Course number: ME564
Course name: Intro to Finite Element Method
Student Name: Mohammad Shafikur Rahman Pashari
BlazerID: mpashari
Email: mpashari@uab.edu
1.1 Define the term finite element.
Ans: It is a numerical/calculative method by which problems of engineering and
mathematical physics can be solved. For example, different problems of structural
analysis, heat transfer, fluid flow, mass transport, electromagnetic are solved by FEM.
Especially, when it is generally not possible to get analytical mathematical solution for a
physical systems to get simulation of it which have complicated geometries, loading or
material properties.
1.2 What does discretization mean in the finite element method?
Ans: In FEM, discretization is the process by which modeling body is divided into
equivalent system of smaller bodies of unites which are called finite elements and these
finite elements are interconnected at common points (nodal points or nodes) with or
without boundary lines and/or surfaces.
1.3 In what year did the modern development of the finite element method begin?
Ans: By using a lattice of line (one-dimensional) elements (bars and beams) for the
solution of stresses in continuous solids, Hrennikoff (in 1941) and McHenry (in 1943),
started the modern development of the FEM. However, during in the 1940s other
people made some contribution. For example, Courant (in 1943) proposed setting up
the solution of stresses in a variational form. In 1947 Levy developed the flexibility or
force method.
1.4 In what year was the direct stiffness method introduced?
Ans: Direct stiffness method was introduced in 1956 by Turner for the first treatment of
two-dimensional elements. Their derivation of stiffness matrices for truss elements,
beam elements, and two-dimensional triangular and rectangular elements in plane
stress and outlined the procedure is commonly known as the direct stiffness method.
1.5 Define the term matrix.
Ans: In FEM, the term matrix is defined by a rectangular array of quantities arranged in
rows and columns which is used as an aid in expressing and solving a system of algebraic
equations.
1.6 What role did the computer play in the use of the finite element method?
Ans: Problem of implementing FEM is that it takes hours and hours to get solution of
thousands of equations by hand calculation for complicated problems. Computer downs
the solution to minutes.
1.7 List and briefly describe the general steps of the finite element method.
Ans: There are seven general steps to solve a problem by finite element.
1st step is to discretize and select the element types.
2nd step is to select a displacement function.
The 3rd step is to define the Strain/Displacement and Stress/Strain relationships
In the 4th step you need to derive the element stiffness matrix and equations
5th step is to assemble the element equations to obtain the global or total equations and
introduce boundary conditions.
6th step is to solve the unknown degress of freedom (or generalized displacements)
The final 7th step is to solve for the element strains and stresses.
1.8 What is the displacement method?
Ans: Displacement method is also called stiffness method. In this method each element
is connected to a commode node. After loading and hence deformation, governing
equations are formulated in terms of nodal displacements using the equations of
equilibrium related to forces and displacements.
1.9 List four common types of finite elements.
Ans: Here is the name of four common types of finite elements with picture.
i. Simple two-noded line elements
ii. Simple two-dimensional elements
iii. Simple three-dimensional elements
iv. Simple axisymmetric triangular and quadrilateral elements
1.10 Name three commonly used method for deriving the element stiffness matrix
and element equations. Briefly describe each method.
Ans:
Direct Equilibrium Method— There are two types approach. In first approach, internal
forces as used to obtain the governing equations. Thereafter, necessary additional
equations are found by producing relevant compatibility equations. As a result a set of
algebraic equations are are formed to determine unknown forces. In the second
approach, displacement of nodes are assumed as unknowns of the problem. Then
governing equations are expressed in terms of nodal displacement using the equations
of equilibrium and an applicable law relating forces of displacements.
Work or Energy Methods— This method is based on principle of minimum potential
energy. In other words, at low temperatures a structure or body shall deform or displace
to a position that (locally) minimizes the total potential energy, with the lost potential
energy being converted into kinetic energy.
Weighted Residuals Methods—In this method, a trial or approximate function is chosen
to approximate the independent variable, such as a displacement or a temperature. As
this trial function does not satisfy the governing differential equation, so substituting the
trial function into the differential equation results in a residual over the whole region of
the problem. In residual method, weighted integral of residuals go to zero.
1.11 To what does the term degrees of freedom refer?
Ans: In Finite Element Analysis each element is separated from the Global System.
Thereafter load is applied to each element. As a result each element rotate and displace.
