Professional Documents
Culture Documents
By
Mr. L.VINOTH ME
ASSISTANT PROFESSOR
DEPARTMENT OF MECHANICAL ENGINEERING
SASURIE COLLEGE OF ENGINEERING
VIJAYAMANGALAM 638 056
QUALITY CERTIFICATE
: ME2353
Subject
Class
being prepared by me and it meets the knowledge requirement of the university curriculum.
L.VINOTH M.E.
Designation:
ASSISTANT PROFESSOR
This is to certify that the course material being prepared by Mr.L.Vinoth is of adequate quality. He has
referred more than five books among them minimum one is from aboard author.
Signature of HD
Name: E.R SIVAKUMAR M.E.,(PhD)
SEAL
CONTENTS
S.NO
TOPIC
PAGE NO
INTRODUCTION
1.2
1.3
1.4
1.5
1.6
13
19
19
19
1.8
24
1.9
24
1.10
24
1.7
25
2.2
28
2.3
BEAM ELEMENT
28
2.4
33
2.5
34
2.6
BEAM ELEMENT
42
45
INTRODUCTION
54
3.2
54
3.3
55
ELEMENT
56
60
61
61
APPROXIMATIONS
ISOPARAMETRIC ELEMENTS
STRUCTURAL MECHANICS APPLICATIONS IN 2
DIMENSIONS
UNIT 4 DYNAMIC ANALYSIS USING ELEMENT METHOD
65
66
71
INTRODUCTION
88
88
88
88
4.2
EQUATION OF MOTION
89
4.3
94
94
98
4.4
99
4.5
MODELLING OF DAMPING
102
102
105
4.1
4.6
106
107
111
111
112
112
112
112
114
5.3
114
5.4
117
QUESTION BANK
120
TEXT BOOKS:
1. P.Seshu, Text Book of Finite Element Analysis, Prentice-Hall of India Pvt. Ltd. New
Delhi, 2007.
2. J.N.Reddy, An Introduction to the Finite Element Method, McGraw-Hill International
Editions(Engineering Mechanics Series), 1993.
3. Cook,Robert.D., Plesha,Michael.E & Witt,Robert.J. Concepts and Applications of
Finite Element Analysis,Wiley Student Edition, 2004.
4. Chandrupatla & Belagundu, Introduction to Finite Elements in Engineering, 3rd
Edition, Prentice-Hall of India, Eastern Economy Editions.
ME2353
L T P C
3 1 0 4
ME2353
UNIT I
FINITE ELEMENT FORMULATION OF BOUNDARY VALUE PROBLEMS
1.1
INTRODUCTION
The finite element method constitutes a general tool for the numerical solution of partial
differential equations in engineering and applied science
The finite element method (FEM), or finite element analysis (FEA), is based on the idea of
building a complicated object with simple blocks, or, dividing a complicated object into small and
manageable pieces. Application of this simple idea can be found everywhere in everyday life as well as
in engineering.
Examples:
Lego (kidsplay) Buildings
Approximation of the area of a circle:
Element Si
i
R
SCE
Discretization of structure
Numbering of Nodes and Elements
Selection of Displacement function or interpolation function
Define the material behavior by using Strain Displacement and Stress Strain
relationships
1
ME2353
i = 1, 2, 3, n.
Problem I
Find the solution for the following differential equation.
SCE
ME2353
qo =0
EI
u(0)=0,
(L)=0,
(0)=0,
(L)=0,
qo =0
a4=
Substitute a4values in u(x)
u(x) =
[x4-4Lx3+6L2x2]
Result:
Final solution u(x) =
[x4-4Lx3+6L2x2]
Problem 2
The differential equation of a physical phenomenon is given by
SCE
ME2353
+ = 4, 0 1
The boundary conditions are: y(0)=0
y(1)=1
Obtain one term approximate solution by using galerkin method
Solution:
Here the boundary conditions are not homogeneous so we assume a trial function as,
y=a1x(x-1)+x
first we have to verify whether the trial function satisfies the boundary condition or not
y=a1x(x-1)+x
when x=0, y=0
x=1, y=1
Resuldual R:
Y=a1x(x-1)+x=a1(x2-x)+x
=a1(2x-1)+1
=2
Substitute
value in given differential equation.
2a1+y=4x
Substitute y vlue
R=2a1+a1x(x-1)+x-4x
In galerkins method
= 0, 0
=0 at x=0 and x=
=
Where, EI
ME2353
After understanding the basic techniques and successfully solved a few problem general
weighted residual statement can be written as
R dx=0 for i= 1,2,..n
Where wi=Ni
The better result will be obtained by considering more terms in polynomial and trigonometric series.
1.4 WEAK FORMULATION OF THE WEIGHTED RESIDUAL STATEMENT.
The analysis in Section as applied to the model problem provides an attractive perspective to the
solution of certain partial differential equations: the solution is identified with a point, which
minimizes an appropriately constructed functional over an admis- sible function space. Weak
(variational) forms can be made fully equivalent to respective strong forms, as evidenced in the
discussion of the weighted residual methods, under certain smoothness assumptions. However, the
equivalence between weak (variational) forms and variational principles is not guaranteed: indeed, there
exists no general method of constructing functionals I [u], whose extremization recovers a desired weak (variational) form. In this
sense, only certain partial differential equations are amenable to analysis and solution by
variational methods.
Vainbergs theorem provides the necessary and sufficient condition for the equivalence of a
weak (variational) form to a functional extremization problem. If such equivalence holds, the functional
is referred to as a potential.
Theorem (Vainberg)
Consider a weak (variational) form
G(u, u) := B(u, u) + (f, u) + (q , u)q = 0 ,
where u U , u U0 , and f and q are independent of u. Assume that G pos- sesses a
Gateaux derivative in a neighborhood N of u, and the Gateaux differen- tial Du1 B(u, u2) is
continuous in u at every point of N .
Then, the necessary and sufficient condition for the above weak form to be derivable from a
potential in N is that
Du1 G(u, u2) = Du2 G(u, u1) ,
Namely that Du1 G(u, u2) be symmetric for all u1, u2 = U0 and all u = N .
Preliminary to proving the above theorem, introduce the following two lemmas:
Lemma 1 Show that Dv I[u] =
SCE
lim
5
ME2353
In the above derivation, note that operations and |=0 are not interchangeable (as they
both refer to the same variable ), while lim0 and |=0 are interchangeable, conditional upon
sufficient smoothness of I [u].
Lemma 2 (Lagranges formula)
Let I [u] be a functional with Gateaux derivatives everywhere, and u, u + u be any points
of U. Then,
I [u + u] I [u] = Du I [u + u]
0 < < 1.
df
d =
lim
0
f ( + ) f ()
lim
0
I [u + u + u] I [u + u]
= Du I [u + u] ,
fi
SCE
ui
uj
fj
ME2353
Two nodes:
Nodal displacements:
Nodal forces:
N/m, N/mm)
i, j
ui, uj (in, m, mm)
fi, fj (lb, Newton) Spring constant (stiffness):
k (lb/in,
Linear
Nonlinear
D
k F/
(> 0) is the force needed to produce a unit stretch.
We only consider linear problems in this introductory course. Consider the equilibrium of
forces for the spring.
At node 1 we have
fi
k(u j
ui )
kui
kuj
and at node j,
fj
k(u j
ui )
kui
kuj
In matrix form,
k
ui
uj
fi
fj
or, where
SCE
ME2353
Problem 4
To find the deformation of the shape
X
K1
u1F1
K2
u2F2
u3F3
For element 1,
k1
k 1 u2
f1 2
k1
k1
f2 2
u3
element 2,
k2
k2
k2
k2
u2
u3
f
f
2
2
where fI at node 2 F2
M is the (internal) force acting on local node i of element Consider the quilibrium of
forces at node
Checking the Results
Deformed shape of the structure
Balance of the external forces
Order of magnitudes of the numbers
Notes about the Spring Elements
Suitable for stiffness analysis
Not suitable for stress analysis of the spring itself
Can have spring elements with stiffness in the lateral direction,
Spring elements for torsion, etc.
1.6 EXAMPLES OF A BAR FINITE ELEMENT
The finite element method can be used to solve a variety of problem types in
engineering, mathematics and science. The three main areas are mechanics of materials, heat
transfer and fluid mechanics. The one-dimensional spring element belongs to the area of
mechanics of materials, since it deals with the displacements, deformations and stresses
involved in a solid body subjected to external loading.
SCE
ME2353
Element dimensionality:
An element can be one-dimensional, two-dimensional or three-dimensional. A spring element
is classified as one-dimensional.
Geometric shape of the element
The geometric shape of element can be represented as a line, area, or volume. The onedimensional spring element is defined geometrically as:
Spring law
The spring is assumed to be linear. Force (f) is directly proportional to deformation () via the
spring constant k, i.e.
Element formulation
There are various ways to mathematically formulate an element. The simplest and limited
approach is the direct method. More mathematically complex and general approaches are energy
(variation) and weighted residual methods.
SCE
ME2353
The direct method uses the fundamentals of equilibrium, compatibility and spring law from a
sophomore level mechanics of material course. We will use the direct method to formulate the onedimensional spring element because it is simple and based on a physical approach.
The direct method is an excellent setting for becoming familiar with such basis concepts of
linear algebra, stiffness, degrees of freedom, etc., before using the mathematical formulation
approaches as energy or weighted residuals.
Assumptions
Spring deformation
The spring law is a linear force-deformation as follows:
f=k
f - Spring Force (units: force)
k - Spring Constant (units: force/length)
- Spring Deformation (units: length)
Spring Behaviour:
A spring behaves the same in tension and compression.
Spring Stiffness:
Spring stiffness k is always positive, i.e., k>0, for a physical linear system.
Nodal Force Direction:
Loading is uniaxial, i.e., the resultant force is along the element. Spring has no resistance to
lateral force.
Weightless Member:
Element has no mass (weightless).
Node Location:
The geometric location of nodes I and J cannot coincide, i.e., xi xj. The length of the element
is only used to visually see the spring.
A column of KE is a vector of nodal loads that must be applied to an element to sustain a
deformed state in which responding nodal DOF has unit value and all other nodal DOF are zero. In
other words, a column of KE represents an equilibrium problem.
SCE
10
ME2353
Example, uI = 1, uJ = 0.
SCE
11
ME2353
SCE
12
ME2353
Force Magnitude - Even if forces are large, a rigid body does not deform. A non-rigid body will
deform even if a force is small. In reality, all bodies deform.
Failure - A rigid body does not fail under any load; while a non-rigid body will result either in
ductile or brittle failure when the applied load causes the normal stress to exceed the breaking
(fracture) stress b of the material. Brittle failure occurs when the applied load on the non-rigid
bar shown above causes the breaking strength of the bar to be exceeded.
Material - The material is not considered in a rigid body. Since a rigid body does not deform (
= 0) this is equivalent to an infinite modulus of elasticity. In contrast the modulus of elasticity
for a non-rigid material is finite, e.g., for steel, Esteel = 30 x 106 psi. (200 GPa). For rigid and
non-rigid bars the material laws are:
SCE
13
ME2353
Rotation About a Fixed Axis - If all the particles of a rigid body move along circular paths,
except the ones which lie on the axis of rotation, it is said to be in rotation about a fixed axis.
