Beruflich Dokumente
Kultur Dokumente
Lecture 1: Introduction
1.1.1 Introduction
The Finite Element Method (FEM) is a numerical technique to find approximate solutions of partial
differential equations. It was originated from the need of solving complex elasticity and structural
analysis problems in Civil, Mechanical and Aerospace engineering. In a structural simulation, FEM
helps in producing stiffness and strength visualizations. It also helps to minimize material weight
and its cost of the structures. FEM allows for detailed visualization and indicates the distribution of
stresses and strains inside the body of a structure. Many of FE software are powerful yet complex
tool meant for professional engineers with the training and education necessary to properly interpret
the results.
Several modern FEM packages include specific components such as fluid, thermal,
electromagnetic and structural working environments. FEM allows entire designs to be constructed,
refined and optimized before the design is manufactured. This powerful design tool has significantly
improved both the standard of engineering designs and the methodology of the design process in
many industrial applications. The use of FEM has significantly decreased the time to take products
from concept to the production line. One must take the advantage of the advent of faster generation
of personal computers for the analysis and design of engineering product with precision level of
accuracy.
• In the 1970s, most commercial FEM software packages (ABAQUS, NASTRAN, ANSYS,
etc.) originated. Interactive FE programs on supercomputer lead to rapid growth of CAD
systems.
• In the 1980s, algorithm on electromagnetic applications, fluid flow and thermal analysis were
developed with the use of FE program.
• Engineers can evaluate ways to control the vibrations and extend the use of flexible,
deployable structures in space using FE and other methods in the 1990s. Trends to solve fully
coupled solution of fluid flows with structural interactions, bio-mechanics related problems
with a higher level of accuracy were observed in this decade.
With the development of finite element method, together with tremendous increases in computing
power and convenience, today it is possible to understand structural behavior with levels of
accuracy. This was in fact the beyond of imagination before the computer age.
The application of finite difference method for engineering problems involves replacing the
governing differential equations and the boundary condition by suitable algebraic equations. For
3
example in the analysis of beam bending problem the differential equation is reduced to be solution
of algebraic equations written at every nodal point within the beam member. For example, the beam
equation can be expressed as:
d 4w q
4
= (1.1.1)
dx EI
To explain the concept of finite difference method let us consider a displacement function variable
namely w = f ( x)
d 4w 1
= wi+4 - wi+3 - 3wi+3 + 3wi+2 + 3wi+2 - 3wi+1 - wi+1 + wi
dx 4 h 4
1
= 4 wi+4 - 4wi+3 +6wi+2 - 4wi+1 + wi (1.1.5)
h
1
= 4 wi+2 - 4wi+1 +6wi - 4wi-1 + wi-2
h
Thus, eq. (1.1.1) can be expressed with the help of eq. (1.1.5) and can be written in finite difference
form as:
q 4
( wi − 2 − 4 wi −1 + 6 wi − 4 wi +1 + wi + 2 ) = h (1.1.6)
EI
Thus, the displacement at node i of the beam member corresponds to uniformly distributed load can
be obtained from eq. (1.1.6) with the help of boundary conditions. It may be interesting to note that,
the concept of node is used in the finite difference method. Basically, this method has an array of
grid points and is a point wise approximation, whereas, finite element method has an array of small
interconnecting sub-regions and is a piece wise approximation.
Each method has noteworthy advantages as well as limitations. However it is possible to
solve various problems by finite element method, even with highly complex geometry and loading
conditions, with the restriction that there is always some numerical errors. Therefore, effective and
reliable use of this method requires a solid understanding of its limitations.
element, plate element, shell element, solid element, etc. All elements are then assembled to obtain
the solution of the entire domain/structure under certain loading conditions. Nodes are assigned at a
certain density throughout the continuum depending on the anticipated stress levels of a particular
domain. Regions which will receive large amounts of stress variation usually have a higher node
density than those which experience little or no stress.
The need of finite element analysis arises when the structural system in terms of its either geometry,
material properties, boundary conditions or loadings is complex in nature. For such case, the whole
structure needs to be subdivided into smaller elements. The whole structure is then analyzed by the
assemblage of all elements representing the complete structure including its all properties.
8
The subdivision process is an important task in finite element analysis and requires some
skill and knowledge. In this procedure, first, the number, shape, size and configuration of elements
have to be decided in such a manner that the real structure is simulated as closely as possible. The
discretization is to be in such that the results converge to the true solution. However, too fine mesh
will lead to extra computational effort. Fig. 1.2.2 shows a finite element mesh of a continuum using
triangular and quadrilateral elements. The assemblage of triangular elements in this case shows
better representation of the continuum. The discretization process also shows that the more accurate
representation is possible if the body is further subdivided into some finer mesh.
Classical
Actual Structure
∂σ x ∂σ y ∂σ z
+ + + FΩ =
0
∂x ∂y ∂z
(Partial Differential Equations)
Assumptions
Equilibrium Equations
Stress-Strain Law
Compatibility Conditions
(Principles of Virtual Work)
FEM
Structural Model
F = K d
(Algebraic Equations)
Varieties of engineering problem like solid and fluid mechanics, heat transfer can easily be solved by
the concept of finite element technique. The basic form of the equation will become as follows
where action, property and response parameter will vary for case to case as outlined in Table 1.2.1.
10
{F } [=
K ]{d } OR {d } [ K ] {F }
−1
=
4. For some problems, there may be a considerable amount of input data. Errors may creep up
in their preparation and the results thus obtained may also appear to be acceptable which
indicates deceptive state of affairs. It is always desirable to make a visual check of the input
data.
5. In the FEM, many problems lead to round-off errors. Computer works with a limited number
of digits and solving the problem with restricted number of digits may not yield the desired
degree of accuracy or it may give total erroneous results in some cases. For many problems
the increase in the number of digits for the purpose of calculation improves the accuracy.
6. Simplification of the boundary is another source of error. The domain may be reduced to the
shape of polygon. If the mesh is refined, then the error involved in the discretised boundary
may be reduced.
7. During arithmetic operations, the numbers would be constantly round-off to some fixed
working length. These round–off errors may go on accumulating and then resulting accuracy
of the solution may be greatly impaired.
13
∂σ x ∂τ yx ∂τ
dxdydz + dxdydz + zx dxdydz + FΩx dxdydz =
0 (1.3.2)
∂x ∂y ∂z
Where, FΩx is the component of body force along x direction. Now, dividing dxdydz on the above
expression, following equilibrium condition is obtained.
∂σ x ∂τ yx ∂τ zx
+ + =
− fx (1.3.3)
∂x ∂y ∂z
Similarly, applying equilibrium condition along Y and Z directions, one can find the following
relations.
∂τ xy ∂σ y ∂τ zy
+ + = − fy (1.3.4)
∂x ∂y ∂z
∂τ xz ∂τ yz ∂σ z
+ + =− fz (1.3.5)
∂x ∂y ∂z
Here, FΩy and FΩz are the component of body forces along Y and Z directions respectively. Satisfying
moment equations (i.e.,= ∑ M x 0;= ∑ M y 0 and= ∑ M z 0; ), one can obtain the following
relations.
=τ xy τ=
yx ; τ yz τ zy and
= τ xz τ zx (1.3.6)
Using eq. (1.3.6), the equilibrium equations (1.3.3 to 1.3.5), can be rewritten in the following form.
∂σ x ∂τ xy ∂τ zx
+ + = − FΩx
∂x ∂y ∂z
∂τ xy ∂σ y ∂τ yz
+ + = − FΩy (1.3.7)
∂x ∂y ∂z
∂τ zx ∂τ yz ∂σ z
+ + = − FΩz
∂x ∂y ∂z
Let assume an element of area ∆Γ on the surface of the solid in equilibrium (Fig.1.3.2) and FΓx, FΓy
and FΓz are the components of external forces per unit area and are acting on the surface.
Consideration of equilibrium along the three axes directions gives the following relations.
σ x l + τ xy m + τ zx n =
FΓx
τ xy l + σ y m + τ yz n =
FΓy (1.3.8)
τ zx l + τ yz m + σ z n =
FΓz
Here, l, m and n are the direction cosines of the normal to the boundary surface. Eq. (1.3.8) is known
as static boundary condition.
15
Let consider points P, Q and R are before deformation and points P’,Q’ and R’ are after deformation
as shown in Fig. 1.3.4 below. Now for small deformation, rotation of PQ will become
∆v ∂v
=θ1 lim
=
∆x →0 ∆x ∂x
∆u ∂u
be: θ 2 lim
Similarly, rotation of PR due to deformation will = =
∆y →0 ∆y ∂y
Thus, the total change of angle between PQ and PR after deformation is as follows which is defined
as shear strain in X-Y plane.
∂v ∂u
γ xy = γ yx =θ1 + θ 2 = +
∂x ∂y
εx = + + + (1.3.9)
∂x 2 ∂x ∂x ∂x
∂v 1 ∂u ∂v ∂w
2 2 2
ε y = + + + (1.3.10)
∂y 2 ∂y ∂y ∂y
∂w 1 ∂u ∂v ∂w
2 2 2
εz = + + + (1.3.11)
∂z 2 ∂z ∂z ∂z
∂v ∂u ∂u ∂u ∂v ∂v ∂w ∂w
γ xy = + + + +
∂x ∂y ∂x ∂y ∂x ∂y ∂x ∂y
(1.3.12)
∂w ∂v ∂u ∂u ∂v ∂v ∂w ∂w
γ yz = + + + +
∂y ∂z ∂y ∂z ∂y ∂z ∂y ∂z
(1.3.13)
∂u ∂w ∂u ∂u ∂v ∂v ∂w ∂w
γ zx = + + + + (1.3.14)
∂z ∂x ∂z ∂x ∂z ∂x ∂z ∂x
The eqs.(1.3.9 to 1.3.14) are known as Green-Lagrange strain displacement equation. The
components of the strain εx, εy, εz, γxy, γyz and γzx define the state of strains in the deformed body, and
can be written in a matrix form as
The relations given in eqs.(1.3.9 to 1.3.14) are non-linear partial differential equations in the
unknown component of the displacements. In case of small deformations, the products and squares
of the first derivatives are assumed to be negligible compared with the derivatives themselves in
many problems of stress analysis. Thus the strain-displacement relations in eqs. (1.3.9 to 1.3.14 )
reduce to linear relations as follows.
