probabilistic Design Analysis • 71
3.1.1 T
raditional (Deterministic) VERSUS
Probabilistic Design Analysis Methods
In traditional deterministic analyses, uncertainties are either ignored or

accounted for by applying conservative assumptions. Uncertainties are
typically ignored if the analyst knows for certain that the input parameter
has no effect on the behavior of the component under investigation. In
this case, only the mean values or some nominal values are used in the
analysis. However, in some situations the influence of uncertainties exists
but is still neglected; for example, the Young’s modulus mentioned ear-
lier or the thermal expansion coefficient, for which the scatter is usually
ignored. Let’s assume that you are performing a thermal analysis and
you want to evaluate the thermal stresses (thermal stresses are directly
proportional to the Young’s modulus as well as to the thermal expansion
coefficient of the material). The equation is:
σtherm = E α ΔT
If the Young’s modulus alone has a Gaussian distribution with a 5

percent standard deviation, then there is almost a 16 percent chance that
the stresses are more than 5 percent higher than what you would think
they are in a deterministic case. This figure increases if you also take into
account that, typically, the thermal expansion coefficient also follows a
Gaussian distribution.
Probability that the Probability that

thermal stresses the thermal
are more than stresses are more
Random input variables 5% higher than than 10% higher
taken into account expected than expected
Young’s modulus (Gauss-
ian distribution with 5% ~16% ~2.3%
standard deviation)
Young’s modulus and
thermal expansion coef-
ficient (each with Gauss- ~22% ~8%
ian distribution with 5%
standard deviation)
When a conservative assumption is used, this actually tells you that

uncertainty or randomness is involved. Conservative assumptions are usu-
ally expressed in terms of safety factors. Sometimes regulatory bodies
demand safety factors in certain procedural codes. If you are not faced with
such restrictions or demands, then using conservative assumptions and
safety factors can lead to inefficient and costly overdesign. You can avoid
overdesign by using probabilistic methods while still ensuring the safety of
the component. Probabilistic methods even enable you to quantify the safety
of the component by providing a probability that the component will survive
operating conditions. Quantifying a goal is the necessary first step toward
achieving it. Probabilistic methods can tell you how to achieve your goal.
3.1.2 Reliability and Quality Issues
Use probabilistic design when issues of reliability and quality are para-
mount. Reliability is usually always a concern because product or com-
ponent failures have significant financial consequences (costs of repair,
replacement, warranty, or penalties); worse, a failure can result in injury
or loss of life. Although perfection is neither physically possible nor
financially feasible, probabilistic design helps you to design safe and reli-
able products while avoiding costly overdesign and conserve manufac-
turing resources (machining accuracy, efforts in quality control, and so
on). Quality is the perception by a customer that the product performs as
expected or better. In a quality product, the customer rarely receives unex-
pected and unpleasant events where the product or one of its components
fails to perform as expected. By nature, those rare “failure” events are
driven by uncertainties in the design. Here, probabilistic design methods
help you to assess how often “failure” events may happen. In turn, you can
improve the design for those cases where the “failure” event rate is above
your customers’ tolerance limit.
3.1.3 Probabilistic Design Terminology
PDS term Description

Quantities that influence the result of an analysis.
In probabilistic design, RVs are often called
Random input “drivers” because they drive the result of an
variables (RVs) analysis. You must specify the type of statistical
distribution the RVs follow and the parameter
values of their distribution functions.
Two (or more) RVs that are statistically dependent
Correlation
on each other.
The results of a finite element analysis.

Random output
The RPs are typically a function of the RVs; that is,
parameters
changing the values of the RVs should change the
(RPs)
value of the RPs.
Probabilistic The RVs and RPs are collectively known as
design variables probabilistic design variables.
A unique set of parameter values that represents
Sample a particular model configuration. A sample is
characterized by random input variable values.
The collection of all samples that are required or
Simulation
that you request for a given probabilistic analysis.

The simulation contains the information used to
determine how the component would behave under
real-life conditions (with all the existing uncertainties
and scatter); therefore, all samples represent the simu-
lation of the behavior.
The combination of definitions and specifications for
the deterministic model (in the form of the analysis
file). The model has these components:
• RVs.
• Correlations.
• RPs.
• The selected settings for probabilistic method and
its parameters.
Probabilistic If you change any part of the probabilistic model, then
model you will generate different results for the probabilistic
analysis (that is, different results values and/or a dif-
ferent number of results). For example, modifying the
analysis file may affect the results file. If you add or
take away an RV or change its distribution function,
you solve a different probabilistic problem (which
again leads to different results). If you add an RP, you
will still solve the same probabilistic problem, but
more results are generated.
A measure of location often used to describe the
general location of the bulk of the scattering data of a
Mean value
random output parameter or of a statistical distribution
function.
(Continued )
(Continued )
Mathematically, the mean value is the arithmetic aver-
age of the data. The mean value also represents the
center of gravity of the data points. Another name for
the mean value is the expected value.
The statistical point where 50% of the data is below the
median value and the 50% is above.
Median For symmetrical distribution functions (Gaussian, uni-
value form, etc.) the median value and the mean value are
identical, while for nonsymmetrical distributions they
are different.
A measure of variability (dispersion or spread) about
the arithmetic mean value, often used to describe the
width of the scatter of a random output parameter or
Standard
of a statistical distribution function.
deviation
The larger the standard deviation, the wider the scatter
and the more likely it is that there are data values
further apart from the mean value.
3.1.4 S
teps for Probabilistic Design Analysis using
ANSYS
The usual process for probabilistic design consists of the following gen-
eral steps:
1. Create an analysis file for use during looping. The file should
represent a complete analysis sequence and must do the following:
• Build the model parametrically (PREP7).
• Obtain the solution(s) (SOLUTION).
• Retrieve and assign to parameters the quantities that will be
used as RVs and RPs (POST1/POST26).
2. Establish parameters in the ANSYS database, which corre-
spond to those used in the analysis file. This step is typical, but not
required (Begin or PDS); however, if you skip this step, then the
parameter names are not available for selection in interactive mode.
3. Enter PDS and specify the analysis file (PDS).
4. Declare random input variables (PDS).
5. Visualize random input variables (PDS). Optional.
6. Specify any correlations between the RVs (PDS).
7. Specify random output parameters (PDS).
8. Choose the probabilistic design tool or method (PDS).

9. Execute the loops required for the probabilistic design analysis
(PDS).
10. Fit the response surfaces (if you did not use a Monte Carlo Sim-
ulation method) (PDS).
11. Review the results of the probabilistic analysis (PDS).
Since analyzing complex problems can be time-consuming, ANSYS

offers you the option of running a probabilistic analysis on a single proces-
sor or distributing the analyses across multiple processors. By using the
ANSYS PDS parallel run capabilities, you can run many analysis loops
simultaneously and reduce the overall run time for a probabilistic analysis.
The following Figure 3.1 shows the flow of information during a
probabilistic design analysis.
Figure 3.1. The flow of information during a probabilistic design analysis.
3.2 Probability Distributions
Probability distributions are a fundamental concept in statistics. They are

used both on a theoretical level and a practical level. Some practical uses
of probability distributions are:
• To calculate confidence intervals for parameters and to calculate

critical regions for hypothesis tests.
• For univariate data, it is often useful to determine a reasonable dis-

tributional model for the data.
• Statistical intervals and hypothesis tests are often based on spe-
cific distributional assumptions. Before computing an interval or
test based on a distributional assumption, we need to verify that
the assumption is justified for the given data set. In this case, the
distribution does not need to be the best-fitting distribution for the
data, but an adequate enough model so that the statistical technique
yields valid conclusions.
• Simulation studies with random numbers generated from using a
specific probability distribution are often needed.
The mathematical definition of a continuous probability function,

f(x), is a function that satisfies the following properties.
1. The probability that x is between two points a and b is:
b
p [ a ≤ x ≤ b ] = ∫ f ( x ) dx
a
2. It is non-negative for all real x.

3. The integral of the probability function is one, that is:
∞
∫−∞ f ( x) dx = 1
What does this actually mean? Since continuous probability

functions are defined for an infinite number of points over a continuous
interval, the probability at a single point is always zero. Probabilities are
measured over intervals, not single points. That is, the area under the
curve between two distinct points defines the probability for that inter-
val. This means that the height of the probability function can in fact be
greater than one. The property that the integral must equal one is equiva-
lent to the property for discrete distributions that the sum of all the prob-
abilities must equal one.
Probability distributions are typically defined in terms of the proba-
bility density function (pdf). However, there are a number of probability
functions used in applications. For a continuous function, the pdf is the
probability that the variate has the value x. Since for continuous distri-
butions the probability at a single point is zero, this is often expressed in
terms of an integral between two points.
b
∫a f ( x) dx = pr [a ≤ X ≤ b]
For a discrete distribution, the pdf is the probability that the variate
takes the value x.
f ( x ) = Pr [ X = x ]
Figure 3.2 shows the normal pdf.

The cumulative distribution function (cdf) is the probability that
the variable takes a value less than or equal to x. That is:
F ( x ) = Pr [ X ≤ x ] = ∞
For a continuous distribution, this can be expressed mathematically as:

∞
F ( x) = ∫ f ( m) d m
−∞
For a discrete distribution, the cdf can be expressed as:
x
F ( x ) = ∑ f (i )
i=0
Figure 3.3 shows the normal cdf.

The horizontal axis is the allowable domain for the given probability
function. Since the vertical axis is a probability, it must fall between zero
and one. It increases from zero to one as we go from left to right on the
horizontal axis.
0.4
0.35
0.3
0.25
Probability
0.2
0.15
0.1
0.05
0
-5 -4 -3 -2 -1 0 1 2 3 4 5
X
Figure 3.2. The normal probability density function (pdf).

0.5
1
Probability
0.5
0.5
0
-4 -3 -2 -1 0 1 2 3 4
X
Figure 3.3. The normal cumulative distribution function (cdf).
3.2.1 Gallery of Common Continuous Distributions
There are a large number of distributions used in statistical applications.

Detailed information on a few of the most common distributions is avail-
able as follows.
0.4
0.35
0.3
0.25
Probability
0.2
0.15
0.1
0.05
0
-5 -4 -3 -2 -1 0 1 2 3 4 5
X
Normal distribution
1.8
1.6
1.4
Probability
1.2
0.8
0.6
0.4
0.2
0
0 1 2 3 4 5 6 7 8 9 10
X
Uniform distribution
0.5
0.45
0.4
0.35
Probability
0.3
0.25
0.2
0.15
0.1
0.05
0
0 1 2 3 4 5 6 7 8 9 10
X
Exponential distribution
1.4
1.2
1
Probability
0.8
0.6
0.4
0.2
0
0 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
X
Weibull distribution
0.7
0.6
0.5
Probability
0.4
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9 10
x
Lognormal distribution
0.7
0.6
0.5
Probability
0.4
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7 8
x
Fatigue life distribution
0.4
0.2
0
Probability
0.6
0.4
0.2
0
700 800 900 1000 1100 1200 1300
X
Gamma distribution
0.2
0.18
0.16
0.14
Probability
0.12
0.1
0.08
0.06
0.04
0.02
0
–15 –10 –5 0 5 10 15
x
Double exponential distribution
0.4
0.35
0.3
0.25
Probability
0.2
0.15
0.1
0.05
0
-6 -4 -2 0 2 4 6
X
Power normal distribution
0.7
0.6
0.5
Probability
0.4
0.3
0.2
0.1
0
0 2 4 6 8 10 12
x
Power lognormal distribution
3.2.2 Normal Distribution
3.2.2.1 Probability Density Function
The general formula for the probability density function of the normal
distribution is:
( )
e −( x − m) / 2s 2
2
f ( x) =
s 2p
where m is the location parameter and is the scale parameter.

The case where m = 0 and s = 1 is called the standard normal
distribution. The equation for the standard normal distribution is:
e− x / 2
2
f ( x) =
2p
Figure 3.2 shows the standard normal pdf.
3.2.2.2 Cumulative Distribution Function
The formula for the cdf of the normal distribution does not exist in a s imple
closed formula. It is computed numerically. Figure 3.3 shows the normal cdf.
3.2.2.3 Common Statistics
Mean The location parameter m

Median The location parameter m
Mode The location parameter m
Range Infinity in both directions
Standard Deviation The scale parameter s
Coefficient of Variation s/m
Skewness 0
Kurtosis 3
3.2.2.4 Parameter Estimation
The location and scale parameters of the normal distribution can be esti-
mated with the sample mean and sample standard deviation, respectively.
3.2.2.5 Comments
The Gaussian or normal distribution is a very fundamental and commonly

used distribution for statistical matters. It is typically used to describe the
scatter of the measurement data of many physical phenomena. Strictly speak-
ing, every random variable follows a normal distribution if it is generated by
a linear combination of a very large number of other random effects, regard-
less of which distribution these random effects originally follow. The Gauss-
ian distribution is also valid if the random variable is a linear combination of
two or more other effects if those effects also follow a Gaussian distribution.
For both theoretical and practical reasons, the normal distribution is
probably the most important distribution in statistics. For example:
• Many classical statistical tests are based on the assumption that the
data follow a normal distribution. This assumption should be tested
before applying these tests.
• In modeling applications, such as linear and nonlinear regression,
the error term is often assumed to follow a normal distribution with
fixed location and scale.
• The normal distribution is used to find significance levels in many
hypothesis tests and confidence intervals.
3.2.3 Uniform Distribution
The general formula for the probability density function of the uniform
distribution is:
1
f ( x) = � for
� A≤ x≤ B
B− A
where A is the location parameter and B - A is the scale parameter.

The case where A = 0 and B = 1 is called the standard uniform
distribution. The equation for the standard normal distribution is:
f ( x ) = 1 for 0 ≤ x ≤ 1
Since the general form of probability functions can be expressed in

terms of the standard distribution, all subsequent formulas in this section
are given for the standard form of the function.
Figure 3.4 shows the uniform pdf.
1.8
1.6
1.4
Probability
1.2
0.8
0.6
0.4
0.2
0
0 1 2 3 4 5 6 7 8 9 10
X
Figure 3.4. The uniform probability density function.

The formula for the cdf of the uniform distribution is:
F ( x ) = x for 0 ≤ x ≤ 1
Figure 3.5 shows the uniform cdf.
0.9
0.8
0.7
Probability
0.6
0.5
0.4
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9 10
X
Figure 3.5. The uniform cumulative distribution function.
Mean ( A + B) / 2
Median ( A + B) / 2
Mode B−A
Range ( B − A )2
12
B− A
Standard Deviation
3 ( B + A)
Coefficient of Variation s/m
Skewness 0
Kurtosis 9/5
3.2.3.4 Parameter Estimation
The method of moments estimators for A and B are:
A = x − 3S
B = x + 3S
The maximum likelihood estimators for A and B are:
A = midrang (Y1 , Y2 , …, Yn ) − 0.5 rang (Y1 , Y2 , …, Yn )
B = midrang (Y1 , Y2 , …, Yn ) + 0.5 rang (Y1 , Y2 , …, Yn )
3.2.3.5 Comments
The uniform distribution is a very fundamental distribution for cases

where no other information apart from a lower and an upper limit
exists. It is very useful to describe geometric tolerances. It can also be
used in cases where there is no evidence that any value of the random
variable is more likely than any other within a certain interval. In this
sense, it can be used for cases where “lack of engineering knowledge”
plays a role.
The uniform distribution defines equal probability over a given range for
a continuous distribution. For this reason, it is important as a reference dis-
tribution. One of the most important applications of the uniform distribution
is in the generation of random numbers. That is, almost all random number
generators generate random numbers on the (0, 1) interval. For other dis-
tributions, some transformation is applied to the uniform random numbers.
3.2.4 Lognormal Distribution
A variable X is lognormally distributed if Y = LN(X) is normally distrib-

uted with “LN” denoting the natural logarithm. The general formula for
the pdf of the lognormal distribution is:
( ) / 2s 2
− ln (( x − q ) / m )
( )
2
e
f ( x) = � x ≥ q; m, s > 0
( x − q) s 2p
where s is the shape parameter, q is the location parameter, and m is the

scale parameter. The case where q = 0 and m = 1 is called the standard log-
normal distribution. The equation for the standard normal distribution is:
e ( ) / 2s 2
− ln( x)
( ) � x ≥ 0, s > 0
2
f ( x) =
x s 2p

Figure 3.6 shows the lognormal pdf for four values of s.
There are several common parameterizations of the lognormal distri-
bution. The form given here is from Evans, Hastings, and Peacock.
lognormal PDF (σ = 0.5) lognormal PDF (σ = 1)

1 0.8
0.6
Probability
Probability
0.5 0.4
0.2
0 0
0 2 4 6 8 10 0 2 4 6 8 10
x x
lognormal PDF (σ = 2) lognormal PDF (σ = 5)
1.5 3
Probability
Probability
1 2
0.5 1
0 0
0 2 4 6 8 10 0 2 4 6 8 10
x x
Figure 3.6. The lognormal probability density function for four values of s.
The formula for the cdf of the lognormal distribution is:
 ln ( x ) 
F ( x) = Φ  � x ≥ 0, s > 0
 s 
where F is the cdf of the normal distribution.

Figure 3.7 shows the lognormal cdf with the same values of s as the
pdf plots above.
lognormal PDF (σ = 0.5) lognormal PDF (σ = 1)

1 1
Probability
Probability
0.5 0.5
0 0
0 2 4 6 8 10 0 2 4 6 8 10
x x
lognormal PDF (σ = 2) lognormal PDF (σ = 5)
1 0.8
0.6
Probability
Probability
0.5 0.4
0.2
0 0
0 2 4 6 8 10 0 2 4 6 8 10
x x
Figure 3.7. The lognormal cumulative distribution function for four values of s.
Mean e0.5s
2
Median Scale parameter m (= 1 if scale parame-

ter not specified).
Mode 1
es
2
Range Zero to positive infinity

Standard Deviation
es es − 1
2
( 2
)
Coefficient of Variation
( e + 2) e − 1
s2 s2
Skewness
(e ) + 2 (e ) + 3 (e ) − 3
4 3 2
s2 s2 s2
Kurtosis
es − 1
2
The maximum likelihood estimates for the scale parameter, m, and

the shape parameter s are:
∧ ∧
m = exp ( m ) and
∑ i =1( ln ( X i ) − m )
N ∧ 2
∧
s =
N
Where:
N�
∧
m=
∑ i =1 lnX i
N
If the location parameter is known, it can be subtracted from the origi-

nal data points before computing the maximum likelihood estimates of the
shape and scale parameters.
3.2.4.3 Comments
The lognormal distribution is a basic and commonly used distribution.

