Teaching Finite Element Analysis For Mechanical Undergraduate Students
Teaching Finite Element Analysis For Mechanical Undergraduate Students
Teaching Finite Element Analysis For Mechanical Undergraduate Students
American
c Society for Engineering Education, 2019
TEACHING FINITE ELEMENT ANALYSIS FOR MECHANICAL
UNDERGRADUATE STUDENTS
ABSTRACT
Since Finite Element Analysis (FEA) has become a daily tool for design engineers in industries,
mechanical engineering students must learn this tool during their education. MECH4200-
Simulation-Based Design teaches how to apply FEA commercial software for mechanical
undergraduate students. Teaching how to use the FEA commercial software is the main
contents of this type of course. There are lots of approaches to teaching this type of course.
Authors believed that understanding of the fundamental concepts of FEA theory would
significantly help students to have a better understanding of FEA theory and to utilize the
application of an FEA commercial software. In 2018 spring semester, authors proposed and
implemented a teaching approach which included teaching fundamental concepts of FEA theory,
teaching its commercial software and implementation of it in class design projects. Although
some students complained about the complexity of fundamental concepts of FEA theory and
tedious theoretical calculations, 92.3% of students agreed that teaching the fundamental concepts
of FEA theory helped them to have a better understanding of the FEA commercial software.
92.3% of students agreed that teaching the fundamental concepts of FEA theory should be kept
as part of the course. At the end of the course, we asked students to take the CSWA-S
Certification, 60% of students in the section with the proposed approach passed the certification
exam while other sections had an average 35.9% of pass rates. Our experience indicated that
teaching the fundamental concepts of FEA theory significantly helped students to have a better
understanding of FEA application and to facilitate them to use FEA commercial software.
1. INTRODUCTION
No. 1 design criteria for mechanical design is designing safe components. It is well known that
theoretical calculations of stress and strain can be conducted only on components with simple
geometries under simple loading conditions. But, FEA (Finite Element Analysis) simulation can
calculate the stress/strain of components and assemblies with complicated geometries under
complicated loading conditions. In the current industry practice of mechanical design, the
prototype of design may not be manufactured for testing unless the design has been approved to
be safe through FEA analysis because such prototype with failure through FEA simulation has an
extremely high possibility of failure and the cost of manufacturing and testing prototype might
be very high. Because FEA is a daily tool for design engineers in many industries [1.2],
mechanical engineering students must know how to use this tool during their education. Many
engineering colleges have offered FEA-related courses to their engineering students [3,4,5,6,7]
FEA commercial software is a version of solid mechanics and is expressed in the form of
computer codes. It is commonly accepted that teaching pure theory and programming of the
1
Finite Element Method (FEM) is for Master-Degree or Ph.D. candidate students and has not been
taught for undergraduate engineering students. Even though the FEM theory and FEA
commercial software are extremely complicated, most FEA commercial software has user-
friendly interfaces and is easy to use. There were many ways to introduce FEA software to
students. One of the common ways was to let students gradually learn FEA software, including
introducing the FEA software during engineering statics in their sophomore year [8,9], mechanics
of materials in their sophomore year [9, 10,11] and design of machine elements in their junior year
[12, 13,14]
. Finally, a full course in finite element analysis was offered for junior or senior
engineering students [1~7].
There were lots of approaches to teaching the FEA course. One frequently-used approach was
focusing on teaching the application of FEA commercial software through case studies. In this
approach, various important interface settings of the FEA commercial software were explained
in detail and practiced through homework and case studies. This was the approach we used in the
past. Through direct observations and conversation with students, we found that: some students
had difficulty choosing proper settings through the software interfaces; some had no clue why
the settings were required, and some had some difficulty properly interpreting the simulation
results. They could not clearly explain that reducing element size would increase the accuracy,
the error of FEA simulation results might mainly come from boundary conditions (loading and
restraints) and the relative errors of the displacement was much smaller than that of the stress in
a convergence study. Although these had been clearly explained during lecturing, students did
not resonate with the explanations because they didn’t understand them. Another frequently-
used approach was a combination of the introduction of the fundamental FEA theory and
utilization of commercial FEA software. It is a common belief that understanding some
fundamental concepts of FEM theory will significantly improve students’ ability to apply FEA
software in engineering design [4, 5, 15, 16, 17]. In 2018 spring semester, authors adapted this
approach to teaching our FEA course. The feedback and results were very encouraged. This
paper will present what we did by using this teaching approach.
Our mechanical engineering program offers a required FEA course in the senior year. We
strongly believed that understanding the fundamental concepts of FEM theory would
significantly help students to have a better understanding of FEA simulation and to facilitate the
application of FEA commercial software. In the spring semester of 2018, the teaching approach
in this course included three key contents: (1) teaching fundamental concepts of finite element
analysis theory, (2) application of FEA commercial software, and (3) implementation of gained
skills in class design projects. This paper will discuss and present what we did in teaching an
FEA course for undergraduate mechanical students. The main focus presented in this paper will
be for teaching fundamental concepts of FEM theory utilized in the first 3 weeks of the semester.
At the end of the course in 2018 spring semester, students participated in a class survey and took
the CSWA-S certification exam (Certified SOLIDWORKS Associate – Simulation). Analysis of
the collected data indicated that teaching the fundamental concepts of FEA theory significantly
helped students to have a better understanding of FEA and facilitated them to use FEA
commercial software.