18
∂u
εx =
∂x
∂v
εy =
∂y
∂w
εz =
∂z
(1.3.16)
∂v ∂u
γ= +
∂x ∂y
xy
∂w ∂v
γ= +
∂y ∂z
yz
∂u ∂w
γ= +
∂z ∂x
zx
Eq. (1.3.16) is known as Von-Karman strain displacement equation. The above equation can be
expressed in a matrix form as given below.
∂
∂x 0 0
∂
ε x 0 0
ε ∂y
y ∂ u
ε z 0 ∂z v
0
=∂ ∂ (1.3.17)
γ xy 0 w
γ yz ∂y ∂x
∂ ∂
γ zx 0 ∂z ∂y
∂ 0
∂
∂z ∂x
The above assumption will be incorrect in case of large deformation problems. In these cases,
geometric nonlinearity has to be considered.
or {σ } = [C ]{ε } (1.3.19)
Where [C] is constitutive matrix. If the material has three orthogonal planes of symmetry, it is said
to be orthotropic. In this case only nine constants are required for describing constitutive relations as
given below.
An isotropic is one for which every plane is a plane of symmetry of material behavior and only two
constants (Young Modulus, E and Poisson ratio µ) are required to describe the constitutive relation.
The following equation includes the effect due to temperature changes as may be necessary in
certain cases of stress analysis.
x 1 0 x 1
0 0
1 0 0 0
y
.
1
1 1 0 0 0 .
1
T
z
or (1.3.22)
xy E 2 1 0 0 .
0
yz
2 1 0
.
0
zx
2 1
zx
0
20
T and α in eq. (1.3.22) denote the difference of temperature and coefficient of thermal expansion
respectively.
The inverse relation of stresses in terms of strain components can be expressed as
(1 − µ ) µ µ 0 0 0
σ x (1 − µ ) µ 0 0 0 ε x 1
σ 1
y (1 − µ ) 0 0 0 ε y
1 − 2µ ε EαT 1
σ z 0 z −
= E 0 (1.3.23)
τ
2 γ xy 1 − 2µ 0
1 − 2µ
xy
τ yz 0 γ yz 0
2
τ zx 1 − 2µ γ zx 0
2
E
where E =
(1 + µ )(1 − 2µ )
+ 2 = (1.3.25)
∂y 2 ∂x ∂x∂y
Eq. (1.3.25) is the compatibility equation since it states the geometric requirements. This condition
will ensure adjacent elements to remain free from discontinuities such as gaps and overlaps.
component varies across the thickness and the stress components σz, τxz and τyz are zero. The state of
stress is specified by σx, σy and τxy only and is called plane stress.
Fig. 1.3.5 Plane stress example: Thin plate with in-plane loading
x 1
1 0 x
E ET
1 0 y 1 (1.3.26)
y 1 2 1
xy
0 0
1 xy
0
2
The strain components can also be expressed in terms of the stress, which is given below.
x 1 0 x 1
1
y 1
0 y T 1 (1.3.27)
E 0
xy
0 2 1
xy
0
It can also be shown that
−µ
εz = (ε x + ε y ) + 1 + µ αT and γ yz = γ zx = 0 (1.3.28)
1− µ 1− µ
translation and displacements linear in z corresponding to rigid body rotation do not result in strain.
As a result, the following relations arise.
ε z = γ yz = γ zx = 0 (1.3.29)
The constitutive relation for elastic isotropic material for this case may be given by,
σ x 1 − µ µ 0 ε x 1
E E α T
σ y = µ 1− µ 0 ε y − 1 (1.3.30)
τ (1 + µ )(1 − 2 µ )
1 − 2µ γ 1 − 2 µ 0
xy 0 0 xy
2
x 0
1 1
x
1
1 0 y 1 T 1 (1.3.32)
y
E 0
xy
0 2
xy
0
1.3.4.3Axisymmetric Problem
Many problems in stress analysis which are of practical interest involve solids of revolution subject
to axially symmetric loading. A circular cylinder loaded by a uniform internal or external pressure,
circular footing resting on soil mass, pressure vessels, rotating wheels, flywheels etc. The strain-
displacement relations in these type of problems are given by
∂u
εx =
∂x
u
εθ =
x
(1.3.33)
∂v
εy =
∂y
∂u ∂v
γ= +
∂y ∂x
xy
The two components of displacements in any plane section of the body along its axis of symmetry
define completely the state of strain and therefore the state of stress. The constitutive relations are
given below for such types of problems.
σ x (1 − µ ) µ µ 0 ε
σ µ − µ µ x
y E 1 0 ε y
= µ µ 1− µ 0 ε (1.3.34)
σ θ (1 + µ )(1 − 2µ ) 1 − 2µ
θ
τ xy 0 0 0 γ
2 xy
24
Considering above aspects, types of engineering analysis are decided. FEA is capable of using
multiple materials within the structure such as:
• Isotropic (i.e., identical throughout)
0
• Orthotropic (i.e., identical at 90 )
• General anisotropic (i.e., different throughout)
The Equilibrium Equations for different cases are as follows:
1. Linear-Static:
Ku = F (1.4.1)
2. Linear-Dynamic
Mu(t ) + Cu (t ) + Ku (t ) =
F (t ) (1.4.2)
3. Nonlinear - Static
Ku + FNL =
F (1.4.3)
1. Nonlinear-Dynamic
Mu(t ) + Cu (t ) + Ku (t ) + F (t ) NL =
F (t ) (1.4.4)
Here, M, C, K, F and U are mass, damping, stiffness, force and displacement of the structure
respectively. Table 1.4.1 shows various types of analysis which can be performed according to
engineering judgment.
10. Calculation of stresses or stress-resultants: Nodal displacements are utilized for the
calculation of stresses or stress-resultants. This may be done for all elements of the
continuum or it may be limited to some predetermined elements. Results may also be
obtained by graphical means. It may desirable to plot the contours of the deformed shape of
the continuum.
The basic steps for finite element analysis are shown in the form of flow chart below:
The configuration of structural elements depends upon the geometry of the structural system and the
number of independent space coordinates (i.e., x, y and z) required to describe the problem. Thus, the
element can be categorized as one, two or three dimensional element. One dimensional element can
be represented by a straight line whose ends will be nodal points. The skeletal structures are
generally modeled by this type of elements. The pin jointed bar or truss element is the simplest
structural element. This element undergoes only axial deformation. The beam element is another
type of element which undergoes in-plane transverse displacements and rotations. The frame
element is the combination of truss and beam element. Thus, the frame element has axial and in-
plane transverse displacements and rotations. This element is generally used to model 1D, 2D and
3D skeletal structural systems. Two-dimensional elements are generally used to model 2D and 3D
continuum. These elements are of constant thickness and material properties. The shapes of these
elements are triangular or rectangular and it consists of 3 to 9 or even more nodes. These elements
are used to solve many problems in solid mechanics such as plane stress, plane strain, plate bending.
Three-dimensional element is the most cumbersome which is generally used to model the 3-D
continuum. The elements have 6 to 27 numbers of nodes or more. Because of large degrees of
freedom, the analysis is time consuming using 3-D elements and difficult to interpret its results.
However, for accurate analysis of the irregular continuum, 3-D elements are useful. To analyze any
real structure, appropriate elements are to be assigned for the finite element analysis. In standard
FEA software, following types of element library are used to discretize the domain.
• Truss element
• Beam element
• Frame element
• Membrane/ Plate/Shell element
• Solid element
• Composite element
• Shear panel
• Spring element
29
• Rigid/Link element
• Viscous damping element
The different types of elements available in standard finite element software are shown in Fig. 1.4.2.
3D Elements
2.1.1 Introduction
Finite element formulation can be constructed from governing differential equations over a domain.
This can be formulated by various ways like Virtual work method, Variational method, Weighted
Residual Method etc.
(2.1.2)
Here, d u v w .
T
For the three dimensional stress-strain condition, there are six
components of stresses ( x , y , z , xy , yz , zx ) and six components of strains in virtual
x
y
U x y z xy yz zx z d
xy (2.1.3)
yz
zx
Or
U d
T
(2.1.4)
According to principle of virtual work, the work done by external forces due to the virtual
displacement of a structure in equilibrium is equal to the work done by the internal forces for the
virtual internal displacement. Therefore, WE U Thus eqs. (2.1.2) and (2.1.4) can be made equal
and can be related as follows:
(2.1.5)
displacement relationship can be expressed as {ε } = [ B ]{d } , where {d} is the displacement vector in
x, y and z directions and [B] is called as the strain displacement relationship matrix. Again, the
stress can be represented in terms of its constitutive relationship matrix: {σ } = [ D ]{ε } . Here [ D ]
is called as the constituent relationship matrix. Using the above relationship in the strain energy
equation one can arrive
1
[ B ]{d } [ D ]{ B}{d } d Ω
T
U ∫∫∫
2 Ω (2.1.7)
Applying the variational principle one can express
3
∂U
{F }
= = ∫∫∫ [ B ] [ D ][ B ] d Ω {d }
T
∂ {d } Ω (2.1.8)
Now, from the relationship of { F } = [ K ]{d } , one can arrive at the element stiffness matrix as:
=[K ] ∫∫∫ [ B ] [ D ][ B ] d Ω
T
Ω (2.1.9)
Thus, by the use of variational principle, the stiffness matrix of a structural element can be obtained
as expressed in the above equation.
w i R d 0 for i 1,2,3,...,n
or (2.1.12)
w Du f d 0
i
Where weighting function wi = wi(x) is chosen from the approximate basis function used for
constructing approximated solution u .
5
2.2.1 Introduction
Galerkin method is the most widely used among the various weighted residual methods. Galerkin
method incorporates differential equations in their weak form, i.e., before starting integration by
parts it is in strong form and after by parts it will be in weak form, so that they are satisfied over a
domain in an integral. Thus, in case of Galerkin method, the equations are satisfied over a domain in
an integral or average sense, rather than at every point. The solution of the equations must satisfy the
boundary conditions. There are two types of boundary conditions:
• Essential or kinematic boundary condition
• Non essential or natural boundary condition
4y
For example, in case of a beam problem ( EI q 0 ) differential equation is of forth order.
x 4
As a result, displacement and slope will be essential boundary condition where as moment and shear
will be non-essential boundary condition.
xy y
Fy 0 (2.2.2)
x y
Where, Fx and Fy are the body forces in X and Y direction respectively. Let assume,
x and y are surface forces in X and Y direction and as angle made by normal to surface
with X– axis (Fig. 2.2.1). Therefore, force equilibrium of element can be written as:
Fx PQ t x OP t xy OQ t
OP OQ
Fx x xy x cos xy sin x cos xyCos 90
PQ PQ
Thus, Fx x xy m (2.2.3)
Where, ℓ and m are direction cosines of normal to the surface. Similarly,
Fy xy y m (2.2.4)
6
Where u and v are weighting functions i.e elemental displacements in X and Y directions
respectively. Now one can expand above equation by using Green’s Theorem.