It is typically used to describe the scatter of the measurement data
of physical phenomena, where the logarithm of the data would fol-
low a normal distribution. The lognormal distribution is very suitable
for phenomena that arise from the multiplication of a large number of
error effects. It is also correct to use the lognormal distribution for a
random variable that is the result of multiplying two or more random
effects (if the effects that get multiplied are also lognormally distrib-
uted). If is often used for lifetime distributions; for example, the scatter
of the strain amplitude of a cyclic loading that a material can endure
until low-cycle-fatigue occurs is very often described by a lognormal
distribution.
The lognormal distribution is used extensively in reliability appli-
cations to model failure times. The lognormal and Weibull distribu-
tions are probably the most commonly used distributions in reliability
applications.
3.2.5 Weibull Distribution
The formula for the pdf of the general Weibull distribution is:
(g −1)
f ( x) =
g  x − m

a a 
 ( g
)
exp − (( x − m) / a ) x ≥ m; g, a > 0
where g is the shape parameter, m is the location parameter, and a is

the scale parameter. The case where m = 0 and a = 1 is called the standard
lognormal distribution. The case where m = 0 and a = 1 is called the
2-parameter Weibull distribution. The equation for the standard Weibull
distribution reduces to:
f ( x) = g ( x)
(g −1)
( )
exp − xg x ≥ 0; g > 0
Weibull PDF (g = 0.5) Weibull PDF (g = 1)

1 0.8
0.6
Probability
Probability
0.5 0.4
0.2
0 0
0 1 2 3 4 5 0 1 2 3 4 5
x x
Weibull PDF (g = 2) Weibull PDF (g = 5)
0.8 1
0.6
Probability
Probability
0.4 0.5
0.2
0 0
0 1 2 3 4 5 0 1 2 3 4 5
x x
Figure 3.8. The Weibull probability density function for four values of g.

Figure 3.8 shows the Weibull pdf for four values of g.
The formula for the cdf of the Weibull distribution is:
F ( x) = 1 − e
( ) x ≥ 0; g > 0
− xg
Figure 3.9 shows the Weibull cdf with the same values of as the pdf
plots above.
Weibull CDF (g = 0.5) Weibull CDF (g = 1)

0.8 1
0.6
Probability
Probability
0.4 0.5
0.2
0 0
0 1 2 3 4 5 0 1 2 3 4 5
x x
Weibull CDF (g = 2) Weibull CDF (g = 5)
1 1
Probability
Probability
0.5 0.5
0 0
0 1 2 3 4 5 0 1 2 3 4 5
x x
Figure 3.9. The Weibull cumulative distribution function for four values of g.
 g + 1
Γ ,
 g 
Mean ∞
Γ ( a ) = ∫t a −1e − t dt
a
where Γ is the gamma function
ln ( 2)
1
Median g
1
 1 g
1 − g  �g > 1
Mode
0 g ≤1
Range Zero to positive infinity.
2
 g + 2    g + 1 
Standard Deviation Γ − Γ
 g    g  
 g + 2
Γ
 g 
Coefficient of Variation 2
−1
  g + 1 
 Γ  g  
3.2.5.4 Comments
The Weibull distribution is used extensively in reliability applications to

model failure times. In engineering, the Weibull distribution is most often
used for strength or strength-related lifetime parameters, and it is the stan-
dard distribution for material strength and lifetime parameters for very brit-
tle materials (for these very brittle materials the “weakest-link-theory” is
applicable).
3.3 Choosing a Distribution for a Random

Variable
The type and source of the data you have determines which distribution
functions can be used or are best suited to your needs.
3.3.1 Measured Data
If you have measured data then you first have to know how reliable that
data is. Data scatter is not just an inherent physical effect, but also includes
inaccuracy in the measurement itself. You must consider that the person
taking the measurement might have applied a “tuning” to the data. For
example, if the data measured represents a load, the person measuring the
load may have rounded the measurement values; this means that the data
you receive are not truly the measured values. Depending on the amount
of this “tuning,” this could provide a deterministic bias in the data that
you need to address separately. If possible, you should discuss any bias
that might have been built into the data with the person who provided that
data to you.
If you are confident about the quality of the data, then how to proceed
depends on how much data you have. In a single production field, the amount
of data is typically sparse. If you have only few data then it is reasonable to
use it only to evaluate a rough figure for the mean value and the standard
deviation. In these cases, you could model the random input variable as a
Gaussian distribution if the physical effect you model has no lower and upper
limit, or use the data and estimate the minimum and maximum limit for a
uniform distribution. In a mass production field, you probably have a lot of
data, in which case you could use a commercial statistical package that will
allow you to actually fit a statistical distribution function that best describes
the scatter of the data.
3.3.2 M
ean Values, Standard Deviation, Exceedence
Values
The mean value and the standard deviation are most commonly used to
describe the scatter of data. Frequently, information about a physical quan-
tity is given in the form that its value is; for example, “100±5.5.” Often, but
not always, this form means that the value “100” is the mean value and “5.5”
is the standard deviation. To specify data in this form implies a Gaussian
distribution, but you must verify this (a mean value and standard deviation
can be provided for any collection of data regardless of the true distribution
type). If you have more information (e.g., you know that the data must be
lognormal distributed), then the PDS allows you to use the mean value and
standard deviation for a definition of a lognormal distribution.
Sometimes the scatter of data is also specified by a mean value and an
exceedence confidence limit. The yield strength of a material is sometimes given
in this way; for example, a 99 percent exceedence limit based on a 95 percent
confidence level is provided. This means that derived from the measured data
we can be sure by 95 percent that in 99 percent of all cases the property values
will exceed the specified limit and only in 1 percent of all cases they will drop
below the specified limit. The supplier of this information is using mean value,
the standard deviation, and the number of samples of the measured data to derive
this kind of information. If the scatter of the data is provided in this way, the best
way to pursue this further is to ask for more details from the data supplier. Since
the given exceedence limit is based on the measured data and its statistical assess-
ment, the supplier might be able to provide you with the details that were used.
If the data supplier does not give you any further information, then you
could consider assuming that the number of measured samples was large.
If the given exceedence limit is denoted with and the given mean value is
denoted with then the standard deviation can be derived from the equation:
X1− a / 2 − X m
s=
C
where the values for the coefficient C are:
Exceedence Probability C
99.5% 2.5758
99.0% 2.3263
97.5% 1.9600
95.0% 1.6449
90.0% 1.2816
3.3.3 No Data
In situations where no information is available, there is never just one

right answer. Below are hints about which physical quantities are usually
described in terms of which distribution functions. This might help you
with the particular physical quantity you have in mind. Also below is a list
of which distribution functions are usually used for which kind of phenom-
ena. Keep in mind that you might need to choose from multiple options.
3.3.3.1 Geometric Tolerances
• If you are designing a prototype, you could assume that the actual
dimensions of the manufactured parts would be somewhere within
the manufacturing tolerances. In this case it is reasonable to use a
uniform distribution, where the tolerance bounds provide the lower
and upper limits of the distribution function.
• Sometimes the manufacturing process generates a skewed distri-
bution; for example, one half of the tolerance band is more likely
to be hit than the other half. This is often the case if missing half
of the tolerance band means that rework is necessary, while falling
outside the tolerance band on the other side would lead to the part
being scrapped. In this case a Beta distribution is more appropriate.
• Often a Gaussian distribution is used. The fact that the normal distribu-
tion has no bounds (it spans minus infinity to infinity) is theoretically
a severe violation of the fact that geometrical extensions are described
by finite positive numbers only. However, in practice this is irrelevant
if the standard deviation is very small compared to the value of the
geometric extension, as is typically true for geometric tolerances.
3.3.3.2 Material Data
• Very often the scatter of material data is described by a Gaussian

distribution.
• In some cases the material strength of a part is governed by the
“weakest-link-theory.” The “weakest-link-theory” assumes that
the entire part would fail whenever its weakest spot would fail.
For material properties where the “weakest-link” assumptions are
valid, then the Weibull distribution might be applicable.
• For some cases, it is acceptable to use the scatter information from
a similar material type. Let’s assume that you know that a material
type very similar to the one you are using has a certain material
property with a Gaussian distribution and a standard deviation of
±5 percent around the measured mean value; then let’s assume that
for the material type you are using, you only know its mean value.
In this case, you could consider using a Gaussian distribution with
a standard deviation of ±5 percent around the given mean value.
• For temperature-dependent materials it is prudent to describe the
randomness by separating the temperature dependency from the
scatter effect. In this case, you need the mean values of your mate-
rial property as a function of temperature in the same way that you
need this information to perform a deterministic analysis. If M(T)
denotes an arbitrary temperature-dependent material property then
the following approaches are commonly used:
Multiplication equation:
M (T )rand = Crand M (T )
Additive equation:
M (T )rand = M (T ) + ∆M rand
Linear equation:
M (T )rand = Crand M (T ) + ∆M rand
• Here, M(T ) denotes the mean value of the material property as a

function of temperature. In the “multiplication equation” the mean
value function is scaled with a coefficient Crand and this coefficient is
a random variable describing the scatter of the material property. In
the “additive equation” a random variable ∆Mrand is added on top of
the mean value function M(T ). The “linear equation” combines both
approaches and here both Crand and ∆Mrand are random variables. How-
ever, you should take into account that in general for the “linear equa-
tion” approach Crand and ∆Mrand are, correlated.
• Deciding which of these approaches is most suitable to describe
the scatter of the temperature-dependent material property requires
that you have some raw data about this material property. Only by
reviewing the raw data and plotting it versus temperature you can
tell which approach is the better one.
3.3.3.3 Load Data
• For loads, you usually only have a nominal or average value. You
could ask the person who provided the nominal value the following
questions: If we have 1,000 components that are operated under

real-life conditions, what would the lowest load value be that only
one of these 1,000 components is subjected to and all others have
a higher load? What would the most likely load value be, that is,
the value that most of these 1,000 components have (or are very
close to)? What would the highest load value be that only one of the
1,000 components is subjected to and all others have a lower load?
To be safe you should ask these questions not only of the person
who provided the nominal value, but also to one or more experts
who are familiar with how your products are operated under real-
life conditions. From all the answers you get, you can then consol-
idate what the minimum, the most likely, and the maximum value
probably is. As verification you can compare this picture with the
nominal value that you would use for a deterministic analysis. If the
nominal value does not have a conservative bias to it then it should
be close to the most likely value. If the nominal value includes a
conservative assumption (is biased), then its value is probably close
to the maximum value. Finally, you can use a triangular distribution
using the minimum, most likely, and maximum values obtained.
• If the load parameter is generated by a computer program then
the more accurate procedure is to consider a probabilistic analysis
using this computer program as the solver mechanism. Use a prob-
abilistic design technique on that computer program to assess what
the scatter of the output parameters are, and apply that data as input
to a subsequent analysis. In other words, first run a probabilistic
analysis to generate an output range, and then use that output range
as input for a subsequent probabilistic analysis.
3.3.4 Choosing Random Output Parameters
Output parameters are usually parameters such as length, thickness, diam-

eter, or model coordinates. The ANSYS PDS does not restrict you with
regard to the number of RPs, provided that the total number of probabi-
listic design variables (i.e., RVs and RPs together) does not exceed 5,000.
ANSYS recommends that you include all output parameters that you
can think of and that might be useful to you. The additional computing
time required to handle more RPs is marginal when compared to the time
required to solve the problem. It is better to define RPs that you might not
consider important before you start the analysis. If you forgot to include
a random output parameter that later turns out to be important, you must
redo the entire analysis.
3.4 Probabilistic Design Techniques
The Monte Carlo Simulation method is the most common and traditional
method for a probabilistic analysis. This method lets you simulate how
virtual components behave the way they are built. One simulation loop
represents one manufactured component that is subjected to a particular
set of loads and boundary conditions. For Monte Carlo simulations, you
can employ either the Direct Sampling method or the Latin Hypercube
Sampling method.
When you manufacture a component, you can measure its geometry
and all of its material properties (although typically, the latter is not done
because this can destroy the component). In the same sense, if you started
operating the component then you could measure the loads it is subjected
to. Again, to actually measure the loads is very often impractical. But the
bottom line is that once you have a component in your hand and start
using it, then all the input parameters have very specific values that you
could actually measure. With the next component you manufacture you
can do the same; if you compared the parameters of that part with the pre-
vious part, you would find that they vary slightly. This comparison of one
component to the next illustrates the scatter of the input parameters. The
Monte Carlo Simulation techniques mimic this process. With this method
you “virtually” manufacture and operate components or parts one after
the other.
The advantages of the Monte Carlo Simulation method are:
• The method is always applicable regardless of the physical effect

modeled in a finite element analysis. It not based on assump-
tions related to the RPs that if satisfied would speed things up
and if violated would invalidate the results of the probabilistic
analysis. Assuming the deterministic model is correct and a very
large number of simulation loops are performed, then Monte
Carlo techniques always provide correct probabilistic results.
Of course, it is not feasible to run an infinite number of simu-
lation loops; therefore, the only assumption here is that the lim-
ited number of simulation loops is statistically representative
and sufficient for the probabilistic results that are evaluated. This
assumption can be verified using the confidence limits, which the
PDS also provides.
• Because of the reason mentioned previously, Monte Carlo simula-
tions are the only probabilistic methods suitable for benchmarking
and validation purposes.
• The individual simulation loops are inherently independent; the

individual simulation loops do not depend on the results of any
other simulation loops. This makes Monte Carlo Simulation tech-
niques ideal candidates for parallel processing.
The Direct Sampling Monte Carlo technique has one drawback: it is

not very efficient in terms of required number of simulation loops.
3.4.1 Direct Sampling
Direct Monte Carlo Sampling is the most common and traditional form of
a Monte Carlo analysis. It is popular because it mimics natural processes
that everybody can observe or imagine and is therefore easy to understand.
For this method, you simulate how your components behave based on the
way they are built. One simulation loop represents one component that is
subjected to a particular set of loads and boundary conditions. The Direct
Monte Carlo Sampling technique is not the most efficient technique, but it is
still widely used and accepted, especially for benchmarking and validating
probabilistic results. However, benchmarking and validating requires many
simulation loops, which is not always feasible. This sampling method is
also inefficient due to the fact that the sampling process has no “memory.”
For example, if we have two RVs X1 and X2 both having a uniform
distribution ranging from 0.0 to 1.0, and we generate 15 samples, we could
get a cluster of two (or even more) sampling points that occur close to
each other if we graphed the two variables (as shown in Figure 3.10).
While in the space of all RVs, it can happen that one sample has input
values close to another sample, this does not provide new information and
1
X2
0
0 X1 1
Figure 3.10. The graph of X1 and X2 illustrating

bad sample distribution.
insight into the behavior of a component in a computer simulation if the

same (or almost the same) samples are repeated.
3.4.2 Latin Hypercube Sampling
The Latin Hypercube Sampling (LHS) technique is a more advanced and

efficient form for Monte Carlo Simulation methods. The only difference
between LHS and the Direct Monte Carlo Sampling technique is that LHS
has a sample “memory,” meaning it avoids repeating samples that have
been evaluated before (it avoids clustering samples). It also forces the tails
of a distribution to participate in the sampling process. Generally, the LHS
technique requires 20 percent to 40 percent fewer simulations loops than
the Direct Monte Carlo Simulation technique to deliver the same results
with the same accuracy. However, that number is largely problem depen-
dent. Figure 3.11 shows the graph of X1 and X2 illustrating Good Sample
Distribution.
1
X2
0
0 X1 1
Figure 3.11. The graph of X1 and X2 illustrating
good sample distribution.
3.5 Postprocessing Probabilistic
Analysis Results
There are two groups of postprocessing functions in the PDS: statistical

and trend. A statistical analysis is an evaluation function performed on a
single probabilistic design variable; for example, a histogram plot of a
random output parameter. A trend analysis typically involves two or more
probabilistic design variables; for example, a scatter plot of one probabi-
listic design variable versus another.
3.5.1 Statistical Postprocessing
Statistical postprocessing allows you several options for reviewing your

results.
3.5.1.1 Sample History
The most fundamental form of postprocessing is directly reviewing the

simulation loop results as a function for the number of simulation loops.
Here, you can review the simulation values, the mean, minimum, or max-
imum values, or the standard deviations.
It is most helpful to review the mean values and standard deviation
history for Monte Carlo Simulation results if you want to decide if the
number of simulation loops was sufficient. If the number of simulation
loops was sufficient, the mean values and standard deviations for all RPs
should have converged. Convergence is achieved if the curve shown in the
respective plots approaches a plateau. If the curve shown in the diagram
still has a significant and visible trend with increasing number of simula-
tion loops then you should perform more simulation loops. In addition, the
confidence bounds plotted for the requested history curves can be inter-
preted as the accuracy of the requested curve. With more simulation loops,
the width of the confidence bounds is reduced.
3.5.1.2 Histogram
A histogram plot is most commonly used to visualize the scatter of a prob-

abilistic design variable. A histogram is derived by dividing the range
between the minimum value and the maximum value into intervals of
equal size. Then the PDS determines how many samples fall within each
interval, that is, how many “hits” landed in the intervals.
Most likely, you will use histograms to visualize the scatter of your RPs.
The ANSYS PDS also allows you to plot histograms of your RVs so you
can double check that the sampling process generated the samples according
to the distribution function you specified. For RVs, the PDS not only plots
the histogram bars, but also a curve for values derived from the distribution
function you specified. Visualizing histograms of the RVs is another way to
make sure that enough simulation loops have been performed. If the number
of simulation loops is sufficient, the histogram bars will:
• Be close to the curve that is derived from the distribution f unction.

• Be “smooth” (without large “steps”).
• Not have major gaps.
The cdf is a primary review tool if you want to assess the reliability or the
failure probability of your component or product. Reliability is defined as
the probability that no failure occurs.
Hence, in a mathematical sense reliability and failure probability are
two sides of the same coin and numerically they complement each other
(are additive to 1.0). The cdf value at any given point expresses the prob-
ability that the respective parameter value will remain below that point.
Figure 3.12 shows the cdf of the random property X:
The value of the cdf at the location x0 is the probability that the values
of X stay below x0. Whether this probability represents the failure proba-
bility or the reliability of your component depends on how you define fail-
ure; for example, if you design a component such that a certain deflection
should not exceed a certain admissible limit then a failure event occurs if
the critical deflection exceeds this limit. Thus for this example, the cdf is
interpreted as the reliability curve of the component. On the other hand,
if you design a component such that the eigenfrequencies are beyond a
certain admissible limit then a failure event occurs if an eigenfrequency
drops below this limit. Thus for this example, the cdf is interpreted as the
failure probability curve of the component.
The cdf also lets you visualize what the reliability or failure probabil-
ity would be if you chose to change the admissible limits of your design.
Often you are interested in visualizing low probabilities and you want to
assess the more extreme ends of the distribution curve. In this case, plot-
ting the cdf in one of the following ways is more appropriate:
• As a Gauss plot (also called a “normal plot”). If the probabilis-

tic design variable follows a Gaussian distribution then the cdf is
displayed as a straight line in this type of plot.
100% F(x)
90%
80%
70%
60%
50%
40%
30%
20%
10% x1 x2 xi X
0%
Figure 3.12. The cumulative distribution function of the random property X.