2
2.1 THE BRIEF DESCRIPTION OF MECH4200-SIMULATION-BASED DESIGN
The MECH4200-Simulation Based Design course utilizes lectures, demonstrations and case
studies for the teaching how to use FEA commercial software for stress-strain analysis of
components and assemblies. It is a required course for senior mechanical engineering students
and has 2-lecture hours and 4-lab hours for a total of 4 credits. SolidWorks Simulation is the
FEA commercial software for this course. SolidWorks is the computer-aided engineering
computer program used in our mechanical engineering program and offers a complete platform
for all purposes of mechanical engineering needs such as 3D modeling, drawing, finite element
analysis, and computer-aided manufacturing. The mechanical engineering program at
Wentworth has 1000 educational licenses to ensure every mechanical student has full access to
every module in SolidWorks academic suit. Prerequisite courses teach the use SolidWorks to
create models and drawings starting in their freshman year including simple analysis using FEA
in courses such as Engineering Statics, Mechanics of Materials and Design of Machine
Elements. In Mechanics of Material course during their sophomore year, they conducted three
finite element simulation labs [10] in which they treated the SolidWorks simulation as a virtual
instrument and used it to simulate and visualize stress/strain of bars under axial loading, shafts
under torsion and beams under lateral force and bending. In Design of Machine Elements course
in their junior year, they used SolidWorks simulation to simulate and visualize the stress and
strain of components with irregular shapes. In these two cases, instructors provide detailed
procedures and demonstrate with explained examples, students mainly followed instructors’
demonstrations to conduct their FEA simulations.
In the spring semester of 2018, the newly-proposed teaching approach for MECH4200-
Simulation Based Design included three key components: (1) teaching fundamental concepts of
FEM theory, (2) teaching how to utilize commercial FEA software, SolidWorks Simulation, for
stress /strain analysis on components and assemblies and (3) the implementation of SolidWorks
simulation for two design projects.
The first three weeks out of a total of 15 weeks in the Simulation-Based Design course was
devoted to teaching fundamental concepts of the FEM. The purpose of this was to help students
gain a basic understanding of FEM theory and allow students to have a better understanding of
the various settings in the FEA software.
The basic concepts or a general procedure for FEM theory [18] can be summarized as followings:
(1) The FEM is a numerical approximation technique for solving deformation of an object. In
this technique, after the x, y and z components of the deformation: 𝑢𝑢(𝑥𝑥, 𝑦𝑦, 𝑧𝑧), 𝑣𝑣(𝑥𝑥, 𝑦𝑦, 𝑧𝑧) and
w(𝑥𝑥, 𝑦𝑦, 𝑧𝑧) of an object are determined, stress/strain of the object can be calculated through
applicationthe of Hooke’s law.
(2) The object is divided into a lot of small divisions (finite elements).
(3) Each element has several nodes along its edges or on its outer surfaces. The adjacent finite
elements are joined by the shared nodes. Therefore, the object is replaced by or represented
as an assembly of finite elements connected at the shared nodes.
3
(4) Approximate deformation or shape functions inside a finite element between nodes are
hypothesized and then determined in terms of nodal deformations.
(5) Different techniques such as the method of least squares and the minimum potential energy
principle are used to make the errors minimum and thus to convert a continuous physical
problem (an object) into a set of linear algebraic equations, which are expressed as a linear
function of nodal deformation.
(6) The loading conditions and restraint conditions of the object are applied to the set of linear
algebraic equations. Then the set of the linear algebraic equations can be solved to obtain the
nodal deformation of each node.
(7) Once the nodal deformations of each node inside an element are known, Hooke’s law can be
applied to each element using the element’s shape functions to calculate stress /strain of any
point of the object.
There are 6 key fundamental concepts of the FEM theory are: (1) Types of elements, (2)
Approximate deformation or shape functions of an element, (3) Element stiffness property which
links together forces, nodal deformations, material properties and element geometry, (4) The
global linear algebraic equations in which the shared nodes are joined, (5) Application of the
loading and restraints in the global linear algebraic equation, and (6) Stress/strain calculation.
The best way to teach and to demonstrate these 6 key fundamental concepts of the finite element
method is through simple examples.
Since the FEM theory can be considered extremely complicated for undergraduate students, a
simplified treatment was warranted for students to familiarize themselves with some
fundamental concepts of FEM theory. The authors collaborated to develop a simplified and
straight forward approach to do this in three weeks. It took a long time and lots of work to
summarize these examples for teaching the fundamental concepts of FEM theory. These
examples will be included and displayed in Appendixes of this paper and might be useful for a
reference. In 2018 spring semester, we developed three lectures and three labs to discuss the
fundamental concepts of the FEM theory.
• Lecture one was about the various types of elements and how to formulate approximate
deformation functions and then how to determine the shape functions.
• The second lecture was about using the minimum potential energy principle to make the
errors minimum and thus to link internal forces, deformation, material property, and
geometry together and how to form global linear algebraic equations.
• The third lecture provided a demonstration example of a 1D stepped bar under axial loading
including determining boundary conditions, stiffness characteristics and all steps required to
solve theoretical equations using hand calculations.
After each lecture, we conducted one lab section dedicated to working on examples together with
students prior to students working on homework assignments during these faculty-guided labs.
In the fourth week, an exam was administered to assess students’ understanding of the
fundamental concepts of FEM theory. The date for this exam was announced on the first date of
the class to ensure students focused on the topics presented in lectures and labs. The following
are concise descriptions of the key topics presented to the students for teaching the fundamental
concepts of FEM theory.