Formation of rotation and displacement of each element due to loading is called
degrees of freedom.
1.12 List five typical areas of engineering where the finite element method is
applied.
Ans: Here are five areas of engineering where finite element method is applied
1. Stress analysis
2. Vibration analysis
3. Heat transfer analysis
4. Fluid flow analysis
5. Crash or impact analysis
1.13 List five advantages of the finite element method.
Ans: Here are five advantages of finite element method—
1. One of the main benefits of FEM is that it can be applied to irregular shaped
bodies without difficulty.
2. With FEM model, you can apply different load conditions quite easily.
3. FEM can handle bodies composed of several different materials.
4. Unlimited numbers and different kinds of boundary conditions can be applied.
5. Size of elements can be changed. This is very useful when implementing small
elements is necessary.
2.1
First we need to discretize the problem and for each element we need to define free
body diagram.
Then for each element we will apply hook’s law. Hence, we get the following equations--
So, we get stiffness matrix for each element.
[k(1)]= k1 0 -k1 [k(2)]= 0 0 0 [k(3)]= k3 0 -k3
0 0 0 0 k2 -k2 0 0 0
-k1 0 k1 0 -k2 k2 -k3 0 k3
As we have to superimpose the stiffness matrix and the superimposed matrix will be 4x4
matrix, so we will transfer all our element matrix from 3x3 to 4x4 matrix form:
[k(1)]=k1 0 -k1 0 [k(2)]= 0 0 0 0 [k(3)]= 0 0 0 0
0 0 0 0 0 0 0 0 0 k3 0 -k3
-k1 0 k1 0 0 0 k2 -k2 0 0 0 0
0 0 0 0 0 0 -k2 k2 0 -k3 0 k3
According the superimpose method,
[K]= [k(1)]+ [k(2)]+ [k(3)]
a. Hence, we get global stiffness matrix by superimposing local stiffness matrix.
[K]= k1 0 -k1 0
0 k3 0 -k3
-k1 0 k1+k2 -k2
0 -k3 -k2 k2+k3
b. We now write the Global nodal force matrix, global nodal displacement matrix
and global stiffness matrix in the {F}= {K}{d} format,
F1x = k1 0 -k1 0 u1
F2x 0 k3 0 -k3 u2
F3x -k1 0 k1+k2 -k2 u3
F4x 0 -k3 -k2 k2+k3 u4
As nodes 1 and 2 are fixed so applying boundary condition, we get u 1= 0 and u2=0
respectively.
Also force P acts on the node 4 in the positive direction, hence f 4x= F4x=P
As a consequence, if we put all these value in matrix equation, then our matrix will
become—
F3x = k1+k2 -k2 u3
F4x -k2 k2+k3 u4
Again as Force at node is null, so F3x=0. That’s why above matrix forms becomes
0 = k1+k2 -k2 u3
P -k2 k2+k3 u4
Hence equation …….. and …… becomes
Therefore, displacements of node 3 is
And node 4 is
2.3
After discretizing the problem in four elements, we get the following free body diagram
—
Therefore, local stiffness for each element is ---
[k1]= k -k U1
-k k U2
[k2]= k -k U2
-k k U3
[k3]= k -k U3
-k k U4
[k4]= k -k U4
-k k U5
a. To get the global stiffness matrix from local stiffness matrices, we have to add as
below, [K]= [k1+[k2]+ [ k3]+[k4]
But each local stiffness matrix is 2x2, whereas Global stiffness matrix is 4x4. So,
we first transform the 2x2 matrices to 4x4 format.
U1 U2 U3 U4 U5
[k1]= k -k 0 0 0 U1
-k k 0 0 0 U2
0 0 0 0 0 U3
0 0 0 0 0 U4
0 0 0 0 0 U5
U1 U2 U3 U4 U5
[k2]= 0 0 0 0 0 U1
0 k -k 0 0 U2
0 -k k 0 0 U3
0 0 0 0 0 U4
0 0 0 0 0 U5
U1 U2 U3 U4 U5
[k3]= 0 0 0 0 0 U1
0 0 0 0 0 U2
0 0 k -k 0 U3
0 0 -k k 0 U4
0 0 0 0 0 U5
U1 U2 U3 U4 U5
[k4]= 0 0 0 0 0 U1
0 0 0 0 0 U2
0 0 0 0 0 U3
0 0 0 k -k U4
0 0 0 -k k U5
Hence the global stiffness matrix of the system by superposition is—
U1 U2 U3 U4 U5
[K]= k -k 0 0 0 U1
-k k+k -k 0 0 U2
0 -k k+k 0 0 U3
0 0 0 k -k U4
0 0 0 -k k U5
b. As node 1 and 5 are fixed and force P is applied at node 3, so applying boundary
condition equation ……. becomes ---
c. As we know now displacement for node points, so reaction for
Node 1 is---
And node 5 is---
2.13
For the problem given above, we have to find out
a) Nodal displacements in each element
b) The forces in each element
c) The reaction forces
by using the direct stiffness method.