• As a lognormal plot. If the probabilistic design variable follows a

lognormal distribution then the cdf is displayed as a straight line in
this type of plot.
• As a Weibull plot. If the probabilistic design variable follows a
Weibull distribution then the cdf is displayed as a straight line in
this type of plot.
3.5.1.4 Print Probabilities
The PDS offers a function where you can determine the cdf at any point
along the axis of the probabilistic design variable, including an interpola-
tion function so you can evaluate the probabilities between sampling points.
This feature is most helpful if you want to evaluate the failure probability
or reliability of your component for a very specific and given limit value.
3.5.1.5 Print Inverse Probabilities
The PDS offers a function where you can probe the cdf by specifying a cer-
tain probability level; the PDS tells you at which value of the probabilistic
design variable this probability will occur. This is helpful if you want to
evaluate what limit you should specify to not exceed a certain failure prob-
ability, or to specifically achieve certain reliability for your component.
3.5.2 Trend Postprocessing
Trend postprocessing allows you several options for reviewing your results.
3.5.2.1 Sensitivities
Probabilistic sensitivities are important in allowing you to improve your

design toward a more reliable and better quality product, or to save money
while maintaining the reliability or quality of your product. You can
request a sensitivity plot for any random output parameter in your model.
There is a difference between probabilistic sensitivities and determin-
istic sensitivities. Deterministic sensitivities are mostly only local gradient
information. For example, to evaluate deterministic sensitivities you can
vary each input parameters by ±10 percent (one at a time) while keeping all
other input parameters constant, then see how the output parameters react to
these variations. As illustrated in the following figure, an output parameter
would be considered very sensitive with respect to a certain input parame-

ter if you observe a large change of the output parameter value.
Steep gradient=higher sensitivity

Y Y1
ΔX
ΔY1
Y2
ΔY2
Lower gradient=lower sensitivity
X
These purely deterministic considerations have various disadvantages

that are taken into consideration for probabilistic sensitivities, namely:
• A deterministic variation of an input parameter that is used to deter-

mine the gradient usually does not take the physical range of vari-
ability into account. An input parameter varied by ±10 percent is not
meaningful for the analysis if ±10 percent is too much or too little
compared with the actual range of physical variability and random-
ness. In a probabilistic approach the physical range of variability
is inherently considered because of the distribution functions for
input parameters. Probabilistic sensitivities measure how much the
range of scatter of an output parameter is influenced by the scatter
of the RVs. Hence, two effects have an influence on probabilistic
sensitivities: the slope of the gradient, plus the width of the scatter
range of the RVs. This is illustrated in the following figures. If a
random input variable has a certain given range of scatter, then the
scatter of the corresponding random output parameter is larger, and
the larger the slope of the output parameter curve is (first illus-
tration). But remember that an output parameter with a moderate
slope can have a significant scatter if the RVs have a wider range of
scatter (second illustration).
Y Range of Y Range of
scatter X Y1 scatter X
Range of
scatter Y1
Y2
Scatter
Range of
range
scatter Y2
Y2 Y2
X X
• Gradient information is local information only. It does not take

into account that the output parameter may react more or less with
respect to variation of input parameters at other locations in the
input parameter space. However, the probabilistic approach not
only takes the slope at a particular location into account, but also all
the values the random output parameter can have within the space
of the RVs.
• Deterministic sensitivities are typically evaluated using a finite-dif-
ferencing scheme (varying one parameter at a time while keeping
all others fixed). This neglects the effect of interactions between
input parameters. An interaction between input parameters exists if
the variation of a certain parameter has a greater or lesser effect if
at the same time one or more other input parameters change their
values as well. In some cases interactions play an important or even
dominant role. This is the case if an input parameter is not sig-
nificant on its own but only in connection with at least one other
input parameter. Generally, interactions play an important role in
10 percent to 15 percent of typical engineering analysis cases (this
figure is problem dependent). If interactions are important, then a
deterministic sensitivity analysis can give you completely incorrect
results. However, in a probabilistic approach, the results are always
based on Monte Carlo simulations, either directly performed using
your analysis model or using response surface equations. Inher-
ently, Monte Carlo simulations always vary all RVs at the same
time; thus if interactions exist then they will always be correctly
reflected in the probabilistic sensitivities.
To display sensitivities, the PDS first groups the RVs into two groups:
those having a significant influence on a particular random output param-
eter and those that are rather insignificant, based on a statistical signif-
icance test. This tests the hypothesis that the sensitivity of a particular
random input variable is identical to zero and then calculates the proba-
bility that this hypothesis is true. If the probability exceeds a certain sig-
nificance level (determining that the hypothesis is likely to be true), then
the sensitivity of that random input variable is negligible. The PDS will
plot only the sensitivities of the RVs that are found to be significant. How-
ever, insignificant sensitivities are printed in the output window. You can
also review the significance probabilities used by the hypothesis test to
decide which group a particular random input variable belonged to the
PDS allows you to visualize sensitivities either as a bar chart, a pie chart,
or both. Sensitivities are ranked so the random input variable having the
highest sensitivity appears first.
In a bar chart the most important random input variable (with the
highest sensitivity) appears in the leftmost position and the others fol-
low to the right in the order of their importance. A bar chart describes
the sensitivities in an absolute fashion (taking the signs into account);
a positive sensitivity indicates that increasing the value of the random
input variable increases the value of the random output parameter for
which the sensitivities are plotted. Likewise, a negative sensitivity indi-
cates that increasing the random input variable value reduces the ran-
dom output parameter value. In a pie chart, sensitivities are relative to
each other.
In a pie chart the most important random input variable (with the
highest sensitivity) will appear first after the 12 o’clock position, and the
others follow in clockwise direction in the order of their importance.
Using a sensitivity plot, you can answer the important questions.
How can I make the component more reliable or improve its quality? If
the results for the reliability or failure probability of the component do
not reach the expected levels, or if the scatter of an output parameter is
too wide and therefore not robust enough for a quality product, then you
should make changes to the important input variables first. Modifying an
input variable that is insignificant would be waste of time.
Of course you are not in control of all random input parameters.
A typical example where you have very limited means of control are mate-
rial properties. For example, if it turns out that the environmental tempera-
ture (outdoor) is the most important input parameter then there is probably
nothing you can do. However, even if you find out that the reliability or
quality of your product is driven by parameters that you cannot control,
this has importance—it is likely that you have a fundamental flaw in your
product design! You should watch for influential parameters like these.
If the input variable you want to tackle is a geometry-related param-
eter or a geometric tolerance, then improving the reliability and quality
of your product means that it might be necessary to change to a more
accurate manufacturing process or use a more accurate manufacturing
machine. If it is a material property, then there is might be nothing you
can do about it. However, if you only had a few measurements for a mate-
rial property and consequently used only a rough guess about its scatter
and the material property turns out to be an important driver of product
reliability and quality, then it makes sense to collect more raw data. In this
way, the results of a probabilistic analysis can help you spend your money
where it makes the most sense—in areas that affect the reliability and
quality of your products the most.
3.5.2.2 Scatter Plots
While the sensitivities point indicate which probabilistic design param-

eters you need to modify to have an impact on the reliability or failure
probability, scatter plots give you a better understanding of how and how
far you should modify the input variables. Improving the reliability and
quality of a product typically means that the scatter of the relevant RPs
must be reduced.
The PDS allows you to request a scatter plot of any probabilistic
design variable versus any other one, so you can visualize the relation-
ship between two design variables (input variables or output parameters).
This allows you to verify that the sample points really show the pattern of
correlation that you specified (if you did so). Typically, RPs are correlated
because they are generated by the same set of RVs. To support the process
of improving the reliability or quality of your product, a scatter plot show-
ing a random output parameter as a function of the most important random
input variable can be very helpful.
When you display a scatter plot, the PDS plots the sampling points
and a trendline. For this trendline, the PDS uses a polynomial function
and lets you choose the order of the polynomial function. If you plot a
random output parameter as a function of a random input variable, then
this trendline expresses how much of the scatter on the random output
parameter (Y-axis) is controlled by the random input variable (X-axis).
The deviations of the sample points from the trendline are caused and
controlled by all the other RVs. If you want to reduce the scatter of the
random output parameter to improve reliability and quality, you have
two options:
• Reduce the width of the scatter of the most important random input
variable(s) (that you have control over).
• Shift the range of the scatter of the most important random input
variable(s) (that you have control over).
The effect of reducing and shifting the scatter of a random input vari-
able is illustrated in the following figures. “Input range before” denotes
the scatter range of the random input variable before the reduction or shift-
ing, and “input range after” illustrates how the scatter range of the random
input variable has been modified. In both cases, the trendline tells how
much the scatter of the output parameter is affected and in which way the
range of scatter of the random input variable is modified.
Input range after

Random output parameter
Random output parameter

Input range
after
Output range before

Output range
before
Output
range
after
Output range
after
Input range before Input range
before
Random input variable Random input variable
It depends on your particular problem if either reducing or shifting

the range of scatter of a random input variable is preferable. In general,
reducing the range of scatter of a random input variable leads to higher
costs. A reduction of the scatter range requires a more accurate process
in manufacturing or operating the product—the more accurate, the more
expensive it is. This might lead you to conclude that shifting the scatter
range is a better idea, because it preserves the width of the scatter (which
means you can still use the manufacturing or operation process that you
have). The following points are some considerations if you want to do that:
• Shifting the scatter range of a random input variable can only lead
to a reduction of the scatter of a random output parameter if the
trendline shows a clear nonlinearity. If the trendline indicates a
linear trend (if it is a straight line), then shifting the range of the
input variables anywhere along this straight line doesn’t make any
difference. For this, reducing the scatter range of the random input
variable remains your only option.
• It is obvious from the second illustration that shifting the range
of scatter of the random input variable involves an extrapolation
beyond the range where you have data. Extrapolation is always dif-
ficult and even dangerous if done without care. The more sampling
points the trendline is based on, the better you can extrapolate. Gen-
erally, you should not go more than 30–40 percent outside of the
range of your data. But the advantage of focusing on the important
RVs is that a slight and careful modification can make a difference.
3.5.2.3 Correlation Matrix
Probabilistic sensitivities are based on a statistical correlation analysis

between the individual probabilistic design variables. The PDS lets you
review the correlation data that has been used to derive sensitivities and
decide if individual sensitivity values are significant or not. This infor-
mation is collected in the correlation matrix of the RPSs versus the RVs.
The PDS also lets you review the correlations that have been sampled
between RVs, which are stored in the RVs correlation matrix. The cor-
relations between RPs are important if you want to use the probabilistic
results of your probabilistic analysis as input for another probabilistic
analysis.
3.6 Tutorial 4: Probabilistic Design

Analysis of Circular Plate Bending
In this tutorial, a circular plate of thickness t with a center hole is rigidly

attached along the inner edge and unsupported along the outer edge. The
plate is subjected to bending by a moment Ma applied uniformly along
the outer edge. The input parameters are subject to uncertainty. Measure-
ments show that the plate dimensions can vary significantly. Specimen
tests show that the material properties can also vary. The applied force is
also subject to uncertainty. You will determine the variation of the output
parameters given the uncertainty of the plate dimensions, material proper-
ties, and applied force. The output parameters that you will study are the
maximum deflection of the plate and the maximum equivalent stress at the
clamped edge.
Y Z
Ma Ma
X
b
b t
a
Geometric properties Value Distribution

Inner radius (b) 100.0±0.1 mm Uniform
Outer radius (a) 300.0±0.1 mm Uniform
Thickness (t) 1.0±0.1 mm Uniform
Material properties Nominal value Distribution

Gaussian distribution
with a mean value
equal to the nominal
Young’s modulus (E) 207,000 N/mm2
value and a standard
deviation of 5% of its
mean value.
Gaussian distribution
with a mean value
equal to the nominal
Poisson’s ratio (ν) 0.3
value and a standard
deviation of 5% of its
mean value.
Loading Nominal value Distribution

Lognormal distribution
with a mean value equal
Applied Moment (Ma) 100 N-mm/mm to the nominal value and
a standard deviation of
10% of its mean value.
Since the problem is axisymmetric only a small sector of elements is

needed. A small angle θ = 30° is used for approximating the circular bound-
ary. The calculated load is equally divided and applied to the outer nodes.
1. Enter the PDS and execute the file for the deterministic model
(CIRCPLATE.txt).
2. Specify CIRCPLATE.txt as the analysis file for the probabilistic
analysis.
3. Define the RVs for your probabilistic analysis. For this problem,
you define the inner radius, outer radius and thickness of the plate,
Young’s modulus and poisson’s ratio, and applied moment as RVs
with various distribution functions.
4. Define the random output variables. You define the maximum
deflection and maximum equivalent stress at the fixed edges as the
output parameters.
5. Obtain solution. Define Monte Carlo as the probabilistic analysis

method and execute the Monte Carlo probabilistic simulations.
6. Perform statistical postprocessing to visualize and evaluate the
Monte Carlo results.
7. Perform trend postprocessing.
8. Generate HTML report and exit.
3.6.3 Step-by-Step Analysis
1. Enter PDS and specify analysis file.

You begin by entering the PDS and executing the file CIRCPLAE.
txt for the deterministic model. The CIRCPLAE.txt file contains
a complete analysis sequence for a simple plate with a single force
load. It uses parameters to define all inputs and outputs.
Main Menu > Prob Design
Utility Menu >File>Read Input from

Choose the file CIRCPLATE.txt from your working directory.
OK.
Close.
File: CIRCPLATE.txt
/TITLE, Bending of Circular Plate (PDS by Al-Tabey)
/PREP7
RADIN=100
RADOUT=300
THICK=1
MOMENT=157.08 ! Total moment on a 30 deg segment
YOUNG=207000
PRATIO=0.3
ET,1,SHELL63
R,1,THICK MP,EX,1,YOUNG MP,NUXY,1,PRATIO
CSYS,1 ! DEFINE CYLINDRICAL C.S.
CYL4,0,0,RADIN,0,RADOUT,30
LESIZE,1, , ,30, , , , ,1
LESIZE,3, , ,30, , , , ,1
LESIZE,2, , ,20,0.5, , , ,1
LESIZE,4, , ,20,0.5, , , ,1
MSHKEY,0
AMESH,1
NROTAT,ALL DL,3,1,ALL,
DL,2,1,UY, DL,2,1,ROTX, DL,2,1,ROTZ, DL,4,1,UY,
DL,4,1,ROTX, DL,4,1,ROTZ, LSEL,S,,,1
NSLL,R,1
F,ALL,MY,-(MOMENT)/31
ALLSEL,ALL FINISH
/SOLUTION
SOLVE FINISH
/POST1
RSYS,1
PLNSOL,U,Z, LSEL,S,,,1
NSLL,S,1
NSORT,U,Z
*GET, DEFMAX, SORT, 0, MAX LSEL,S,,,3
NSLL,S,1
NSORT,S,EQV
*GET, STRMAX, SORT, 0, MAX ALLSEL, ALL
2. Specify analysis file.
Now define CIRCPLATE as the analysis file for the probabilis-
tic analysis. PDS uses CIRCPLATE to create a file for perform-
ing analysis loops. PDS uses the CIRCPLATE input and output
parameters as RVs and RPs.
Main Menu > Prob Design>-Analysis File-Assign
Choose the file CIRCPLATE from your working directory. Use
the Browse button if you choose.
Main Menu > Prob Design>- Analysis File-Assign
Choose the file CIRCPLATE from your working directory. Use

the Browse button if you choose.
3. Define input variables.

You will now define the input parameters and their distribution
functions.
Main Menu > Prob Design>Prob Definitns>Random Input
Define Inner Radius as an input variable:

Add a random input variable.
Choose RADIN as the parameter.

Choose UNIF as the distribution type.
OK.
Enter RADIN-0.1 and RADIN+0.1 for the lower and upper bound-
ary, respectively.
OK.
The following figure will appear.
Define Outer Radius as an input variable.

Choose RADOUT as the parameter.
OK.
Enter RADOUT-0.1 and RADOUT+0.1 for the lower and

upper boundary, respectively.
OK.

Similarly, define Thickness as an input variable.

Define thickness as an input variable.
Choose THICK as the parameter.
OK.
Enter THICK-0.1 and THICK+0.1 for the lower and upper bound-
ary, respectively.
OK.
Define Young’s Modulus as an input variable.

Choose YOUNG as the parameter.
Choose GAUS as the distribution type.
OK.
Enter YOUNG for the mean value and 0.05*YOUNG for the stan-
dard deviation.
OK.
Similarly, define Poisson’s Ratio as an input variable.

Choose PRATIO as the parameter.
Choose GAUS as the distribution type.
OK.
Enter PRATIO for the mean value and 0.05*PRATIO for the stan-
dard deviation.
OK.
Define Applied Moment as an input variable.

Choose MOMENT as the parameter.
Choose LOG1 as the distribution type.
OK.
Enter MOMENT for the mean value and 0.1*MOMENT for the
standard deviation.
OK.
Close.
4. Define output parameters.

You will now define the maximum deflection of the plate and
the maximum equivalent stress at the clamped edges as output
parameters.
Main Menu> Prob Design> Prob Definitns>
Random Output
Add a random output variable.

Choose DEFMAX as the parameter.
OK.
Add a random output variable.

Choose STRMAX as the parameter.
OK.
Close.
5. Execute Monte Carlo simulations to obtain solution.

You will now specify the Monte Carlo simulation method and vari-
ous options. You will specify the Latin Hypercube sampling tech-
nique. For the same accuracy, it generally requires fewer simulation

loops than the Direct Monte Carlo sampling technique. You will set
the number of simulations to 40 and the number of repetitions to
1 to give 40 analysis loops. Forty loops will be a sufficient number
for demonstration purposes.
Main Menu> Prob Design>-Prob Method-
Monte Carlo Sims
Choose the Latin Hypercube sampling method.

OK.
Type in 40 for the Number of Simulations and 1 for the Num-

ber of Repetitions to give a total of 40 simulations.
Choose Random Sampling for the Interval Sampling Option.
Choose Execute ALL Sims for the Autostop Option.
Choose Use Continue CONT for the Random Seed Option.
OK.
Main Menu> Prob Design>Run>-Exec Serial-

Run Serial
Enter Result1 for the Solution Set Label. It is the name for the
results set.
OK.
Review the information in the dialog box and then choose OK

to initiate the solution.
The solution is complete when the statement “LOOP 40 OUT

OF 40—CYCLE 1 OUT OF 1 IS FINISHED” appears in the
ANSYS Output Window.
6. Perform Statistical Postprocessing.

(a) Sample history
You will now review simulation results and mean value plots for
maximum deflection. If the number of simulations is sufficient, the
mean value plots for random output variables converge (the curve
flattens out).
Main Menu> Prob Design> Prob Results-Statistics>Sampl
History
Choose DEFMAX for the Prob Design Variable.

Choose Samples for the Plot Type.
OK.
Main Menu> Prob Design> Prob Results-Statistics>

Sampl History

Choose Mean Values for the Plot Type.
OK.
The curve is relatively flat, indicating that the number of sim-

ulations is sufficient.
(b) Histogram
You will also review a histogram plot for the maximum deflection
in order to visualize the scatter of this random output parameter.
To double check that the number of simulation loops is sufficient,
you will plot a histogram of the Young’s modulus random input
variable. The ANSYS PDS not only plots the histogram bars, but
also plots a curve for values derived from the distribution function
that you specified. If the number of simulations is sufficient, the
histogram for Young’s modulus will have bars close to the curve
derived for the distribution function, which are smooth and without
any major gaps.

Histogram

OK.

Histogram
Choose YOUNG for the Prob Design Variable.

OK.