General Motion - Any motion of a rigid body that consists of the combination of both
translations
There are six rigid body modes in general three-dimensional situation; three translational
along the x, y, and z axes and three rotational about x, y, and z axes. Illustrations of these rigid body
modes are presented as follows:
SCE
14
ME2353
Translational
Rigid Body Modes
Rotational
Rigid Body Modes
x-direction
about x-axis
y-direction
about y-axis
z-direction
about z-axis
x0
2 u2
l
2
3 u3
xl
at x 0,
l
at x ,
2
at x l ,
u u1
u u2
u u3
u1 a
a u1
l
l2
u2 a b +c
2
4
2
u1 a bl cl
u3 4u2 3u1
l
solving
2
c 2 u3 2u2 u1
l
b
Substituting the values of a, b and c in equation (1) and collecting the coefficients of u1 , u2 and u3
SCE
15
ME2353
N1
3
x
4 2
x
l
l
x
x2
N2 4 4 2
l
l
x 2x
1 1
l
l
x x
4 1
l l
x
x2
N3 2 2
l
l
x 2x
1
l
l
N 3
1
x
4 2
x
l
l
x
x2
N1 1 3 2 2
l
l
Where
N 2 4
x
8 2
x
l
l
x
3
l 4 l 2
x
T
4
B 8 2
l
l
x
1
l 4 l2
D E for a bar element (1-D case - only axial stress x and strain x exist x E x )
volume
dV Adx A dx
since the cross-sectional area A is constant for the total length of the bar.
x
3
4
l
l2
l
x 3
x
4
k A 8 2 E 4 2
l
l l
l
0
x
1
l 4 l2
3
x 3
x
l 4 l 2 l 4 l 2
l
4
x
3
x
k AE 8 2
4 2
l
l l
l
0
1
x 3
x
4 2 4 2
l l
l
l
x 1
x
4
8 2 4 2 dx
l l
l
l
x 4
x 3
x 1
x
3
4 2 8 2 4 2 4 2
l l
l l
l l
l
l
x 4
x
x 1
x
4
4
8 2 8 2
8 2 4 2 dx
l l
l
l l
l
l
l
x 4
x 1
x 1
x
1
4 2 8 2 4 2 4 2
l l
l l
l l
l
l
k B D B dV
T
Volume
SCE
N1
x
N 2
x
N 3
x 4
x 1
x
3
4 2 8 2 4 2
x
l l
l l
l
l
16
ME2353
To determine K 11 :
l
9 12 x 12 x 16 x 2
3 4 x 3 4 x
K11 AE 2 2 dx AE 2 3 3 4 dx
l
l l
l
l
l
l
l
0
0
l
9 24 x 16 x 2
9 x 24 x 2 16 x 3
9l 24l 2 16l 3
K11 AE 2 3 4 dx AE 2
4 AE 2 3 4
l
l
l
2l 3
3l 0
2l
3l
l
l
0
l
9 12 16
27 36 16 AE
K11 AE AE
3l 7
3l
l l 3l
K11
7 AE
3l
To determine K 12
and K 21 :
l
l
12 24 x 16 x 32 x 2
3 4 x 4 8 x
K12 AE 2 2 dx AE 2 3 3 4 dx
l l l l
l
l
l
l
0
0
l
12 x 40 x 2 32 x 3
K12 AE 2
4
2l 3
3l 0
l
l
12 x 40 x 2 32 x 3
12l 40l 2 32l 3
12 20 32
36 60 32
K12 AE 2
AE
2 3 4 AE AE
3
4
2l
3l 0
2l
3l
l
3l
3l
l
l
l
AE
K12
8 K 21
3l
To determine K 13
and K 31 :
l
l
3 12 x 4 x 16 x 2
3 4 x 1 4 x
K13 AE 2 2 dx AE 2 3 3 4 dx
l l l l
l
l
l
l
0
0
l
3 16 x 16 x 2
3 x 16 x 2 16 x 3
3l 16l 2 16l 3
3 8 16
K13 AE 2 3 4 dx AE 2 3 4 AE 2 3 4 AE
l
l
l
2l
3l 0
2l
3l
l l 3l
l
l
0
9 24 16 AE
K13 AE
3l 1 K 31
3l
SCE
17
ME2353
To determine K 22
l
l
16 32 x 32 x 64 x 2
4 8 x 4 8 x
K 22 AE 2 2 dx AE 2 3 3 4 dx
l l l l
l
l
l
l
0
0
l
16 64 x 64 x 2
16 x 64 x 2 64 x 3
K 22 AE 2 3 4 dx AE 2
4
3
l
l
l
l
2
l
3l 0
0
l
K 22
16 AE
3l
To determine K 23
and K 32
l
4 24 x 32 x 2
4 x 24 x 2 32 x 3
K 23 AE 2 3 4 dx AE 2
4
l
l
l
2l 3
3l 0
l
0
l
4l 24l 2 32l 3
4 12 32
12 36 32 AE
K 23 AE 2 3 4 AE AE
3l 8
2l
3l
l 3l
3l
l
K 23
8 AE
K 32
3l
To determine K 33
l
l
1 4 x 4 x 16 x 2
1 4 x 1 4 x
K33 AE 2 2 dx AE 2 3 3 4 dx
l l l l
l
l
l
l
0
0
l
l
4 16 x 8 x 32 x 2
4 8 x 1 4 x
K 23 AE 2 2 dx AE 2 3 3 4 dx
l l l l
l
l
l
l
0
0
l
1 8 x 16 x 2
x 8 x 2 16 x 3
K 33 AE 2 3 4 dx AE 2 3 4
l
l
l
2l
3l 0
l
0
l
l 8l 2 16l 3
1 4 16
3 12 16 AE
K33 AE 2 3 4 AE AE
3l 7
2l
3l
3l
l l 3l
K33
SCE
7 AE
3l
18
ME2353
7 8 1
AE
8 16 8
Assembling, we get k
3L
1 8 7
xu
= K u2
Total potential energy in the spring
= K u2 - P u
=0
Hence we can conclude that a deformable body is in equilibrium when the potential energy is
having stationary value.
SCE
19
ME2353
Hence the principle of minimum potential energy states among all the displacement equations
that internal compatibility and the boundary condition those that also satisfy the equation of
equilibrium make the potential energy a minimum is a stable system
Problem 7
k1
k2
x
2
Given:
A
p
p
l
y
i
n
g
P 500 N,
u1
0
u4
0
Find:
Solution:
t
h
e
100
200
k 2
200
100
k 3
SCE
100
(N/mm)
100
(1)
200
(N/mm)
200
(2)
100
(N/mm)
100
(3)
100
20
ME2353
u1
100
u2
100
u3
0
u4
100
100
200
200
200
200
100
100
200
0
300
100
100
100
100
100
300
200
u1
F1
200
u2
100
u3
u4
F4
300
100
100
(4)
200 u2
300 u3
0
(5)
SCE
st
u2
P / 250
u3
3P / 500
and 4
th
(6)
(mm)
3
F1
100u 2
200 (N)
F4
100u 3
300 (N)
21
ME2353
ui
fi
200
uj
fj
200
200
Problem 8
k4
F1
k1
2
4
k2
k3
F2
2
For the spring system with arbitrarily numbered nodes and elements, as
shown above, find the global stiffness matrix.
Solution:
First we construct the following
Element Connectivity Table
Element
1
2
3
4
Node i (1)
4
2
3
2
Node j (2)
2
3
5
1
Which specifies the global node numbers corresponding to the local node numbers for
each element? Then we can write the element stiffness matrices as follows
SCE
22
ME2353
u4
u2
k1
k1
k1
u5
k3
k3
k 2
k1
u3
k 3
u2
k3
k 4
k3
u3
k2
k2
k2
k2
u2
u1
k4
k4
k4
k4
Finally, applying the superposition method, we obtain the global stiffness matrix as
follows
We may note that N1 and N2 obey the definition of shape function that is the shape
function will have a value equal to unity at the node to which it belong and zero value at other
nodes.
u1
u2
k4
k4
0
0
k4
k2 k4
SCE
k1
k2
u3
u4
u5
k2
k1
k3
k2
k1
0
k3
0
k3
23
k1
0
0
k3
ME2353
SCE
24
ME2353
UNIT II
ONE DIMENSIONAL FINITE ELEMENT ANALYSIS
2.1 ONE DIMENSIONAL ELEMENTS
Bar and beam elements are considered as One Dimensional elements. These
elements are often used to model trusses and frame structures.
Bar, Beam and Truss
Bar is a member which resists only axial loads. A beam can resist axial,
lateral and twisting loads. A truss is an assemblage of bars with pin joints and a frame
is an assemblage of beam elements.
Stress, Strain and Displacement
Stress is denoted in the form of vector by the variable x as x, Strain is denoted
in the form of vector by the variable x as e x, Displacement is denoted in the form of
vector by the variable x as ux.
Types of Loading
(1) Body force (f)
It is a distributed force acting on every elemental volume of the body. Unit is
Force / Unit volume. Ex: Self weight due to gravity.
(2) Traction (T)
It is a distributed force acting on the surface of the body. Unit is Force / Unit
area. But for one dimensional problem, unit is Force / Unit length. Ex: Frictional
resistance, viscous drag and Surface shear.
(3) Point load (P)
It is a force acting at a particular point which causes displacement.
Finite Element Modeling
It has two processes.
(1) Discretization of structure
(2) Numbering of nodes.
SCE
25
ME2353
CO ORDINATES
(A) Global co ordinates,
(B) Local co ordinates and
(C) Natural co ordinates.
Natural Co Ordinate ()
Shape function
26
ME2353
Types of beam
1. Cantilever beam,
2. Simply Supported beam,
3. Over hanging beam,
4. Fixed beam and
5. Continuous beam.
Types of Transverse Load
1. Point or Concentrated Load,
2. Uniformly Distributed Load and
3. Uniformly
SCE
27
ME2353
SCE
28
ME2353
Assumptions
Nodal Forces and Moments
Forces and moments can only be applied at the nodes of the beam element, not
between the nodes. The nodal forces and moments, , are related to the nodal displacements
and rotations, through the element stiffness matrix,
.
Constant Load
The loads that are applied to the beam element are assumed to be static and not to
vary over the time period being considered, this assumption is only valid if the rate of change
of the force is much less than the applied force (F >> dF/dt). If the loads vary significantly,
(if the variation in load is not much less than the applied force) then the problem must be
considered as dynamic.
Weightless Member
The weight (W) of the beam is neglected, if it is much less than the total resultant
forces (F) acting on the beam. If the weight of the beam is not neglected, then its effects must
be represented as vertical forces acting at the nodes, by dividing up the weight and lumping it
at the nodes, proportionally according to it's placement along the beam.
Prismatic Member
The beam element is assumed to have a constant cross-section, which means that the
cross-sectional area and the moment of inertia will both be constant (i.e., the beam element is
a prismatic member). If a beam is stepped, then it must be divided up into sections of
constant cross-section, in order to obtain an exact solution. If a beam is tapered, then the
beam can be approximated by using many small beam elements, each having the same crosssection as the middle of the tapered length it is approximating. The more sections that are
used to approximate a tapered beam, the more accurate the solution will be.
The moment of inertia is a geometric property of a beam element, which describes the
beams resistance to bending and is assumed to be constant through the length of the element.
The moment of inertia can be different along different axes if the beam element is not
symmetric, we use the moment of inertia (I) of the axis about which the bending of the beam
occurs
SCE
29
ME2353
Where (Iz) refers to the moment of inertia, resisting bending about the "z" axis and (Iy) about
the "y" axis.
The Beam Element is a Slender Member
A beam is assumed to be a slender member, when it's length (L) is more than 5 times
as long as either of it's cross-sectional dimensions (d) resulting in (d/L<.2). A beam must be
slender, in order for the beam equations to apply, that were used to derive our FEM
equations.
30
ME2353
Isotropic Material
A beam element has the same mechanical and physical properties in all directions,
i.e., they are independent of direction. For instance, cutting out three tensile test specimens,
one in the x-direction, one in the y-direction and the other oriented 45 degrees in the x-y
plane, a tension test on each specimen, will result in the same value for the modulus of
elasticity (E), yield strength
and ultimate strength
. Most metals are considered
isotropic. In contrast fibrous materials, such as wood, typically have properties that are
directionaly dependant and are generally considered anisotropic (not isotropic).