1) Types of Elements
4
One of the fundamental concepts of FEM theory is to divide an object into many smaller finite
elements. A node in a finite element is a coordinate point which is on the edges or on the outer
faces of the elements, which are sketched as a round dot in Figure 1. One element can have
multiple nodes but must have at least 2. The variety of element types used in FEA analysis were
presented and discussed. For example, a bar under axial loading can be simplified as 1D
elements because stress and strain are exactly the same in the same cross-section. For a 2-D
plane stress problem, we use 2D elements because the stress and strain are the same in the
thickness or depth direction. Sheet metal problems it can be represented by a shell element
because the stress /strain in the thickness direction is only due to a bending moment. When
considering the general case, the object can be divided into 3D elements. Several element types
shown to the students are depicted in Figure 1.
Deformation function of a finite element is a continuous function of the coordinates and valid
only inside this element and are only solely dependent upon the deformation of nodes inside the
element. In FEM theory, we typically use the shape functions to express the approximate
deformation functions. In the following, we will only use the x-component of deformation
function as an example. The same method can be used for y- and z- component of deformation
functions.
𝑛𝑛
5
Where 𝑢𝑢(𝑥𝑥, 𝑦𝑦, 𝑧𝑧) is the x-component of the deformation at the coordinate point (𝑥𝑥, 𝑦𝑦, 𝑧𝑧). 𝑢𝑢𝑖𝑖 is the
x-component of deformation at the ith node point. n is the number of total nodes of the element.
𝑁𝑁𝑖𝑖 (𝑥𝑥, 𝑦𝑦, 𝑧𝑧) is the ith shape function for the ith node and is an interpolation function. The shape
functions have the following three properties.
We used four examples during lecturing to explain how to build deformation functions and to
explain the shape functions. These four examples are displayed for the reference in Appendix A.
FEM can provide a numerical approximate result. Several different techniques such as the
minimum potential energy principle can be used in FEM to make the errors minimum. Through
the minimum potential energy principle, we can establish the relationship among internal forces
at nodes, nodal deformation, material property and element geometry. This relationship is
typically called as element properties. We used a 1D element with two nodes during lecturing to
show this process. This example is displayed for the reference in Appendix B.
Now that the object has been divided into many independent finite elements. It should be
emphasized that the shared nodes on adjacent elements are joined together and the shared nodes
on adjacent elements have the exact same deformation. So, the FEM does not analyze an
original object but analyzes the assembly of many finite elements which are joined through the
shared nodes on adjacent elements. We used a 1D-element example during lecturing to explain
this process. This example is displayed for the reference in Appendix C.
The best way to provide students with a better understanding of the fundamental concepts of
FEM theory is to manually solve a stress /strain problem by using the FEM theory. We used one
lecture to explain and to demonstrate a 1D-element example. This 1D-element example we did
during the lecture, is displayed for the reference in Appendix D. We asked students to complete
a similar 1D-element example for calculating stress/strain with different boundary conditions
during a lab.
The primary purpose of using SolidWorks simulation in lectures and labs is to demonstrate how
to correctly use FEA software to run stress-strain analysis on components and assemblies under
6
different boundary and loading conditions. The students are presented with a general procedure
which can be applied to almost any commercially available FEA software.
The general procedure for FEA simulation in SolidWorks simulation with a concise description
of each step is as follows:
(1) Pre-processing: The main tasks in this step are to remove unimportant features which have
little effect on stress-strain analysis and to ensure 3D models can mesh.
(2) Define the type of analysis: SolidWorks simulation is an FEA simulation module of the
SolidWorks platform and contains a variety of simulation types such as static analysis,
thermal analysis, fatigue analysis, frequency analysis, linear dynamic analysis and more [19].
For MECH4200-Simulation-Based Design, the main focus was static analysis.
(3) Create a study: This creates a study name for simulation and all saved documents. Many
different types of simulation could be run with different boundary conditions and element
sizes.
(4) Define materials for each component: A proper material name with corresponding
mechanical properties must be assigned to each component in order to run FEA simulation.
(5) Define connections or interfaces: For a single component’s analysis, there is no need to
specify the connections. For assemblies, however, contact relationships among components
must be defined.
(6) Define restrains and loads: This is to specify loading and supporting conditions and
explained as defining the boundary conditions.
(7) Meshing: In this step, by the proper settings for meshing such as meshing control and global
meshing size, the 3D modeling will automatically mesh into thousands of small elements.
(8) Run the analysis: After settings are fully defined, the FEA software can solve the nodal
deformations of all nodes. After the “run the analysis” is completed, all information about
stress, deformation, and strain of the object is available.
(9) Post-processing: The main task of this step is to display stress, strain, and deformation of
objects under consideration.
(10) Verification / Convergence: This step is to compare the FEA simulation results with
theoretical calculation results and the results obtained from testing for the purpose of
verification if available.
In teaching the MECH4200 Simulation Based Design course, we followed the above procedure
to teach students how to appropriately use SolidWorks Simulation and to run the stress-strain
analysis on both components and assemblies. Lectures expanded on the following topics: (1)
Pre-process, (2) Type of elements, (3) FEA errors and convergence, (4) Meshing techniques, (5)
Type of contacts and simplified theoretical connectors, (6) Loading and restraints, (7) Post-
process and (8) Verification. During lectures and labs, detailed demonstrations of why and how
to change every important setting were presented. Students then practiced and implemented
what they learned through weekly case studies and two design projects
Conducting case study assignments was one way to let students practice and implement what
they learned. Typically, each assignment was targeted to a specific skill or setting and most of
the time the required setting for each homework was fully specified so that students could
7
practice a specific skill and obtain the expected results. Conducting class design projects was
another more effective way to practice and to implement what they learned to calculate stress-
strain in an application. When conducting the class design projects, students were required to
choose appropriate settings for the FEA simulation by themselves.