Green Theorem states that if x, yand x, y are continuous functions then their first and
second partial derivatives are also continuous. Therefore,
2 2
dxdy 2 2 dxdy
x x y y x
y
x
y
m ds
(2.2.6)
Assuming, x ; u; 0 one can rewrite with the use of above relation as
x y
x u
x
u dx dy x
x
dx dy x u ds (2.2.7)
Similarly, assuming y ; 0 and v
x y
y v
y
v dx dy y
y
dx dy y m v ds (2.2.8)
7
Again, assuming x y ; v; 0
x y
x y v
y
v dx dy x y
x
dx dy x y v ds (2.2.9)
And assuming, x y ; 0; u
x y
x y u
y
u dx dy x y
y
dx dy x y m u ds
Putting values of eqs. (2.2.7), (2.2.8) and (2.2.9), in eq. (2.2.5), one can get the following relation:
x u y v xy v xy u dx dy
x y x y
Here, Fx and Fy are the body forces and u & v are virtual displacements in X and Y directions
respectively.
Considering first term of eq. (2.2.11), virtual displacement u is given to the element of unit
thickness. Dotted position in Fig. 2.2.2 shows the virtual displacement. Thus, work done by x :
x dy u u dx x dyu x u dxdy (2.2.12)
x x
Similarly, considering second term of eq. (2.2.11), virtual work done by body forces is
F x u Fyvdx dy
8
Putting eqs. (2.2.3) & (2.2.4) in third term of eq. (2.2.11) we get the virtual work done by surface
forces as:
F x uds Fyvds
x x
u dxdy x x dxdy
y y vdxdy y y dxdy (2.2.14)
xy x
v
y
u xy xy dxdy
If external work done is represented by WE and U is the internal work done then,
U w E 0 or U w E (2.2.16)
Thus in elasticity problems, Galerkin’s method turns out to be the principle of virtual work, which
can be stated that “A Deformable body is said to be in equilibrium, if the total work done by external
forces is equal to the total work done by internal forces.” The work done above is virtual as either
forces or deformations are also virtual. Thus, Galerkin’s approach can be followed in all problems
involving solution of a set of equations subjected to specified boundary values.
w i p,ii d 0 (2.2.19)
Integrating by parts of the above expression, the following relation can be obtained.
If the nodal pressure and interpolation functions are denoted by p and N respectively, then the
pressure at any point inside the fluid domain can be expressed as
p N p
Similarly, the weighted function can also be written with the help of interpolation function as
w N w
10
Thus, pi,i L p L N p Bp , where, L = differential operator.
x y
Similarly, w i,i L W L N w Bw
Thus, w i,i p,i d w B B p d
T T
(2.2.21)
p
w i p,i d = w N d
T T
(2.2.22)
n
Here, Γdenotes the surface of the fluid domain and n represents the direction normal to the surface.
Thus, from eq. (2.2.20), one can write the expression as:
p
Thus, w B Bp d w N d
T T T T
n
Or, G p S (2.2.23)
Where,
T T
G B Bd N N N N d
T
x x y y
(2.2.27)
p
and S N d
T
n
Here, n is the direction normal to the surface. Thus, solving the above equation with the prescribed
boundary conditions, one can find out the pressure distribution inside the fluid domain by the use of
finite element technique.
11
Displacement function is the beginning point for the structural analysis by finite element method.
This function represents the variation of the displacement within the element. On the basis of the
problem to be solved, the displacement function needs to be approximated in the form of either
linear or higher-order function. A convenient way to express it is by the use of polynomial
expressions.
a. The displacement function must be continuous within the elements. This can be ensured by
choosing a suitable polynomial. For example, for an n degrees of polynomial, displacement
function in I dimensional problem can be chosen as:
u=α 0 + α1 x + α 2 x 2 + α 3 x3 + α 4 x 4 + ..... + α n x n (2.3.1)
b. The displacement function must be capable of rigid body displacements of the element. The
constant terms used in the polynomial (α0 to αn) ensure this condition.
c. The displacement function must include the constant strains states of the element. As
element becomes infinitely small, strain should be constant in the element. Hence, the
displacement function should include terms for representing constant strain states.
2.3.1.2 Compatibility
Displacement should be compatible between adjacent elements. There should not be any
discontinuity or overlapping while deformed. The adjacent elements must deform without causing
openings, overlaps or discontinuous between the elements.
Elements which satisfy all the three convergence requirements and compatibility condition
are called Compatible or Conforming elements.
The geometric invariance can be ensured by the selection of the corresponding order of terms on
either side of the axis of symmetry.
1
x y
x2 xy y2
x3 x2y xy2 y3
x Ni x i u N i u i (2.3.3)
y N i yi v N vi i
If Ni = N´i, then the element is called isroparametric. Fig. 2.3.2(a) shows the two dimensional 8 node
isoparametric element.
If the geometry of element is defined by shape functions of order higher than that for representing
the variation of displacements, then the elements are called superparametric (Fig. 2.3.2(b)).
If the geometry of element is defined by shape functions of order lower than that for representing the
variation of displacements then the elements are called subparametric (Fig. 2.3.2(c)).
Few of the elements with number of nodes are shown in Fig. 2.3.4.
15
The steps to be followed in the computer program are shown in the form of flow chart in Fig. 2.4.1
for assembling the local stiffness matrix to global stiffness matrix.
18
Here, α refers to unknown displacement; β refers to known displacement (≠0) and γ refers to zero
displacement. Thus, the above equation can be separated and solved independently to find required
unknown results as shown below.
F K d K d K d
or, K d F K d as d 0
Thus, d K
1
F K d
(2.4.7)
For computer programming, several techniques are available for handling boundary conditions. One
of the approaches is to make the diagonal element of stiffness matrix corresponding to zero
displacement as unity and corresponding all off-diagonal elements as zero. For example, let consider
a 3×3 stiffness matrix with following force-displacement relationship.
F1 k11 k12 k13 d1
F2 = k21 k22 k23 d 2
F k k33 d3 (2.4.8)
3 31 k32
Now, if the third node has zero displacement (i.e., d3= 0) then the matrix will be modified as follows
to incorporate the boundary condition.
F1 k11 k12 0 d1
F2 = k21 k22 0 d 2
0 0 1 d3 (2.4.9)
0
{δ } = {d 2 } (2.4.12)
Similarly 3rd row will be:
F3 − k32 × δ = k31 × d1 + k33 × d3 (2.4.13)
Thus above three equations can be written in a combined form as
F1 − k12δ k11 0 k13 d1
δ = 0 1 0 d 2
F − δ k
32 31 0 k33 d3 (2.4.14)
Another approach may also be followed to take care the known restrained displacements by
assigning a higher value δ (say δ =1020) in the diagonal element corresponding to that displacement.
F1 k11 k12 k13 d1
k
δ ×10 × k=
20
22 21 k22 ×1020 k23 d 2
k31 k33 d3
F3 k32
(2.4.15)
1020 k 22 k 21d1 k 22 1020 d 2 k 23 d 3
As d3 is corresponding to zero displacement, the above equation can be simplified to the following.
1020 k 22 k 21d1 k 22 1020 d 2
or 1020 k 22 k 22 1020 d 2
d 2 known displacement is ensured
If the overall stiffness matrix is to be formed in half band form then the numbering of nodes should
be such that the bandwidth is minimum. For this the labels are put in a systematic manner
irrespective of whether the joint displacements are unknowns or restraints. However, if the unknown
displacements are labeled first then the matrix operations can be restricted up to unknown
displacement labels and beyond that the overall stiffness matrix may be ignored.
1
Natural coordinate system is basically a local coordinate system which allows the specification of a
point within the element by a set of dimensionless numbers whose magnitude never exceeds unity.
This coordinate system is found to be very effective in formulating the element properties in finite
element formulation. This system is defined in such that the magnitude at nodal points will have
unity or zero or a convenient set of fractions. It also facilitates the integration to calculate element
stiffness.
Fig. 3.1.2 Linear interpolation function for two node line element
Similarly, for three node line element, the shape function can be derived with the help of natural
coordinate system which may be expressed as follows:
3
3x 2 x 2
1 2
l
N1
l 1 3 2 2
2
(3.1.5)
2
4x 4x
2
N N 4 4
l l 2
N
x 2x
2 2
3
2
l l2
The detailed derivation of the interpolation function will be discussed in subsequent lecture. The
variation of the shape functions over the length of the three node element are shown in Fig. 3.1.3
Fig. 3.1.3 Variation of interpolation function for three node line element
Now, if φ is considered to be a function of L1 and L2, the differentiation of φ with respect to x for
two node line element can be expressed by the chain rule formula as
d L1 L 2
. . (3.1.6)
dx L1 x L 2 x
Thus, eq.(3.1.4) can be written as
L1 1 L2 1
and (3.1.7)
x l x l
Therefore,
4
d 1
dx l L2 L1
(3.1.7)
The integration over the length l in natural coordinate system can be expressed by
p !q !
L L2 q dl
p
l (3.1.9)
l
1
p q 1!
Here, p! is the factorial product p(p-1)(p-2)….(1) and 0! is defined as equal to unity.