The Proportional Limit is not Exceeded
It is assumed that the beam element is initially straight and unstressed. It is also
assumed that the material does not yield, therefore the beam will be straight after the load is
released. These assumptions mean that the beam must be made of an elastic material, one
which will return to it's original size and shape when all loads are removed, if not stressed
past the materials elastic or proportional limit. It is also assumed that the beam is not stressed
past the proportional limit, at which point the beam will take a permanent set and will not
fully return to it's original size and shape, when all loads are removed. Below the proportional
limit an elastic material is in the linear elastic range, where the strain ( ) varies linearly with
the applied load and the stress ( ) varies linearly according to:
, where E is the
modulus of elasticity.
Rigid Body Modes for the One-Dimensional Beam Element
Rigid body motion occurs when forces and/or moments are applied to an unrestrained
mesh (body), resulting in motion that occurs without any deformations in the entire mesh
(body). Since no strains (deformations) occur during rigid body motion, there can be no
stresses developed in the mesh. In order to obtain a unique FEM solution, rigid body motion
must be constrained. If rigid body motion is not constrained, then a singular system of
equations will result, since the determinate of the mesh stiffness matrix is equal to zero (i.e.,
).
There are two rigid body modes for the one-dimensional beam element, a translation
(displacement) only and a rotation only. These two rigid body modes can occur at the same
time resulting in a displacement and a rotation simultaneously. In order to eliminate rigid
body motion in a 1-D beam element (body), one must prescribe at least two nodal degrees of
freedom (DOF), either two displacements or a displacement and a rotation. A DOF can be
equal to zero or a non-zero known value, as long as the element is restrained from rigid body
motion (deformation can take place when forces and moments are applied) .
For simplicity we will introduce the rigid body modes using a mesh composed of a
single element. If only translational rigid body motion occurs, then the displacement at local
node I will be equal to the displacement at local node J. Since the displacements are equal
there is no strain developed in the element and the applied nodal forces cause the element to
move in a rigid (non-deflected) vertical motion (which can be either up as shown below or it
can be in the downward direction depending on the direction of the applied forces).
SCE
31
ME2353
This rigid body mode can be suppressed by prescribing a vertical nodal displacement.
If rotational rigid body motion occurs, then the rotation at local node I will be equal to
the rotation at local node J (i.e., in magnitude and direction). In this situation the nodal forces
and/or moments applied to the element, cause the element to rotate as a rigid body (either
clockwise as shown below or counterclockwise depending on the direction of the applied
forces and/or moments).
This rigid body mode can be suppressed by prescribing a nodal translation or rotation.
If translational and rotational rigid body motion occurs simultaneously then:
Simple
Examples
with and without Rigid Body Motion
Case
of
Problems
Determinant
Stable/Unsta Rigid Body
of
Mesh
ble
Mode(s)
Equations
Stiffness
Structure
Present
Matrix
Unstable
and
Dependent
Equations
Dependent
Equations
Unstable
SCE
Beam
32
ME2353
Dependent
Equations
Unstable
Stable
None
Independen
t Equations
Stable
None
Independen
t Equations
v a bx cx 2 dx 3
The slope
..(1)
dv
b 2cx 3dx 2
dx
..(2)
v v1
v1 a
at x 0,
1 b
at x l ,
at x l ,
v v2
2
a v1
b 1
v2 a bl cl 2 +dl 3
solving
2 b 2cl 3dl 2
3
1
v v1 21 2
2 2
l
l
2
1
d 3 v1 v2 2 1 2
l
l
c
Substituting the values of a, b, c and d in equation (1), and collecting the coefficients of
v1 , 1 , v2 , 2 we obtain
v N1v1 N 21 N 3v3 N 4 2
where
SCE
33
ME2353
x2
x3
2
,
l2
l3
x2
x3
3 2 2 3 ,
l
l
x 2 x3
,
l l2
x 2 x3
2
l l
N1 1 3
N2 x 2
N3
N4
dx R
x
R y R
R
y
d2y
y 2
R
dx
d 2v
d2
y
N1v1 N 21 N3v2 N 4 2
dx 2
dx 2
d N1
x y
2
dx
2
d N2
dx 2
d N3
dx 2
v1
d N 4 1
dx 2 v2
2
2
x B a
We Know that,
SCE
34
ME2353
K B D B dv
T
d 2 N1
2
dx
d 2 N2
2
d 2N
K y dx2 E y 2 1
d N3
dx
volume
dx 2
2
d N4
dx 2
d 2 N 2
1
2
dx
2
2
d N 2 d N1
dx 2 dx 2
l
2
K E 0 y 2
2
v
d N2 3 d N2 1
dx dx
d 2 N 4 d 2 N1
dx 2 dx 2
Where,
d 2 N2
dx 2
d 2 N1 d 2 N 2
2
2
dx dx
d 2 N2
2
dx
d 2 N3
dx 2
d 2 N4
dv
dx 2
d 2 N1 d 2 N 3
2
2
dx dx
d 2 N 2 d 2 N3
2
2
dx dx
d 2 N3 d 2 N 2
2
2
dx dx
d 2 N3
2
dx
d 2 N 4 d 2 N 2 d 2 N 4 d 2 N3
2
2
2
2
dx dx dx dx
d 2 N1 d 2 N 4
2
2
dx dx
d 2 N2 d 2 N4
2
2
dx dx
dAdx
d 2 N3 d 2 N 4
2
2
dx
dx
2
d 2 N4
2
dx
y 2 dA I
d 2 N1
2
dx
d 2 N2
2
v dx 2 d N
EI
0 d 2 N dx 2 1
3
dx 2
2
d N4
dx 2
d 2 N2
dx 2
d 2 N3
dx 2
d 2 N4
dx
dx 2
Where,
dN1 6 x 6 x 2
d 2 N1 6 12 x
3x 2 2 x3
N1 1 2 3
2 3
2 3
l
l
dx
l
l
dx 2
l
l
N2 x
SCE
dN 2
d 2 N 2 4 6 x
2 x 2 x3
4 x 3x 2
2
1
2
2
l
l
dx
l
l
dx 2
l
l
35
ME2353
N3
dN 3 6 x 6 x 2
d 2 N 3 6 12 x
3x 2 3x3
2 3
l2
l3
dx
l2
l3
dx 2
l
l
N4
dN 4 3 x 2 x
d 2 N4 6x 2
x3 x 2
2
l2
l
dx
l2
l
dx 2
l
l
6 12 x
l2 l3
4 6 x
2
l
l
l 6 12 x 4 6 x 6 12 x 6 x 2
K EI 0
2 l 3 l l 2 l 2 l 3 l 2 l dx
6 12 x l
2 3
l
l
6x 2
2
l
l
l 6
12 x 6 12 x
K11 EI 2 3 2 3 dx
0
l l
l
l
l 6l 12 x 6l 12 x
K11 EI
dx
0
l3
l3
2
2
l 36l 72 xl 72 xl 144 x
K11 EI
dx
0
l6
2
l 36l
144 xl 144 x 2
K11 EI 6 6 6 dx
0
l
l
l
36 xl 2 144 x 2l 144 x3
K11 EI 6
2l 6
3l 6 o
l
36 72 48
K11 EI 3 3 3
l
l
l
K11
12EI
l3
l 6
12 x 4 6 x
K12 EI 2 3
2 dx
0
l l
l
l
SCE
36
ME2353
l 6l 12 x 4l 6 x
K12 EI
dx
2
0
l3
2
2
l 24l 48 xl 36 xl 72 x
K12 EI
dx
0
l5
2
l 24l
84 xl 72 x 2
K12 EI 5 5 5 dx
0
l
l
l
2
l 24 xl
84 x 2l 72 x3
K12 EI 5
5
0
2l 5
3l 0
l
24 42 24
K12 EI 2 2 2
l
l
l
K12
6EI
l2
l 6
12 x 6 12 x
K13 EI 2 3 2 3 dx
0
l l
l
l
l 6l 12 x 6l 12 x
K13 EI
dx
3
0
l3
2
2
l 36l 72 xl 72 xl 144 x
K13 EI
dx
0
l6
2
l 36 xl
144 x 2l 144 x3
K13 EI 6
0
2l 6
3l 6 0
l
36 72 48
K13 EI
l3
K 13
12EI
l3
l 6
12 x 6 x 2
K14 EI 2 3 2 dx
0
l l
l
l
l 6l 12 x 6 x 2l
K14 EI
dx
2
0
l3
SCE
37
ME2353
2
2
l 12l 36 xl 24 xl 72 x
K14 EI
dx
0
l5
2
l 12l
60 xl 72 x 2
K14 EI 5 5 5 dx
0
l
l
l
12 xl 2 60 x 2l 72 x3
K14 EI 5
5
0
2l 5
3l 0
l
l
30 12 24
K14 EI
l2
K14
6EI
l2
l 4
6 x 6 12 x
K 21 EI
2 2 3 dx
0
l
l l
l
K 21
6EI
l2
l 4
6 x 4 6 x
K 22 EI
2
2 dx
0
l l
l
l
l 4l 6 x 4l 6 x
K 22 EI
dx
2
2
0
l
l
16l 2 24 xl 24 xl 36 x 2
K 22 EI
dx
0
l4
2
l 16l
48 xl 36 x 2
K 22 EI 4 4 4 dx
0
l
l
l
16 xl 2 48 x 2l 36 x3
K 22 EI 4
4
0
2l 4
3l 0
l
l
16 24 12
K 22 EI
K 22
SCE
4EI
l
38
ME2353
l 4
6 x 6 12 x
K 23 EI
2 2 3 dx
0
l l
l
l
l 4l 6 x 6l 12 x
K 23 EI
dx
2
3
0
l
l
2
2
l 24l 36 xl 48 xl 72 x
K 23 EI
dx
0
l5
2
2
l 24l 84 xl 72 x
K 23 EI
dx
0
l5
24 xl 2 84 x 2l 72 x3
K 23 EI 5
5
0
2l 5
3l 0
l
l
24 42 24
K 23 EI
l2
K 23
6EI
l2
l 4
6x 6x 2
K 24 EI
2 2 dx
0
l l
l
l
l 4l 6 x 6 x 2l
K 24 EI
dx
2
2
0
l
l
2
2
l 8l 24 xl 12 xl 36 x
K 24 EI
dx
0
l4
2
2
l 8l 36 xl 36 x
K 24 EI
dx
0
l4
8 xl 2 36 x 2l 36 x3
K 24 EI 4
4
0
2l 4
3l 0
l
l
18 12 8
K 24 EI
K 24
SCE
2EI
l
39
ME2353
l 6
12 x 6 12 x
K 31 EI 2 3 2 3 dx
0 l
l l
l
K 31
12EI
l3
l 6
12 x 4 6 x
K 32 EI 2 3
2 dx
0 l
l l
l
K 32
6EI
l2
l 6
12 x 6 12 x
K 33 EI 2 3 2 3 dx
0 l
l l
l
l 6l 12 x 6l 12 x
K 33 EI
dx
3
3
0
l
l
2
2
l 36l 72 xl 72 xl 144 x
K 33 EI
dx
0
l6
2
2
l 36l 144 xl 144 x
K 33 EI
dx
0
l6
36 xl 2 144 x 2l 144 x3
K33 EI 6
0
2l 6
3l 6 0
l
l
36 72 48
K 33 EI
l3
K 33
12EI
l3
l 6
12 x 6 x 2
K 34 EI 2 3 2 dx
0 l
l l
l
l 6l 12 x 6 x 2l
K 34 EI
dx
3
2
0
l
l
2
2
l 12l 24 xl 36 xl 72 x
K34 EI
dx
0
l5
SCE
40
ME2353
2
2
l 12l 60 xl 72 x
K34 EI
dx
0
l5
12 xl 2 60 x 2l 72 x3
K34 EI 5
5
0
2l 5
3l 0
l
l
12 30 24
K 34 EI
l2
K 34
6EI
l2
l 6x
2 6 12 x
K 41 EI 2 2 3 dx
0
l l
l
l
K 41
6EI
l2
l 6x
2 4 6 x
K 42 EI 2
2 dx
0
l l
l
l
K 42
2EI
l
l 6x
2 6 12 x
K 43 EI 2 2 3 dx
0
l l
l
l
K 43
6EI
l2
l 6x
2 6x 2
K 44 EI 2 2 dx
0
l l
l
l
l 6 x 2l 6 x 2l
K 44 EI
dx
2
2
0
l
l
2
2
l 4l 12 xl 12 xl 36 x
K 44 EI
dx
0
l4
2
2
l 4l 24 xl 36 x
K 44 EI
dx
0
l4
SCE
41
ME2353
4 xl 2 24 x 2l 36 x3
K 44 EI 4
4
0
2l 4
3l 0
l
l
12 4 12
K 44 EI
K 44
4EI
l
Therefore K is
12
l3
6
2
K EI l
12
l3
6
2
l
6
l2
4
l
6
l2
2
l
12
l3
6
l2
12
l3
6
l2
6
l2
2
l
6
l2
4
v1
v2
Nodal displacements
1st
degree
of
freedom
2nd
degree of
freedom
3rd
SCE
M1
F1
M2
Nodal forces
vertical
displacement at
node i
slope or rotation at
node i
i or 1
vertical
v j or v2
F2
vi or v1
corresponding to
42
Fi or F1
bending moment at
node i
M i or M 1
F j or F2
ME2353
degree of
freedom
4th
degree of
freedom
displacement at
node j
slope or rotation at
node j
bending moment at
node j
4 j or 2
M j or M 2
The stiffness term kij indicates the force (or moment) required at i to produce a unit
deflection (or rotation) at j, while all other degrees of freedom are kept zero.