First we will illustrate free body diagram after discretizing 4 elements from the system.
To get the local stiffness for each element, we consider the below matrix which is based
on the figure below
[k]= k1 - k1
- k1 k1
Therefore, local stiffness matrix for each element are—
[k1]= 60 -60
- 60 60
[k2]= 60 -60
- 60 60
[k3]= 60 -60
- 60 60
[k3]= 60 -60
- 60 60
According to Direct stiffness method, Global Stiffness [K] will be summation of all local
stiffness matrix, i.e.
[K]= [k1]+[k2]+ [k3]+[k4]
However, Global stiffness matrix is 5x5 matrix, whereas local matrices are 2x2. So, we
need to transfer all local matrices to 5x5 matrix, before summing them to form one
matrix.
u1 u2 u3 u4 u5
1
[k ]= 60 -60 0 0 0 u1
- 60 60 0 0 0 u2
0 0 0 0 0 u3
0 0 0 0 0 u4
0 0 0 0 0 u5
u1 u2 u3 u4 u5
2
[k ]= 0 0 0 0 0 u1
0 60 -60 0 0 u2
0 - 60 60 0 0 u3
0 0 0 0 0 u4
0 0 0 0 0 u5
u1 u2 u3 u4 u5
3
[k ]= 0 0 0 0 0 u1
0 0 0 0 0 u2
0 0 60 -60 0 u3
0 0 - 60 60 0 u4
0 0 0 0 0 u5
u1 u2 u3 u4 u5
4
[k ]= 0 0 0 0 0 u1
0 0 0 0 0 u2
0 0 0 0 0 u3
0 0 0 60 -60 u4
0 0 0 - 60 60 u5
So, after summing Global Stiffness Matrix will be,
u1 u2 u3 u4 u5
[K]= 60 -60 0 0 0 u1
- 60 60+60 -60 0 0 u2
0 - 60 60+60 -60 0 u3
0 0 - 60 60+60 -60 u4
0 0 0 - 60 60 u5
u1 u2 u3 u4 u5
=>[K]= 60 -60 0 0 0 u1
- 60 120 -60 0 0 u2
0 - 60 120 -60 0 u3
0 0 - 60 120 -60 u4
0 0 0 - 60 60 u5
Now we will apply boundary condition. As node 1 & 5 are fixed. So, displacement
at these nodes will be zero. That means, u1= & u5=0.
Another important to note that Force at node 3 is 5kN, which means F3x=5kN.
Also, there is no force at node 3 & node 4. So, F2x=0 & F4x=0.
Now to find out the displacements of other nodes, we need to apply the {F}={K}
{U}
U1 U2 U3 U4 U5
F1x 60 -60 0 0 0 U1=0
F2x=0 = - 60 120 -60 0 0 U2
F3x=5k 0 - 60 120 -60 0 U3
N
F4x=0 0 0 - 60 120 -60 U4
F5x 0 0 0 - 60 60 U5=0
Therefore,
0=120*u2-60*u3
2u2=u3
0=-60*u3+120u4
u3=2u4
and 5kN=-60u2+120u3-60u4
Then replacing value of u2 and u4 in terms of u3 in the last equation, we get
5kN=-60*0.5u3+120u3-60*0.5u3
5kN=-30u3+120u3-30u3
5kN=60u3
u3= 0.0833 meter
So, u2=0.5*0.0833=0.04166 meter=u4
(b) As we now know the displacement values of each node, so our {F}={K}{U}
form becomes---
U1 U2 U3 U4 U5
F1x 60 -60 0 0 0 0
0 = - 60 120 -60 0 0 0.04166
5kN 0 - 60 120 -60 0 0.0833
0 0 0 - 60 120 -60 0.04166
F5x 0 0 0 - 60 60 0
Therefore, F1x=-60*0.04166=2.4996 kN
And F5x=-60*0.04166=2.4996 kN
(c) To find out reaction forces at each nodes, however, we have to use local
stiffness for of {f}={k}{d} for each element seperately.