The area of the triangles can be written using Cartesian coordinates considering x, y as coordinates
of an arbitrary point P inside or on the boundaries of the element:
1 𝑥1 𝑦1
1
A = 2
�1 𝑥2 𝑦2 �
1 𝑥3 𝑦3
1 𝑥 𝑦
1
A1 = 2
�1 𝑥2 𝑦2 �
1 𝑥3 𝑦3
1 𝑥 𝑦
1
A2 = � 1 𝑥3 𝑦3 �
2
1 𝑥1 𝑦1
1 𝑥 𝑦
1
A3 = 2
�1 𝑥1 𝑦1 �
1 𝑥2 𝑦2
The relation between two coordinate systems to define point P can be established by their nodal
coordinates as
1 1 1 1 𝐿1
𝑥
� �=� 1𝑥 𝑥2 𝑥3 � �𝐿2 � (3.1.12)
𝑦 𝑦1 𝑦2 𝑦3 𝐿3
Where,
𝑥 = 𝐿1 𝑥1 + 𝐿2 𝑥2 + 𝐿3 𝑥3
6
𝑦 = 𝐿1 𝑦1 + 𝐿2 𝑦2 + 𝐿3 𝑦3
The inverse between natural and Cartesian coordinates from eq.(3.1.12) may be expressed as
𝐿1 𝑥2 𝑦3 −𝑥3 𝑦2 𝑦2 − 𝑦3 𝑥3 − 𝑥2 1
1
�𝐿2 � = 2𝐴 �𝑥3 𝑦1 −𝑥1 𝑦3 𝑦3 − 𝑦1 𝑥1 − 𝑥3 � �𝑥 � (3.1.13)
𝐿3 𝑥1 𝑦2 −𝑥2 𝑦1 𝑦1 − 𝑦2 𝑥2 − 𝑥1 𝑦
The derivatives with respect to global coordinates are necessary to determine the properties of an
element. The relationship between two coordinate systems may be computed by using the chain rule
of partial differentiation as
L1 L 2 L 3
. . .
x L1 x L 2 x L3 x
b1 b b
. 2. 3.
2A L1 2A L 2 2A L3 (3.1.14)
3
bi
.
i1 2A Li
Where, c1 = x3 – x2; c2 = x1 – x3 and c3 = x2 – x1. The above expressions are looked cumbersome.
However, the main advantage is the ease with which polynomial terms can be integrated using
following area integral expression.
p!q!r!
L L 2 q L3r dA
p
2A (3.1.16)
A
1
p q r 2!
u 1L1 2 L 2 3L3
u
T
Or, (3.1.17)
1 0 0 u1 1 0 0
T
u 0 1 0 u 2 0 1 0 u i
T
(3. 1.21)
0 0 1 u 3 0 0 1
The above expression can be written in terms of interpolation function as u N u i
T
Where,
1 0 0
N L1 L2 L3 0 1 0 L1 L 2 L3
T
(3. 1.22)
0 0 1
Similarly, the displacement variation v in Y direction can be expressed as follows.
v N v i
T
(3.1.23)
Thus, for two displacement components u and v of any point inside the element can be written as:
0 ui
u N
T T
d = = T T
(3.1.24)
v 0 N vi
Thus, the shape function of the element will become
L1 L2 L3 0 0 0
N (3.1.25)
0 0 0 L1 L2 L3
It is important to note that the shape function Ni become unity at node i and zero at other nodes of
the element. The displacement at any point of the element can be expressed in terms of its nodal
displacement and the interpolation function as given below.
u = N1u1 + N 2u2 + N 3u3
(3.1.26)
v = N1v1 + N 2v2 + N 3v3
9
The triangular element can be used to represent the arbitrary geometry much easily. On the other
hand, rectangular elements, in general, are of limited use as they are not well suited for representing
curved boundaries. However, an assemblage of rectangular and triangular element with triangular
elements near the boundary can be very effective (Fig. 3.2.1). Triangular elements may also be used
in 3-dimensional axi-symmetric problems, plates and shell structures. The shape function for
triangular elements (linear, quadratic and cubic) with various nodes (Fig. 3.2.2) can be formulated.
An internal node will exist for cubic element as seen in Fig. 3.2.2(c).
Fig. 3.2.1 Finite element mesh consisting of triangular and rectangular element
Similarly, for “m” node element having three degrees of freedom at each node, the displacement
function can be expressed as
𝑢 = 𝛼0 + 𝛼1 𝑥 + 𝛼2 𝑦 + 𝛼3 𝑥 2 + 𝛼4 𝑥𝑦 + 𝛼5 𝑦 2 + . . … . . . +𝛼𝑚−1 𝑦 𝑛
𝑣 = 𝛼𝑚 + 𝛼𝑚+1 𝑥 + 𝛼𝑚+2 𝑦 + 𝛼𝑚+3 𝑥 2 + 𝛼𝑚+4 𝑥𝑦+. . . . . +𝛼2𝑚−1 𝑦 𝑛 (3.2.5)
2 𝑛
𝑤 = 𝛼2𝑚 + 𝛼2𝑚+1 𝑥 + 𝛼2𝑚+2 𝑦 + 𝛼2𝑚+3 𝑥 + 𝛼2𝑚+4 𝑥𝑦+. . . . . +𝛼3𝑚−1 𝑦
11
Where, {d} is the nodal displacements. To simplify the above expression for finding out the shape
function, the displacements in X direction can be separated out which will be as follows:
𝑢1 1 𝑥1 𝑦1 ⍺∝0
{𝑢𝑖 }=� 2 � = �1 𝑥2 𝑦2 � �⍺∝1 �
𝑢 (3.2.8)
𝑢3 1 𝑥3 𝑦3 ⍺∝2
To obtain the polynomial coefficients, {α} the matrix of the above equation are to be inverted. Thus,
1
0 1 x1 y1
u1 x 2 y3 x 3 y 2 x 3 y1 x1 y3 x1 y 2 x 2 y1 u1
y2
1 1 x 2 u 1 y 2 y3
y3 y1 y1 y 2 u 2
2 2A
y3 u
2 1 x 3
u 3 x3 x 2 x1 x 3 x 2 x1 3
a1 a2 a 3 u1
1
b1 b2 b 3 u 2 (3.2.9)
2A
c1 c2 c3 u 3
Where, A is the area of the triangle and can be obtained as follows.
1 𝑥1 𝑦1
1
A = 2 �1 𝑥2 𝑦2 � (3.2.10)
1 𝑥3 𝑦3
Now, eq. (3.2.1) can be written from the above polynomial coefficients.
12
1
u x 2 y3 x 3 y 2 y 2 y3 x x 3 x 2 y u1
2A
1
x 3 y1 x1y3 y3 y1 x x1 x 3 y u 2 (3.2.11)
2A
1
x1y 2 x 2 y1 y1 y 2 x x 2 x1 y u 2
2A
Thus, the interpolation function can be obtained from the above as:
1
x 2 y 3 x 3 y 2 y 2 y 3 x x 3 x 2 y
N1
2A
N N 2
1 x 3 y1 x1y3 y3 y1 x x1 x 3 y
(3.2.12)
2A
N
1
3
x
1 2 y x y
2 1 y1 y 2 x x 2 x 1 y
2A
Such three node triangular element is commonly known as constant strain triangle (CST) as its strain
is assumed to be constant inside the element. This property may be derived from eq. (3.2.1) and
eq.(3.2.2). For example, in case of 2-D plane stress/strain problem, one can express the strain inside
the triangle with the help of eq.(3.2.1) and eq.(3.2.2):
u ( 0 1x 2 y)
x 1
x x
v ( 3 4 x 5 y)
y 5 (3.2.13)
y y
v u
xy 2 2
x y
CST is the simplest element to develop mathematically. As there is no variation of strain inside the
element, the mesh size of the triangular element should be small enough to get correct results. This
element produces constant temperature gradients ensuring constant heat flow within the element for
heat transfer problems.
Higher order elements are useful if the boundary of the geometry is curve in nature. For curved case,
higher order triangular element can be suited effectively while generating the finite element mesh.
Moreover, in case of flexural action in the member, higher order elements can produce more
accurate results compare to those using linear elements. Various types of higher order triangular
13
elements are used in practice. However, most commonly used triangular element is the six node
element for which development of shape functions are explained below.
Fig. 3.2.4 (a) Six node triangular element (b) Lines of constant values of the area coordinates
Using the above field variable function, one can reach the following expression using interpolation
function and the nodal values.
6
x, y N i x, yi (3.2.15)
i1
Here, the every shape function must be such that its value will be unity if evaluated at its related
node and zero if evaluated at any of the other five nodes. Moreover, as the field variable
representation is quadratic, each interpolation function will also become quadratic. Fig. 3.2.4(a)
shows the six node element with node numbering convention along with the area coordinates at three
corners. The six node element with lines of constant values of the area coordinates passing through
the nodes is shown in Fig. 3.2.4(b). Now the interpolation functions can be constructed with the help
of area coordinates from the above diagram. For example, the interpolation function N1 should be
unity at node 1 and zero at all other five nodes. According to the above diagram, the value of L1 is 1
14
at node 1 and ½ at node 4 and 6. Again, L1 will be 0 at nodes 2, 3 and 5. To satisfy all these
conditions, one can propose following expression:
1
N1 x, y N1 L1 ,L 2 ,L3 L1 L1
2 (3.2.16)
Evaluating the above expression, the value of N1 is becoming ½ at node 1 though it must become
unity. Therefore, the above expression is slightly modified satisfying all the conditions and will be as
follows:
1
N1 2L1 L1 L1 2L1 1
2 (3.2.17)
Eq. (3.2.17) assures the required conditions at all the six nodes and is a quadratic function, as L1 is a
linear function of x and y. The remaining five interpolation functions can also be obtained in similar
fashion applying the required nodal conditions. Thus, the shape function for the six node triangle
element can be written as given below.
N1 L1 2L1 1
N 2 L 2 2L 2 1
N 3 L3 2L3 1 (3.2.18)
N 4 4L1L 2
N 5 4L 2 L3
N 6 4L3L1
Such six node triangular element is commonly known as linear strain triangle (LST) as its strain is
assumed to vary linearly inside the element. In case of 2-D plane stress/strain problem, the element
displacement field for such quadratic triangle may be expressed as
u ( x, y ) =α 0 + α1 x + α 2 y + α 3 x 2 + α 4 xy + α 5 y 2
v ( x, y ) =α 6 + α 7 x + α 8 y + α 9 x 2 + α10 xy + α11 y 2 (3.2.19)
So the element strain can be derived from the above displacement field as follows.
u
x 1 2 3 x 4 y
x
v
y 8 10 x 211y (3.2.20)
y
v u
xy 2 4 x 2 5 y 7 2 9 x 10 y
x y
The above expression shows that the strain components are linearly varying inside the element.
Therefore, this six node element is called linear strain triangle. The main advantage of this element is
that it can capture the variation of strains and therefore stresses of the element.