Sign conventions followed
Upward forces are positive and upward displacements are positive.
Counter-clockwise moments are positive and counter-clockwise rotations are positive.
Formulae required cantilever beam subjected to concentrated load and moment.
P
PL2
2 EI
PL3
3EI
ML
EI
ML2
2 EI
k11 L3
upwards. Refer table for
3EI
displacement DoF number and force DoF number. Now the beam configuration is given by
Figure 1. We can observe from the figure that the slope at node i is not zero. To make the
slope at i equal to zero, we need to apply a counter-clockwise moment k21 . Refer Figure 2.
indicates the displacement or rotational DoF). v1
k21 L2
at node i. Refer Figure
2 EI
k11 L3 k21 L2
i.e.,
SCE
k11 L2 k21 L
.(2).
2 EI
EI
Solving these two equations, k11 and k21 are found. The fixed end
43
ME2353
reaction force and the reaction moment are assumed to be acting upwards and
counterclockwise, respectively. Now use force equilibrium equation to find fixed end
reaction force k31 . Fy 0 k11 k31 0 and moment equilibrium equation about
node i to find fixed end reaction moment k41 .... M i 0
k41
k11
k11
Figure 2.
Figure 1.
k11
Figure 3.
k31
Figure 4.
k21
k21
1unit
k11 12 EI
L3
6
EI
k21
L2
12 EI
k31 3
L
6 EI
k
41 L2
SCE
44
ME2353
k12 L3
k L2 k L3
. Refer Figure 2. Equating these two deflections 22 12 (1) But this upward
3EI
2 EI
3EI
k12 L2
force k12 will also produce a negative slope at node i which is
. Refer Figure 3. Hence
2 EI
the rotation produced by k22 should be greater than that produced by k12 so that the resultant
k22 L k12 L2
equation
about
node
to
find
fixed
end
k 22 k32 L k 42 0 .
k22
k22
k12
Figure 1.
k22
k12
reaction
1 rad
Figure 4.
k12
Figure 2.
moment
k42
....
k12 6 EI
L2
k42
k22 4 EI
k32 L
6 EI
k32 2
L
2 EI
k
42 L
Figure 3.
SCE
45
ME2353
Initially you have a horizontal beam element. Since v1 1 0 , we can fix node i. To produce
an upward deflection at node j (i.e., allowing third degree of freedom to occur), apply an
upward force k33 .
k33 L3
upwards. Now the beam configuration is given by Figure 1. We can observe from
3EI
the figure that the slope at node j is not zero. To make the slope at j equal to zero, we need to
apply a clockwise moment k43 . Refer Figure 2. But this moment k43 will produce a downward
v2
k43 L2
at node j. Refer Figure 3. In order to have a resultant unit upward
2 EI
displacement at node j, upward displacement produced by force k33 must be greater than the
deflection
k33 L3 k43 L2
1 ..(1). At the
3EI
2 EI
same time, the positive slope produced at node j by the force k33 must be cancelled by the
k33 L2 k43 L
.(2). Solving these two
2 EI
EI
equations, k33 and k43 are found. The fixed end reaction force and the reaction moment are
assumed to be acting upwards and counterclockwise, respectively. Now use force equilibrium
equation to find fixed end reaction force k13 Fy 0 k13 k33 0 and moment
negative slope produced by the moment k43 . i.e.,
equilibrium equation about node i to find fixed end reaction moment k23 ....
k 23 k33 L k 43 0 .
k43
k23
k13
Figure 1.
Figure 4.
k33
k43
k43
1unit
Figure 2.
k33
k33
k13 12 EI
L3
k23 6 EI
L2
12 EI
k33
3
L
6 EI
k 2
43 L
Figure 3.
46
ME2353
k44 L2
. This upward deflection should be canceled by applying a
2 EI
k L3
downward force k34 at node j. The downward deflection produced by k34 is 34 . Refer Figure
3EI
3
2
k L k L
2. Equating these two deflections 44 34 (1) But this downward force k34 will also
2 EI
3EI
k L2
produce a negative slope at node j which is 34 . Hence the rotation produced by k44 should be
2 EI
k L k L2
greater than that produced by k34 so that the resultant rotation is 1 radians. 44 34 1 .(2)
EI
2 EI
Refer Figure 3. Solving these two equations, k34 and k44 are found. The fixed end reaction force
and the reaction moment are assumed to be acting upwards and counterclockwise, respectively.
Now use force equilibrium equation to find fixed end reaction force k14
produce a upward deflection
0 k14 k34 0 and moment equilibrium equation about node i to find fixed end
k44
k24
k44
k14
Figure 1.
Figure 4.
k34
k34
k44
k34
k14 6 EI
L2
k24 2 EI
L
6 EI
k34 2
L
4 EI
k
44 L
Figure 3.
Figure 2.
Problem
Find the slopes at the supports and support reaction forces and support reaction moments for the
beam shown in Figure. Take E=210 GPa, I = 210-4 m4. Daryl Logan P4-24 page 208.
kN
m
5m
4m
v2
SCE
47
ME2353
Force
kN
q
m
qL
kN
2
Lm
Moment
qL2
kN m
12
qL
kN
2
qL2
kN m
12
for element 1,
12.5 kN
10.416667 kN m
5 m kN
for element 2,
10 kN
kN
5
m
10 kN
4m
EI 210 GPa 2 10
12.5 kN
10.416667 kN m
6.66667 kN m
6.66667 kN m
kN
4
4
m 210 10 2 2 10 m 42000 kN-m2
m
4
(1)
(2)
SCE
48
ME2353
12.5
F1 12.5
F1
M
10.416667 0 10.416667
1
12.5 10
F2 22.5
F2
0 3.75
M 2 10.416667 6.66667
F2
10
F2 10
6.66667 0 6.66667
M 3
support
applied forces
reactions
v1 v2 v3 0
1 ?
2 ? 3 ?
0 3.75 10, 080
16,800 10, 080 15, 750 33, 600 42, 000 15, 750
21, 000 2
F3 10
0
0
7,875
15, 750
7,875 15, 750 0
0
0
15, 750
21, 000 15, 750 42, 000 3
0 6.66667
Eliminating the first, third and fifth rows and columns of the stiffness matrix, the reduced matrix
becomes
0 1
10.416667 33, 600 16,800
3.75
Solving these equations
1 3.59623 10 4 rad
2 9.9206349 10 5 rad
3 1.0912698 10 4 rad
Substituting these values in the assembled matrix to find the support reactions, we find
F3 10
F1 12.5 2.625 kN
F1 9.875 kN
SCE
F3 6.71869 kN
49
ME2353
kN
10.416667 kN-m
9.875 kN
6.66667 kN-m
28.406 kN
6.71869 kN
10.416667 kN-m
2.624832 kN
6.45813 kN-m
2.70816 kN-m
2.624832 kN
6.66667 kN-m
3.2811975 kN
3.2811975 kN
F1(1) 10, 080 3.59623 104 10, 080 9.92 105 2.624832 kN
M 1(1)
F2(1)
M 2(1)
Element 2
SCE
50
ME2353
PROBLEM
Given that E=210 GPa and I=410-4 m4, cross section of the beam is constant.
Determine the deflection and slope at point C. calculate the reaction forces and moments.
DARYL LOGAN P 171-172
1kN
20 kN-m
B
3m
3m
M1
F1
M2
M3
F3
F2
1
v1
2
v2
3
v3
SCE
51
ME2353
K K
1
12 EI
l3
6 EI
l2
12 EI
l3
6 EI
2
l
K K
1
6 EI
l2
4 EI
l
6 EI
2
l
2 EI
l
12 EI
l3
6 EI
2
l
12 EI
l3
6 EI
2
l
6 EI
l2
2 EI
l
6 EI
2
l
4 EI
12 18 12 18
36 18 18
6 18
12 18 12 18
18 18 18 36
Assembling:-
F1
12
M
18
1
F2
6 12
3.1
M2
18
F3
0
M
0
3
18
36
18
18
0
0
12
18
24
0
12
18
18 0
0 v1
18 0
0 1
0
0
0 v2
72 18 18 2
18 12 18 v3
18 18 36 3
F2
6
3 .1 1 0
m2
24
0
v2
7 2 2
0
SCE
52
ME2353
F2
12 18
m
18 18
2
6
3.1 10
F
12
18
m3
18 18
F1=10000N
M1=12500N-m
F3=0
M3= -2500N-m
17,500N-m
12,500N-m
10,000N
10,000N
2,500N-m
0
10kN
12.5kN-m
20kN-m
2.5kN-m
10kN
SCE
53
ME2353
SCE
54
ME2353
UNIT III
TWO DIMENSIONAL FINITE ELEMENT ANALYSIS
3.1 INTRODUCTION
Two dimensional elements are defined by three or more nodes in a two dimensional
plane (i.e., x, y plane). The basic element useful for two dimensional analysis is the
triangular element.
SCE
54
ME2353
u c1 c2 x c3 y
In order to find the three unknowns c1 , c2 and c3 , we apply the boundary conditions
at x1 , y1 , u u1
u c1 c2 x1 c3 y1
at x2 , y2 , u u2
u c1 c2 x2 c3 y2
at x3 , y3 , u u3
u c1 c2 x3 c3 y3
y
3 (x3,y3)
u3
y1 c1
y2 c2
y3 c3
u2
u1
1 (x1,y1)
1 x3
y1
y2
y3
2 (x2,y2)
u1
u2
u3
y1
1 2 3
1
y2
1 2 3
2A
1 2 3
y3
where
2 A 1 2 3 and
i x j yk xk y j
i y j yk
i x j xk
1 x2 y3 x3 y2
1 y2 y3
1 x2 x3
2 x3 y1 x1 y3
2 y3 y1
2 x3 x1
3 x1 y2 x2 y1
3 y1 y2
3 x1 x2
c1
1 2 3 u1
1
1 2 3 u2
c2
c 2 A
3 u3
2
3
1
Substituting the values of c1, c2 and c3 in u c1 c2 x c3 y , we get
SCE
1
i i x i y ,
2A
55
i 1, 2,3
ME2353
y
4 (0,b)
1 (0,0)
3 (a,b)
2 (a,0)
x
u x , y c1 c2 x c3 y c4 x y
(1)
This polynomial contains four linearly independent terms and is linear in x and y, with a
bilinear term in x and y. The polynomial requires an element with four nodes. There are two
possible geometric shapes: a triangle with the fourth node at the centroid of the triangle or a
rectangle with nodes at the vertices.