The histogram bars resemble the pdf, indicating that the num-
ber of simulations is sufficient. However, further increase in num-
ber of simulation to 50 will improve the results.
(c) Cumulative DF and probabilities

You will also determine the probability that the maximum deflec-
tion is below 0.525 mm.
CumulativeDF

OK.
Note: The curve shows that there is a about a 93 percent probability

that the deflection remains below 1.375.
Probabilities

Choose Less than for the Relation Label.
Enter 2.0 for the Limit Value.
OK.
Note: that the probability is 97.8 percent that the maximum deflec-
tion is below the value of 2.0 mm. After reviewing the information,
choose File>Close.
(d) Inverse problem

You will also determine the maximum deflection that will give a 90
percent probability that the deflection is below that value.
Inverse Prob

Enter 0.90 for the Target Probability.
OK.

Note: that there is a 90 percent probability that the maximum deflection

is below 1.438 mm. After reviewing the results, choose File>Close.
7. Perform trend postprocessing.
You will now request sensitivity plots for DEFMAX and STR-
MAX to determine which RVs are most significant.
Main Menu> Prob Design> Prob Results Trends>
Sensitivities
Choose DEFMAX for the Response Param.

OK.
Main Menu > Prob Design> Prob Results Trends> Sensitivities

Choose STRMAX for the Response Param.
OK.
Note that the legend indicates that the RVs THICK and MOMENT
are important for the random output parameter DEFMAX and
STRMAX.
You will then request scatter plots of DEFMAX versus the
most significant random input variable THICK.
Main Menu> Prob Design> Prob Results-Trends>
Scatter Plot
Choose THICK for the first parameter.

Choose DEFMAX for the second parameter.
OK.
Finally, you will determine the correlation coefficients between

the RPs and the RVs.
Main Menu> Prob Design> Prob Results-Trends>
Correl Matrix
Choose Input Output for the Type of Matrix.

OK.
OK.

Note: the correlation coefficients. After reviewing the results,

choose File>Close.
You will now generate an HTML report for your probabilistic
analysis. It includes deterministic model information, probabilistic
model information, and probabilistic analysis results.
Main Menu> Prob Design> Prob Results-Report>
Report Options
OK (use the default options).
Main Menu> Prob Design> Prob Results-Report>

Generate Report
Enter report file name, your first name, and your last name.
OK.
You will find the report and all related files in your current
directory in a subdirectory with the report file name.
Toolbar: Quit
Choose Quit—No Save! OK

Chapter 4
APDL Programming
APDL stands for ANSYS Parametric Design Language, a scripting

language that you can use to automate common tasks or even build your
model in terms of parameters (variables). While all ANSYS commands
can be used as part of the scripting language, the APDL commands
discussed here are the true scripting commands and encompass a wide
range of other features such as repeating a command, macros, if-then-
else branching, do-loops, and scalar, vector, and matrix operations. While
APDL is the foundation for sophisticated features such as design optimi-
zation, probabilistic design analysis, and adaptive meshing, it also offers
many conveniences that you can use in your day-to-day analyses.
APDL is also a macro language to create macros. You can record a
frequently used sequence of ANSYS commands in a macro file (these are
sometimes called command files). Creating a macro enables you to, in
effect, create your own custom ANSYS command. In addition to execut-
ing a series of ANSYS commands, a macro can call graphical user inter-
face (GUI) functions or pass values into arguments.
4.1 Create the Analysis File
The analysis file is a key component and crucial to ANSYS optimization

and probabilistic design analysis. The program uses the analysis file to
form the loop file, which is used to perform analysis loops. Any type of
ANSYS analysis (structural, thermal, magnetic, etc.; linear or nonlinear)
can be incorporated in the analysis file.
There are two ways to create an analysis file:
• Input commands line by line with a system editor.

• Create the analysis interactively through ANSYS and use the
ANSYS command log as the basis for the analysis file.
Both methods have advantages and disadvantages. Creating the file

with a system editor is the same as creating a batch input file for the analy-
sis. This method allows you full control of parametric definitions through
exact command inputs. It also eliminates the need to clean out unneces-
sary commands later. However, if you are not moderately familiar with
ANSYS commands, this method may be inconvenient. You may find it
easier to perform the initial analysis interactively, and then use the result-
ing command log as the basis for the analysis file. In this case, final editing
of the log file may be required in order to make it suitable for optimization
looping.
4.1.1 Analysis File using Log Files
The ANSYS program records every command it executes, whether typed

in directly or executed by a function in the GUI, in two places: the
session log file and the internal database command log.
• The session log file is a text file that is saved in your working
directory.
• The database command log is saved in the ANSYS database. You
can copy this log to a file at any time by choosing Utility Menu>
File> Write DB Log File.
Both files are command logs that can be used as input to the ANSYS
program.
4.1.1.1 Session Log File
Every ANSYS session produces a session log named Jobname.LOG. The

default jobname is FILE or file, depending on the operating system. The
program opens the log file when you first enter the program, and closes
it when you exit the program. The session log file provides a complete
record of your ANSYS session (in terms of commands) and is quite valu-
able as a means of recovering from a system crash or catastrophic user
mistake. By reading in a renamed copy of your log file (or by submitting
it as a batch file), you can re-execute every command in your log file, rec-
reating your database exactly as it existed previously.
Your log file is also useful as a debugging tool that can help to reveal
any mistakes you might have made in an ANSYS session. Should you
APDL Programming • 147
require help from your ASD in debugging an ANSYS session, he or she

will almost certainly ask to see a copy of your log file.
Each new ANSYS session appends commands to the existing file Job-
name.LOG. (That is, the log file is not overwritten during a new ANSYS
session, but added to.) A “time stamp,” consisting of the current date and
time, is included so that you can identify the start of each session. Use /
FILNAME,1 to start a new log file for the session.
You can list your entire log file during an interactive run by picking
Utility Menu> List> Files> Log File. Since this file is in ASCII format,
you can view and edit it readily using an external text editor.
4.1.1.2 Database Command Log
ANSYS captures commands generated (or typed in) during an ANSYS

session not only in the log file but also in memory. This in-memory ver-
sion of the command history is called the internal database log. When you
save the database, the program saves this command log in the database file
(Jobname.DB) along with the other database information.
Use either Utility Menu> File> Write DB Log File or the LGWRITE
command to write the database command log to a named ASCII file. You
can then edit this file, make desired changes, and use the file as command
input to the program. This capability is especially useful if you want to use
the command history that was created during an interactive session, but
have somehow lost or corrupted the session log file (Jobname.LOG) that
was associated with your database.
If you create your database in multiple sessions by saving and resum-
ing the database file, the ANSYS program keeps the database log con-
tinuous by appending each new command that is processed. Therefore,
the internal database log is not fragmented; it will represent the complete
database.
4.1.2 Using a Command Log File as Input
The procedure for re-executing the commands contained in a Jobname.

LOG file or in the database log consists of three main steps:
1. Establish the Command Log File

The method to do this depends on whether you use the session log
file or the database log.
4.1.2.1 Session Log File
To establish a command log file from the session log file (Jobname.
LOG), perform these steps:
1. Rename or copy the session log file to a different name.

2. List the log file by choosing Utility Menu> List> Files> Log File.
3. Choose File> Save as from the Log File window.
4.1.2.2 Database Command Log
To establish a command log file from the database log, pick Utility
Menu> File> Write DB Log File. You can specify a file name or use the
default name, Jobname.LGW. You also have the option (with the Kedit
field) to write all commands (default), essential commands only (Kedit
= REMOVE), or essential commands with nonessential commands com-
mented out (Kedit = COMMENT).
1. Edit the Command Log File

Sometimes, you will need to edit your command log file before
using it as program input. As you edit your log file, you may want
to add comments or indentation to improve its readability. You can
add comments to your log file by using the comment character (!).
2. Read in the Edited Log File
In an interactive session, pick Utility Menu> File> Read Input
from to read in the edited command log file.
4.2 Tutorial 5: Stress Analysis of Bicycle

Wrench
In this tutorial will be found the von Mises stresses for the bicycle wrench
made of steel shown under the given distributed and boundary condition.
Material Properties:
Modulus of elasticity: E = 200 GPa (Steel)
Poisson’s ratio: ν = 0.32
Geometry:
Lengths and radii as shown
Thickness: 3mm
Loading: Distributed load: 88 N/cm
Constraints: ux, uy, uz = 0 at left hexagon
7 mm side R = 1.25 cm
88 N / cm
1.5 cm
1 cm 3 cm 1 cm
9 mm side
Fixed all around

First, solve the problem interactively and then save the database log file.
• Start ANSYS
File -> Save As -> Tutorial 1 -> OK
• Use the structural solid element PLANE82 for FEM modeling:
Preprocessor -> Element Type -> Add/Edit/Delete ->
Add -> Structural Mass-Solid -> Select 8node 82 ->
OK -> Options -> Element Behavior:
Select Plane Stress w/thk -> OK -> Close
• Enter Real Constants for the element type chosen:
Preprocessor -> Real Constants -> Add/Edit/Delete -> Add
-> OK -> Enter Thickness THK = 0.3 -> OK -> Close
• Enter material property data for specified steel:
Preprocessor -> Material Props -> Material Model ->
Structural -> Linear -> Elastic -> Isotropic -> Enter
Young’s modulus EX = 200e9 and Poisson’s ratio
PRXY = 0.32 -> OK -> Close
• Create geometry for two similar rectangles 1.5 cm by 3 cm at
locations (2.25,0.5) and (7.25, 0.5):
Preprocessor -> Modeling -> Create -> Areas-Rectangle
-> By 2 Corners -> In dialogue box enter WP X = 2.25,
WP Y = 0.5, Width = 3, Height = 1.5 -> Apply ->
Enter values for next rectangle: WP X = 7.25,
WP Y = 0.5, Width = 3, Height = 1.5 -> OK
• Create geometry for three circles, all of 1.25 cm radius:
Preprocessor -> Modeling -> Create -> Areas -> Circle ->
Solid Circle -> In dialogue box enter WP X = 1.25, WP Y =
1.25, Radius = 1.25 -> Apply -> Enter values for next circle
-> WP X = 6.25, WP Y = 1.25, Radius = 1.25 -> Apply ->
Enter values for next circle -> WP X = 11.25, WP Y = 1.25,

Radius = 1.25 -> OK
• Perform the Boolean operation add to union the areas together:
Preprocessor -> Modeling -> Operate -> Booleans -> Add
-> Areas -> Pick All
• Create geometry for three hexagons, two of 7 mm side (at the
ends) and one of 9 mm side (in the center):
Preprocessor -> Modeling -> Create -> Areas -> Polygon
-> Hexagon -> In dialogue box enter WP X = 1.25, WP Y =
1.25, Radius = 0.7 -> Theta = 120 -> Apply -> Enter values
for next hexagon -> WP X = 6.25, WP Y = 1.25, Radius =
0.9 -> Theta = 120 -> Apply -> Enter values for next hexa-
gon -> WP X = 11.25, WP Y = 1.25, Radius = 0.7 -> Theta =
120 -> OK
• Perform the Boolean operation subtract to get the hexagonal
holes in the wrench the body:
Preprocessor -> Modeling -> Operate -> Booleans -> Sub-
tract -> Areas -> Click on the solid portion of wrench ->
Apply -> One by one pick the three hexagonal areas -> OK
• Now create a mesh in the final wrench shape, first refining the
mesh size:
Preprocessor -> Meshing -> Size Controls -> ManualSize
-> Global > Size -> Enter Size = 0.1 -> OK
Preprocessor -> Meshing -> Mesh -> Areas -> Free -> Click
on wrench -> OK
• Apply the boundary conditions and the load:
Preprocessor -> Loads -> Analysis Type -> New Analysis
-> Static -> OK Preprocessor -> Loads -> Define Loads ->
Apply -> Structural -> Displacement -> On Key Points ->
Click on the six corner points of the left hexagon -> OK ->
Select All DOF -> OK
Preprocessor -> Loads -> Define Loads -> Apply -> Struc-
tural -> Pressure -> On Lines -> Pick the line indicated in
problem statement (top line of right arm) -> OK -> Enter
VALUE = 88 -> OK
• Perform the solution:
Solution -> Solve -> Current LS -> OK
• Start post-processing: Check the deformed shape:
General Post Proc -> Plot Results -> Deformed Shape ->
Def + undef edge -> OK
• To establish a command log file from the database log, pick
Utility Menu> File> Write DB Log File. You can specify a
file name or use the default name, Jobname.LGW.

Using any text editor, you can edit either one of the following files:
Session log file: Tutorial 1.log
Database command log file: Tutorial 1.lgw
Note: One may prefer to edit database command l of file having essential
commands with nonessential commands commented out using charac-
ter. The nonessential comments have been highlighted as follows.
/BATCH
! /COM,ANSYS RELEASE 11.0 UP20070125
11:04:23
08/30/2008
/input,menust,tmp,’’,,,,,,,,,,,,,,,,1
! /GRA,POWER
! /GST,ON
! /PLO,INFO,3
! /GRO,CURL,ON
! /CPLANE,1
! /REPLOT,RESIZE
WPSTYLE,,,,,,,,0
! SAVE, Tutorial-1,db,C:\A6_ANS~1\
FINOPT~1\OPTIMI~1\
/PREP7
!*
ET,1,PLANE82
!*
KEYOPT,1,3,3
KEYOPT,1,5,0
KEYOPT,1,6,0
!*
! SAVE, Tutorial 1,db,
!*
R,1,0.3,
!*
!*
MPTEMP,,,,,,,,
MPTEMP,1,0
MPDATA,EX,1,,200e9
MPDATA,PRXY,1,,0.32
BLC4,2.25,0.5,3,1.5
BLC4,7.25,.5,3,1.5

CYL4,1.25,1.25,1.25
CYL4,6.25,1.25,1.25
CYL4,11.25,1.25,1.25
FLST,2,5,5,ORDE,2
FITEM,2,1
FITEM,2,-5
AADD,P51X
RPR4,6,1.25,1.25,0.7,120
RPR4,6,6.25,1.25,0.9,120
RPR4,6,11.25,1.25,0.7,120
! /REPLOT,RESIZE
! alist, all
FLST,3,3,5,ORDE,2
FITEM,3,1
FITEM,3,-3
ASBA, 6,P51X
ESIZE,0.1,0,
MSHKEY,0
CM,_Y,AREA
ASEL, , , , 4
CM,_Y1,AREA
CHKMSH,’AREA’
CMSEL,S,_Y
!*
AMESH,_Y1
!*
CMDELE,_Y
CMDELE,_Y1
CMDELE,_Y2
!*
!*
ANTYPE,0
! /ZOOM,1,S
CRN,-0.465951,-0.036231,-0.465951,0.144092
FLST,2,6,3,ORDE,6
FITEM,2,9
FITEM,2,13
FITEM,2,15
FITEM,2,19
FITEM,2,21
FITEM,2,-22
!*
/GO
DK,P51X, , , ,0,ALL, , , , , ,
! /AUTO,1
! /REP,FAST
FLST,2,1,4,ORDE,1
FITEM,2,7
/GO
!*
SFL,P51X,PRES,88,
FINISH
/SOL
! /STATUS,SOLU
SOLVE
FINISH
/POST1
! PLDISP,2
! LGWRITE,’Tutorial
1’,’lgw’,’C:\A6_ANS~1\FINOPT~1\
OPTIMI~1\’,COMMENT
After removing nonessential comments and adding comments:

/PREP7
!* Selection of Element Type
ET,1,PLANE82
KEYOPT,1,3,3 ! Option plane with thickness
KEYOPT,1,5,0
KEYOPT,1,6,0
!* Enter thickness value as a real constant
R,1,0.3,
!* Enter material properties
MPTEMP,,,,,,,,
MPTEMP,1,0
MPDATA,EX,1,,200e9
MPDATA,PRXY,1,,0.32
! Create geometric model
BLC4,2.25,0.5,3,1.5
BLC4,7.25,.5,3,1.5
CYL4,1.25,1.25,1.25
CYL4,6.25,1.25,1.25
CYL4,11.25,1.25,1.25
! Add rectangles with three circular areas
FLST,2,5,5,ORDE,2
FITEM,2,1
FITEM,2,-5
AADD,P51X
! Create three hexagonal areas
RPR4,6,1.25,1.25,0.7,120
RPR4,6,6.25,1.25,0.9,120
RPR4,6,11.25,1.25,0.7,120
! Subtract hexagonals to create final shape
FLST,3,3,5,ORDE,2
FITEM,3,1
FITEM,3,-3
ASBA, 6,P51X
! Meshing of the geometry
ESIZE,0.1,0,
MSHKEY,0
CM,_Y,AREA
ASEL, , , , 4
CM,_Y1,AREA
CHKMSH,’AREA’
CMSEL,S,_Y
AMESH,_Y1
CMDELE,_Y
CMDELE,_Y1
CMDELE,_Y2
!*
ANTYPE,0 ! Set analysis type for Static
Analysis
!
! Apply displacement boundary conditions
FLST,2,6,3,ORDE,6
FITEM,2,9
FITEM,2,13
FITEM,2,15
FITEM,2,19
FITEM,2,21
FITEM,2,-22
/GO
DK,P51X, , , ,0,ALL, , , , , ,
! Apply pressure
FLST,2,1,4,ORDE,1
FITEM,2,7
/GO
SFL,P51X,PRES,88,
FINISH ! Save and Exit the Preprocessor
(PREP7)
!
/SOL ! Enter Solution phase
SOLVE
FINISH
4.3 Tutorial 6: Heat Loss from a Cylindrical

Cooling Fin
In this tutorial, a very long rod 25 mm in diameter has one end maintained
at 100°C. The surface of the rod is exposed to ambient air at 25°C with a
convection heat transfer coefficient of 10 W/m2.K. Determine the heat loss
from the rod constructed of pure copper.
Tb = 100° T∞ = 25°C
h = 10 W/m2.K
Air
D = 2.5 cm
L = 1.32 m

Diameter 25 mm Uniform
Length 1320 mm Uniform
Material properties Nominal value Distribution

At Average Temperature
Thermal Conductivity
398 W/m.k T = (Tb + T∞)/2
Coefficient
= 62.5C°~355k
Convection Heat
10 W/m2.k
Transfer Coefficient
Loading Nominal Value Location

Temperature Thermal
100°C At the base of the fin.
Load
At the cylindrical edge
Convection Thermal
10 W/m2.k and the flat face at the
Load
tip of the fin.