The course has two class design projects: a minor design project and a major design project.
The individual minor project was designed to focus on skills for running an FEA simulation in a
project-based environment. Students not only implemented necessary FEA simulation skills but
also needed to make some decisions for a choice of setting. The ideal topic of a minor project
should an issue which cannot be solved by a simple closed-form theoretical hand calculation. But
there are the empirical formulas/curves or tables for providing acceptable solutions. The minor
project was a 3-week individual project and was released in week 6 after lecturing and case
studies for FEA simulation on components were completed. One minor project utilized the FEA
simulation of stress concentration factors of components with various geometries and different
loading for comparison to standard tables. Even though stress concentration factors cannot be
obtained using a simple theoretical calculation, many handbooks and textbooks provide a set of
curves to verify FEA results are within reasonable limits compared to theory.
The major project was a team-based project with 2~4 team members. It was released in week 9
after lecturing on FEA simulation including assemblies was completed. After the major project
was released, one-hour lecture time per week was used to discuss the major project and to
answer questions. The lecture was not to show how to run the FEA simulation of the major
project but explained some possible approaches since a variety of potential approaches existed.
One two-hour lab per week was devoted to working on the major project in a classroom while
the instructor was available to provide guidance. Each design team was also required to spend at
least an additional two hours outside of the classroom working on the major project. The team-
based major design project was FEA simulation on a real product or a sub-assembly of a real
product. A real product greatly increases students’ interests in conducting FEA simulation and
redesigning the product because the product requires real design constraints and specifications.
The experience they gained through this project could be applied to almost any industrial hands-
on real FEA simulation experience. One example of a major design project was the FEA
simulation on an engine hoist, which is a common tool used in small workshops or homes to lift
and move heavy objects.
At the beginning of the course, students were reluctant to learn the fundamental concepts of
finite element method theory. But they were forced to learn and study it because there was one
exam covering this topic. After they learned fundamental concepts of FEM theory and manually
conducting strain - strain analysis of bars under axial loading by using FEM theory, they seemed
to appreciate learning fundamental concepts of FEM theory and had a better understanding of
fundamental concepts of FEM theory. Based on direct observation, conversation, and students’
feedback, the following were some perceived benefits of learning fundamental concepts of FEM
theory.
• Students had a better understanding of the various types of elements because they went
through several examples to build the approximate deformation functions of elements.
8
• They developed a better understanding of the errors encountered by the use of commercial
FEA software. They knew that the FEA software would not provide exact accurate stress-
strain results but provided approximate displacement results because the deformation
functions of each element are only an approximate function.
• Students had a better understanding of meshing in the FEA software, which is a very
important skill for applying FEA commercial software. They developed an understanding
that the smaller size of the element would obtain much better results. They were able to
apply the H-method and P-method for meshing in SolidWorks Simulation software [19]. The
H-method is an automatic reduction of the size of elements to reach the convergence target.
The P-method is an automatic increase in the order of the deformation functions while the
size of the element is kept the same.
• Students knew that the FEA is actually to solve the deformations of each node or the
deformation of each element. The stress/strain will be just a by-product, that is, stress and
strain can be calculated after the deformation of each element are known.
• Learning the fundamental concepts of FEM theory and conducting manual stress-strain
calculations by using the FEM theory significantly improved their understanding of applying
the appropriate loading conditions and restraints.
• The manual stress-strain calculation of bars under axial loading by using FEM theory
significantly improved their understanding of the procedure of applying FEA commercial
software for stress-strain simulation of a component. They understood the purposes of key
steps of the procedure.
In order to assess the effectiveness of teaching the fundamental concepts of FEM theory in the
FEA course, we conducted a class survey during a lab at the end of the class. The survey data is
presented in Table 1 and Table 2. Per Table 1, 92.3% of students agreed that providing lectures
regarding fundamental concepts of the FEA method facilitated their understanding of the
appropriate implementation of the FEA method in SolidWorks Simulation. Per Table 2, 92.3%
of students agreed that continuing lectures of fundamental concepts of the FEA method should
be kept as part of the course of MECH4200 Simulation Based design. The survey results were
very positive and were in agreement with the finding in lots of previous publications [5, 6, 16, 17],
that understanding some fundamental concepts of FEM theory would significantly improve
students’ ability to apply FEA software in engineering design.
9
At the end of the MECH4200 Simulation Based Design course, students were asked to take the
CSWA-S (The Certified SOLIDWORKS Associate – Simulation) certification exam. The
CSWA-S certification exam is used to evaluate the students’ understanding of the principles of
the Finite Element Method (FEM) and their ability to use the SolidWorks Simulation to conduct
the stress and strain simulation. The CSWA-S certification exam is hosted by the SolidWorks
company. It is a two-and-half-hour certification test with a passing score of 70 out of 100 points.
The administration fee for taking CSWA-S certification exam is $100. But our mechanical
engineering program has 1000 education license and partners with the SolidWorks. Our students
can take this exam with a waiver of this administration fee. In 2018 spring, there were a total of
four sections of MECH4200 class with different instructors. We did not use the textbook for this
course. However, we use the same syllabi to teach the same contents. In the author’s section of
the class, teaching the fundamental concepts of FEM theory was added as an additional method
to help students to have a better understanding of fundamental concepts of FEA theory. The
other sections were still a focus on the application of SolidWorks. The section with this
approach of teaching the fundamental concepts of FEM theory had a pass rate of 60% (15 out of
total 25 students) while the other four sections had an average pass rate of 35.9%, in which the
focus of application of the SolidWorks Simulation was used. This result indicated that teaching
the fundamental concepts of FEM theory facilitated students to use FEA commercial software.