A triangle with a fourth node at the center does not provide a single-valued variation of u at
inter-element boundaries, resulting in incompatible variation of u at inter-element boundaries and is
therefore not admissible.
The linear rectangular element is a compatible element because on any side, the single
variable u varies only linearly and there are two nodes to uniquely define it.
Here we consider an approximation of the form given in eqauation (1) and use a rectangular
element with sides a and b. For the sake of convenience we choose a local coordinate system
at 0, 0 , u u1
u c1
at a, 0 , u u2
u c1 c2 a
at a, b , u u3
u c1 c2 a c3ab
at 0, b , u u3
u c1 c4b
SCE
c4
u1 u2 u3 u4
,
ab
56
ME2353
v3
3 x1 , y1
u3
v1
1 x1 , y1
v2
u1
u2
2 x2 , y2
Figure shows a 2-D two-variable linear triangular element with three nodes and the two dof at each
node. The nodes are placed at the corners of the triangle. The two variables (dof) are displacement
in x-direction (u) and displacement in y-direction (v). Since each node has two dof, a single element
has 6 dof. The nodal displacement vector is given by
u1
v
1
u
U 2
v2
u3
v3
We select a linear displacement function for each dof as
u x, y c1 c2 x c3 y
v x, y c4 c5 x c6 y
where u x, y and v x, y describe displacements at any interior point x, y of the element.
The above two algebraic equations can also be written as
u 1 x
v 0 0
SCE
y 0 0
0 1 x
c1
c
2
0 c3
y c4
c5
c6
57
ME2353
Using steps we had developed for the 2-D single-variable linear triangular element, we can write
c1
1 2 3 u1
1
1 2 3 u2
c2
c 2 A
3 u3
2
3
1
c4
1 2 3 v1
1
v
c5
1
2
3
2
c 2 A
3 v3
2
6
1
and using the interpolation functions we had developed for the 2-D single-variable linear triangular
element, we can write
1
i i x i y ,
2A
i 1, 2,3
u x, y N1
0
N1
U N a
N2
0
0
N2
N3
0
u1
v
1
0 u2
N 3 v2 The
u3
v3
SCE
58
ME2353
x x
v
y
y
xy u v
y x
N i
i
x
and note that
N i
i
y
N1
u
N1u1 N 2u2 N 3u3
x x
x
N 2
x
N1
v
N1v1 N 2 v2 N 3v3
y y
y
N 2
y
u1
v
1
u1
N 3
N 3
N1
N 2
u
0
0
0 2
u2
x
x
x
x
v2
u3
u3
v3
u1
v
1
v1
N 3
N 3 u2
N1
N 2
0
0
v2 0
y
y
y
y v2
v
3
u3
v3
N1
u v
N1
x x
y 0
xy N
1
0
N1
y
N1
x
N 2
x
0
N 2
y
0
N 2
y
N 2
x
N 3
x
0
N 3
y
u1
0 v
1 1
N 3 u2
0
y v2
1
N 3 u3
x v3
N1
x
N 2
y
N 2
x
N 3
y
N 3
x
u1
v
1
u2
v2
u3
v3
u1
v
0 1
u
3 2
v
3 2
u3
v3
31 B 36 a61
SCE
59
ME2353
D
x
x
y D y
xy
xy
D B a
B D B
T
63
33
dV t B D B dxdy .
T
36
Volume
Area
K tA B D B
T
D matrix is the material constitutive matrix, either for the plane-stress case or for the plane-strain
case depending on the problem in hand.
and substituting them back in u, we get
y
x
z
E
E
E
y
y x
z
E
E
E
y z
z x
E
E
E
x
SCE
60
ME2353
E
E
y
y x
E
E
solving the above two equations
for x and y , we get
x
E
x y and
1 2
E
x y
y
1 2
x
xy G xy
E
xy
21
1 E
1 21
2
2
xy
1 1
1
2
xy
E
xy
21
E 1
xy
1 2
2
1
1
0 0
x
0 y
1
xy
2
0
z x x
SCE
61
ME2353
x
y 2 x
E
E
E
y
y x y 2 x
E
E
E
rearranging the terms we get
x
y
x
1 2 1
E
E
y
y x 1
1 2
E
E
mutiplying by X by and Y by 1 -
y 2
x x 1 2
1
E
E
1 y x 1 1 y 1 1 2
E
E
y
1 1 2
x 1 2
E
E
adding the above two equations to eliminate x
x 1 y
x 1 y
x 1 y
x 1 y
y E
y
E
y
E
y
E
y
1 1 1 2
1 1 1 1
1 2 1 1
1 2 2
x 1 y
1 1 2
similarly
x E
1 x y
1 1 2
andas before
E
xy
21
writing x , y and xy in matrixform
xy
SCE
62
ME2353
x
1
E
y
0
xy
1
0
x
0 x
1 2 xy
2
0
It is difficult to represent the curved boundaries by straight edges element a large number of
element may be used to obtain reasonable resembalance between original body and the
assemblage
Two-Dimensional Problems
Review of the Basic Theory
In general, the stresses and strains in a structure consist of six components:
and
sx , sy , s z , txy , t yz , tzx
for stresses,
for strains.
sy
t yz
t xy
sx
tzx
sz
x
z
Under contain conditions, the state of stresses and strains can be simplified. A general
3-D structure analysis can, therefore, be reduced to a 2-D analysis.
t zx
(e z
0)
(1)
Plane stress:
sz
SCE
tyz
63
ME2353
Plane strain:
A long structure with a uniform cross section and transverse loading along its
length (z-direction).
1/ E
ey
n /E
gxy
n /E
1/ E
0
sx
ex
sy
ey
1/ G
txy
gxy 0
where e0 is the initial strain, E the Youngs modulus, n the Poissons ratio and G the
shear modulus. Note that,G =
(
SCE
64
ME2353
which means that there are only two independent materials constants for homogeneous and
isotropic materials.
We can also express stresses in terms of strains by solving the above equation,
The above relations are valid for plane stress case. For plane strain case, we
need to replace the material constants in the above equations in the following fashion,
n
n
1- n
For example, the stress is related to strain by
Initial strains due to temperature change (thermal loading) is given by,
where a is the coefficient of thermal expansion, T the change of temperature. Note that
if the structure is free to deform under thermal loading, there will be no (elastic) stresses in the
structure.
3.6 GENERALIZED COORDINATES APPROACH TO NODEL APPROXIMATIONS
ty
p
tx
y
St
Su
x
The boundary S of the body can be divided into two parts, Su and St. The boundary
conditions (BCs) are described as, in which tx and ty are traction forces (stresses on the boundary)
and the barred quantities are those with known values.
In FEM, all types of loads (distributed surface loads, body forces, concentrated forces
and moments, etc.) are converted to point forces acting at the nodes.
Exact Elasticity Solution
The exact solution (displacements, strains and stresses) of a given problem must satisfy the
equilibrium equations, the given boundary conditions and compatibility conditions (structures
should deform in a continuous manner, no cracks and overlaps in the obtained displacement field)
SCE
65
ME2353
The element connectivity table for the above domain is explained as table.
Element (e)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Nodes
123
234
435
536
637
738
839
931
Shape function for the CST element Shape function N1 = (p1 + q1x + r1y) / 2A
SCE
66
ME2353
u(x, y) N1
Displacement function u =
v(x, y) 0
N2
N3
N1
N2
u1
v1
0 u2
X
N 3 v2
u3
v3
q1
1
0
Strain Displacement matrix [B] =
2A
r1
Where, q1 = y2 y3
r1 = x3 x2
q2 = y3 y1
r2 = x1 x3
q3 = y1 y2
r3 = x2 x1
q2
r1
q1
0
r2
q3
0
r2
q2
r3
r3
q3
Stress Strain relationship matrix (or) Constitutive matrix [D] for two
dimensional element
SCE
1 v
1 v
1 v
67
ME2353
Stress Strain relationship matrix for two dimensional plane stress problems
The normal stress z and shear stresses xz, yz are zero.
1
v
0
v 1
0
[D] = 1 v 2
1 v
0
0
q1
1
0
[B] =
2A
r1
q2
r1
q1
0
r2
q3
0
r2
q2
r3
r3
q3
1
v
0
0
v 1
[D] = 1 v 2
1 v
0
0
Temperature Effects
Distribution of the change in temperature (T) is known as strain. Due to the
change in temperature can be considered as an initial strain e0.
= D (Bu - e0)
Galerkin Approach
Stiffness matrix [K] e = [B]T [D][B] A t.
Force Vector {F}e = [K]e {u}
SCE
68
ME2353
Problem (I set)
1. Determine the shape functions N1, N2 and N3 at the interior point P for the
triangular element for the given figure.
The two dimensional propped beam shown in figure. It is divided into two CST
elements. Determine the nodal displacement and element stresses using plane stress
conditions. Body force is neglected in comparison with the external forces.
Take, Thickness (t) = 10mm,
Youngs modulus (E) = 2x105 N/mm2,
Poissons ratio (v) = 0.25.
SCE
69
ME2353
3. A thin plate is subjected to surface traction as in figure. Calculate the global stiffness matrix.
SCE
70
ME2353
Elasticity equations are used for solving structural mechanics problems. These
equations must be satisfied if an exact solution to a structural mechanics problem is to be
obtained. Thest are four basic sets of elasticity equations they are
TRUSS ELEMENT
A truss element is defined as a deformable, two-force member that is subjected to
loads in the axial direction. The loads can be tensile or compressive. The only degree of
freedom for a one-dimensional truss (bar) element is axial (horizontal) displacement at each
node.
If a truss structure is tapered, then it can be approximated by using many small truss
elements, each having the same cross-section as the middle of the tapered length it is
approximating. The more sections that are used to approximate a tapered truss, the more
accurate the solution will be.
SCE
71
ME2353
Weightless Member
The weight (W) of the truss is neglected since it is assumed to be much less than the
total resultant forces (F) acting on the truss. If the weight of the truss is not neglected, then
its effects must be represented as vertical forces acting at the nodes. But since truss element
is defined as two-force member it cannot have any vertical (shear) force, thus the member
weight has to be neglected. If shear forces exist, then a beam element must be used to model
the structure.
Nodal Forces
For one-dimensional truss element, forces (loads) can only be applied at the nodes of
the element, but not between the nodes. This is consistent with the FEM equations which
relate nodal forces to nodal displacements through the stiffness matrix.
Axially Loaded
For one-dimensional truss element, forces (loads) can only be applied at the centroid
of the element cross-sectional area.
Buckling Effect not Considered
A bar element can be subjected to either tensile or compressive forces. Tensile forces
can be applied to a bar of any cross-sectional area or member length, and failure is
associated with sudden fracture or general yielding. When compressive forces are applied to
a member, it can either fail due to crushing or buckling. Buckling is present when the
member bends and laterally deflects as shown on the right figure below.
Buckling is not accounted for in the formulation of the truss element. Members that
do not buckle are classified as short columns and members that buckles are classified as
long columns. The structural response of a short column can be predicted with a truss
element.
To determine if buckling will occur the reader should refer to a mechanics of
material textbook. We will now introduce a simple geometric guideline to determine if
buckling might occur. If the ratio between the member length and the least dimension of the
cross-section is equal or less than 10, the member is considered a short column and buckling
will not occur, i.e.,
SCE
72
ME2353
In the second case if a bar element is subjected to a compressive force, the element will not
predict the buckling response. One should note that the above geometric rule is a simple
guideline, however, in reality buckling depends not only on the member length and crosssectional area, but material properties and support conditions.