First, solve the problem interactively and then save the database
log file.
• Set preference for thermal analysis:
Preference > Thermal > OK
• Use the thermal solid element Quad4node55 for FEM modeling:
Preprocessor -> Element Type -> Add/Edit/Delete ->
Add-> Thermal Mass-Solid -> Select 4node 55 -> OK ->
Options -> Element Behavior:
Select Axisymmetric -> OK -> Close
• Enter material property data for specified steel:
Preprocessor -> Material Props -> Material Model ->
Thermal > Conductivity > Isotropic > Enter conductivity
value KXX = 398 > OK -> Close
• Create geometry for one rectangles 12.5E-3 m by 1.32 m at
locations (0,0):
Preprocessor -> Modeling -> Create -> Areas-Rectangle
-> By 2 Corners -> In dialogue box enter WP X = 0, WP
Y = 0, Width = 12.5E-3, Height = 1.3 -> OK
• Now, to create a mesh, first refining the mesh size by speci-
fying the number of elements along the length (line numbers
2 and 4) = 600 and along radius (line numbers 1 and 3) = 5:
Preprocessor -> Meshing -> Size Controls -> ManualSize

-> Lines > Picked Lines > type 2 then press Enter > type
4 then press Enter-> OK > Enter # of divisions = 600
> Apply > type 1 then press Enter > type 3 then press
Enter-> OK > Enter # of divisions = 5 -> OK
• Now create the mesh:
Preprocessor -> Meshing -> Mesh -> Areas -> Free ->
Click on the area of select Pick All -> OK
• Select type of analysis:
Solution > Analysis Type -> New Analysis -> Steady-State
-> OK
• Apply the temperature boundary condition:
Solution > Define Loads > Apply -> Thermal > Tem-
perature > On Line > Enter line No. = 1 > Select TEMP
and enter value = 100 > OK.
• Apply the convective boundary conditions:
Solution > Define Loads > Apply -> Thermal > Convection
> On Lines > Enter line Numbers = 2, 3 and 4 (after typ-
ing each number press ENTER key) > OK > Enter Film
Coeff = 10 and Bulk Temperature = 25 > OK.
• Perform the solution:
Solution -> Solve -> Current LS -> OK
• To establish a command log file from the database log, pick
Utility Menu> File> Write DB Log File. You can specify a
file name or use the default name, Jobname.LGW.
Using any text editor, edit the Database command log file: Tutorial 2.lgw
/BATCH
! /COM,ANSYS RELEASE 11.0 UP20070125
13:36:28 08/30/2008
/input,start110,ans,’C:\Program Files\
ANSYS
Inc\v110\ANSYS\apdl\’,,,,,,,,,,,,,,,,1
!*
/NOPR
/PMETH,OFF,0
KEYW,PR_SET,1
KEYW,PR_STRUC,0
KEYW,PR_THERM,1
KEYW,PR_FLUID,0
KEYW,PR_ELMAG,0
KEYW,MAGNOD,0
KEYW,MAGEDG,0
KEYW,MAGHFE,0
KEYW,MAGELC,0
KEYW,PR_MULTI,0
KEYW,PR_CFD,0
/GO
!*
! /COM,
! /COM,Preferences for GUI filtering have
been set to display:
! /COM, Thermal
!*
/PREP7
!*
ET,1,PLANE55
!*
KEYOPT,1,1,0
KEYOPT,1,3,1
KEYOPT,1,4,0
KEYOPT,1,8,0
KEYOPT,1,9,0
!*
!*
MPTEMP,,,,,,,,
MPTEMP,1,0
MPDATA,KXX,1,,398
BLC4,0,0,12.5e-3,1.32
! /PNUM,KP,0
! /PNUM,LINE,1
! /PNUM,AREA,0
! /PNUM,VOLU,0
! /PNUM,NODE,0
! /PNUM,TABN,0
! /PNUM,SVAL,0
! /NUMBER,0
!*
! /PNUM,ELEM,0
! /REPLOT
!*
! /DIST,1,0.729,1
! /REP,FAST
! /DIST,1,0.729,1
! /REP,FAST
! /DIST,1,0.729,1
! /REP,FAST
! /DIST,1,0.729,1
! /REP,FAST
! /DIST,1,0.729,1
! /REP,FAST
! /DIST,1,0.729,1
! /REP,FAST
! /DIST,1,0.729,1
! /REP,FAST
! /DIST,1,0.729,1
! /REP,FAST
! /DIST,1,0.729,1
! /REP,FAST
! /DIST,1,0.729,1
! /REP,FAST
! /DIST,1,0.729,1
! /REP,FAST
FLST,5,2,4,ORDE,2
FITEM,5,2
FITEM,5,4
CM,_Y,LINE
LSEL, , , ,P51X
CM,_Y1,LINE
CMSEL,,_Y
!*
LESIZE,_Y1, , ,600, , , , ,1
!*
FLST,5,2,4,ORDE,2
FITEM,5,1
FITEM,5,3
CM,_Y,LINE
LSEL, , , ,P51X
CM,_Y1,LINE
CMSEL,,_Y
!*
LESIZE,_Y1, , ,5, , , , ,1
!*
MSHKEY,0
CM,_Y,AREA
ASEL, , , , 1
CM,_Y1,AREA
CHKMSH,’AREA’
CMSEL,S,_Y
!*
AMESH,_Y1
!*
CMDELE,_Y
CMDELE,_Y1
CMDELE,_Y2
!*
FINISH
/SOL
!*
ANTYPE,0
! /AUTO,1
! /REP,FAST
! /ZOOM,1,S
CRN,0.362883,-0.778990,0.337117,-0.538560
FLST,2,1,4,ORDE,1
FITEM,2,1
!*
/GO
DL,P51X, ,TEMP,100,0
FLST,2,3,4,ORDE,2
FITEM,2,2
FITEM,2,-4
/GO
!*
SFL,P51X,CONV,10, ,25,
! /STATUS,SOLU
SOLVE
FINISH
/POST1
!*
! /EFACET,1
! PLNSOL, TEMP,, 0
LGWRITE,’Tutorial 2’,’lgw’,’C:\A6_ANS~1\
!
FINOPT~1\OPTIMI~1\’,COMMENT
After editing
/PMETH,OFF,0 ! Setting preference
KEYW,PR_SET,1
KEYW,PR_STRUC,0
KEYW,PR_THERM,1
KEYW,PR_FLUID,0
KEYW,PR_ELMAG,0
KEYW,MAGNOD,0
KEYW,MAGEDG,0
KEYW,MAGHFE,0
KEYW,MAGELC,0
KEYW,PR_MULTI,0
KEYW,PR_CFD,0
/PREP7 ! Enter into preprocessor (PREP7)
!*
ET,1,PLANE55 ! Set element type with
axisymmetric option
!*
KEYOPT,1,1,0
KEYOPT,1,3,1
KEYOPT,1,4,0
KEYOPT,1,8,0
KEYOPT,1,9,0
MPTEMP,,,,,,,, ! Enter material properties
MPTEMP,1,0
MPDATA,KXX,1,,398
BLC4,0,0,12.5e-3,1.32 ! Geometric modeling
FLST,5,2,4,ORDE,2 ! Divide lines 2 and 4
into 600 divisions
FITEM,5,2
FITEM,5,4
CM,_Y,LINE
LSEL, , , ,P51X
CM,_Y1,LINE
CMSEL,,_Y
!*
LESIZE,_Y1, , ,600, , , , ,1
!*
into 5 divisions
FITEM,5,1
FITEM,5,3
CM,_Y,LINE
LSEL, , , ,P51X
CM,_Y1,LINE
CMSEL,,_Y
!*
LESIZE,_Y1, , ,5, , , , ,1
!*
MSHKEY,0 ! Mesh the area
CM,_Y,AREA
ASEL, , , , 1
CM,_Y1,AREA
CHKMSH,’AREA’
CMSEL,S,_Y
!*
AMESH,_Y1
!*
CMDELE,_Y
CMDELE,_Y1
CMDELE,_Y2
!*
FINISH ! Save and exit preprocessor
/SOL ! Enter Solution module
!*
ANTYPE,0 ! Set analysis type as steady-state
FLST,2,1,4,ORDE,1 ! Specify temperature
boundary conditions on line 1
FITEM,2,1
FLST,2,3,4,ORDE,2 ! Specify convective b/
conditions on lines, 2,3 & 4
FITEM,2,2
FITEM,2,-4
SOLVE ! Solve the problem
FINISH
Further editing for design optimization:
• Define fin length (L) and fin radius (R) as PARAMETERS to be

used as design variable.
• Using APDL commands get the fin heat transfer rate (THEAT)
and fin tip temperature (Tempend) to be used in the objective
function and state variable.
• Define objective function as ObjFunct=C-TOTALHEAT
/PMETH,OFF,0 ! Setting preference
KEYW,PR_SET,1
KEYW,PR_STRUC,0
KEYW,PR_THERM,1
KEYW,PR_FLUID,0
KEYW,PR_ELMAG,0
KEYW,MAGNOD,0
KEYW,MAGEDG,0
KEYW,MAGHFE,0
KEYW,MAGELC,0
KEYW,PR_MULTI,0
KEYW,PR_CFD,0
!
! *******************************PREPRO-
CESSING
!
/PREP7 ! Enter into preprocessor (PREP7)
!*
ET,1,PLANE55 ! Set element type with
axisymmetric option
!*
KEYOPT,1,1,0
KEYOPT,1,3,1
KEYOPT,1,4,0
KEYOPT,1,8,0
KEYOPT,1,9,0
MPTEMP,,,,,,,, ! Enter material properties
MPTEMP,1,0
MPDATA,KXX,1,,398
L=1.32 ! Set fin length as parameter L
R=12.5e-3 ! Set fin radius as parameter R
BLC4,0,0,R,L ! Geometric modeling in terms
of R and L
into 600 divisions
FITEM,5,2
FITEM,5,4
CM,_Y,LINE
LSEL, , , ,P51X
CM,_Y1,LINE
CMSEL,,_Y
!*
LESIZE,_Y1, , ,600, , , , ,1
!*
into 5 divisions
FITEM,5,1
FITEM,5,3
CM,_Y,LINE
LSEL, , , ,P51X
CM,_Y1,LINE
CMSEL,,_Y
!*
LESIZE,_Y1, , ,5, , , , ,1
!*
MSHKEY,0 ! Mesh the area
CM,_Y,AREA
ASEL, , , , 1
CM,_Y1,AREA
CHKMSH,’AREA’
CMSEL,S,_Y
!*
AMESH,_Y1
!*
CMDELE,_Y
CMDELE,_Y1
CMDELE,_Y2
!*
FINISH ! Save and exit preprocessor
!
!
**************************************SOLU-
TION
!
/SOL ! Enter Solution module
!*
ANTYPE,0 ! Set analysis type as steady-state

FLST,2,1,4,ORDE,1 ! Specify temperature
boundary conditions on line 1
FITEM,2,1
FLST,2,3,4,ORDE,2 ! Specify convective b/
conditions on lines, 2,3 & 4
FITEM,2,2
FITEM,2,-4
SOLVE ! Solve the problem
FINISH
! ****************************** POST
PROCESSING
!
/POST1
FSUM
*GET,THEAT,FSUM,HEAT
*SET,TOTALHEAT,-THEAT
ObjFunct=10000-TOTALHEAT
ALLSEL
*GET,Tempend,NODE,607,TEMP,
FINISH
Chapter 5
Design Optimization
Optimizing a structure or manufactured product refers to a sequence of

analyses aimed at reducing the weight or cost of the structure or product
while maintaining the functionality of the structure or product, and that
meets all engineering requirements, and that can be manufactured and
assembled. This process may include modifications to the structural or
product topology, component thicknesses, and material properties. The
optimization process is iterative where the specified design variables are
altered, in the course of the optimization process, in a manner that mod-
ifies the objective function (structural or product weight or cost) in an
advantageous manner. “The term optimization, or mathematical program-
ming, refers to the study of problems in which one seeks to minimize or
maximize a real function by systematically choosing the values of real
or integer variables from within an allowed set.” Using the computer to
achieve the most efficient design of a product, finite element analysis
(FEA) and other methods are used. The finite element method (FEM) is
among the most powerful tool for nonlinear simulations.
In the past, design engineers have performed a combination of man-
ual and automated methods to accomplish design optimization. They have
applied FEA to parts of a CAD drawing, determined what components
needed work and then redrew the object manually. Increasingly, software
that is tightly integrated with the CAD program can perform the analysis
and automatically redraw the object.
5.1 Optimum Design
The optimum design is the best design in some predefined sense. Among
many examples, the optimum design for a frame structure may be the
one with minimum weight or maximum frequency; in heat transfer, the
minimum temperature; or in magnetic motor design, the maximum peak
torque. In many other situations minimization of a single function may not

be the only goal, and attention must also be directed to the satisfaction of
predefined constraints placed on the design (e.g., limits on stress, geome-
try, displacement, heat flow).
The independent variables in an optimization analysis are the design
variables. The vector of design variables is indicated by:
X = [ X1 X2 X 3  X n ] (5.1)
Design variables are subject to n constraints with upper and lower

limits, that is,
X i ≤ X i ≤ X i (i = 1, 2, 3, ……, n) (5.2)
Where: n = number of design variables.

The design variable constraints are often referred to as side con-
straints and define what is commonly called feasible design space.
Now, minimize: f = f ( x ) (5.3)
Subject to:
gi ( x ) ≤ gi (i = 1, 2, 3, ……, m1 ) (5.4)
hi ( x ) ≤ hi ( x ) (i = 1, 2, 3, ……, m2 ) (5.5)
Wi ≤ Wi ( x ) ≤ Wi (i = 1, 2, 3, ……, m3 ) (5.6)
Where:
f = objective function
gi, hi, wi = state variables containing the design, with under bar and
over bars representing lower and upper bounds respectively.
m1 + m2 + m3 = number of state variables constraints with various
upper and lower limit values. The state variables can also be referred
to as dependent variables in that they vary with the vector x of design
variables.
(Equation 5.3) through (Equation 5.6) represent a constrained min-
imization problem whose aim is the minimization of the objective func-
tion f under the constraints imposed by (Equations 5.2, 5.4, 5.5, and 5.6).
Design configurations that satisfy all constraints are referred to as feasible
designs. Design configurations with one or more violations are termed
Design Optimization • 169
infeasible. In defining feasible design space, a tolerance is added to each

state variable limit. So if X* is a given design set defined as:
X * =  X *1 X *2 X *3  X *n  (5.7)
The design is deemed feasible only if:
( )
g *i = gi x* ≤ gi + ai (i = 1, 2, 3, ……, m1 ) (5.8)
( )
hi − bi ≤ h*i = hi x* (i = 1, 2, 3, ……, m2 ) (5.9)
( )
Wi − gi ≤ W *i = Wi x* ≤ Wi − gi (i = 1, 2, 3, ……, m3 ) (5.10)
Where: ai, bi and gi = tolerances

And
X i ≤ X *i ≤ X i (i = 1, 2, 3, ……, n) (5.11)
(Equation 5.8) to (Equation 5.11) are the defining statements of a fea-

sible design set. As design sets are generated by methods or tools (dis-
cussed as follows) and if an objective function is defined, the best design
set is computed and its number is stored. The best set is determined under
one of the following conditions.
1. If one or more feasible sets exist the best design set is the feasi-
ble one with the lowest objective function value. In other words,
it is the set that most closely agrees with the mathematical goals
expressed by (Equation 5.3) to (Equation 5.6).
2. If all design sets are infeasible, the best design set is the one closest
to being feasible, irrespective of its objective function value.
5.1.1 Optimum Design Fundamentals
Some of the fundamental concepts needed to understand optimization pro-

cess are explained as follows:
Problem formulation: Problem formulation is normally the most
difficult part of the process. It is the selection of design variables, con-
straints, objectives, and models of the disciplines. A further consideration
is the strength and breadth of the interdisciplinary coupling in the problem.
Design variables: A design variable is a specification that is control-

lable from the point of view of the designer. For instance, the thickness of
a structural member can be considered a design variable. Another might
be the material the member is made out of. Design variables can be con-
tinuous (such as a wing span), discrete (such as the number of ribs in a
wing), or boolean (such as whether to build a monoplane or a biplane).
Design problems with continuous variables are normally solved more eas-
ily. Design variables are often bounded, that is, they often have maximum
and minimum values. Depending on the solution method, these bounds
can be treated as constraints or separately.
Constraints: A constraint is a condition that must be satisfied in order
for the design to be feasible. An example of a constraint in aircraft design
is that the lift generated by a wing must be equal to the weight of the
aircraft. In addition to physical laws, constraints can reflect resource lim-
itations, user requirements, or bounds on the validity of the analysis mod-
els. Constraints can be used explicitly by the solution algorithm or can be
incorporated into the objective using Lagrange multipliers.
Objectives: An objective is a numerical value that is to be maximized
or minimized. For example, a designer may wish to maximize profit or
minimize weight. Many solution methods work only with single objec-
tives. When using these methods, the designer normally weighs the vari-
ous objectives and sums them to form a single objective. Other methods
allow multi-objective optimization.
Models: The designer must also choose models to relate the con-
straints and the objectives to the design variables. These models are depen-
dent on the discipline involved. They may be empirical models, such as
a regression analysis of aircraft prices, theoretical models, such as from
computational fluid dynamics, or reduced-order models of either of these.
In choosing the models the designer must trade off fidelity with analysis
time. The multidisciplinary nature of most design problems complicates
model choice and implementation. Often several iterations are necessary
between the disciplines in order to find the values of the objectives and
constraints. As an example, the aerodynamic loads on a wing affect the
structural deformation of the wing. The structural deformation in turn
changes the shape of the wing and the aerodynamic loads. Therefore, in
analyzing a wing, the aerodynamic and structural analyses must be run a
number of times in turn until the loads and deformation converge.
Standard form: Once the design variables, constraints, objectives,
and the relationships between them have been chosen, the problem can
be expressed in the following form: find x that minimizes J(x) subject to
g(x) ≤ 0, h(x) = 0 and � xlb ≤ x ≤ xub where J is an objective, x is a vector
of design variables, g is a vector of inequality constraints, h is a vector of

equality constraints, xlb and xub are vectors of lower and upper bounds on
the design variables. Maximization problems can be converted to min-
imization problems by multiplying the objective by constraints can be
reversed in a similar manner. Equality constraints can be replaced by two
inequality constraints.
Optimization Tree: The optimization is divided into three types,
namely, the continuous, discrete, and multi-objective optimization. The
various ways for optimization along with the procedures to be followed
are listed in Figure 5.1.
Problem solution: The problem is normally solved using appropriate
techniques from the field of optimization. These include gradient-based
algorithms, population-based algorithms, or others. Very simple problems
can sometimes be expressed linearly; in that case, the techniques of linear
programming are applicable.
Gradient-based
methods Population-based Other methods
• Newton’s method
• Random search • Genetic algorithms
• Steepest descent
• Grid search • Memetic
• Conjugate
• Simulated annealing algorithms
gradient sequential
harmony search direct • Particle swarm
quadratic
search optimization
programming
Figure 5.1. Optimization tree listing the optimization methods.