15
Fig. 3.2.5 contains four types of element. Type 1 has only three nodes, type 2 element has five
nodes, type 3 has four nodes and type 4 has six nodes. The shape function for 3-node and 6-node
triangular elements has already been derived. The shape functions of 6-node element can suitably be
degraded to derive shape functions of other two types of triangular elements.
u N1u1 N 2 u 2 N 3u 3 N 4 u 4 N 5u 5 N 6 u 6 (3.2.21)
Where, N1, N2, …, N6 are the shape functions and is given in eq.(3.2.18). If there is no node between
2 and 3, the displacement along line 2-3 is considered to vary linearly. Thus the displacement at an
assumed node 5´ may be written as
u 2 u3
u '5
2 (3.2.22)
Substituting, the value of u’5 for u5 in eq.(3.2.21) the following expression will be obtained.
u 2 u3
u N1u1 N 2 u 2 N 3u 3 N 4 u 4 N 5 N6u 5 (3.2.23)
2
Thus, the displacement function can be expressed by five nodal displacements as:
N N
u N1u1 N 2 5 u 2 N 3 5 u 3 N 4 u 4 N 6 u 5
2 2 (3.2.24)
However, the displacement function for the five node triangular element can be expressed as
u N1u1 N 2 u 2 N 3 u 3 N 4 u 4 N 5 u 5 (3.2.25)
Comparing eq.(3.2.24) and eq.(3.2.25) and observing node 6 of six node triangle corresponds to
node 5 of five node triangle, we can write
N5 N
N '1 N1, N 2 N 2 , N 3 N 3 5 , N 4 N 4 an dN 5 N 5
2 2 (3.2.26)
Hence, the shape function of a five node triangular element will be
17
Applying nodal conditions, the above expression may be written in matrix form as
f1 1 x1 y1 x1 y1 a0
f2 1 x2 y2 x2 y2 a1
(3.3.2)
f3 1 x3 y3 x3 y3 a2
f4 1 x4 y4 x4 y4 a3
The unknown polynomial coefficients may be obtained from the above equation with the use of
nodal field variables.
1
a0
1 x1 y1 x1 y1
f1
a1 1 x2 y2 x2 y2
f2
(3.3.3)
a 1
x3 y3 x3 y3
f3
a
f
2
1 x4 y4
3
x4 y4 4
Thus, the field variable at any point inside the element can be described in terms of nodal values as
1
a0 1 x1 y1 x1 y1 f1
a 1 x2 y2
f x, y 1 x y xy 1 1 x y xy
x2 y2
f2
a2
1 x3 y3 x3 y3 f3
1
a3 x4 y4 x4 y4 f4
(3.3.4)
f1
f
N1 N2 N3 N 4 2
f3
f4
From the above expression, the shape function Ni can be derived and will be as follows.
20
x − x2 y − y4
N1 =
x1 − x2 y1 − y4
x − x1 y − y3
N2 =
x2 − x1 y2 − y3
(3.3.5)
x − x4 y − y2
N3 =
x3 − x4 y3 − y2
x − x3 y − y1
N4 =
x4 − x3 y4 − y1
Now, substituting the nodal coordinates in terms of (x1, y1) as (–a, –b) at node 1; (x2, y2) as (a, –b) at
node 2; (x3, y3) as (a, b) at node 3 and (x4, y4) as (–a, b) at node 4 the above expression can be re-
written as:
1
N1 = ( x − a )( y − b )
4ab
1
N2 = ( x + a )( y − b )
4ab
(3.3.6)
1
N3 = ( x + a )( y + b )
4ab
1
N4 = ( x − a )( y + b )
4ab
Thus, the shape function N can be found from the above expression in Cartesian coordinate system.
From the figure, the relation between two coordinate systems can be expressed as
x−x y− y
=ξ = and η (3.3.7)
a b
Here, 2a and 2b are the width and height of the rectangle. The coordinate of the center of the
rectangle can be written as follows:
x1 + x2 y1 + y4
=x = and y (3.3.8)
2 2
Thus, from eq. (3.3.7) and eq.(3.3.8), the nodal values in natural coordinate systems can be derived
which is shown in Fig. 3.3.4(b). With the above relations variations of x & h will be from -1 to +1.
Now the interpolation function can be derived in a similar fashion as done in section 3.3.1.1. The
filed variable can be written in natural coordinate system ensuring inter-element continuity as:
f x, h aa
0 1x a2 h a3xh (3.3.9)
The coordinates of four nodes of the element in two different systems are shown in Table 3.3.1 for
ready reference for the derivation purpose. Applying the nodal values in the above expression one
can get
f1 1 1 1 1 a0
f2 1 1 1 1 a1 (3.3.10)
f3 1 1 1 1 a2
f4 1 1 1 1 a3
22
Table 3.3.1 Cartesian and natural coordinates for four node element
1 x1 h
4
N
1 x 1 h
1
N2
4
Ni (3.3.13)
N3
1 x1 h
N4
4
1 x1 h
4
23
The nodal field variables can be obtained from the above expression after putting the coordinates at
nodes.
f1 1 1 1 1 1 1 1 1 a0
f2 1 1 1 1 1 1 1 1 a1
f 1 1
3 1 1 1 1 1 1 a2
f 1 1 1
1 1 1 1 1 a3
fi 4 Aai (3.3.15)
f5 1 0 1 0 0 1 0 0 a4
f6 1 1 0 1 0 0 0 0 a5
f7 1 0 1 0 0 1 0 0 a6
f 1 1 0
8 1 0 0 0 0 a7
Replacing the unknown coefficient αi in eq.(3.3.14) from eq.(3.3.15), the following relations will be
obtained.
24
1 1 1 1 2 2 2 2 f1
0 0 0 0 0 2 0 2 f2
0
0 0 0 2 0 2 0
f3
2 1 1 1 1 1 2 0 2 0 f4
1 x h x xh h x h xh
2 2 2
0
4 1 1 1 1 0 0 0 f5
1 1 1 1
0 2 0 2 f6
1 1 1 1 2 0 2 0 f7
1 1 1 1 0 2 0 f8
2
f1
f2
f3
f
N1 N2 N3 N4 N5 N6 N7 N8 4
f5
f6
f7
f
8 (3.3.16)
Thus, the interpolation function will become
The shape functions of rectangular elements with higher nodes can be derived in similar manner
using appropriate polynomial satisfying all necessary criteria. However, difficulty arises due to the
inversion of large size of the matrix because of higher degree of polynomial chosen. In next lecture,
the shape functions of rectangular element with higher nodes will be derived in a much simpler way.
25
coordinate of jth node in the element. In the above equation if we put ξ = ξj, and j ≠ i, the value of the
function fi(ξ) will be equal to zero. Similarly, putting ξ = ξi, the numerator will be equal to
denominator and hence fi(ξ) will have a value of unity. Since, Lagrange interpolation function for ith
node includes product of all terms except jth term; for an element with n nodes, fi(ξ) will have n-1
degrees of freedom. Thus, for one-dimensional elements with n-nodes we can define shape function
as N i () f i () .
2 x x1
1
l (3.4.2)
The shape function for two node bar element as shown in Fig. 3.4.1 can be derived from eq.(3.4.1) as
follows:
2 1 1
N1 f1 1
1 2 1 (1) 2
1 1 1 (3.4.3)
N 2 f1 1
2 1 1 (1) 2
2
3 ( ) 1 1
N1 f1 1
2 1 3
1 1(2) 2
1
3 1 1
N 2 f 2 1 2 (3.4.4)
2
1 2 3 1 1
1 2 1() 1
N 3 f3 1
3
1 3 2 2(1) 2
27
Fig. 3.4.3 Quadratic shape functions for three node bar element
Similarly, other interpolation functions can be derived which are given below.
1
N 2 f1 1 1
, f 2
4
1
` N 3
, f 2
f 2 1 1 (3.4.7)
4
1
N 4
, f1
f 2 1 1
4
These shape functions are exactly same as eq.(3.3.13) which was derived earlier by choosing
polynomials.
28
Thus, it is observed that the two dimensional Lagrange element contains internal nodes (Fig. 3.4.6)
which are not connected to other nodes.
(Module 3, lecture 3). The interpolation function can be derived by inspection in terms of natural
coordinate system as follows:
(a) Linear element
1
N i
, 1 i 1 i (3.4.10)
4
(b) Quadratic element
(i) For nodes at 1, 1
1
N i
, 1 i 1
i i i 1 (3.4.11a)
4
(ii) For nodes at 1, 0
, 1 i 1 2
1
N i (3.4.11b)
2
(iii) For nodes at 0, 1
1 2 1 i
1
N i
, (3.4.11c)
2
(c) Cubic element
(i) For nodes at 1, 1
i 9
N i
,
1
1 i 1 2 2 10
32 (3.4.12a)
1
(ii) For nodes at 1,
3
Thus, the nodal conditions must be satisfied by each interpolation function to obtain the functions
serendipitously. For example, let us consider an eight node element as shown in Fig. 3.4.8 to derive
its shape function. The interpolation function N1 must become zero at all nodes except node 1, where
its value must be unity. Similarly, at nodes 2, 3, and 6, ξ = 1, so including the term ξ − 1 satisfies the
zero condition at those nodes. Similarly, at nodes 3, 4 and 7, η = 1 so the term η − 1 ensures the zero
condition at these nodes.
Again, at node 5, (ξ, η) = (0, −1), and at node 8, (ξ, η) = (−1, 0). Hence, at nodes 5 and 8, the term (ξ
+ η + 1) is zero. Using this reasoning, the equation of lines are expressed in Fig. 3.4.9. Thus, the
32
Fig. 3.4.9 Equations of lines for two dimensional eight node element
N 5 1 2 1 ,
1 1
N1 1 1 1 ,
4 2
N 6 1 1 2 ,
1 1
N 2 1 1 1 ,
4 2
(3.4.13)
N 7 1 2 1 ,
1 1
N 3 1 1 1 ,
4 2
N 4 1 1 1 and N8 1 1 2
1 1
4 2
33
It may be observed that the Lagrange elements have a better degree of completeness in polynomial
function compare to serendipity elements. Therefore, Lagrange elements produce comparatively
faster and better accuracy.
34
There are two basic families of three-dimensional elements similar to two-dimensional case.
Extension of triangular elements will produce tetrahedrons in three dimensions. Similarly,
rectangular parallelepipeds are generated on the extension of rectangular elements. Fig. 3.5.1 shows
few commonly used solid elements for finite element analysis.
Derivation of shape functions for such three dimensional elements in Cartesian coordinates are
algebraically quite cumbersome. This is observed while developing shape functions in two
dimensions. Therefore, the shape functions for the two basic elements of the tetrahedral and
parallelepipeds families will be derived using natural coordinates.
The polynomial expression of the field variable in three dimensions must be complete or incomplete
but symmetric to satisfy the geometric isotropy requirements. Completeness and symmetry can be
35
ensured using the Pascal pyramid which is shown in Fig. 3.5.2. It is important to note that each
independent variable must be of equal strength in the polynomial.