Isotropic Material
A truss element has the same mechanical and physical properties in all directions, i.e., they
are independent of direction. For instance, cutting out three tensile test specimens, one in the
x-direction, one in the y-direction and the other oriented 45 degrees in the x-y plane, a
tension test on each specimen, will result in the same mechanical values for the modulus of
elasticity (E), yield strength y and ultimate strength u. Most metals are considered
isotropic. In contrast fibrous materials, such as wood, typically have properties that are
directionally dependant and are generally considered anisotropic (not isotropic).
Constant (Static) Load
The loads that are applied to the truss element are assumed to be static and not to vary over
the time period being considered. This assumption is only valid if the rate of change of the
force is much less than the applied force (F >> dF/dt), i.e., the loads are applied slowly. If
the loads vary significantly, (if the variation in load is not much less than the applied force)
then the problem must be considered as dynamic.
Poisson's Effect not Considered
Poisson's ratio is a material parameter. Poisson's effect is when a uniform cross-section bar
is subject to a tensile load, and the axial stretching is accompanied by a contraction in the
lateral dimension. For one-dimensional truss element., this effect is neglected for simplicity,
i.e., v = 0.
Cross Section Remains Plane
For one-dimensional element, although the force(s) are acting on only the centroid of the
truss (bar) element, it is assumed that it has a uniform effect to the plane. Thus the cross
section will move uniformly and remain plane and normal to the axial axis before and after
loading.
SCE
73
ME2353
Homogenous Material
A truss element has the same material composition throughout and therefore the same
mechanical properties at every position in the material. Therefore, the modulus of elasticity
E is constant throughout the truss element. A member in which the material properties varies
from one point to the next in the member is called inhomogenous (non-homogenous). If a
truss is composed of different types of materials, then it must be divide up into elements that
are each of a single homogeneous material, otherwise the solution will not be exact.
The left figure shows a composite bar composed of brass and aluminum. This structure can
be divided into two elements as shown on the right, one element for the brass with E 1 = 15 x
106 psi and one for the aluminum with E2 = 10 x 106 psi.
Frame
SCE
74
ME2353
vj
uj
Fiy
ui
vi
Fix
ui
u j
F jx
y
vi
vj
F jy
Note that the deformations occurring in the truss members are so small that they are only
axial. The axial displacement of the truss can be resolved along horizontal x-axis and
vertical y-axis. But in our derivation, let us resolve the horizontal and vertical displacements
(in xy-axes) of a joint along and perpendicular to the truss member (in xy -axes). Refer to
the Figure in the next page. Note ui sin component acting towards negative y -direction
and all other components acting towards in +ve x - and y -directions.
ui cos
ui
ui ui cos vi sin
vi ui sin vi cos
ui sin
u j u j cos v j sin
vi cos
vi
vj u j sin v j cos
vi sin
u cos
v
i sin
u j 0
vj 0
u T u
sin
cos
0
0
0
0
cos
sin
0 ui
0 vi
sin u j
cos v j
It is important to note that the displacements vi and vj are both zero since there can be no
displacements perpendicular to the length of the member. Also T T
1
SCE
75
ME2353
Similarly, we resolve forces along the length of the member (positive x direction)
and perpendicular to the length of the member (positive y direction)
Fix cos
F
iy sin
Fjx 0
Fjy 0
F T F
sin
cos
0
0
0 Fix
0 Fiy
sin Fjx
cos Fjy
0
0
cos
sin
The arbitrarily inclined truss member can be thought of as a simple bar element
oriented at the same angle . Hence, we can write the finite element equation for this
inclined bar element (in xy coordinate system) as
Fix
1
F
iy AE 0
F jx L 1
F jy
0
0 1
0 0
0 1
0 0
0 ui
0 vi
0 u j
0 vj
F k u
Substituting F and u from the previous equations, we can write
T F k T u
Pre-multiplying the above equation by T ,
1
T T F T k T u
1
But T
Fix
c2
cs c 2 cs ui
F
s 2 cs s 2 vi
iy AE cs
2
2
cs u j
Fjx L c cs c
2
Fjy
cs
s 2 v j
cs s
where c cos 2 and s sin 2 .
Computation of strain and stress in the truss element
SCE
76
ME2353
The change in length of the truss member is equal to the change in axial displacement of the
truss member in the xy co-ordinate system
u j ui
cos
Strain
in
the
truss
sin
element
cos
ui
v
i
sin
u j
v j
is
given
by
i.e.,
ui
cos sin cos sin vi
e
L
u j
v j
Stress in the truss element is given by e E e , i.e.,
e E
cos
sin
cos
ui
sin vi
u j
v j
Problem
The two-element truss is subjected to external loading as shown in figure. Using the
same node and element numbering as shown in figure, determine the displacement
components at node 3, the reaction components at nodes 1 and 2, and the element
displacement, stresses and forces. The elements have modulus of elasticity E1 = E2 = 1010
lb
and cross-sectional areas A1 = A2 = 1.5 in 2
2
in
(0, 40)
300 lb 500 lb
(40, 40)
2
(0,0)
SCE
77
ME2353
1 4
Degree of freedom
6 3
3
5
2
2
1
For element 1
3
For element 2
vj
uj
vi
vj
ui
uj
vi
b. Nodal forces
element 1
element 2
Fiy
Fjy
Fix
i
SCE
78
Fjx
ME2353
FINITEFELEMENT EQUATION
ix
Fix
c2
F
iy
AE
cs
F
L c2
jx
F
cs
jy
cs
s2
cs
2
s
c2
cs
c2
cs
cs
s2
cs
s2
ui
vi
uj
v
j
For element 1
AE
1 .5 1 0 1 0 6
lb
2 .6 5 1 6 5 1 0 5 2
L
5 6 .5 6 8 5
in
4 5,
(1)
10
5 1.325826 1.325826 1.325826 1.325826
L 0.5 0.5 0.5 0.5
0,
A E 1 .5 1 0 1 0 6
3 .7 5 1 0 5
L
40
For element 2
(2)
AE 0
L 1
0
0
105
0
3 3 .7 5
4
0
5 3 .7 5
6
0
3 .7 5
3 .7 5
0
0
0
1
F1x
1 1.325
2 1.325
F1y
F
3 0
2 x 105
4 0
F2 y
5 1.325
F3x
F
6
1.325
3y
SCE
1.325
1.325
3.75
1.325 3.75
1.325
79
1.325 1.325 u1
1.325 1.325 v1
3.75
0 u2
quation
0
0 v2
5.0751 1.325 u3
1.325 1.325 v3
ME2353
knowns
F3 x 500 lb
F3 y 300 lb
unknowns
F3 x 5 0 0 lb
F1 x
F3 y 3 0 0 lb
F1 y
u1 0
F2 x
u2 0
v1 0
F2 y
u2 0
u3
v1 0
v2 0
v3
u1 0
v2 0
F1x
1.325826 1.325826
F1 y
1.325826 1.325826
F
0
0
2 x 105
0
0
F2 y
1.325826 1.325826
500
300
1.325826 1.325826
u3 5.33 10
0 1.325826 1.325826 0
0 1.325826 1.325826 0
0
3.75
00
0
0
00
0 5.0751826 1.325826 u3
0 1.325826 1.325826 v3
0
0
3.75
0
3.75
0
u3 and v3
F1x 300 lb
F1 y 300 lb
in
F2 x 200 lb
v3 1.731 10 in
F2 y 0
SCE
80
ME2353
E
10 10
lb
[u5 cos 0 u3 sin 0 ]
(0.5333 1) (0 0) 133.325 2
L
40
in
E
(u 5 cos 45 u 6 sin 45 ) ( u1 cos 45 u 2 sin 45 )
L
10 10 6
(0.5333 10 3 cos 45 ) (1.731 10 3 sin 45 ) 0
56.57
10 10 6
lb
PROBLEM
To illustrate how we can combine spring and bar element in one structure, we can
solve the two-bar truss supported by a spring as shown below. Both bars have E = 210 GPa
and A = 5.0 x10-4 m2. Bar one has a length of 5 m and bar two a length of 10 m. the spring
stiffness is k = 2000 kN/m.
25kN
2
5m
3
1
45
10m
k=2000 kN/m
3
4
Solution :
Given : E = 210 GPa
N
m2
N
NOTE: A spring is considered as a bar element whose stiffness is 2 x 106 2
m
A = 5.0 x10-4 m2 , L1 = 5 m,
SCE
L2 = 10 m,
81
K = 2 x 106
ME2353
v2
Displacements
u2
2
v63
3
u3 5
v12
u1 1
Forces
v4
u4
F2Y
8
7
F2X
2
F3Y
3
1
F3X
F1Y
F1X
F4Y
F4X
=1350
l2=cos2 =0.5
m2=sin2 =0.5
lm=cos sin =-0.5
SCE
2
1
135O
82
ME2353
K
(1)
5 10
210 10 kN/m
6
5m
1
(1)
180O
2
3
5 10
210 10 kN/m
6
(2)
10 m
1
1 2
(2)
1 1 1 1 1
1 1 1
5 2 1
105 10
3 1 1 1 1
4 1 1 1 1
Element 2:
=1800
l2=cos2 =1
m2=sin2 =0
lm=cos sin =0
1 1
5 2 0
105 10
5 1
6 0
1 1
2 0
3 1
4 0
5 6
0 1 0
0 0 0
0 1 0
0 0 0
0 1 0
0 0 0
0 1 0
0 0 0
Element 3:
=2700
l2=cos2 =0
m2=sin2 =1
lm=cos sin =0
SCE
83
ME2353
(3)
0
0 1
0 0
0
1
6 2
10
7
0 1
F2 X
3 105 105 105 105
0
10
5 105
0
0
0 105
F3 X
F3Y
6
0
0
0
0
0
7
0
0
0
0
0
F4 X
F
8
0 20
0
0
0
4Y
0
20
0
0
0
0
0
20
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
u1
v
1
u2
v2
u3
v3
u4
v
4
0
25
F2 X
F2Y
5
10
F3 X
F3Y
F
4X
F
4Y
SCE
0
105 108 105 105
105
0
0
0 105
0
0
0
0
0
0
0
0
0
0
0 20
0
0
0
84
0
20
0
0
0
0
0
20
u1
v
1
0
0
0
0
0
0
ME2353
0
210 105 u1
5
10
25
105 125 v1
On solving,
u1=-1.724 x 10-3 m
v1=-3.4482 x 10-3 m
F3y = 0
F4x = 0
F4y = 6.89 kN
(1)
1.724
E
3 3.4482
l m l m10
0
L
(1)
1.724
E
3 3.4482
l m l m10
0
L
85
ME2353
12 in
3 in
50000 lb
x
24 in
Solution
The beam structure looks very different from a spring. However, its behavior is
very similar. Deflection occurs along the x-axis only. The only significant difference
between the beam and a spring is that the beam has a variable cross-sectional area. An
exact solution can be found if the beam is divided into an infinite number of elements,
then, each element can be considered as a constant cross-section spring element, obeying
the relation F = ku, where k is the stiffness constant of a beam element and is given by
k = AE/L.
In order to keep size of the matrices small (for hand- calculations), let us divide the
beam into only three elements. For engineering accuracy, the answer obtained will be in
an acceptable range. If needed, accuracy can be improved by increasing the number of
elements.