Our experience with teaching FEA to undergraduate mechanical students indicated that teaching
the fundamental concepts of FEM theory significantly helped them develop a better
understanding of fundamental concepts of FEM theory and facilitated them to apply commercial
FEA software for stress/strain simulations on components and assemblies.
4. REFERENCES
[1] Lee, R., Capece, V. R., and Baker, J., “Integration Of Finite Element Software In
Undergraduate Engineering Courses,” 2001 ASEE Annual Conference and Exposition, June
16-19, Albuquerque, New Mexico.
[2] Zecher, J., “Teaching Finite Element Analysis In An Met Program,” 2002 ASEE Annual
Conference and Exposition, June 16-29, Montreal, Canada.
[3] Le, X., Moazed, A. R., and Duva, A. W., “The Design Projects for the Simulation-Based
Design Course,” 2016 ASEE Annual Conference and Exposition, June 26-29, New Orleans,
Louisiana.
[4] Papadopoulos, J. M., Papadopoulos, C., and Prantil, V. C., "Philosophy of Integrating FEA
Practice Throughout the Undergraduate CE/ME Curriculum," 2011 ASEE Annual
Conference and Exposition, June 26-29, Vancouver, BC, 2011.
[5] Salamon, N., Wagle, G., and Lissenden, C., “Applications Of Finite Element Analysis For
Undergraduates,” 2002 ASEE Annual Conference and Exposition, June 16-19, Montreal,
Canada.
[6] Prantil, V., Musto, J., and Howard, W. (2001, June), “Finite Element Analysis In A
Mechanics Course Sequence,” 2011 ASEE Annual Conference and Exposition, June 26-29,
Vancouver, BC, 2011.
[7] Shih, R., “The Applied Finite Element Analysis Course At Oregon Institute Of Technology,”
2007 ASEE Annual Conference and Exposition, June 24-27, Honolulu, Hawaii.
10
[8] Pike, M., “Introducing Finite Element Analysis In Statics,” 2001 ASEE Annual Conference
and Exposition, June 16-19, Albuquerque, New Mexico.
[9] Chaphalkar, P., and Blekhman, D., “Introducing Finite Element Analysis In The First Course
Of Statics And Solid Mechanics,” 2007 ASEE Annual Conference and Exposition, June 24-
27, Honolulu, Hawaii.
[10] Le, X., Olia, M., Moazed, A., and Roberts, R. L., “Design a New Set of Strength Labs for
the Course, 'Mechanics of Materials',” 2016 ASEE Annual Conference and Exposition, June
26-29, New Orleans, Louisiana.
[11] Steif, P., “Use of Simplified FEA To Enhance Visualization In Mechanics Of Materials,”
2004 ASEE Annual Conference and Exposition, June 20-23, Salt Lake City, Utah.
[12] Le, X., Duva, A. W., and Jackson, M., “The Balance of Theory, Simulation, and Projects
for Mechanical Component Design Course,” 2014 ASEE Annual Conference and
Exposition, Indianapolis, June 15-18, Indiana.
[13] Reffeor, W. S., “Using FEA as a Pedagogical Tool for Teaching Machine Component
Design,” 2018 ASEE Annual Conference and Exposition, June 24-27, Salt Lake City, Utah.
[14] Ryan, R., “Use of A Finite Element Stress Analysis Program To Enhance Learning In A
Machine Design Course,” 2004 ASEE Annual Conference and Exposition, June 20-23, Salt
Lake City, Utah.
[15] Coyle, M., and Keel, C., “Teaching Finite Element Analysis To Second Year Students,”
2001 ASEE Annual Conference and Exposition, June 16-19, Albuquerque, New Mexico.
[16] N. Smith and J. Davis, "Connecting Theory and Software: Experience with an
Undergraduate Finite Element Course," 2015 ASEE Annual Conference and Exposition,
June 14-17, Seattle
[17] Bearden, K., and Bandstra, J., “Finite Element Analysis for The Mechanical Engineering
Technology Student,” 2001 ASEE Annual Conference and Exposition, June 16-19,
Albuquerque, New Mexico.
[18] O.C. Zienkiewicz, R.L. Taylor, J.Z. Zhu, “Finite element method: its basis and
fundamentals,” Amsterdam; Boston: Butterworth-Heinemann, 2013
[19] Paul Kuroski, “Engineering Analysis with SOLIDWORKS Simulation 2017”, 2017, SDC
publications
Example 1: Build the deformation function of a 1D element with two nodes and element
length 𝐿𝐿𝑝𝑝 as shown in Figure 2.