The following 3-D quadratic polynomial with complete terms can be applied to an element having
10 nodes.
f x, h, z aa
0 1x a2 h a3z a4 x a5 h a6z a7 xh a8 hz a9zx
2 2 2
(3.5.1)
However, the geometric isotropy is not an absolute requirement for field variable representation to
derive the shape functions.
The linear shape function for this element can be expressed as,
{N} = [ L1 L4 ]
T
L2 L3
(3.5.2)
Here, L1 , L2 , L3 , L4 are the set of natural coordinates inside the tetrahedron and are defined as follows
Vi
Li =
V (3.5.3)
Where Vi is the volume of the sub element which is bound by point P and face i and V is the total
volume of the element. For example L1 may be interpreted as the ratio of the volume of the sub
element P234 to the total volume of the element 1234. The volume of the element V is given by the
determinant of the nodal coordinates as follows:
1 1 1 1
1 x1 x2 x3 x4
V=
6 y1 y2 y3 y4 (3.5.4)
z1 z2 z3 z4
The relationship between the Cartesian and natural coordinates of point P may be expressed as
1 1 1 1 1 L1
x x x4 L2
1 x2 x3
=
y y1 y 4 L3
y2 y3 (3.5.5)
z z1 z2 z3
z 4 L4
It may be noted that the identity included in the first row ensure the matrix invertible.
37
L1 + L2 + L3 + L4 = 1
(3.5.6)
The inverse relation is given by
L1 V1 a1 b1 c1 1
L
2 1 V2 a2 b2 c2 x
=
L3 6V V3 a3 b3 c3 y (3.5.7)
L4 V4 a4 b4 c4 z
Here, Vi is the volume subtended from face i and terms ai , bi , and ci represent the projected area of
face i on the x, y, z coordinate planes respectively and are given as follows:
ai = ( z j y k − z k y j ) + ( z k y l − z l y k ) + ( z l y j − z j y l )
bi = ( z j xk − z k x j ) + ( z k xl − z l xk ) + ( z l x j − z j xl )
(3.5.8)
ci = ( y j x k − y k x j ) + ( y k xl − y l x k ) + ( y l x j − y j xl )
i, j , k , l will be in cyclic order (i.e., 1 2 3 4 1). The volume coordinates fulfil all nodal
conditions for interpolation functions. Therefore, the field variable can be expressed in terms of
nodal values as
f x, y, z L1f1 L2f2 L3f3 L4f4
(3.5.9)
Though the shape functions (i.e., the volume coordinates) in terms of global coordinates is
algebraically complex but they are straightforward. The partial derivatives of the natural coordinates
with respect to the Cartesian coordinates are given by
∂Li ai ∂Li b ∂Li c
= , = i , = i
∂x 6V ∂y 6V ∂z 6V (3.5.10)
Similar to area integral, the general integral taken over the volume of the element is given by,
p!q!r!s!
∫ L L L L dV = ( p + q + r + s + 3)!.6V
p q r s
1 2 3 4
(3.5.11)
v
The four node tetrahedral element is a linear function of the Cartesian coordinates. Hence, all the
first partial derivatives of the field variable will be constant. The tetrahedral element is a constant
strain element as the element exhibits constant gradients of the field variable in the coordinate
directions.
Higher order elements of the tetrahedral family are shown in Fig. 3.5.1. The shape functions for such
higher order three dimensional elements can readily be derived in volume coordinates, as for higher-
order two-dimensional triangular elements. The second element of this family has 10 nodes and a
cubic form for the field variable and interpolation functions.
+++++++++++
3.5.2 Brick Elements
38
Various orders of elements of the parallelepiped family are shown in Fig. 3.5.1. Fig. 3.5.4 shows the
eight-node brick element with reference to a global Cartesian coordinate system and then with
reference to natural coordinate system. The natural coordinates for the brick element can be relate
Cartesian coordinate system by
x−x y− y z−z
=ξ = , η = and ζ (3.5.12)
a b c
Here, 2a, 2b and 2c are the length, height and width of the element. The coordinate of the center of
the brick element can be written as follows:
x1 + x2 y1 + y4 z1 + z5
=x = , y = and z (3.5.13)
2 2 2
Thus, from eq.(3.5.12) and eq.(3.5.13), the nodal values in natural coordinate systems can be derived
which is shown in Fig. 3.5.4(b). With the above relations variations of x, h & z will be from -1 to +1.
Now the interpolation function can be derived in several procedures as done in case of two
dimensional rectangular elements. For example, the interpolation function can be derived by
inspection in terms of natural coordinate system as follows:
1
N i
, , 1 i 1 i 1 i (3.5.14)
8
By using field variable the following terms of the polynomial may be used for deriving the shape
function for eight-node brick element.
f x, h, z aa
0 1x a2 h a3z a4 xh a5 hz a6zx a7zhx
(3.5.14)
The above equation is incomplete but symmetric. However, such representations are quite often used
and solution convergence is achieved in the finite element analysis. Again, the shape functions for
39
three dimensional 8-node or 27-node brick elements can be derived using Lagrange interpolation
function. For this we need to introduce interpolation function in the ζ-direction. Thus, for example,
the Lagrange interpolation function for a three dimensional 8 node brick element can be obtained
from the product of appropriate interpolation functions in the ξ, η and ζ directions. Therefore, the
shape function will become
N i
, , f i
fi fi Where, i = 1,2,3, …., n-node (3.5.15)
Thus using the Lagrange interpolation function the shape function at node 1 can be expressed as
2
2 2
N1
, , f1
f1 f1
2 1
1 2 1 2 (3.5.16)
1 1 1 1
1 1 1
1 (1) 1 (1) 1 (1) 4
Using any of the above concepts, the interpolation function for 8-node brick element can be found as
follows:
1 1
N1 1 1 1 , N1 1 1 1 ,
4 4
1 1
N 3 1 1 1 , N 4 1 1 1 ,
4 4 (3.5.17)
1 1
N 5 1 1 1 , N 6 1 1 1 ,
4 4
1 1
N 7 1 1 1 , N8 1 1 1
4 4
The shape functions of rectangular parallelepiped elements with higher nodes can be derived in
similar manner satisfying all necessary criteria.
40
(a) The Continuum to be discritized (b) Discritization using Triangular Elements (c)
Discritization using rectangular elements (d) Discritization using a combination of
rectangular and quadrilateral elements
41
Figure 3.6.1(b) presents a possible mesh using triangular elements. Though, triangular elements can
suitable approximate the circular boundary of the continuum, but the elements close to the center
becomes slender and hence affect the accuracy of finite element solutions. One possible solution to
the problem is to reduce the height of each row of elements as we approach to the center. But,
unnecessary refining of the continuum generates relatively large number of elements and thus
increases computation time. Alternatively, when meshing is done using rectangular elements as
shown in Fig 3.6.1(c), the area of continuum excluded from the finite element model is significantly
adequate to provide incorrect results. In order to improve the accuracy of the result one can generate
mesh using very small elements. But, this will significantly increase the computation time. Another
possible way is to use a combination of both rectangular and triangular elements as discussed in
section 3.2. But such types of combination may not provide the best solution in terms of accuracy,
since different order polynomials are used to represent the field variables for different types of
elements. Also the triangular elements may be slender and thus can affect the accuracy. In Fig.
3.6.1(d), the same continuum is discritized with rectangular elements near center and with four-node
quadrilateral elements near boundary. This four-node quadrilateral element can be derived from
rectangular elements using the concept of mapping. Using the concept of mapping regular triangular,
rectangular or solid elements in natural coordinate system (known as parent element) can be
transformed into global Cartesian coordinate system having arbitrary shapes (with curved edge or
surfaces). Fig. 3.6.2 shows the parent elements in natural coordinate system and the mapped
elements in global Cartesian system.
42
43
n
(3.6.1)
y= N1 y1 + N 2 y2 + ... + N n y=
n ∑N y
i =1
i i
n
z= N1 z1 + N 2 z2 + ... + N n z=
n ∑N z
i =1
i i
Where,
n=No.of Nodes
N i =Interpolation Functions
x i ,yi ,zi =Coordinates of Nodal Points of the Element
One can also express the field variable variation in the element as
44
n
φ (ξ ,η , ζ ) = ∑ N i (ξ ,η , ζ ) φi (3.6.2)
i =1
As the same shape functions are used for both the field variable and description of element
geometry, the method is known as isoparametric mapping. The element defined by such a method is
known as an isoparametric element. This method can be used to transform the natural coordinates of
a point to the Cartesian coordinate system and vice versa.
Example 3.6.1
Determine the Cartesian coordinate of the point P (ξ= 0.8, η= 0.9) as shown in Fig. 3.6.3.
Solution:
As described above, the relation between two coordinate systems can be represented through their
interpolation functions. Therefore, the values of the interpolation function at point P will be
(1 − ξ )(1 − n) (1 − 0.8)(1 − 0.9)
=N1 = = 0.005
4 4
(1 + ξ )(1 − n) (1 + 0.8)(1 − 0.9)
=N2 = = 0.045
4 4
(1 + ξ )(1 + n) (1 + 0.8)(1 + 0.9)
=N3 = = 0.855
4 4
(1 − ξ )(1 + n) (1 − 0.8)(1 + 0.9)
=N4 = = 0.095
4 4
4
x = ∑ N i xi = 0.005 ×1 + 0.045 × 3 + 0.855 × 3.5 + 0.095 ×1.5 = 3.275
i =1
4
y = ∑ N i yi = 0.005 ×1 + 0.045 ×1.5 + 0.855 × 4.0 + 0.095 × 2.5 = 3.73
i =1
Thus the coordinate of point P (ξ= 0.8, η= 0.9) in Cartesian coordinate system will be 3.275, 3.73.
Solid isoparametric elements can easily be formulated by the extension of the procedure followed for
2-D elements. Regardless of the number of nodes or possible curvature of edges, the solid element is
just like a plane element which is mapped into the space of natural co-ordinates, i.e,
ξ = ±1,η = ±1, ζ = ±1 .
∂ξ
= ∑ ∂ξ
i =1
xi
Similarly one can calculate the other terms J12, J21 and J22 of the Jacobian matrix. Hence,
46
n N i n
N i
x x yi
i1 x i
J n i 1
(3.6.5)
N i xi
n
N i
i1 h
i 1 h
yi
x x
J
1
(3.6.6)
y
x
J11* J12*
Considering * are the elements of inverted [J] matrix, we may arise into the following
J *
J 22
21
relations.