As mentioned earlier in this chapter, spring, truss, and beam elements are lineelements and the shape of the cross section of an element is irrelevant. Only the crosssectional area is needed (also, moment of inertia for a beam element undergoing a
bending load need to be defined). The beam elements and their computer models are
shown
Here, the question of which cross-sectional area to be used for each beam section
arises. A good approximation would be to take the diameter of the mid-section and use
that to approximate the area of the element.
k1
k2
k3
k1
k3
3
1
SCE
k2
2
86
ME2353
Cross-sectional area
The average diameters are: d1 = 10.5 in., d2 = 7.5 in., d3 = 4.5. (diameters are taken at the
mid sections and the values are found from the height and length ratio of the triangles
shown in figure 2.10), which is given as
12/L = 3/(L-24),
L = 32
A2 = 56.25 in2
A3 = 15.9 in2
24 in
12 in
d1
d2
d3
3 in
Original
Averaged
L- 24
L
Stiffness
k1 = A1 E/L1 = (86.59)(4 106/8) = 4.3295 107 lb./in., similarly,
k2 = A2 E/L2 = 2.8125 107 lb./in.
k3 = A3 E/L3 = 7.95 106 lb./in.
Element Stiffness Equations
[K(1)] = 43.295 107 1 -1
-1 1
Similarly,
SCE
1 -1
-1 1
1 -1
-1 1
87
ME2353
43.295
-43.295
0
0
-43.295 43.295+28.125
-28.125
0
0
-28.125
28.125+7.95 -7.95
0
0
-7.95
7.95
[Kg] =
106
10
43.295 -43.295
0
0
-43.295 71.42 -28.125 0
0
-28.125 36.075 -7.95
0
0
-7.95
7.95
u1
u2 =
u3
u4
F1
F2
F3
F4
10
71.42
-28.125
0
-28.125 0
36.075 -7.95
-7.95
7.95
u2
u3
u4
0
0
5000
Solving we get,
u2
u3 =
u4
0.0012
0.0029
0.0092
in.
The deflections u2, u3, and u4 are only the approximate values, which can be
improved by dividing the beam into more elements. As the number of elements increases,
the accuracy will improve.
SCE
88
ME2353
UNIT IV
DYNAMIC ANALYSIS USING ELEMENT METHOD
4.1 INTRODUCTION
It provides the basic equations necessary for structural dynamical analysis and developed
both the lumped and the consistent mass matrix involved in the analysis of bar beam and spring
elements.
4.1.1 Fundamentals of Vibration
Any motion which repeats itself after an interval of time is called vibration or oscillation or
periodic motion
All bodies possessing mass and elasticity are capable of producing vibration.
4.1.2 Causes of Vibrations
o Unbalanced forces in the machine. These force are produced from within the machine
itself
o Elastic nature of the system.
o Self excitations produced by the dry friction between the two mating surfaces.
o External excitations applied on the system.
o Wind may causes vibrations
o Earthquakes may causes vibrations
4.1.3 Types of Vibrations
1.According to the actuating force
Free or natural vibrations
Forced vibrations
Damped vibrations
Undamped vibrations
2.According to motion of system with respect to axis
Longitudinal vibrations
Transverse vibrations
Torsional vibrations
SCE
88
ME2353
Mid surface
3
x
2
1
w1 ,
w
,
x 1
w
y
w2 ,
w, v
w
.
,
y
Ni wi
i 1
N xi (
w
x
)i
N yi (
w
y
)i
where Ni, Nxi and Nyi are shape functions. This is an incompatible element! The
stiffness matrix is still of the form
k = BTEBdV ,
where B is the strain-displacement matrix, and E the stress- strain matrix.
Minding Plate Elements:
4-Node Quadrilateral
SCE
8-Node Quadrilateral
89
ME2353
3
6
4
1
w
,
, x,
; y
x, y.
0, etc., at selected
xz
yz
nodes to reduce the DOF (using relations in (15)). Obtain:
1
At each node: w, x
w
,
x
.
y
SCE
90
ME2353
Test Problem:
z
P
y
L
x
L/t = 10,
= 0.3
wc ( PL2/D)
0.00593
0.00598
0.00574
0.00565
:
0.00560
Mesh
2 2
4 4
8 8
16 16
:
Exact Solution
SCE
91
ME2353
Example:
Sea shell, egg shell (the wonder of the nature); Containers, pipes, tanks;
Car bodies;
Roofs, buildings (the Superdome), etc.
Forces in shells:
Membrane forces + Bending Moments
(cf. plates: bending only)
internal forces:
p
membrane stresses
dominate
SCE
92
ME2353
Shell Theory:
Thin shell theory
Shell theories are the most complicated ones to formulate and analyze in mechanics
(Russians contributions).
Engineering
Shell Elements:
v
u
Q4 or Q8 shell element.
Curved shell elements:
v
i
SCE
93
ME2353
L/2
F
A
R
80o
F
Pinched Cylinder
Roof
F2
R
F
b
F1
F
Pinched Hemisphere
F(t)
m
c
SCE
f=f(t)
m - mass
k - stiffness
c - damping
94
ME2353
Free Vibration:
f(t) = 0 and no damping (c = 0)
Eq. (1) becomes
mu ku
(meaning: inertia force + stiffness force = 0)
Assume:
u(t)
where
U sin (wt) ,
is the frequency of oscillation, U the amplitude.
i.e.,
2
U m sin( t)
2
w m k U 0.
kU sin( t) 0
0,
which yields
w
m
This is the circular natural frequency of the single DOF system (rad/s). The cyclic frequency
(1/s = Hz) is
w
f
,
2p
SCE
96
ME2353
u = U s in w t
U
t
U
T=1/f
Undamped Free Vibration
With non-zero damping c, where
0
cc
2mw
2 km
wd
where x
c
(damping ratio).
cc
0.15
(usually 1~5%)
w.
SCE
97
ME2353
Mu
Cu
in which:
f (t) ,
Ku
(8)
mass matrix,
damping matrix,
stiffness matrix,
forcing vector.
rAL 1
2
r,A,L
u1
2
u2
rAL
2
2
0
0
rAL
2
diagonal atrix
SCE
98
ME2353
v1
q
v2
r, A, L
q2
rNT NdV
156
rAL 22L
420 54
13L
22L
54
13L V1
4L2
13L
13L
156
3L2 Q1
22L v2
3L2
22L
4L2
Q2
t (time)
L (length)
m (mass)
a (accel.)
f (force)
r (density)
Choice I
s
m
kg
m/s2
N
kg/m3
Choice II
s
mm
Mg
mm/s2
N
Mg/mm3
SCE
99
ME2353
Mu
Ku
u sin(
w t ),
w u cos( w t ),
(t )
u
u sin(
w t ),
M u
w i M ui
0.
u i (i=1,2,,n) are the normal modes (or natural modes, mode shapes, etc.).
SCE
100
ME2353
u iT
Ku
u iT M u
j 0,
0,
for i j,
if wi w j . That is, modes are orthogonal (or independent) to each other with respect to K and
M matrices.
Note:
Magnitudes of displacements (modes) or stresses in normal mode analysis have no physical
meaning.
For normal mode analysis, no support of the structure is necessary.
i = 0 there are rigid body motions of the whole or a part of the structure. apply this to check
the FEA model (check for mechanism or free elements in the models).
Lower modes are more accurate than higher modes in the FE calculations (less spatial variations
in the lower modes fewer elements/wave length are needed).
Example:
v2
q2
r, A, EI
1
2
L
2
K wM
K
SCE
EI 12
3
L
101
ME2353
EVP:
12 156l
6L
2
4L
6L 22Ll
in which l
rAL
22Ll
0,
2
4L l
/ 420 EI .
3.533
EI
4
rAL
#2
#3
w2
#1
34.81
EI
v2
q2
1.38
1
,
4
rAL
v2
q2
7.62
.
L
Exact solutions:
w1
EI
4
rAL
3.516
1
2
,
w2
22.03
EI
4
rAL
1
2
.
with
SCE
K
&
(17)
are found from
102
Damping ratio
ME2353
Modal Damping
Incorporate the viscous damping in modal equations.
Modal Equations
Use the normal modes (modal matrix) to transform the coupled system of
dynamic equations to uncoupled system of equations.
We have
2
M
u
i
0i ,
u
u
T
i
T
i
1,2,..., n
(18)
satisfies:
K u
M u
0,
0,
1,
for i
j,
and
u iT M u
u iT K u
SCE
2
i
i
103
for i = 1, 2, , n.
ME2353
u 1
n )
u 2 L u n
&z&
z&
Pre-multiply by
&z&
f ( t ).
p ( t ),
where C
I
T
(proportional damping),
f (t ) .
2 i
2
i z i
i
i z& i
( t ),
i = 1,2,,n. (24)
Equations in (22) or (24) are called modal equations. These are uncoupled, secondorder differential equations, which are much easier to solve than the original dynamic
equation (coupled system).
To recover u from z, apply transformation (21) again, once z is obtained from (24).
SCE
104
ME2353
Notes:
Only the first few modes may be needed in constructing the modal matrix
(i.e.,
could be an n m rectangular matrix with m<n). Thus, significant reduction in the
size of the system can be achieved.
Modal equations are best suited for problems in which higher modes are not
important (i.e., structural vibrations, but not shock loading).
E u
Harmonicloading
(25)
2 i
i Zk i
zi
pi
sin t,
i=1,2,,m.
(26)
zi
zi (t)
(1
2
p
i
2 2
i
i
(2i
sin( t
2
where
ci
/i i
ci
cc
, damping
ratio
2m i
105
ME2353
t
u(t)
t
Compute responses by integrating through time:
u1
u n u n+1
u2
t0 t1 t2
t n t n+1
B. Modal Method
First, do the transformation of the dynamic equations using the modal matrix before the
time marching:
Then, solve the uncoupled equations using an integration method. Can use, e.g.,
10%, of the total modes (m= n/10).
Uncoupled system, Fewer equations,
No inverse of matrices,
More efficient for large problems.
SCE
106
ME2353
k1
k2
o
2
k3
o
3
o
4
Figure 2.4
Solution:
Again apply the three steps outlined previously.
Step 1: Find the Element Stiffness Equations
Element 1:
1
[K(1)] = 10 -10 1
-10 10 2
Element 2:
Element 3:
3
(3)
[K ] =
SCE
[K(2)] = 15 -15 2
-15 15 3
20 -20 3
-20 20 4
107
ME2353
10 -10
0
0
-10 10 + 15 -15
0
0
-15
15 + 20 -20
0
0
-20
20
10 -10 0
0
= -10 25 -15 0
0 -15 35 -20
0 0 -20 20
10 -10 0 0
F1
F2 = -10 25 -15 0
F3
0 -15 35 -20
F4
0 0 -20 20
u1
u2
u3
u4
PROBLEM
In the spring structure shown, k1 = 10 N/mm, k2 = 15 N/mm, k3 = 20 N/mm, k4 = 25 N/mm, k5 = 30
N/mm, k6 = 35 N/mm. F2 = 100 N. Find the deflections in all springs.
k1
k3
k2
F2
k6
k4
k5
SCE
108
ME2353
Solution:
Here again, we follow the three-step approach described earlier, without specifically
mentioning at each step.
Element 1:
Element 2:
Element 3:
[K(2)] =
15 -15 1
-15 15 2
[K(3)] =
20 -20 2
-20 20 3
Element 4:
2
(4)
[K ] =
Element 5:
Element 6:
[K(1)] = 10 -10 1
-10 10 4
25 -25 2
-25 25 3
2
(5)
[K ] =
30 -30 2
-30 30 4
[K(6)] =
35 -35 3
-35 35 4
[Kg] =
10+15
-15
0
-10
-15 15+20+25+30 -20-25
-30
0
-20-25
20+25+35 -35
-10
-30
-35
10+30+35
1
2
3
4
25 -15 0 -10
[Kg] = -15 90 -45 -30
0 -45 80 -35
-10 -30 -35 75
SCE
109
ME2353
F1
F2 =
F3
F4
25
-15
0
-10
-15 0 -10
90 -45 -30
-45 80 -35
-30 -35 75
u1
u2
u3
u4
Which
100
0
90 -45
= -45 80
u2
u3 gives
Deflections:
Spring 1:
u4 u1 = 0
Spring 2:
u2 u1 = 1.54590
Spring 3:
u3 u2 = -0.6763
Spring 4:
u3 u2 = -0.6763
Spring 5:
u4 u2 = -1.5459
Spring 6:
u4 u3 = -0.8696
SCE
110
ME2353
UNIT V
APPLICATIONS IN HEAT TRANSFER &FLUID MECHANICS
5.1 ONE DIMENSIONAL HEAT TRANSFER ELEMENT
In structural problem displacement at each nodel point is obtained. By
using these displacement solutions, stresses and strains are calculated for each
element. In structural problems, the unknowns are represented by the
components of vector field. For example, in a two dimensional plate, the
unknown quantity is the vector field u(x,y),where u is a (2x1)displacement
vector.