5.1.2 Applications and Examples
5.1.2.1 Applications
1. Design of structures and manufactured products to reduce weight

while maintaining required structural, functional, and manufactur-
ing characteristics.
2. Optimization of component thickness and shape based on user-
defined objective and constraints.
3. Structural and product design constrained by size, weight, fre-
quency, and so on.
4. Feasible structural and product designs subject to size and shape
constraints.
5. Design of structures and manufactured products for minimum cost,
weight, or size.
5.1.2.2 Examples
A few examples of optimization are listed as follows:
1. Optimization of structural crashworthiness behavior, for example,

automotive industry, aviation and aeronautical industry, transporta-
tion safety, and so on.
2. Optimization of sheet metal forming processes, for example, opti-
mization of tool geometry with respect to springback compensation.
3. Identification and optimization of material parameters in nonlinear
material models.
5.2 Design Optimization Using ANSYS
The optimization module is an integral part of the ANSYS program that

can be employed to determine the optimum design. The ANSYS program
can determine an optimum design, a design that meets all specified require-
ments yet demands a minimum in terms of expenses such as weight, s urface
area, volume, stress, cost, and other factors. An optimum design is one that
is as effective as possible. Virtually any aspect of design can be optimized:
dimensions (such as thickness), shape (such as fillet radii), placement of

supports, cost of fabrication, natural frequency, material property, and so
on. Among many examples, the optimum design for a frame structure may
be the one with minimum weight or maximum frequency; in heat trans-
fer, the minimum temperature; or in magnetic motor design, the maximum
peak torque. Any ANSYS item that can be expressed in terms of parame-
ters is a candidate for design optimization. In many other situations mini-
mization of a single function may not be the only goal, and attention must
also be directed to the satisfaction of predefined constraints placed on the
design (e.g., limits on stress, geometry, displacement, heat flow).
5.2.1 D
esign Optimization Terminology and
Information Flow
While working toward an optimum design, the ANSYS optimization rou-

tines employ three types of variables that characterize the design process:
design variables, state variables, and the objective function. These vari-
ables are represented by scalar parameters in ANSYS Parametric Design
Language (APDL). The use of APDL is an essential step in the optimiza-
tion process. The independent variables in an optimization analysis are
the design variables. To understand the terminology involved in design
optimization, consider the following problem:
Find the minimum-weight design of a beam of rectangular cross-sec-
tion subject to the following constraints:
• Total stress σ should not exceed σmax [s < smax].

• Beam deflection Δ should not exceed Δmax [∆ < ∆ max].
• Beam height h should not exceed hmax [h < hmax].
Design Variables (DVs): Independent quantities varied to achieve

the optimum design. Upper and lower limits are specified to serve as “con-
straints” on the DVs. These limits define the range of variation for the DV.
In the above beam example, width b and height h are obvious candidates
for DVs. Both b and h cannot be zero or negative, so their lower limit
would be b,h > 0.0. Also, h has an upper limit of hmax. Up to 60 DVs may
be defined in an ANSYS design optimization problem.
State Variables (SVs): Quantities that constrain the design. Also
known as “dependent variables,” they are typically response quantities that
are functions of the DVs. A state variable may have a maximum and min-
imum limit, or it may be “single sided,” having only one limit. Our beam
example has two SVs: σ (the total stress) and Δ (the beam deflection). You
can define up to 100 SVs in an ANSYS design optimization problem.
Objective Function: The dependent variable that you are attempting

to minimize. It should be a function of the DVs (i.e., changing the values
of the DVs should change the value of the objective function). In the beam
example, the total weight of the beam could be the objective function (to
be minimized). You may define only one objective function in an ANSYS
design optimization problem.
Optimization Variables: Collectively, the design variables, state
variables, and the objective function. In an ANSYS optimization, these
variables are represented by user-named variables called parameters. You
must identify which parameters in the model are DVs, which are SVs, and
which is the objective function.
Design Set or Design: A unique set of parameter values representing
a given model configuration. Typically, a design set is characterized by the
optimization variable values; however, all model parameters (including
those not identified as optimization variables) are included in the set.
Feasible Design: A design that satisfies all specified constraints
(those on the SVs as well as on the DVs). If any one of the constraints is
not satisfied, the design is considered infeasible. The best design is the one
that satisfies all constraints and produces the minimum objective function
value. (If all design sets are infeasible, the best design set is the one closest
to being feasible, irrespective of its objective function value.)
Analysis File: An ANSYS input file containing a complete analy-
sis sequence (preprocessing, solution, and postprocessing). The file must
contain a parametrically defined model, using parameters to represent all
inputs and outputs to be used as DVs, SVs, and the objective function.
Loop File: An optimization file (named Jobname.LOOP), created
automatically via the analysis file. The design optimizer uses the loop file
to perform analysis loops.
Loop: A single pass through the analysis file. Output for the last loop
performed is saved in file Jobname.OPO. An optimization iteration (or
simply iteration) is one or more analysis loops that result in a new design
set. Typically, an iteration equates to one loop; however, for the first order
method, one iteration represents more than one loop.
Optimization Database: This contains the current optimization envi-
ronment, which includes optimization variable definitions, parameters, all
optimization, specifications, and accumulated design sets. This database
can be saved (to Jobname.OPT) or resumed at any time in the optimizer.
The following Figure 5.2 illustrates the flow of information during an
optimization analysis.
The analysis file must exist as a separate entity. The optimization
database is not part of the ANSYS model database.
File.DB
ANSYS
database
file
RESUME
SAVE
ANSYS
Model
database
File.LOOP
OPEXE Loop file
Analysis file Optimization
(parametrically OPEXE database
defined model) OPEXE
File.OPO
Last loop
OPRESU
OPSAVE
output
File.OPT
optimization
data file
Figure 5.2. Optimization data flow.
5.2.2 Optimization Methods
The ANSYS optimization procedure offers several methods and tools that
in various ways attempt to address the mathematical problem stated ear-
lier. ANSYS optimization methods perform actual minimization of the
objective function of Equation 5.3. It will be shown that they transform
the constrained problem into an unconstrained one that is eventually min-
imized. Design tools, on the other hand, do not directly perform minimi-
zation. Use of the tools offer alternate means for understanding design
space and the behavior of the dependent variables. Methods and tools are
discussed in the following sections.
The ANSYS program uses two optimization methods to accommo-
date a wide range of optimization problems:
• The subproblem approximation method is an advanced zero-order

method that can be efficiently applied to most engineering problems.
• The first-order method is based on design sensitivities and is more

suitable for problems that require high accuracy.
For both the subproblem approximation and first-order methods, the

program performs a series of analysis-evaluation-modification cycles.
That is, an analysis of the initial design is performed, the results are evalu-
ated against specified design criteria, and the design is modified as neces-
sary. The process is repeated until all specified criteria are met.
5.2.2.1 Subproblem Approximation Method
The subproblem approximation method can be described as an advanced

zero-order method in that it requires only the values of the dependent vari-
ables, and not their derivatives. There are two concepts that play a key
role in the subproblem approximation method: the use of approximations
for the objective function and state variables, and the conversion of the
constrained optimization problem to an unconstrained problem.
Approximations: For this method, the program establishes the relation-
ship between the objective function and the DVs by curve fitting. This is done
by calculating the objective function for several sets of DV values (i.e., for
several designs) and performing a least squares fit between the data points.
The resulting curve (or surface) is called an approximation. Each optimization
loop generates a new data point, and the objective function approximation is
updated. It is this approximation that is minimized instead of the actual objec-
tive function. State variables are handled in the same manner. An approxima-
tion is generated for each state variable and updated at the end of each loop.
Conversion to an Unconstrained Problem: State variables and limits
on design variables are used to constrain the design and make the optimiza-
tion problem a constrained one. The ANSYS program converts this problem
to an unconstrained optimization problem because minimization techniques
for the latter are more efficient. The conversion is done by adding penal-
ties to the objective function approximation to account for the imposed
constraints.
Convergence Checking: At the end of each loop, a check for convergence
(or termination) is made. The problem is said to be converged if the current, pre-
vious, or best design is feasible and any of the following conditions are satisfied:
• The change in objective function from the best feasible design to

the current design is less than the objective function tolerance.
• The change in objective function between the last two designs is
less than the objective function tolerance.
• The changes in all design variables from the current design to the
best feasible design are less than their respective tolerances.
• The changes in all design variables between the last two designs are
less than their respective tolerances.
Sometimes the solution may terminate before convergence is reached.

This happens if one of the following conditions is true:
• The number of loops specified has been performed.

• The number of consecutive infeasible designs has reached the spec-
ified limit. The default number is 7.
5.2.2.2 First-Order Method
Like the subproblem approximation method, the first-order method con-

verts the problem to an unconstrained one by adding penalty functions to
the objective function. However, unlike the subproblem approximation
method, the actual finite element representation is minimized and not an
approximation. The first-order method uses gradients of the dependent
variables with respect to the design variables. For each iteration, gradient
calculations (which may employ a steepest descent or conjugate direction
method) are performed in order to determine a search direction, and a
line search strategy is adopted to minimize the unconstrained problem.
Thus, each iteration is composed of a number of sub-iterations that include
search direction and gradient computations. That is why one optimization
iteration for the first-order method performs several analysis loops.
Convergence Checking: First-order iterations continue until either
convergence is achieved or termination occurs. The problem is said to be
converged if, when comparing the current iteration design set to the previ-
ous and best sets, one of the following conditions is satisfied:
• The change in objective function from the best design to the current
design is less than the objective function tolerance.
• The change in objective function from the previous design to the
current design is less than the objective function tolerance.
5.2.3 Optimization Design Tools
In addition to the two optimization techniques, the ANSYS program offers

a set of strategic tools that can be used to enhance the efficiency of the
design process. For example, a number of random design iterations can be
performed. The initial data points from the random design calculations can
serve as starting points to feed the optimization methods.
5.2.3.1 Single Loop Analysis Tool
This is a simple and very direct tool for understanding design space. It is
not necessary but it may be useful to compute values of state variables or
the objective function. The design variables are all explicitly defined by
the user. A single loop is equivalent to one complete FEA. At the beginning
of each iteration, the user defines design variable values, X=X*=design
variables defined by the user, and executes a single loop or iteration. If
either state variables or the objective function are defined, corresponding
state variable and objective function values will result.
5.2.3.2 Random Tool
Multiple loops are performed, with random design variable values at each
loop. This design tool will fill the design variable vector with randomly
generated values for each iteration. Each random design iteration is equiv-
alent to one complete analysis loop. A maximum number of loops and a
desired number of feasible loops can be specified. This tool is useful for
studying the overall design space, and for establishing feasible design sets
for subsequent optimization analysis. Random iterations continue until
either one of the following conditions is satisfied:
n � r� = N r
n f =� N f
Where:
nr = number of random iterations performed per each execution
nf = total number of feasible design sets (including feasible sets from
previous executions)
Nr = maximum number of iterations
Nf = desired number of feasible design sets.
5.2.3.3 Sweep Tool
The sweep tool is used to scan global design space that is centered on a
user-defined, reference design set. Upon execution, a sweep is made in
the direction of each design variable while holding all other design vari-
ables fixed at their reference values. The state variables and the objective
function are computed and stored for subsequent display at each sweep
evaluation point.
A sweep execution will produce ns design sets calculated from:
n � s� = nN s
Where:
n = number of design variables
Ns = number of evaluations to be made in the direction of each design
variable.
5.2.3.4 Factorial Tool
This is a statistical tool that is used to generate design sets at all

extreme combinations of design variable values. This technique is
related to the technology known as design of experiment that uses a
two-level, full and fractional factorial analysis. The primary aim is to
compute main and interaction effects for the objective function and the
state variables.
5.2.3.5 Gradient Tool
The gradient tool computes the gradient of the state variables and the
objective function with respect to the design variables. A reference design
set is defined as the point of evaluation for the gradient. Using this tool,
you can investigate local design sensitivities.
5.2.4 General Process for Design Optimization
One can approach an ANSYS optimization in two ways: as a batch run or

interactively via the graphical user interface (GUI). If you are familiar with
ANSYS commands, you can perform the entire optimization by creating an
ANSYS command input file and submitting it as a batch job. This may be
a more efficient method for complex analyses (e.g., nonlinear) that require
extensive run time. Alternatively, the interactive features of optimization
offer greater flexibility and immediate feedback for review of loop results.
When performing optimization through the GUI, it is important to first
establish the analysis file for your model. Then all operations within the
optimizer can be performed interactively, allowing the freedom to probe
your design space before the actual optimization is done. The insights
you gain from your initial investigations can help to narrow your design
space and achieve greater efficiency during the optimization process.
(The interactive features can also be used to process batch optimization
results.)
The process involved in design optimization consists of the following
general steps. The steps may vary slightly, depending on whether you are
performing optimization interactively (through the GUI), in batch mode,
or across multiple machines.
1. Create an analysis file to be used during looping. This file should

represent a complete analysis sequence and must do the following:
• Build the model parametrically (PREP7).
• Obtain the solution(s) (SOLUTION).
• Retrieve and assign to parameters the response quantities that
will be used as state variables and objective functions (POST1/
POST26).
2. Enter OPT and specify the analysis file (OPT).
3. Declare optimization variables (OPT).
4. Choose optimization tool or method (OPT).
5. Specify optimization looping controls (OPT).
6. Initiate optimization analysis (OPT).
7. Review the resulting design sets data (OPT) and postprocess results
(POST1/POST26).
1. Create the analysis file

The analysis file is a key component and crucial to ANSYS opti-
mization. The program uses the analysis file to form the loop file,
which is used to perform analysis loops. Any type of ANSYS anal-
ysis (structural, thermal, magnetic, etc.; linear or nonlinear) can
be incorporated in the analysis file. In this file, the model must be
defined in terms of parameters (which are usually the DVs), and
results data must be retrieved in terms of parameters (for SVs and
the objective function). Only numerical scalar parameters are used
by the design optimizer. There are two ways to create an analysis
file:
• Input commands line by line with a system editor.
• Create the analysis interactively through ANSYS and use the
ANSYS command log as the basis for the analysis file.
No matter how you intend to create the analysis file, the basic
information that it must contain is the same. The steps it must
include are explained next.
5.2.4.1 Build the Model Parametrically
PREP7 is used to build the model in terms of the DV parameters. For our
beam example, the DV parameters are B (width) and H (height), so the
element real constants are expressed in terms of B and H:
/PREP7
! Initialize DV parameters:
B=2.0 ! Initialize width
H=3.0 ! Initialize height
!
ET,1,BEAM3 ! 2-D beam element
AREA=B*H ! Beam cross-sectional area
IZZ=(B*(H**3))/12 ! Moment of inertia about Z
R,1,AREA,IZZ,H ! Real constants in terms of DV
parameters
!
! Rest of the model:
MP,EX,1,30E6 ! Young’s modulus
N,1 ! Nodes
N,11,120
FILL
E,1,2 ! Elements
EGEN,10,1,-1
FINISH ! Leave PREP7
5.2.4.2 Obtain the Solution
The SOLUTION processor is used to define the analysis type and analysis
options, apply loads, specify load step options, and initiate the finite element
solution. The SOLUTION input for the beam example could look like this:
/SOLU
ANTYPE,STATIC ! Static analysis (default)
D,1,UX,0,,11,10,UY ! UX=UY=0 at the two ends
of the beam
SFBEAM,ALL,1,PRES,100 ! Transverse pressure

(load per unit length) = 100
SOLVE
FINISH ! Leave SOLUTION
5.2.4.3 Retrieve Results Parametrically
This is where we retrieve results data and assign them to parameters.

These parameters usually represent SVs and the objective function. The
*GET command (Utility Menu> Parameters> Get Scalar Data), which
assigns ANSYS calculated values to parameters, is used to retrieve the
data. POST1 is typically used for this step, especially if the data are to be
stored, summed, or otherwise manipulated.
In our beam example, the weight of the beam is the objective func-
tion (to be minimized). Since weight is directly proportional to volume,
and assuming uniform density, minimizing the total volume of the beam
is the same as minimizing its weight. Therefore, we can use volume as
the objective function. The SVs for this example are the total stress and
deflection. The parameters for these data may be defined as follows:
/POST1
SET,...
NSORT,U,Y ! Sorts nodes based on UY deflection
*GET,DMAX,SORT,,MAX ! Parameter DMAX = maximum

deflection
!
! Derived data for line elements are accessed
through ETABLE:
ETABLE,VOLU,VOLU ! VOLU = volume of each element
ETABLE,SMAX_I,NMISC,1 ! SMAX_I = max. stress at
end I of each element
ETABLE,SMAX_J,NMISC,3 ! SMAX_J = max. stress at
end J of each element
!
SSUM ! Sums the data in each column of the ele-
ment table
*GET,VOLUME,SSUM,,ITEM,VOLU ! Parameter VOLUME

= total volume
ESORT,ETAB,SMAX_I,,1 ! Sorts elements based on

absolute value of SMAX_I
*GET,SMAXI,SORT,,MAX ! Parameter SMAXI = max.

value of SMAX_I
ESORT,ETAB,SMAX_J,,1 ! Sorts elements based on

absolute value of SMAX_J
*GET,SMAXJ,SORT,,MAX ! Parameter SMAXJ = max.

value of SMAX_J
SMAX=SMAXI>SMAXJ ! Parameter SMAX = greater of

SMAXI and SMAXJ
FINISH
...
2. Enter OPT and specify the analysis file
The remaining steps are performed within the OPT processor.
When you first enter the optimizer, any parameters that exist in the
ANSYS database are automatically established as design set num-
ber 1. To enter the optimizer, use one of these methods:
Command(s): /OPT
GUI: Main Menu> Design Opt
In interactive mode, you must specify the analysis file name.
To specify the analysis file name, use one of these methods:
Command(s): OPANL
GUI: Main Menu> Design Opt> Analysis File> Assign
3. Declare optimization variables
The next step is to declare optimization variables, that is, specify
which parameters are DVs, which ones are SVs, and which one is
the objective function. As mentioned earlier, up to 60 DVs and up
to 100 SVs are allowed, but only one objective function is allowed.
To declare optimization variables, use one of these methods:
Command(s): OPVAR
GUI:
Main Menu> Design Opt> Design Variables
Main Menu> Design Opt> State Variables
Main Menu> Design Opt> Objective
Minimum and maximum constraints can be specified for SVs
and DVs. Constraints are needed for the objective function. Each
variable has a tolerance value associated with it, which you may
input or let default to a program-calculated value.
4. Choose optimization tool or method
In the ANSYS program, several different optimization tools and
methods are available. Single loop is the default. To specify a tool
or method to be used for subsequent optimization looping, use one
of these methods:
Command(s): OPTYPE
GUI: Main Menu> Design Opt> Method/Tool
5. Specify optimization looping controls
Each method and tool has certain looping controls associated with
it, such as maximum number of iterations, and so on. All of the
commands that you use to set these controls are accessed by the
menu path Main Menu> Design Opt> Method/Tool.
There are also a number of general controls that affect how
data is saved during optimization. They are as follows:
• To specify the file where optimization data is to be saved

(defaults to Jobname.OPT):
Command(s): OPDATA
GUI: Main Menu> Design Opt> Controls
• To determine whether information from the best design set is
saved (by default, the database and results files are saved only
for the last design set):
Command(s): OPKEEP
GUI: Main Menu> Design Opt> Controls
6. Initiate optimization analysis
After all appropriate controls have been specified, you can initiate
looping:
Command(s): OPEXE
GUI: Main Menu> Design Opt> Run
Upon execution of OPEXE, an optimization loop file (Job-
name.LOOP) is written from the analysis file. This loop file, which
is transparent to the user, is used by the optimizer to perform analy-
sis loops. Looping will continue until convergence, termination (not
converged, but maximum loop limit or maximum sequential infea-
sible solutions encountered), or completion (e.g., requested number
of loops for random design generation) has been reached.
The values of all optimization variables and other parameters at
the end of each iteration are stored on the optimization data file (Job-
name.OPT). Up to 130 such sets are stored. When the 130th set is
encountered, the data associated with the “worst” design are discarded.
7. Review design sets data
After optimization looping is complete, you can review the result-
ing design sets in a variety of ways using the commands described
in this section. These commands can be applied to the results from
any optimization method or tool.
To list the values of parameters for specified set numbers:
Command(s): OPLIST
GUI: Main Menu> Design Opt> Design Sets> List
To graph specified parameters versus set number so you can
track how a variable changed from iteration to iteration:
Command(s): PLVAROPT
GUI: Main Menu> Design Opt> Graphs/Tables
In addition to reviewing the optimization data, you may wish to
postprocess the analysis results using POST1 or POST26. By default,
results are saved for the last design set in file Jobname.RST (or .RTH,
etc., depending on the type of analysis). The results and the database
for the best design set will also be available if OPKEEP,ON was
issued before looping. The “best results” will be in file Jobname.BRST
(.BRTH, etc.), and the “best database” will be in Jobname.BDB.
5.2.5 Guidelines for Performing Optimization