J11* J12*
x x h
(3.6.7)
J 21
*
J 22
*
y x h
Similarly, for three dimensional case, the following relation exists between the derivative operators
in the global and the natural coordinate system.
∂ ∂x ∂y ∂z ∂ ∂
∂ξ ∂ξ ∂ξ ∂ξ ∂x
∂x
∂ ∂x ∂y ∂z ∂ ∂
= = [J ] (3.6.8)
∂η ∂η ∂η ∂η ∂y ∂y
∂ ∂x ∂y ∂z ∂ ∂
∂ζ ∂ζ ∂ζ ∂ζ ∂z ∂z
Where,
47
∂x ∂y ∂z
∂ξ ∂ξ ∂ξ
∂x ∂y ∂z
[J ] = (3.6.9)
∂η ∂η ∂η
∂x ∂y ∂z
∂ζ ∂ζ ∂ζ
[J] is known as the Jacobian Matrix for three dimensional case. Putting eq. (3.6.1) in eq. (3.6.9) and
after simplifying one can get
∂N i ∂N i ∂N i
∂ξ xi ∂ξ
yi
∂ξ
zi
n
∂N ∂N i ∂N i
[ J ] = ∑ i xi (3.6.10)
∂η
yi zi
i =1
∂η ∂η
∂N i ∂N i ∂N i
∂ζ xi ∂ζ
yi
∂ζ
zi
w. Thus,
u N N n
1 0 0
x x x u1
u
N1 N n
h
h
h
0 0
un
(3.7.3)
N1 N n
v1
v
0 0
x x x
N n
v
0 0
N1
n
v
h
h
h
As a result, eq. (3.7.2) can be written using eq. (3.7.3) which will be as follows.
49
N N n u1
1 0 0
:
J11 N N n :
0 1
(3.7.4)
0
* *
J12 0 0
u n
0 0 J*21 J*22
J* * N1 N n v1
J*22 0
*
21 J11 J12 0
:
N1 N n :
0
v n
0
Or,
Bd (3.7.5)
Where {d} is the nodal displacement vector and [B] is known as strain displacement relationship
matrix and can be obtained as
N1 N n
0
0
J11
* * N N n
J12 0 0 1
0 0 (3.7.6)
*
B 0 0 J 21 J 22
*
J* J* J* J* 0 N1 N n
21 22 11 12 0
N1 N n
0 0
It is necessary to transform integrals from Cartesian to the natural coordinates as well for calculation
of the elemental stiffness matrix in isoparametric formulation. The differential area relationship can
be established from advanced calculus and the elemental area in Cartesian coordinate can be
represented in terms of area in natural coordinates as:
dA dx dy J d
d (3.7.7)
Here J is the determinant of the Jacobian matrix. The stiffness matrix for a two dimensional
element may be expressed as
A
Here, [B] is the strain-displacement relationship matrix and t is the thickness of the element. The
above expression in Cartesian coordinate system can be changed to the natural coordinate system as
follows to obtain the elemental stiffness matrix
50
1 1
Example 3.7.1:
Calculate the Jacobian matrix and the strain displacement matrix for four node two dimensional
quadrilateral elements corresponding to the gauss point (0.57735, 0.57735) as shown in Fig. 3.6.4.
Solution:
The Jacobian matrix for a four node element is given by,
n N i n
N i
x x yi
i1 x i
J n i 1
N i xi
n
N i
i1 h
i 1 h
yi
For the four node element one can find the following relations.
1 1 N1 1 N1 1
N1 , ,
4 4 4
1 1 N 2 1 N 2 1
N2 , ,
4 4 4
1 1 N 3 1 N 3 1
N3 , ,
4 4 4
51
1 1 N 4 1 N 4 1
N4 , ,
4 4 4
Now, for a four node quadrilateral element, the Jacobian matrix will become
N1 N 2 y1
N 3 N 4 x 1
y2 x 2
J
N N 2 y3
N 3
N 4 x 3
1
y 4
x 4
1 1 1 1 x1 y1
x2 y2
4 4 4 4
1 1 1 1 x 3 y3
4 4 4 4 x 4 y 4
Putting the values of ξ & η as 0.57735 and 0.57735 respectively, one will obtain the following.
N1 N1
0.10566 0.10566
N 2 N 2
0.10566 0.39434
N 3 N 3
0.39434 0.39434
N 4 N 4
0.39434 0.10566
4
N i
Hence, J11 xi 0.105661 0.105663 0.394343.5 0.394341.5 1.0
i 1 x
Similarly, J12 =0.64632, J21 =0.25462 and J22 =1.14962.
Hence,
1.00000 0.64632
J
0.25462 1.14962
Thus, the inverse of the Jacobian matrix will become:
* J12* 1.1671 0.6561
J * J11
J * *
21 J 22 0.2585 1.0152
Hence strain displacement matrix is given by,
52
N1 N n
0 0
J11
* * N N n
J12 0 0 1 0 0
B 0 0 J*21 J*22
J* * N1 N n
J*22 0
*
21 J11 J12 0
N1 N n
0 0
1.1671 0.6561 0 0
0 0 0.2585 1.0152
0.2585 1.0152 1.1671 0.6561
0.10566 0.10566 0.39434 0.39434 0 0 0 0
0.10566 0.39434 0.39434 0.10566 0 0 0 0
0 0 0 0 0.10566 0.10566 0.39434 0.39434
0 0 0 0 0.10566 0.39434 0.39434 0.10566
∂u
∂x
∂u
∂y
∂u
∂z
ε x 1 0 0 0 0 0 0 0 0 ∂v
ε 0
y 0 0 0 1 0 0 0 0 ∂x
ε z 0 0 0 0 0 0 0 0 1 ∂v
=
γ xy 0 1 0 1 0 0 0 0 0 ∂y
γ yz 0 0 0 0 0 1 0 1 0 ∂v
γ zx 0 0 1 0 0 0 1 0 0 ∂z
∂w
∂x
∂w
∂y
∂w (3.7.10)
∂z
8
For an 8 node brick element u can be represented as, u N i ui and similarly for v & w.
i 1
∂u 8
∂N ∂u ∂N ∂u ∂N
8 8
= ∑ i ui , = ∑ i ui & = ∑ i ui
∂ξ i =1 ∂ξ ∂η i =1 ∂η ∂ζ i =1 ∂ζ
∂v 8
∂N ∂v 8
∂N ∂v 8
∂N
= ∑ i vi , = ∑ i vi & = ∑ i vi (3.7.12)
∂ξ i =1 ∂ξ ∂η i =1 ∂η ∂ζ i =1 ∂ζ
∂w 8 ∂N i ∂w 8 ∂N i ∂w 8
∂N
=∑ wi , =∑ wi & = ∑ i wi
∂ξ i =1 ∂ξ ∂η i =1 ∂η ∂ζ i =1 ∂ζ
∂N i
∂ξ 0 0
∂N i
J *
J*
J*
0 0 0 0 0 0 0 ∂η
0
11 12 13
∂N i
* * *
0 0 0 J21 J22 J23 0 0 0
0 0
8
0 0 0 0 0 0 *
J 31 *
J 32 *
J 33 ∂ζ
[ B] * ×∑ (3.7.14)
J 21
*
J 22 *
J 23 J11* J12* J13* 0 0 0 i =1 ∂N i ∂N i
∂η 0
0 0 0 *
J 31 *
J 32 *
J 33 *
J 21 *
J 22 *
J 23 ∂ξ
*
J 31
*
J 32 *
J 33 0 0 0 J11* J12* J13
*
∂N i ∂N i
0 ∂ζ ∂η
∂N i ∂N i
∂ζ 0
∂ξ
The stiffness matrix may be found by using the following expression in natural coordinate system.
1 1 1
V 1 1 1
56
The integrations, we generally encounter in finite element methods, are quite complicated and it is
not possible to find a closed form solutions to those problems. Exact and explicit evaluation of the
integral associated to the element matrices and the loading vector is not always possible because of
the algebraic complexity of the coefficient of the different equation (i.e., the stiffness influence
coefficients, elasticity matrix, loading functions etc.). In the finite element analysis, we face the
problem of evaluating the following types of integrations in one, two and three dimensional cases
respectively. These are necessary to compute element stiffness and element load vector.
d ; , d d ; , , d d d ; (3.8.1)
Approximate solutions to such problems are possible using certain numerical techniques. Several
numerical techniques are available, in mathematics for solving definite integration problems,
including, mid-point rule, trapezoidal-rule, Simpson’s 1/3rd rule, Simpson’s 3/8th rule and Gauss
Quadrature formula. Among these, Gauss Quadrature technique is most useful one for solving
problems in finite element method and therefore will be discussed in details here.
The concept of Gauss Quadrature is first illustrated in one dimension in the context of an integral in
1 x2
the form of I d from
f (x)dx . To transform from an arbitrary interval of x1≤ x ≤ x2
1 x1
to an interval of -1 ≤ ξ ≤ 1, we need to change the integration function from f(x) to ϕ(ξ) accordingly.
Thus, for a linear variation in one dimension, one can write the following relations.
1 1
x x1 x 2 N1x1 N 2 x 2
2 2
1 1 1 1
so for 1, x x1 x 2 x1
2 2
1, x x 2
x2 1
I f (x)dx d
x1 1
Numerical integration based on Gauss Quadrature assumes that the function ϕ(ξ) will be evaluated
over an interval -1 ≤ ξ ≤ 1. Considering an one-dimensional integral, Gauss Quadrature represents
the integral ϕ(ξ) in the form of
1 n
I
1
d w i
i w1
1 w 2
2 .. w
(3.8.2)
i1
57
Where, the ξ1, ξ2, ξ3, ..., ξn represents n numbers of points known as Gauss Points and the
corresponding coefficients w1, w2, w3, …, wn are known as weights. The location and weight
coefficients of Gauss points are calculated by Legendre polynomials. Hence this method is also
sometimes referred as Gauss-Legendre Quadrature method. The summation of these values at n
sampling points gives the exact solution of a polynomial integrand of an order up to 2n-1. For
example, considering sampling at two Gauss points we can get exact solution for a polynomial of an
order (2×2-1) or 3. The use of more number of Gauss points has no effect on accuracy of results but
takes more computation time.