For element 1,
f(1)1x = 60 -60 U1=0
f(1)2x -60 60 U2=0.04166
So, F1x=
F2x=
For element 2,
F2x = 60 -60 U2=0.04166
F3x -60 60 U3=0.0833
For element 3,
F3x = 60 -60 U3=0.0833
F4x -60 60 U4=0.04166
For element 4,
F4x = 60 -60 U4=0.04166
F5x -60 60 U5=0
2.15
For the problem given above, we have to find out
d) Nodal displacements in each element
e) The forces in each element
f) The reaction forces
by using the direct stiffness method.
To solve the problem, at fist let’s draw free body diagram after discretizing the system
by 3 elements
Element 3
Element 1 Element 2
Then we have to find out local stiffness matrix for each element separately.
However, we already know that for an element like below,
Local stiffness matrix will be,
[k]= k1 - k1
- k1 k1
Therefore, local stiffness matrix for each element will be,
[k1]= 500 - 500
-500 500
[k2]= 500 -500
- 500 500
[k3]= 1000 - 1000
- 1000 1000
According to direct stiffness method, we know that Global stiffness matrix will be
[K]= [k1]+[k2]+[k3]
However, Global Stiffness matrix is 3x3 matrix, whereas our local matrix are 2x2. Hence,
we have to transfer our local stiffness matrix to 4x4 before adding them together.
Hence,
u1 u2 u3 u4
1
[k ]= 500 0 -500 0 u1
0 0 0 0 u2
-500 0 500 0 u3
0 0 0 0 u4
u1 u2 u3 u4
2
[k ]= 0 0 0 0 u1
0 500 -500 0 u2
0 -500 500 0 u3
0 0 0 0 u4
u1 u2 u3 u4
3
[k ]= 0 0 0 0 u1
0 0 0 0 u2
0 0 1000 -1000 u3
0 0 -1000 1000 u4
So, according to direct stiffness method, Global stiffness is
[K]= 500 0 -500 0 u1
0 500 -500 0 u2
-500 -500 500+500+1000 -1000 u3
0 0 -1000 1000 u4
=>[K]= 500 0 -500 0 u1
0 500 -500 0 u2
-500 -500 2000 -1000 u3
0 0 -1000 1000 u4
Let’s now apply boundary condition.
As nodes 1, 2 and 4 are fixed, so there is no displacement in these three nodes.
Therefore, displacement for these nodes will be u1=0; u2=0; & u4=0
Similarly, as forces are applied to nodes 3, so force component of these nodes will be—
F3x=2 kN
So, substituting all these values in the {F}= {K}{U} format, we get
F1x 500 0 -500 0 0
F2x = 0 500 -500 0 0
2kN -500 -500 2000 -1000 u3
F4x 0 0 -1000 1000 0
Considering row 3, we get—
2000=2000 * u3
u3= 0.001 m
(a) Hence, we can state that node 1, 2 & 4 have zero displacement and node 3 has
displacement of 0.001 meter.
(b) As we now know the displacement values, so forces on each element will be,
F1x=-500*u3=-500*0.001=-0.5kN
F2x=-500*u3=-500*0.001=-0.5kN
F3x==-500*u3=-500*0.001=-0.5kN
(c) Now we are going to calculate reaction forces for each node. But for that we
have to consider {f}={k}{u} for each element separately.
For element 1,
f1x = 500 -500 u1=0
f3x -500 500 u3=0.001
Hence, f1x= -500*0.001=-0.5 kN
f3x=500*0.001= 0.5kN
Similarly, for element 2,
f2x = 500 -500 u2=0
f3x -500 500 u3=0.001
Hence, f1x= -500*0.001=-0.5 kN
f3x=500*0.001= 0.5kN
and for element 3,
f3x = 1000 -1000 u3=0.001
f4x -1000 1000 u4=0
Hence, f3x= 1000*0.001=1 kN
f4x=-1000*0.001= -1kN