11
At node 1, 𝑥𝑥 = 0 and 𝑢𝑢(𝑥𝑥 = 0) = 𝑢𝑢1, so 𝑎𝑎1 + 𝑎𝑎2 × 0 = 𝑎𝑎1 = 𝑢𝑢1 (6)
At node 2, 𝑥𝑥 = 𝐿𝐿𝑒𝑒 and 𝑢𝑢(𝑥𝑥 = 𝐿𝐿𝑒𝑒 ) = 𝑢𝑢2 , so 𝑎𝑎1 + 𝑎𝑎2 × 𝐿𝐿𝑒𝑒 = 𝑎𝑎1 + 𝑎𝑎2 𝐿𝐿𝑒𝑒 = 𝑢𝑢2 (7)
From Equations (6) and (7), we have: 𝑎𝑎1 = 𝑢𝑢1 and 𝑎𝑎2 = (𝑢𝑢2 − 𝑢𝑢1 )/𝐿𝐿𝑝𝑝 . Per Equation (5), the
deformation function for this 1D 2 node element is:
(𝑢𝑢2 − 𝑢𝑢1 )
𝑢𝑢(𝑥𝑥) = 𝑢𝑢1 + 𝑥𝑥 0 ≤ 𝑥𝑥 ≤ 𝐿𝐿𝑝𝑝 (8)
𝐿𝐿𝑝𝑝
Per Equation (1) and Equation (8), we can determine the shape functions:
𝑢𝑢(𝑥𝑥) = 𝑁𝑁1 (𝑥𝑥)𝑢𝑢1 + 𝑁𝑁2 (𝑥𝑥)𝑢𝑢2 (9)
𝑥𝑥 𝑥𝑥
𝑁𝑁1 (𝑥𝑥) = 1 − , 𝑁𝑁2 (𝑥𝑥) = (10)
𝐿𝐿𝑝𝑝 𝐿𝐿𝑝𝑝
Example 2: Build the deformation function of a 1D element with three nodes. In a local
coordinate system, the coordinate values for node 1, node 2 and node 3 are -1, 0, and 1,
respectively as shown in Figure 3.
1 1
𝑁𝑁1 (𝜉𝜉) = − 𝜉𝜉(1 − 𝜉𝜉); 𝑁𝑁2 (𝜉𝜉) = (1 − 𝜉𝜉)(1 + 𝜉𝜉); 𝑁𝑁3 (𝜉𝜉) = 𝜉𝜉(1 + 𝜉𝜉)
2 2
Example 3: Build the deformation function and determine the corresponding shape functions of
a 2D triangle element with three nodes as shown in Figure 4. The coordinates for the node 1,
node 2 and node 3 are (0,1), (0,0) and (-1,0), respectively.
12
Since the 2D triangle element has three nodes, the deformation function can be assumed to be a
linear function: 𝑢𝑢(𝑥𝑥, 𝑦𝑦) = 𝑎𝑎1 + 𝑎𝑎2 𝑥𝑥 + 𝑎𝑎3 𝑦𝑦. Per three equations at the node points and per
Equation (1), we can have the following deformation functions and the corresponding shape
functions.
𝑢𝑢(𝑥𝑥, 𝑦𝑦) = 𝑢𝑢2 + (𝑢𝑢2 − 𝑢𝑢3 )𝑥𝑥 + (𝑢𝑢1 − 𝑢𝑢2 )𝑦𝑦
𝑢𝑢(𝑥𝑥, 𝑦𝑦) = 𝑁𝑁1 (𝑥𝑥, 𝑦𝑦)𝑢𝑢1 + 𝑁𝑁2 (𝑥𝑥, 𝑦𝑦)𝑢𝑢2 + 𝑁𝑁3 (𝑥𝑥, 𝑦𝑦)𝑢𝑢3
𝑁𝑁1 (𝑥𝑥, 𝑦𝑦) = 𝑦𝑦; 𝑁𝑁2 (𝑥𝑥, 𝑦𝑦) = 1 + 𝑥𝑥 − 𝑦𝑦; 𝑁𝑁3 (𝑥𝑥, 𝑦𝑦) = −𝑥𝑥
Example 4: Display the forms of deformation function for a tetrahedral element with four nodes
and ten nodes. Derivation of deformation equations of a 3D element will be too tedious and is
not necessary. However, it was worthy of mentioning that the deformation function for a 3D
tetrahedral element with four nodes as shown in Figure 5 a) will be a linear function as:
𝑢𝑢(𝑥𝑥, 𝑦𝑦, 𝑧𝑧) = 𝑎𝑎1 + 𝑎𝑎2 𝑥𝑥 + 𝑎𝑎3 𝑦𝑦 + 𝑎𝑎4 𝑦𝑦
The deformation function for a 3D tetrahedral element with ten nodes as shown in Figure 5 b)
will be a quadratic function as:
𝑢𝑢(𝑥𝑥, 𝑦𝑦, 𝑧𝑧) = 𝑎𝑎1 + 𝑎𝑎2 𝑥𝑥 + 𝑎𝑎3 𝑦𝑦 + 𝑎𝑎4 𝑦𝑦+𝑎𝑎5 𝑥𝑥 2 +𝑎𝑎6 𝑦𝑦 2 +𝑎𝑎7 𝑧𝑧 2 +𝑎𝑎8 𝑥𝑥𝑥𝑥+𝑎𝑎9 𝑦𝑦𝑦𝑦+𝑎𝑎10 𝑧𝑧𝑧𝑧
Example 5: Establish element properties of a 1D element with two nodes as shown in Figure 6.
The constant cross-section, length, and material Young’s modulus of the 1D bar element are 𝐴𝐴𝑝𝑝 ,
𝐿𝐿𝑝𝑝 and 𝐸𝐸𝑝𝑝 , respectively. The two nodes are the node i and the node j with corresponding nodal
deformation 𝑢𝑢𝑖𝑖 and 𝑢𝑢𝑗𝑗 , respectively. The internal forces on the nodes i and j are 𝑓𝑓𝑖𝑖𝑖𝑖 and 𝑓𝑓𝑗𝑗𝑗𝑗 ,
where the first subscript is the node number and the second subscript is the element number.