Analysis
5.2.5.1 Choosing Design Variables
DVs are usually geometric parameters such as length, thickness, diameter,

or model coordinates. They are restricted to positive values. Some points
to remember about DVs are:
• Use as few DVs as possible. Obviously, more DVs demand more itera-
tions and, therefore, more computer time. One way to reduce the num-
ber of design variables is to eliminate some DVs by expressing them
in terms of others, commonly referred to as design variable linking.
• Specify a reasonable range of values for the design variables (MIN
and MAX). Too wide a range may result in poor representation
of design space, whereas too narrow a range may exclude “good”
designs. Remember that only positive values are allowed, and that
an upper limit must be specified.
• Choose DVs such that they permit practical optimum designs.
5.2.5.2 Choosing State Variables
SVs are usually response quantities that constrain the design. Examples of
SVs are stresses, temperatures, heat flow rates, frequencies, deflections,
absorbed energy, elapsed time, and so on. A state variable need not be an
ANSYS-calculated quantity; virtually any parameter can be defined as a state
variable. Some points to keep in mind while choosing state variables are:
• When defining SVs (OPVAR command), a blank input in the MIN

field is interpreted as “no lower limit.” Similarly, a blank in the
MAX field is interpreted as “no upper limit.” A zero input in either
of these fields is interpreted as a zero limit.
• Choose enough SVs to sufficiently constrain the design. In a stress anal-
ysis, for example, choosing the maximum stress as the only SV may not
be a good idea because the location of the maximum stress may change
from loop to loop. Also avoid the other extreme, which would be to
choose the stress in every element as a state variable. The preferred
method is to define the stresses at a few key locations as state variables.
• For the subproblem approximation method, if possible, choose SVs
that have a linear or quadratic relationship with the DVs.
• If a state variable has both an upper and lower limit, specify a rea-
sonable range of limit values. Avoid very small ranges, because
feasible designs may not exist. A stress range of 500 to 1,000 psi,
for example, is better than 900 to 1,000 psi.
• If an equality constraint is to be specified (such as frequency =
386.4 Hz), define two state variables for the same quantity and
bracket the desired value, illustrated as follows:
*GET,FREQ,ACTIVE,,SET,FREQ ! Parameter FREQ
= calculated
frequency
FREQ1=FREQ
FREQ2=FREQ
/OPT
OPVAR,FREQ1,SV,,387 ! Upper limit on FREQ1 =
387
OPVAR,FREQ2,SV,386 ! Lower limit on FREQ2 =
386
• Avoid choosing SVs near singularities (such as concentrated
loads) by using selecting before defining the parameters.
5.2.5.3 Choosing the Objective Function
The objective function is the quantity that you are trying to minimize or max-
imize. Some points to remember about choosing the objective function are:
• The ANSYS program always tries to minimize the objective func-

tion. If you need to maximize a quantity x, restate the problem and
minimize the quantity x1 = C − x or x1 = 1/x, where C is a number
much larger than the expected value of x. C − x is generally a bet-
ter way to define the objective function than 1/x because the latter,
being an inverse relationship, cannot be as accurately represented by
the approximations used in the subproblem approximation method.
• The objective function should remain positive throughout the opti-
mization, because negative values may cause numerical prob-
lems. To prevent negative values from occurring, simply add a
sufficiently large positive number to the objective function (larger
than the highest expected objective function value).
5.2.5.4 Restarting an Optimization Analysis
To restart an optimization analysis, simply resume the optimization data-

base file (Jobname.OPT):
Command(s): OPRESU
GUI: Main Menu> Design Opt> Opt Database> Resume
Once the data is read in, you can respecify optimization type, con-
trols, and so on, and initiate looping. (The analysis file corresponding to
the resumed database must be available in order to perform optimization.)
To initiate looping:
Command(s): OPEXE
GUI: Main Menu> Design Opt> Run
You can use (Main Menu> Design Opt> Opt Database> Resume)
in an interactive session to resume optimization data. If there is data in
the optimization database at the time you want to resume, you should
first clear the optimization database. To clear the optimization database:
Command(s): OPCLR
GUI: Main Menu> Design Opt> Opt Database> Clear & Reset
Since the ANSYS database is not affected by the OPCLR command,
it may also be necessary to clear the ANSYS database if the resumed
optimization problem is totally independent of the previous one. To clear
the ANSYS database:
Command(s): /CLEAR
GUI: Utility Menu> File> Clear & Start New
5.2.6 Sample Optimization Analysis
In the following example, you will perform an optimization analysis of a

hexagonal steel plate using the GUI for the analysis.
Problem Description: You will build a parametric model of a hexagonal
steel plate, using thickness t1 and fillet radius fil as the parameters. All other dimen-
sions are fixed. This example uses a 2-D model and takes advantage of symmetry.
Problem Specifications: The loading for this example is tensile pres-
sure (traction) of 100 MPa at the three flat faces.
The following material properties are used for this analysis: Thickness
= 10 mm
Young’s modulus (E) = 2.07e5 MPa
Poisson’s ratio (υ) = 0.3
1. Test analysis file

To test the analysis file, you clear the database and then read input
from the hexplate.lgw file.
1. Choose menu path Utility Menu> File> Clear & Start New.
Click on OK.
2. When the Verify dialog box appears, click Yes.
3. Change the jobname. Choose menu path Utility Menu> File>
Change Jobname. The Change Jobname dialog box appears.
4. Change the jobname to hexplate and click on OK.
5. Choose menu path Utility Menu> File> Read Input from. In
the Files list, click on hexplate.lgw. Then click on OK. You see
a replay of the entire analysis. Click on Close when the “Solu-
tion is done!” message appears.
! *******************************
! First Pass: Create analysis file.
! ********************************
*create,hexplate
!
! GEOMETRY (in mm)
!-----------------
*afun,deg !Degree units for trig. functions
inrad=200*cos(30)-20 ! Inner radius
t1=30 ! Thickness
fil=10 ! Fillet radius
/prep7
! Create the three bounding annuli
cyl4,-200,,inrad,-30,inrad+t1,30
cyl4,200*cos(60),200*sin(60),inrad,-
90,inrad+t1,-150
cyl4,200*cos(60),200*sin(-
60),inrad,90,inrad+t1,150
aplot
aadd,all
adele,all ! Delete area, keep lines
lplot
! Fillets on inner slot
lsel,,radius,,inrad+t1 ! Select inner arcs
l1 = lsnext(0) ! Get their line numbers
l2 = lsnext(l1)
l3 = lsnext(l2)
lfillet,l1,l2,fil ! Fillets
lfillet,l2,l3,fil
lfillet,l3,l1,fil
lsel,all
lplot
! Keep only symmetric portion
wprot,,90
lsbw,all
wprot,,,60
lsbw,all
csys,1
lsel,u,loc,y,0,60
ldele,all,,,1
lsel,all
ksll
ksel,inve
kdele,all ! Delete unnecessary keypoints
ksel,all
lplot
! Create missing lines and combine right
edge lines
csys,0
ksel,,loc,y,0
lstr,kpnext(0),kpnext(kpnext(0))
! Bottom symmetry edge
ksel,all
csys,1
ksel,,loc,y,60
lstr,kpnext(0),kpnext(kpnext(0))
! 60-deg. symm. edge
ksel,all
csys,0
lsel,,loc,x,100
lcomb,all ! Add lines at the right edge
lsel,all
! Create the area
al,all
aplot
! MESHING
! -------
et,1,82,,,3 ! Plane stress with thickness
r,1,10 ! Thickness
mp,ex,1,2.07e5 ! Young’s modulus of
steel, MPa
mp,nuxy,1,0.3 ! Poisson’s ratio

smrt,3
amesh,all
eplot
finish
! LOADING
! -------
/solu
csys,1
lsel,u,loc,y,1,59
dl,all,,symm ! Symmetry b.c.
csys,0
lsel,,loc,x,100
sfl,all,pres,-50 ! Pressure load (MPa)
lsel,all
lplot
! SOLUTION
! --------
eqslv,pcg
solve
! POSTPROCESSING
! --------------
/post1
plnsol,s,eqv ! Equivalent stress contours
/dscale,,off ! Displacement scaling off
/expand,6,polar,half,,60
!Symmetry expansion
/replot
/expand
!Retrieve max equivalent stress & volume
nsort,s,eqv
*get,smax,sort,,max
! smax = max. equivalent stress
etable,evol,volu
ssum
*get,vtot,ssum,,item,evol
! vtot = total volume
finish
2. Enter the optimizer and identify analysis file
In the next several steps of this problem, you optimize the design. The
overdesigned steel plate under tension loading of 100 MPa needs to
be optimized for minimum weight subject to a maximum von Mises

stress limit of 150 MPa. You are allowed to vary the thickness t1 and
fillet radius fil. First, enter the optimizer and identify the analysis file.
1. Choose menu path Main Menu> Design Opt> Analysis File>
Assign. The Assign Analysis File dialog box appears.
2. In the Files list, click once on hexplate.lgw and then click on
OK.
3. Identify the optimization variables
1. Choose menu path Main Menu> Design Opt> Design Vari-
ables. The Design Variables dialog box appears.
2. Click on Add. The Define a Design Variable dialog box
appears.
3. In the list of parameter names, click on T1. Type 20.5 in the
MIN field and 40 in the MAX field. Click on Apply.
4. In the list of parameter names, click on FIL. Type 5 in the MIN
field and 15 in the MAX field. Click on OK.
5. Click on Close to close the Design Variables dialog box.
6. Choose menu path Main Menu> Design Opt> State Vari-
ables. The State Variables dialog box appears.
7. Click on Add. The Define a State Variable dialog box appears.
8. In the list of parameter names, click on SMAX. Type 150 in the
MAX field. Click on OK.
9. Click on Close to close the State Variables dialog box.
10. Choose menu path Main Menu> Design Opt> Objective. The
Define Objective Function dialog box appears.
11. In the list of parameter names, click on VTOT. Set the TOLER
field to 1.0. Click on OK.
4. Run the optimization
This step involves specifying run time controls and the optimization
method, saving the optimization database, and executing the run.
1. Choose menu path Main Menu> Design Opt> Controls. The
Specify Run Time Controls dialog box appears.
2. Change the OPKEEP setting from “Do not save” to “Save.”
Click on OK.
3. Specify an optimization method. Choose menu path Main
Menu> Design Opt> Method/Tool. The Specify Optimization
Method dialog box appears.
4. Choose Subproblem. Click on OK. Click OK again.
5. Save the optimization database. Choose menu path Main
Menu> Design Opt> Opt Database> Save. In the Selection
field, type hexplate.opt0. Click on OK.
6. Execute the run. Choose menu path Main Menu> Design Opt>
Run. Review the settings and click on OK. (If you receive any
warning messages during the run, close them.)
7. Notes will appear to let you know which design set ANSYS is
currently running. When the run converges, review the Execu-
tion Summary. Click on OK.
5. Review the results
In this step, you start by listing design sets, then graph the objective
function and state variables versus set number.
1. Choose menu path Main Menu> Design Opt> Design Sets>
List. The List Design Sets dialog box appears.
2. Verify that the ALL Sets option is selected. Click on OK.
3. Review the information that appears in the window. Click on
Close.
4. Choose menu path Main Menu> Design Opt> Graphs/
Tables. The Graph/List Tables of Design Set Parameters dialog
box appears.
5. For X-variable parameter, select Set number. For Y-variable
parameters, select VTOT. Click on OK. Review the graph.
6. Choose menu path Main Menu> Design Opt> Graphs/
Tables. The Graph/List Tables of Design Set Parameters dialog
box appears.
7. For X-variable parameter, select Set number. For Y-variable
parameters, select SMAX. Unselect VTOT by clicking on it.
Click on OK. Review the graph.
6. Restore the best design
In this step, you restore the best design. First, however, save the
optimization database to a file.
1. Choose menu path Main Menu> Design Opt> Opt Data-
base> Save. The Save Optimization Data dialog box appears.
2. In the Selection field, type hexplate.opt1. Then click on OK.
3. Choose menu path Main Menu> Finish.
4. Issue the following commands in the ANSYS Input window.
After you type each command in the window, press ENTER
resume, hexplate, bdb/post1 file,hexplate, brst lplot.
5. Choose menu path Main Menu> General Postproc> Read
Results> First Set.
6. Choose menu path Main Menu> General Postproc> Plot
Results> Contour Plot> Nodal Solu. The Contour Nodal
Solution Data dialog box appears.
7. Choose Stress from the list on the left. Choose von Mises
SEQV from the list on the right. Click on OK. Review the plot.
8. Choose menu path Utility Menu> PlotCtrls> Style> Displace-

ment Scaling. For DMULT, select 0.0 (off). Click on OK.
9. Choose menu path Utility Menu> PlotCtrls> Style> Symme-
try Expansion> User-Specified Expansion. The Expansion
by Values dialog box appears.
10. Fill in the 1st Expansion of Symmetry section. For NREPEAT,
type 6. For TYPE, choose Polar. For PATTERN, choose Alter-
nate Symm. Type 0, 60, and 0 in the DX, DY, and DZ fields,
respectively. Click on OK.
7. Exit ANSYS
Click on Quit in the ANSYS Toolbar. Select an option to save, then
click on OK.
5.3 Tutorial 7: Design Optimization

Tutorial
In this tutorial will be addressed the optimization of heat transfer rate from
a cylindrical pin fin. A very long rod 25 mm in diameter has one end main-
tained at 100°C. The surface of the rod is exposed to ambient air at 25°C
with a convection heat transfer coefficient of 10 W/m2.K. Determine the
heat loss from the rod constructed of pure copper.

Diameter 25 mm Uniform
Length 1320 mm Uniform
Nominal
Material properties value Distribution
Thermal Conductivity At Average Temperature
398 W/m.K
Coefficient T = (Tb + T∞)/2 = 62.5°C ~ 355k
Convection Heat
10 W/m2.K
Transfer Coefficient
Nominal
Loading value Location
Temperature Thermal
100°C At the base of the fin.
Load
Convection Thermal At the cylindrical edge and the
10 W/m2.K
Load flat face at the tip of the fin.
• Since the problem is axisymmetric only a small sector of elements

is needed. A rectangular strip is used for approximating the cylin-
drical fin.
• The analysis is carried out for steady-state conditions.
• Material properties are considered to be constant.
• One-dimensional conduction along the rod.
• Negligible radiation exchange with surroundings.
• Uniform heat transfer coefficient.
5.3.2 Solutions/Results
Results FEA Book (Incropera)

Units Watts (W) Watts (W)
Total Heat Loss 29.09103 29.4
5.3.3 Input File Using Log File
You can perform the example optimization analysis of a cylindrical fin

using the ANSYS commands shown as follows. Items prefaced with an
exclamation point (!) are comments.
The analysis happens in essentially two passes. In the first pass, you
create the analysis file. In the second pass, you create the optimization
input. If you prefer, you can perform the second pass of the example anal-
ysis using the GUI method rather than ANSYS commands.
/BATCH
! ***************************************
/PREP7
!
ET,1,PLANE55
!*
KEYOPT,1,1,0
KEYOPT,1,3,1
KEYOPT,1,4,0
KEYOPT,1,8,0
KEYOPT,1,9,0
!
! ************ Material Properties (Conduction

and Convection Coefficients):
Pure Copper
!* At Degree : 355 kelvin
MPTEMP,,,,,,,,
MPTEMP,1,0
MPDATA,KXX,1,,398
MPTEMP,,,,,,,,
MPTEMP,1,0
MPDATA,HF,1,,10
!
! ********************************************
Geometry
*SET,L,1.32 ! LENGTH IN M
*SET,R,12.5E-3 ! RADIUS IN M
BLC4,0,0,R,L
!
! ********************************************
Meshing
MSHKEY,0
CM,_Y,AREA
FLST,5,2,4,ORDE,2
FITEM,5,2
FITEM,5,4
CM,_Y,LINE
LSEL, , , ,P51X
CM,_Y1,LINE
CMSEL,,_Y
!*
LESIZE,_Y1, , ,600, , , , ,1
!*
FLST,5,2,4,ORDE,2
FITEM,5,1
FITEM,5,3
CM,_Y,LINE
LSEL, , , ,P51X
CM,_Y1,LINE
CMSEL,,_Y
!*
LESIZE,_Y1, , ,5, , , , ,1
!*
CM,_Y,AREA
ASEL, , , , 1
CM,_Y1,AREA
CHKMSH,’AREA’
CMSEL,S,_Y
!*
MSHKEY,1
AMESH,_Y1
MSHKEY,0
!*
CMDELE,_Y
CMDELE,_Y1
CMDELE,_Y2
FINISH
! ********************************************
SOLUTION
/SOLU
!
ANTYPE,0 ! Steady-State Thermal Analysis
!
FLST,2,1,4,ORDE,1
FITEM,2,1
DL,P51X, ,TEMP,100,0 ! Prescribed temperature
at the left end
!
FLST,2,2,4,ORDE,2
FITEM,2,2
FITEM,2,-3
SFL,P51X,CONV,10, ,25, ! Convective boundary
condition
!
SOLVE
FINISH
!
! ********************************************
***** POST
PROCESSING
/POST1
FSUM
*GET,THEAT,FSUM,HEAT
*SET,TOTALHEAT,-THEAT
ObjFunct=10000-TOTALHEAT
ALLSEL
*GET,Tempend,NODE,607,TEMP,
FINISH
5.3.4 Optimization Problem
Using a BATCH file for the analysis describes this example optimization
analysis as consisting of two passes. In the first you create an analysis file,
and in the second you create the optimization input. It is better to avoid
graphical picking operations when defining a parametric model. Thus,
the GUI method is not recommended for performing the first pass of the
example analysis and will not be presented here. However, it is acceptable
to perform the optimization pass of the cylindrical fin example using the
GUI method instead of the ANSYS commands shown earlier. The GUI
procedure for performing the optimization pass follows.
1. Enter the Design Optimization (Design Opt) and execute the file for
the deterministic model (Fin.txt).
2. Specify Fin.txt as the analysis file for the design optimization analysis.
3. Define the design variables, state variables, and objective function
for your design optimization analysis. For this problem, you define
the length and radius of the cylindrical fin as design variables, tem-
perature at the tip of the fin as state variable, and total heat dissi-
pated as the objective function.
4. Run the optimization. Select a suitable method for optimization.
5. Obtain solution.
6. Perform postprocessing to visualize and evaluate the design sets of
optimization results.
7. Save the optimization database.