Heat transfer can be defined as the transmission of energy from one
region another region due to temperature difference. A knowledge of the
temperature distribution within a body is important in many engineering
problems. There are three modes of heat transfer.
They are:
(i) Conduction
(ii) Convection
(iii) Radiation
SCE
111
ME2353
0 on
E(n-l)
SCE
( k-u) u 0 at
112
x l;
ME2353
where k is the spring constant. If the spring stiffness is set to a very large
value, the above boundary condition enforces u; if we let k 0, the above
boundary condition corresponds to a prescribed traction boundary. In practice,
such generalized boundary conditions are often used to model the influence of
the surroundings. For example, if the bar is a simplified model of a building and its
foundation, the spring can represent the stiffness of the soil.
5.2.2 Two-Point Boundary Value Problem With Generalized Boundary
Conditions
u(l)
uk
- ku(l)
SCE
113
STRONG
AND WEAK FORMS FOR ONE-DIMENSIONAL PROBLEMS
ME2353
In the partition approach, the total boundary is partitioned into the natural
boundary, and the complementary essential boundary, The natural boundary condition has
the generalized form defined by The resulting strong form for the partition method is
summarized in.
5.2.3
In this section, we will derive the general weak form for two-point boundary value
problems. Both the penalty and partition methods described in will be considered. To
obtain the general weak form for the penalty method, we multiply the two equations in the
strong by the weight function and integrate over the domains over which they hold: the
domain
for the differential equation and the domain
for the generalized boundary
condition.
5.3 SCALE VARIABLE PROBLEM IN 2 DIMENSIONS
x N1
u
y 0
N2
N3
N4
N1
N2
N3
x1
y
2
x1
0 y2
N 4 x3
y3
x4
y 4
J 11 J 12
J
;
J 21 J 22
SCE
114
STRONG
AND WEAK FORMS FOR ONE-DIMENSIONAL PROBLEMS
ME2353
J 22 J 12
0
J 0
J 21 J 11
B 1
0
J 21
J 22
N1
0 N1
J 11
J 12 0
0
0
N1
N1
N 2
N 2
0
0
0
0
N 2
N 2
N 3
N 3
0
0
0
0
N 3
N 3
1
v
v
0
v
[D]
1
0
2
(1 v )
1 v , for plane stress conditions;
0
0
v
1 v
E
v 1 v
[D]
(1 v)(1 2v)
0
0
2
0
x
F e [N ] ;
Fy
T
i 1
f (x)dx w f (x )
i
SCE
115
N 4
N 4
0
0
N 4
N 4
ME2353
Number of
Location
Corresponding Weights
Points
xi
wi
x1 = 0.000
2.000
x1, x2 =
1
0.577350269189
3
1.000
x1, x3
3
0.774596669241
5
5
0.555555
9
x2=0.000
8
0.888888
9
x1, x4 = 0.8611363116
0.3478548451
x2, x3 = 0.3399810436
0.6521451549
Problem (I set)
1
1. Evaluate I cos
1
x
2
1
x
2
dx , using one point and two point
I
3e
2. Evaluate
1
x 2
1
SCE
116
ME2353
E=2x105N/mm2,=
0.25,
u=[0,0,0.003,0.004,0.006, 0.004,0,0]T,
0,
=0.
n = t
SCE
in I ,
on q I ,
u = u
on u I ,
in ,
in ,
117
ME2353
where u = u(x1 , x2 , x3 , t) is the unknown displacement field, is the mass density, and I
= (0, T ) with T being a given time. Also, u0 and v0 are the prescribed initial displacement
and velocity fields. Clearly, two sets of boundary conditions are set on u and q ,
respectively, and are assumed to hold throughout the time interval I . Likewise, two sets of
initial conditions are set for the whole domain at time t = 0. The strong form of the
resulting initial/boundary- value problem is stated as follows: given functions f , t , u
, u0 and
v0 , as well as a constitutive equation for , find u in I , such that the equations are
satisfied.
A Galerkin-based weak form of the linear elastostatics problem has been derived in Section In the elastodynamics case, the only substantial difference involves the inclusion
R
of the term w u
d, as long as one adopts the semi-discrete approach. As a result, the
weak form at a fixed time can be expressed as
Z
Z
Z
Z
w u
d +
w f d +
w t d .
s w : d =
Following the development of Section 7.3, the discrete counterpart of can be written as
Z
Z
Z w t d .
h
wh uh d + (wh ) D(uh ) d
wh f d +
=
Following a standard procedure, the contribution of the forcing vector Fint,e due to
interele- ment tractions is neglected upon assembly of the global equations. As a result, the
equations is give rise to their assembled counterparts in the form
Mu + Ku
= F,
where u
is the global unknown displacement vector1 . The preceding equations are, of course,
subject to initial conditions that can be written in vectorial form as u
(0) = u
0 and v
(0) = v
0
.
The most commonly employed method for the numerical solution of the system of coupled linear second-order ordinary differential equations is the Newmark method. This
:= v
method is based on a time series expansion of u and u
. Concentrating on the time
interval (tn ,tn+1 ], the Newmark method is defined by the equations
1
2
u
n+1 = u
n + v
n tn + [(1 2)a
n + 2a
n+1 ]tn ,
2
v
+ [(1 )a
n + a
n+1 ]tn ,
v
n+1 = n
SCE
118
ME2353
It is clear that the Newmark equations define a whole family of time integrators.
It is important to distinguish this family into two categories, namely implicit and explicit
integrators, corresponding to > 0 and = 0, respectively.
The overhead hat symbol is used to distinguish between the vector field u and the
solution vector uemanating from the finite element approximation of the vector field u.
The general implicit Newmark integration method may be implemented as follows: first,
solve (9.18)1 for a
n+1 , namely write
a
n+1 =
t2n
n
(u
n+1 u
n v
n tn ) a
Then, substitute (9.19) into the semi-discrete form (9.17) evaluated at tn+1 to find that
M +K
n
u
n+1 = Fn+1
.
n+1 = K(u
n + v
n tn +a
n tn ) + Fn+1 .
Ma
If M is rendered diagonal (see discussion in Chapter 8), then a
n+1 can be determined
without solving any coupled linear algebraic equations. Then, the velocities bv n+1 are
immediately computed from (9.18)2 . Also, the displacements u
n+1 are computed from
indepen-dently of the accelerations a
n+1 .
SCE
119
ME2353
QUESTION BANK
PART A QUESTIONS WITH ANSWERS
UNIT 1
1.
2.
3.
4.
5.
6.
7.
1D element
2D element
3D element
8.
9.
SCE
120
ME2353
10.
10. What are the methods are generally associated with the finite element analysis?
Force method
Displacement or stiffness method.
11. Explain stiffness method.
Displacement or stiffness method, displacement of the nodes is considered as the
unknown of the problem. Among them two approaches, displacement method is
desirable.
12. What is meant by post processing?
Analysis and evaluation of the solution result is referred to as post processing.
Postprocessor computer program help the user to interpret the result by displaying
them in graphical form.
13. Name the variation methods.
Ritz method.
Ray-Leigh Ritz method.
14. What is meant by degrees of freedom?
When the force or reaction act at nodal point node is subjected to deformation. The
deformation includes displacement rotation, and or strains. These are collectively
known as degrees of freedom
15. What is meant by discretization and assemblage?
The art of subdividing a structure in to convenient number of smaller components
is known as discretization. These smaller components are then put together. The
process of uniting the various elements together is called assemblage.
16. What is Rayleigh-Ritz method?
It is integral approach method which is useful for solving complex structural
problem, encountered in finite element analysis. This method is possible only if a
suitable function is available.
17. What is Aspect ratio?
It is defined as the ratio of the largest dimension of the element to the smallest
dimension. In many cases, as the aspect ratio increases the in accuracy of the
solution increases. The conclusion of many researches is that the aspect ratio
SCE
121
UNIT 2
21. List the two advantages of post processing.
Required result can be obtained in graphical form. Contour diagrams can be used to
understand the solution easily and quickly.
22. During discretization, mention the places where it is necessary to place a
node?
Concentrated load acting point
Cross-section changing point
Different material interjections
Sudden change in point load
23. What is the difference between static and dynamic analysis?
Static analysis: The solution of the problem does not vary with time is known as
static analysis
Example: stress analysis on a beam
Dynamic analysis: The solution of the problem varies with time is known as
dynamic analysis
SCE
122
ME2353
element.
Global axes are defined for the entire system. They are same in direction for all the
elements even though the elements are differently oriented.
26.
energy is expressed as
f x , y ,
dy d y
....dx
27.
dx dx
28.
29.
30.
SCE
123
ME2353
31.
What are the basic steps involved in the finite element modeling.
Discretization of structure.
Numbering of nodes.
32.
F K u
33.
What is discretization?
The art of subdividing a structure in to a convenient number of smaller components is
known as discretization.
34.
35.
36.
37.
38.
39.
The accuracy of the result can be improved by increasing the order of the
polynomial. It is easy to formulate and computerize the finite element equations
SCE
124
ME2353
40.
41.
42.
Write down the expression of stiffness matrix for one dimensional bar element.
AE 1 1
K
l 1 1
State the properties of stiffness matrix
It is a symmetric matrix
The sum of elements in any column must be equal to zero
It is an unstable element. So the determinant is equal to zero.
43.
u 0 0l
0 u2
mu
u
4
Write down the expression of shape function N and displacement u for one
2
44.
u
1
45.
(W)
46.
47.
Write down the finite element equation for one dimensional two noded bar element.
AE 1 1
l 1 1
48.
What is truss?
A truss is defined as a structure made up of several bars, riveted or welded together.
SCE
125
ME2353
49.
States the assumption are made while finding the forces in a truss.
All the members are pin jointed.
The truss is loaded only at the joint
The self weight of the members is neglected unless stated.
50.
51.
52.
53.
54.
55.
56.
58.
SCE
126
ME2353
60.
UNIT 4
61.
Write down the stiffness matrix equation for two dimensional CST elements.
Stiffness matrix K B
D B At
62.
1 12 0
0
12
63.
64.
65.
66.
D B 2rA
67.
Write down the shape functions for 4 noded rectangular elements using natural
coordinate system.
SCE
127
ME2353
N1 1 11
4
68.
N3 1 1 1
N4 1 1 1
4
4
Write down Jacobian matrix for 4 noded quadrilateral elements.
J
J
69.
N2 1 11
4
11
12
21
22
Write down stiffnes matrix equation for 4 noded isoparametric quadrilateral elements.
Stiffness matrix K t
B T D B J
1 1
70.
71.
73.
74.
75.
76.
L2 = A2/A
L3 = A3/A
77.
SCE
128
ME2353
78.
79.
80.
UNIT 5
81.
82.
83.
Contour diagrams can be used to understand the solution easily and quickly.
85.
SCE
129
ME2353
86.
87.
88.
89.
90.
to unity as possible.
91.
92.
93.
SCE
130
ME2353
plane. The basic element useful for two dimensional analysis is the triangular
element.
94.
96.
97.
98.
99.
SCE
131