13
𝑑𝑑𝑢𝑢𝑝𝑝 (𝑥𝑥) (𝑢𝑢𝑗𝑗 − 𝑢𝑢𝑖𝑖 )
𝜀𝜀𝑝𝑝 (𝑥𝑥) = = 0 ≤ 𝑥𝑥 ≤ 𝐿𝐿𝑝𝑝 (12)
𝑑𝑑𝑥𝑥 𝐿𝐿𝑝𝑝
𝐸𝐸𝑃𝑃 (𝑢𝑢𝑗𝑗 − 𝑢𝑢𝑖𝑖 )
𝜎𝜎𝑝𝑝 (𝑥𝑥) = 𝐸𝐸𝑃𝑃 𝜀𝜀𝑝𝑝 (𝑥𝑥) = 0 ≤ 𝑥𝑥 ≤ 𝐿𝐿𝑝𝑝 (13)
𝐿𝐿𝑝𝑝
From Equation (12) and (13), the 1D element with two nodes will have a constant strain and
stress inside this element.
1 𝐿𝐿𝑝𝑝
1 (𝑢𝑢𝑗𝑗 − 𝑢𝑢𝑖𝑖 )2
Π𝑝𝑝 = � 𝜎𝜎𝑃𝑃 (𝑥𝑥)𝜀𝜀𝑝𝑝 (𝑥𝑥)𝑑𝑑𝑑𝑑 − �𝑓𝑓𝑖𝑖𝑖𝑖 𝑢𝑢𝑖𝑖 + 𝑓𝑓𝑗𝑗𝑗𝑗 𝑢𝑢𝑗𝑗 � = � 𝐸𝐸 𝐴𝐴𝑝𝑝 𝑑𝑑𝑑𝑑 − �𝑓𝑓𝑖𝑖𝑖𝑖 𝑢𝑢𝑖𝑖 + 𝑓𝑓𝑗𝑗𝑗𝑗 𝑢𝑢𝑗𝑗 �
𝑉𝑉𝑝𝑝 2 0 2 𝑃𝑃 𝐿𝐿2𝑝𝑝
2
1 𝐸𝐸𝑃𝑃 𝐴𝐴𝑝𝑝 �𝑢𝑢𝑗𝑗 − 𝑢𝑢𝑖𝑖 �
= − �𝑓𝑓𝑖𝑖𝑖𝑖 𝑢𝑢𝑖𝑖 + 𝑓𝑓𝑗𝑗𝑗𝑗 𝑢𝑢𝑗𝑗 � (14)
2 𝐿𝐿𝑝𝑝
Using the minimum potential energy principle, we can use Equation (14) to get followings:
𝜕𝜕Π𝑝𝑝 𝐸𝐸𝑃𝑃 𝐴𝐴𝑝𝑝 �𝑢𝑢𝑖𝑖 − 𝑢𝑢𝑗𝑗 � 𝜕𝜕Π𝑝𝑝 𝐸𝐸𝑃𝑃 𝐴𝐴𝑝𝑝 �𝑢𝑢𝑗𝑗 − 𝑢𝑢𝑖𝑖 �
=0= − 𝑓𝑓𝑖𝑖𝑖𝑖 , =0= − 𝑓𝑓𝑗𝑗𝑗𝑗
𝜕𝜕𝑢𝑢𝑖𝑖 𝐿𝐿𝑝𝑝 𝜕𝜕𝑢𝑢𝑗𝑗 𝐿𝐿𝑝𝑝
Therefore, we have
𝐸𝐸𝑃𝑃 𝐴𝐴𝑝𝑝
�𝑢𝑢𝑖𝑖 − 𝑢𝑢𝑗𝑗 � = 𝑓𝑓𝑖𝑖𝑖𝑖 (15)
𝐿𝐿𝑝𝑝
𝐸𝐸𝑃𝑃 𝐴𝐴𝑝𝑝
− �𝑢𝑢𝑖𝑖 − 𝑢𝑢𝑗𝑗 � = 𝑓𝑓𝑗𝑗𝑗𝑗 (16)
𝐿𝐿𝑝𝑝
Equations (15) and (16) are called as element properties. They link together internal forces on
the nodes 𝑓𝑓𝑖𝑖𝑖𝑖 and 𝑓𝑓𝑗𝑗𝑗𝑗 , the nodal deformation 𝑢𝑢𝑖𝑖 and 𝑢𝑢𝑗𝑗 , the material property 𝐸𝐸𝑃𝑃 and the element
geometrical properties 𝐴𝐴𝑝𝑝 and 𝐿𝐿𝑝𝑝 . Therefore, through the minimum potential energy principle,
the stress/strain analysis problems are converted into a set of linear algebraic equations.
Example 6: Build the global linear algebraic equations of a problem with two 1D elements as
shown in Figure 7. 𝐹𝐹𝑖𝑖 represents external force on the node i. 𝑓𝑓𝑖𝑖𝑖𝑖 is the internal force on the
node i of the element p. 𝑢𝑢𝑖𝑖 is the nodal deformation of the node i. 𝐸𝐸𝑝𝑝 , 𝐴𝐴𝑝𝑝 and 𝐿𝐿𝑝𝑝 are the
Young’s modulus, cross-section area and the length of the element P, respectively.