5.3.6 Step-by-Step Analysis
1. Test analysis file

To test the analysis file, you clear the database and then read input
from the Fin.lgw file.
Choose menu path Utility Menu> File> Read Input from. In
the Files list, click on Fin.lgw. Then click on OK. You see a replay
of the entire analysis. Click on Close when the “Solution is done!”
message appears.
Main Menu > Design Opt
Utility Menu >File>Read Input from.

Choose the file Fin.txt from your working directory.
OK.
2. Enter the optimizer and identify analysis file

In the next several steps of this problem, you optimize the design.
The previously designed cylindrical fin under thermal loading of
100°C at the base of the fin needs to be optimized for maximum
heat loss through the surface subjected to a maximum temperature
limit at the tip of the fin to 40°C. You are allowed to vary the length
L and the radius R.
First, enter the optimizer and identify the analysis file.
Choose menu path Main Menu> Design Opt> Analysis File>
Assign. The Assign Analysis File dialog box appears.
In the Files list, click once on Fin.lgw and then click on OK.
Main Menu> Design Opt> Analysis File> Assign
Choose the file Fin from your working directory.

Use the Browse button if you choose.
3. Identify the optimization variables

You will now define the Design Variables (DV) and their limits.
Main Menu> Design Opt> Design Variables
Define Length and Radius as design variable.

Add a Design variable.
Choose L as the parameter.

Type 0.3 m in the MIN field.
Type 1.5 m in the MAX field
Apply
Choose R as the parameter.

Type 0.01 m in the MIN field.
Type 0.05 m in the MAX field.
OK.
Close.
Define State Variables and the constraining limits.

Main Menu> Design Opt> State Variables.
Add a State variable.
Click on TEMPEND.
Type 40°C in the MAX field.
OK.
Define Objective Function and its tolerance.

Main Menu> Design Opt> Objective.
Click on TOTALHEAT.
OK.
4. Run the optimization

This step involves specifying run time controls and the optimiza-
tion method, saving the optimization database, and executing
the run.
Main Menu> Design Opt> Controls.
Change OPKEEP setting from “Do not save” to “Save.”

OK.
Main Menu> Design Opt> Method/Tool.

Choose Sub-Problem
OK
OK.
Main Menu> Design Opt> Opt Database> Save

To save the optimization database, type Fin.opt0

OK.
Main Menu> Design Opt> Run

OK.
Solution Converges
Click OK.
5. Review the results

In this step, you start by listing design sets, then graph the objec-
tive function and state variables versus set number.
Main Menu> Design Opt> Design Sets> List
Select ALL Sets.

OK.
Review results and Close.

Main Menu> Design Opt> Graphs/Tables
For X-variable select Set number

For Y-variable select ObjFunct
OK.
Review the Graph for Objective function
Main Menu> Design Opt> Graphs/Tables
For X-variable select Set number.

For Y-variable select TOTALHEAT.
Unselect OBJFUNCT.
OK.
Review the Graph.
6. Restore the best design

Main Menu> Design Opt> Opt Database> Save.
Type Fin.opt1.
OK.
Main Menu> Finish.

Type in Ansys Input Window.

Resume,Fin,bdb/post1 File,Fin,brth lplot.
Utility Menu> PlotCtrls> Style> Symmetry Expansion>

User-Specified Expansion.
Main Menu> General Postproc> Plot Results >Contour Plot>

Nodal Solu.
Review results.
7. Exit ANSYS.
Toolbar: Quit
Choose Quit—Select an option to Save! OK.
Bibliography
Cook, R.D. 1995. Finite Element Modeling for Stress Analysis, 1st ed. H
oboken,
NJ: John Wiley & Sons.
Incropera, F.P. 1985. Fundamentals of Heat and Mass Transfer, Example 3.7,
2nd ed., 104. Singapore: John Wiley & Sons.
Reddy, J.N. 1993. An Introduction to the Finite Element Method, New York:
McGrawHill.
Logan, D.L. 2001. A First Course in the Finite Element Method, 3rd ed. Mason,
OH: Thomas Learning Publishing.
Reddy, J.N. 1972. “Exact Solutions of Moderately Thick Laminated Shells.”
Journal Engineering Mechanics 110, no. 5, pp. 794–805.
Timoshenko, S. 1956. Strength of Material, Part II, Elementary Theory and
Problems, 3rd ed., 111. New York, NY: D. Van Nostrand Co., Inc.
About the Authors
Wael A. Altabey is an assistant professor in the department of mechanical

engineering, faculty of engineering, Alexandria University, Alexandria,
Egypt and has been a postdoctoral researcher at the International Institute
for Urban Systems Engineering, Southeast University, Nanjing, Jiangsu,
China. Dr Altabey’s research has been in the area of composite structures
and utilization of artificial neural networks for damage detection.
Email: wael.altabey@gmail.com

Systems Engineering, Southeast University, Nanjing, China. Dr Noori’s
research has been in the area of mechanical and random vibrations and
structural health monitoring.
Email: mnoori@outlook.com
Libin Wang is a professor and the dean of the school of civil engineering
at Nanjing Forestry University, in Nanjing, China. He has been an e ducator
and scholar, for over 20 years, and has taught the subject of finite element
analysis both at the undergraduate and graduate level. Dr Wang’s research
is in the field of engineering mechanics, hazard mitigation, and bridge
engineering, as well as timber structures.
Email: jhwlb@163.com
Index
A Continuous probability function,

Analysis file, ADPL programming 76
advantages and disadvantages, Convergence checking, 177
145–146 Correlation, 72
using command log files, Correlation matrix, 106–107
147–148 Cumulative distribution function
using log files, 146–147 (cdf), 77–78
ANSYS Parametric Design
Language (APDL) D
programming Design optimization, ANSYS
creating analysis file, 145–148 program
heat loss from cylindrical cooling general process for, 179–185
fin, 155–165 guidelines for performing,
stress analysis of bicycle wrench, 185–187
148–155 optimization design tools,
177–179
B optimization methods, 175–177
BEAM188, 29 sample optimization analysis,
BEAM189, 29 187–193
terminology and information
C flow, 173–174
cdf. See Cumulative distribution Design optimization tutorial
function approach and assumptions, 194
Ceramic matrix composites description, 193–194
(CMCs), 20 input file using log file, 194–197
Circular plate bending, 107–143 optimization problem, 197
CMCs. See Ceramic matrix solutions/results, 194
composites step-by-step analysis, 198–215
Compression, 22 summary of steps, 197–198
Constitutive matrices, 31 Design set, 174
Constraints, 170 Design variables (DVs), 170, 173,
Continuous distributions, 78–81 185
222 • Index
Direct Monte Carlo sampling, Linear layered structural shell

97–98 element (SHELL99), 27
DVs. See Design variables Lognormal distribution, 85–88
Dynamic analysis Loop, 174
description of, 1 Loop file, 174
harmonic analysis of structure,
1–11 M
modal analysis of structure, Mean value, 73
11–17 Median value, 74
static analysis vs., 1 Metal matrix composites (MMCs),
Dynamic loads, 1 19
MMCs. See Metal matrix
E composites
Exceedence values, 92 Modal analysis of structure
assigning loads and solving,
F 11–13
Factorial tool, 179 description of, 11
Feasible design, 174 postprocessing-viewing results,
Finite strain shell (SHELL181), 27 13–15
First-order method, 177 preprocessing-defining problem,
Flexural loads, 23 11
using reduced method, 16–17
G Modeling composites using
Gradient tool, 179 ANSYS
failure criteria, 33–35
H layered configuration, 29–33
Harmonic analysis of structure modeling and postprocessing
assigning loads and solving, 3–6 guidelines, 35–38
command log file, 3 proper element type selection,
description of, 1–2 26–29
postprocessing-viewing results, Monte Carlo techniques
6–11 advantages, 96–97
preprocessing-defining problem, direct sampling, 97–98
2 Latin hypercube sampling, 98
Heat loss from cylindrical cooling
fin, 155–165 N
Node offset, 32–33
I Nonlinear layered structural shell
Individual layer properties, 29–31 element (SHELL91), 27
Normal distribution, 81–83
L
Latin hypercube sampling, 98 O
Layered structural solid element Objective function, 174, 186
(SOLID191), 28–29 Optimization database, 174
Index • 223
Optimization data flow, 175 terminology, 72–74

Optimization design tools traditional (deterministic) vs.,
factorial tool, 179 71–72
gradient tool, 179 variables, 73
random tool, 178 Probabilistic design techniques
single loop analysis tool, 178 advantages, 96–97
sweep tool, 178–179 direct Monte Carlo sampling,
Optimization methods 97–98
description of, 175–176 Latin hypercube sampling, 98
first-order method, 177 Probabilistic model, 73
subproblem approximation Probabilistic sensitivities, 101–104
method, 176–177 Probability density function (pdf),
Optimization tree, 171 76–77
Optimization variables, 174 Probability distributions
Optimum design gallery of common continuous
applications, 172 distributions, 78–81
description, 167–169 lognormal distribution, 85–88
examples, 172 normal distribution, 81–83
fundamental concepts, 169–171 practical uses of, 75–76
uniform distribution, 83–85
P Weibull distribution, 88–91
pdf. See Probability density Problem formulation, 169
function Problem solution, 171
PMCs. See Polymer matrix
composites Q
Polymer matrix composites Quality, probabilistic design, 72
(PMCs)
definition of, 19 R
description of, 20–22 Random input variables, 72
loading of, 22–23 Random output variables, 73
with other structural materials, Random tool, 178
23–26 Random variable distribution
properties of, 20 exceedence values, 92
Postprocessing probabilistic mean values, 92
analysis measured data, 91
statistical postprocessing, no data, 93–95
99–101 output parameters, 95
trend postprocessing, 101–107 standard deviation, 92
PREP7, 181 Reduced method, modal analysis,
Probabilistic design analysis 16–17
circular plate bending, 107–143 Reliability, probabilistic design, 72
definition of, 69–70
reliability and quality issues, 72 S
steps using ANSYS, 74–75 Sample, 73
224 • Index
Sample optimization analysis, Static analysis vs. dynamic

187–193 analysis, 1
Scatter plots, 105–106 Statistical postprocessing
Shear loads, 22–23 technique
SHELL63, 28 cumulative distribution function,
SHELL91 (nonlinear layered 100–101
structural shell element), 27 histogram, 99
SHELL99 (linear layered print inverse probabilities, 101
structural shell element), 27 print probabilities, 101
SHELL181 (finite strain shell), 27 sample history, 99
Simply supported laminated plate Stress analysis of bicycle wrench,
under pressure 148–155
applying loads, 56–68 Subproblem approximation
approach and assumptions, 39 method, 176–177
step-by-step analysis, 40–56 SVs. See State variables
summary of steps, 39–40 Sweep tool, 178–179
Simulation, 73
Single loop analysis tool, 178 T
SOLID46 (3-D layered structural Tension, 22
shell element), 27–28 3-D layered structural shell
SOLID65, 29 element (SOLID46), 27–28
SOLID95, 28 Trend postprocessing technique
SOLID191 (layered structural correlation matrix, 106–107
solid element), 28–29 scatter plots, 105–106
Standard deviation, 74 sensitivities, 101–104
Standard lognormal distribution,
86, 89 U
Standard normal distribution, 81 Uniform distribution, 83–85
Standard uniform distribution, 83
State variables (SVs), 174, W
185–186 Weibull distribution, 88–91
Other titles in our Sustainable Structural
Systems COLLECTION
Mohammad Noori, Editor
Numerical Structural Analysis

by Steven O’Hara and Carisa H. Ramming
A Systems Approach to Modeling Community Development Projects

by Bernard Amadei
Seismic Analysis and Design Using the Endurance Time Method,

Volume I: Concepts and Development
by H.E. Estekanchi and H.A. Vafai
Seismic Analysis and Design Using the Endurance Time Method,

Volume II: Advanced Topics and Application
by H.E. Estekanchi and H.A. Vafai
Momentum Press is one of the leading book publishers in the field of engineering,
mathematics, health, and applied sciences. Momentum Press offers over 30 collections,
including Aerospace, Biomedical, Civil, Environmental, Nanomaterials, Geotechnical,
and many others.
Momentum Press is actively seeking collection editors as well as authors. For more
information about becoming an MP author or collection editor, please visit
http://www.momentumpress.net/contact
Announcing Digital Content Crafted by Librarians

Momentum Press offers digital content as authoritative treatments of advanced engineering top-
ics by leaders in their field. Hosted on ebrary, MP provides practitioners, researchers, faculty,
and students in engineering, science, and industry with innovative electronic content in sensors
and controls engineering, advanced energy engineering, manufacturing, and materials science.
Momentum Press offers library-friendly terms:
• perpetual access for a one-time fee

• no subscriptions or access fees required
• unlimited concurrent usage permitted
• downloadable PDFs provided
• free MARC records included
• free trials
The Momentum Press digital library is very affordable, with no obligation to buy in future years.
For more information, please visit www.momentumpress.net/library or to set up a trial in the US,
please contact mpsales@globalepress.com.
EBOOKS Using ANSYS for Finite Element
ALTABEY • NOORI • WANG

FOR THE Analysis SUSTAINABLE STRUCTURAL
ENGINEERING
Dynamic, Probabilistic Design and Heat
SYSTEMS COLLECTION
LIBRARY Mohammad Noori, Editor
Transfer Analysis, Volume II
Create your own
Wael A. Altabey • Mohammad Noori • Libin Wang
Customized Content
Bundle—the more Over the past two decades, the use of finite element method as a
books you buy, design tool has grown rapidly. Easy to use commercial software, such as
Using ANSYS
the greater your ANSYS, have become common tools in the hands of students as well as
practicing engineers. The objective of this book is to demonstrate the
discount!
use of one of the most commonly used Finite Element Analysis software,
THE CONTENT
• Manufacturing
ANSYS, for linear static, dynamic, and thermal analysis through a series
of tutorials and examples. Some of the topics covered in these tutorials
include development of beam, frames, and Grid Equations; 2-D elastic-
for Finite
Engineering
• Mechanical
ity problems; dynamic analysis; composites, and heat transfer problems.
Element
Using ANSYS for Finite Element Analysis, Volume II

& Chemical
Analysis
These simple, yet, fundamental tutorials are expected to assist the users
Engineering with the better understanding of finite element modeling, how to control
• Materials Science modeling errors, and the use of the FEM in designing complex load
& Engineering bearing components and structures. These tutorials would supplement
• Civil & a course in basic finite element or can be used by practicing engineers
Environmental who may not have the advanced training in finite element analysis. Dynamic, Probabilistic
Engineering
• Advanced Energy Wael A. Altabey is an assistant professor in the department of Design and Heat
Technologies mechanical engineering, faculty of engineering, Alexandria University,
Alexandria, Egypt and has been a postdoctoral researcher at the Transfer Analysis
THE TERMS International Institute for Urban Systems Engineering, Southeast
• Perpetual access for University, Nanjing, China.
a one time fee
Volume II
• No subscriptions or
access fees
• Unlimited
concurrent usage
• Downloadable PDFs
• Free MARC records
Systems Engineering, Southeast University, Nanjing, China. Wael A. Altabey
For further information,
Libin Wang is a professor and the dean of the school of civil engineer- Mohammad Noori
Libin Wang
ing at Nanjing Forestry University, in Nanjing, China. He has been an
a free trial, or to order,
contact: educator and scholar, for over 20 years, and has taught the subject of
sales@momentumpress.net finite element analysis both at the undergraduate and graduate level.
ISBN: 978-1-94708-322-6

Using ANSYS For Finite Element Analysis, Volume II Dynamic, Probabilistic Design and Heat Transfer Analysis PDF

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Using ANSYS For Finite Element Analysis, Volume II Dynamic, Probabilistic Design and Heat Transfer Analysis PDF

Hochgeladen von

Copyright:

Verfügbare Formate

EBOOKS Using ANSYS for Finite Element

ALTABEY • NOORI • WANG

Using ANSYS for Finite Element Analysis, Volume II

Wael A. Altabey, Mohammad Noori,

MOMENTUM PRESS, LLC, NEW YORK

Copyright © Momentum Press®, LLC, 2018.

All rights reserved. No part of this publication may be reproduced, stored

First published by Momentum Press®, LLC

ISBN-13: 978-1-94708-322-6 (print)

Momentum Press Sustainable Structural Systems Collection

Collection ISSN: 2376-5119 (print)

Cover and interior design by Exeter Premedia Services Private Ltd.,

Printed in the United States of America

Finite Element Method (FEM) is a well-established technique for analyz-

ANSYS, composite materials, Dynamics, Failure analysis, Fatigue loads,

5 Design Optimization 167

Figure 2.1. Illustrating the combined effect on modulus of the

Figure 3.3. The normal cumulative distribution function (cdf). 78

Finite element method (FEM) is a well-established technique for analyz-

The main objective of this book is to serve as a practical tutorial to

In addition, it could be used for computer sessions of short courses on

Structural analysis is mainly concerned with finding out the behavior of a

1.1 Tutorial 1: Harmonic Analysis

In this tutorial, the harmonic analysis of a cantilever beam will be

structure to harmonically time-varying loads. This tutorial was created

Modulus of elasticity (E) = 206800(106) N/m2

Density = 7830 kg/m3

We will now conduct a harmonic forced response test by applying a

ANSYS provides three methods for conducting a harmonic analy-

1.1.1 Step-by-Step ANSYS Solution

1.1.1.1 Preprocessing: Defining the Problem

Student should be able to make simple cantilever model himself or other-

1.1.1.2 The Command Log File

/TITLE, Dynamic Analysis

1.1.1.3 Solution: Assigning Loads and Solving

1. Define analysis type (Harmonic)

The following window will appear. Use the default settings

Constrain all DOF as shown in the above window.

being applied to the structure as these loads could be in or out

5. Set the frequency range

You should now have the following in the ANSYS Graphics

6. Solve the system

1.1.1.4 Postprocessing: Viewing the Results

1. Open the TimeHist Processing (POST26) Menu

And the following window should pop up:

We are interested in the Nodal Solution > DOF Solution >

3. List Stored Variables

4. Plot UY vs. frequency

Note that we get peaks at frequencies of approximately 8.3

As marked by an “A” in the above window, change the Y-axis

For ANSYS version lower than 7.0, the “Variable Viewer”

1.2 Tutorial 2: Modal Analysis of

In this tutorial, the modal analysis of a cantilever beam will be addressed.

Modulus of elasticity (E) = 206800(106) N/m2

Density = 7830 kg/m3

1.2.1 Step-by-Step ANSYS Solution

1.2.1.1 Preprocessing: Defining the Problem

The simple cantilever beam is same as used in the previous tutorial.

1.2.1.2 Solution: Assigning Loads and Solving

1. Define analysis type

The following window will appear:

As shown, select the Subspace method and enter 5 in the “No.

For a better understanding of these options see the Commands

5 Design Optimization 167

Figure 2.1. Illustrating the combined effect on modulus of the

Figure 3.3. The normal cumulative distribution function (cdf). 78

1.1 Tutorial 1: Harmonic Analysis

1.2 Tutorial 2: Modal Analysis of

250 uniform-thickness layers per element. Alternately, the element allows