Use the results shown in Equations (15) and (16) in example 4, we obtain the element properties
for element 1 and element 2. For element 1, we have:
𝐸𝐸1 𝐴𝐴1
(𝑢𝑢1 − 𝑢𝑢2 ) = 𝑓𝑓11 (17)
𝐿𝐿1
𝐸𝐸1 𝐴𝐴1
− (𝑢𝑢1 − 𝑢𝑢2 ) = 𝑓𝑓21 (18)
𝐿𝐿1
For the element 2, we have:
14
𝐸𝐸2 𝐴𝐴2
(𝑢𝑢2 − 𝑢𝑢3 ) = 𝑓𝑓22 (19)
𝐿𝐿2
𝐸𝐸2 𝐴𝐴2
− (𝑢𝑢2 − 𝑢𝑢3 ) = 𝑓𝑓32 (20)
𝐿𝐿2
Since only node 2 is shared, per loading conditions as shown in Figure 7, we have the following
equations:
𝐹𝐹1 = 𝑓𝑓11 , 𝐹𝐹2 = 𝑓𝑓21 + 𝑓𝑓22 , 𝐹𝐹3 = 𝑓𝑓32 (21)
Using Equation (21) and Equations (17) ~ (20), we can get the global linear algebraic equations
for two-1D-element system as followings:
𝐸𝐸1 𝐴𝐴1
(𝑢𝑢1 − 𝑢𝑢2 ) = 𝑓𝑓11 = 𝐹𝐹1
𝐿𝐿1
𝐸𝐸 𝐴𝐴 𝐸𝐸2 𝐴𝐴2
− 1𝐿𝐿 1 (𝑢𝑢1 − 𝑢𝑢2 ) + 𝐿𝐿 (𝑢𝑢2 − 𝑢𝑢3 ) = 𝑓𝑓21 + 𝑓𝑓22 = 𝐹𝐹2
1 2 (22)
𝐸𝐸2 𝐴𝐴2
− 𝐿𝐿 (𝑢𝑢2 − 𝑢𝑢3 ) = 𝑓𝑓32 = 𝐹𝐹3
2
Equation (22) is the global linear algebraic equations. When loading conditions and restraints
are given, we can use Equation (22) to solve nodal deformations, to fully define deformation
functions of every element, and to further calculate strain and stress by using Hooke’s law.
Example 7: An axial bar with two different constant-sections is under axial loading. A 100-N
force is applied to the stepped cross-section as shown in Figure 8. Both ends of the bar are fixed.
15
𝑁𝑁
The information for this bar is:𝐸𝐸 = 210000(𝑚𝑚𝑚𝑚2 ), 𝐹𝐹 = 100 (𝑁𝑁), 𝐿𝐿1 = 100(𝑚𝑚𝑚𝑚), 𝐿𝐿2 =
𝐸𝐸𝐴𝐴1 𝐸𝐸𝐴𝐴2
100(𝑚𝑚𝑚𝑚), = 3000(𝑁𝑁. 𝑚𝑚𝑚𝑚) and = 1000(𝑁𝑁. 𝑚𝑚𝑚𝑚). Use the FEM theory to calculate
𝐿𝐿1 𝐿𝐿2
strain, stress and reaction forces
16
−3000𝑢𝑢2 = 𝐹𝐹1
4000𝑢𝑢2 = 100 (𝑔𝑔)
−1000𝑢𝑢2 = 𝐹𝐹3
Step 6: Solve the global linear algebraic equations
We can solve the Equation (g) to determine the nodal deformation and the reaction forces on the
restraints as followings:
𝑢𝑢2 = 0.025(𝑚𝑚𝑚𝑚), 𝐹𝐹1 = −70(𝑁𝑁), 𝐹𝐹3 = −25(𝑁𝑁) (h)
Therefore, the nodal deformations of each node in this example are:
𝑢𝑢1 = 0, 𝑢𝑢2 = 0.025(𝑚𝑚𝑚𝑚), 𝑢𝑢3 = 0 (i)
Step 7: Calculate the strain and stress of the step-bar system
Using equations (a), (b) and (i), the deformation functions of element 1 and element are:
For element 1:
𝑢𝑢1 (𝑥𝑥) = 0.00025𝑥𝑥 0 ≤ 𝑥𝑥 ≤ 100(𝑚𝑚𝑚𝑚) (j)
For element 2:
𝑢𝑢2 (𝑥𝑥) = 0.025 − 0.00025𝑥𝑥 0 ≤ 𝑥𝑥 ≤ 100(𝑚𝑚𝑚𝑚) (k)
Using the definition of strain and the Hooke’s law, we can calculate the strain and stress for the
step-bar system.
Per Equation (j), the strain and stress of the element 1 will be:
𝜕𝜕𝑢𝑢 (𝑥𝑥)
𝜀𝜀1 (𝑥𝑥) = 1 = 0.00025
𝜕𝜕𝜕𝜕 0 ≤ 𝑥𝑥 ≤ 100(𝑚𝑚𝑚𝑚)
𝜎𝜎1 (𝑥𝑥) = 𝐸𝐸𝜀𝜀1 (𝑥𝑥) = 210000 × 0.00025 = 52.5(𝑀𝑀𝑀𝑀𝑀𝑀)
Per Equation (k), the strain and stress of the element 2 will be:
𝜕𝜕𝑢𝑢 (𝑥𝑥)
𝜀𝜀2 (𝑥𝑥) = 2 = −0.00025
𝜕𝜕𝜕𝜕 0 ≤ 𝑥𝑥 ≤ 100(𝑚𝑚𝑚𝑚)
𝜎𝜎2 (𝑥𝑥) = 𝐸𝐸𝜀𝜀2 (𝑥𝑥) = 210000 × (−0.00025) = −52.5(𝑀𝑀𝑀𝑀𝑀𝑀)
17