1
( )d w1
( 1) (3.8.3)
Since there are two parameters w1 and 1 , we need a first order polynomial for ϕ(ξ) to evaluate the
a 0 a1 ,
eq.(3.8.3) exactly. For example, considering,
1
Error a 0 a1
d w1
1 0
1
2a 0 w1 a 0 a11 0
a 0 2 w1 w1a11 0 (3.8.4)
Thus, the error will be zero if w1 2 and 1 0 . Putting these in eq.(3.8.3), for any general ϕ, we
have
1
I
d 2 0 (3.8.5)
1
1
d w1
1 w 2
2 (3.8.6)
This means we have four parameters to evaluate. Hence we need a 3rd order polynomial for ϕ(ξ) to
exactly evaluate eq.(3.8.6).
a 0 a1 a 22 a 33
Considering,
1
Error a 0 a1 a 2 2 a 3
3
d w1 2
1 w 2
1
58
points are symmetrically placed with respect to origin and those symmetrical points have the same
weights. For accuracy in the calculation maximum number digits for gauss point and gauss weights
should be taken. The Location and weights given in the Table 3.8.1 must be used when the limits of
integration ranges from -1 to 1. Integration limits other than [-1, 1], should be appropriately changed
to [-1, 1] before applying these values.
Example 1:
1 2x
Evaluate I e x 2 dx using one, two and three point gauss Quadrature.
0 x 2
Solution:
Before applying the Gauss Quadrature formula, the existing limits of integration should be changed
from [0, 1] to [-1, +1]. Assuming, a bx , the upper and lower limit can be changed. i.e., at x =
0, ξ = -1 and at x = 1, ξ = +1. Thus, putting these conditions and solving for a & b, we get a = -1 and
b = 2. The relation between two coordinate systems will become 2x 1 and d 2dx .
Therefore the initial equation can be written as
60
1
2
1
1 2
2
I e dx
1
2
1 2
2
1 1 2 4 1
1
2 1
Or, I e d
1 8
2
Numerical integrations using Gauss Quadrature method can be extended to two and three
dimensional cases in a similar fashion. Such integrations are necessary to perform for the analysis of
plane stress/strain problem, plate and shell structures and for the three dimensional stress analysis.
n n n
, d d w i
i , d w j w i i , j
1 1 1
I
1 1 1
i1 i1 i1
Or,
n n
i , j
I w i w j (3.9.1)
i1 j1
Example 1:
yd 4 x b3
yc4 x a 2
Solution:
Before applying the Gauss Quadrature formula, the above integral should be converted in terms of
and and the existing limits of y should be changed from [-4,4] to [-1, 1] and that of x is from
[2,3] to [-1,1].
62
b a b a 5 d
x ; dx
2 2 2 2
d c d c
y 4; dy 4d
2 2
1 1 1 1
3
2
I 2
d d , d d
2
2 2 4
1 1 1 1
3
2
, 2
2 4 2 3 1 2
2 2 2
where
2
1 1 1 1
1 ; 1 ; 2 ; 2
3 3 3 3
1 2
2 2
1 , 1 1 , 2 w1
I w 1 w 2
2 , 1 2 , 2 w 2
In a similar way one can extend the gauss Quadrature for three dimensional problems also and the
integral can be expressed by.
n n n
i , j, k
, , d d d w i w j w k
1 1 1
I (3.9.3)
1 1 1
i1 j1 k 1
The above equation will produce exact value for a polynomial integrand if the sampling points are
selected as described earlier sections.
64
Here, [B] and [D] are the strain displacement relationship matrix and constitutive matrix respectively
and integration is performed over the domain. As the element stiffness matrix will be calculated in
natural coordinate system, the strain displacement matrix [B] and Jacobian matrix [J] are functions
of
, and . In case of two dimensional isoparametric element, the stiffness matrix will be
simplified to
1 1
k t [B] [D][B]d
T
d J (3.9.5)
1 1
This is actually an 8×8 matrix containing the integrals of each element. We do not need to integrate
elements below the main diagonal of the stiffness matrix as it is symmetric. Considering,
, t[B]T [D][B] J , the element stiffness matrix will become after numerical integration as
n n
i , j
k w i w j (3.9.6)
i1 j1
Where, L terms are the triangular area coordinates and the wi terms are the weights associated with
those coordinates. The locations of integration points are shown in Fig. 3.9.2.
65
The sampling points and their associated weights are described below:
For sampling point =1 (Linear triangle)
1
w1 1 L11 L12 L13 (3.9.9)
3
For sampling points =3 (Quadratic triangle)
1 1
w1 L11 L12 , L13 0
3 2
1 1
w2 L21 0, L22 L23 (3.9.10)
3 2
1 1 1
w3 L31 , L32 0, L33
3 2 2
For sampling point = 7 (Cubic triangle)
27 1
w1 L11 L12 L13
60 3
8 1
w2 L21 L22 , L23 0
60 2
8 1
w3 L31 0, L32 L23
60 2
8 1
w4 L41 L43 , L42 0 (3.9.11)
60 2
3
w5 L51 1, L52 L53 0
60
3
w6 L61 L63 0, L62 1
60
3
w7 L71 L72 0, L73 1
60
66
i1
A (3.9.12)
Where, L terms are the volume coordinates and the wi terms are the weights associated with those
coordinates. The locations of Gauss points are shown in Fig. 3.9.3.
The sampling points and their associated weights are described below:
For sampling point = 1 (Linear tetrahedron)
1
w1 1 L11 L12 L13 L14 (3.9.13)
4
For sampling points = 4 (Quadratic tetrahedron)
1
w1 L11 0.5854102, L12 L13 L14 0.1381966
4
1
w2 L22 0.5854102, L21 L23 L24 0.1381966
4
(3.9.14)
1
w3 L33 0.5854102, L31 L32 L34 0.1381966
4
1
w4 L44 0.5854102, L41 L42 L43 0.1381966
4
For sampling points = 5 (Cubic tetrahedron)
67
4 1
w1 L11 L12 L13 L14
5 4
9 1 1
w2 L21 , L22 L23 L24
20 3 6
9 1 1
w3 L32 , L31 L33 L34 (3.9.15)
20 3 6
9 1 1
w4 L43 , L41 L42 L44
20 3 6
9 1 1
w5 L54 , L51 L52 L43
20 3 6
68
1 𝑥1 𝑦1 1 2 3
1 1 1 8
A = 2 �1 𝑥2 𝑦2 � = �1 5 4� =
2 2
[(30-12) - (12-9) + (8-15)] = 2 = 4
1 𝑥3 𝑦3 1 3 6
1 𝑥 𝑦 1 3 4
1 1 1 4
𝐴1 = 2
�1 𝑥2 𝑦2 �= �1 5
2
4� = 2 [(30-12) - (18-12) + (12-20)] = 2 = 2
1 𝑥3 𝑦3 1 3 6
1 𝑥 𝑦 1 3 4
1 1 1 2
𝐴2 = 2
�1 𝑥3 𝑦3 �= �1 3
2
6� = 2 [(9-12) - (9-8) + (18-12)] = 2 = 1
1 𝑥1 𝑦1 1 2 3
69
1 𝑥 𝑦 1 3 4
1 1 1 2
𝐴3 = 2
�1 𝑥1 𝑦1 �= �1 2
2
3� = 2 [(8-15) - (12-20) + (9-8)] = 2 = 1
1 𝑥2 𝑦2 1 5 4
𝐴1 2
𝑁1 = = 4 = 0.5
𝐴
𝐴2 1
𝑁2 = = 4 = 0.25
𝐴
𝐴3 1
𝑁3 = = 4 = 0.25
𝐴
u = 𝑁1 𝑢1 +𝑁2 𝑢2 +𝑁3 𝑢3
= 0.5 x 11 + 0.25 x 14 + 0.25 x 17 = 13.25 mm
70
The procedure for four node triangular element is the same as five node triangular element to derive
its interpolation functions. Here, node 5 and 6 are omitted and therefore displacements in these
nodes can be expressed in terms of the displacements at their corner nodes. Hence,
u 2 u3 u1 u 3
u '5 an d u '6
2 2 (3.11.1)
Substituting the values of u’5 and u’6 in eq.(3.3.8), the following relations can be obtained.
u 2 u 3 u 3 u1
u N1u1 N 2 u 2 N 3 u 3 N 4 u 4 N 5 N6
2 2
(3.11.2)
N N
N N6
N1 6 u1 N 2 5 u 2 N 3 5 u N4u 4
2
2 2 3
Now, the displacement at any point inside the four node element can be expressed by its nodal
displacement with help of shape function.
u N1u1 N 2 u 2 N 3 u 3 N 4 u 4 (3.11.3)
Comparing eq. (3.11.2) and eq. (3.11.3), one can find the following relations.
N6 4L L
N1 N1 L1 2L1 1 3 1 L1 1 2L 2
2 2
N 4L L
N 2 N 2 5 L 2 2L 2 1 2 3 L 2 1 2L1
2 2 (3.11.4)
N N6 4L L 4L3 L1
N 3 N 3 5 L3 2L3 1 2 3 L3
2 2
N 4 N 4 4L1L 2
Thus, the shape functions for the four node triangular element are
71
N1 L1 1 2L 2
N 2 L 2 1 2L1
(3.11.5)
N 3 L3
N 4 4L1L 2
Evaluate the integral: I x 11x 32 dx using one, two and three point gauss Quadrature.
2
2
Solution:
The existing limits of integration should be changed from [-2, +3] to [-1, +1]. Assuming, a +bx ,
the upper and lower limit can be changed. i.e., at x1 2, 1 1 and at x2 3, 2 1 . Thus,
putting these limits and solving for a & b, we get a = -0.2 and b = 0.4. The relation between two
coordinate systems will become:
5 1
0.2 0.4x or x and dx 2.5d
2
1
2
I x 2 11x 32 dx
2
x3 11x 2
3
32 x
3
2 2
99 8
9 96 22 64
2 3
37.5 83.33333 120.83333
Thus, Iexact = -120.83333
72
I
d w1
1
1
5 0 12 5 0 1
I 3 0.8889 2.5 11 32
2
2
5 0.7746 12
0.5556 2.5 11 5 0.7746 1 32
2 2
5 0.7746 12
0.5556 2.5 11 5 0.7746 1 32
2 2
I 3 0.8889 2.50.25 5.5 32
0.5556 2.55.9365 26.8015 32
0.5556 2.5 2.0635 15.8015 32
2.523.3336 0.4100 25.4120
2.5 48.3356 120.839
Thus, % of error = (120.83333-120.839)×100/120.83333 = 4.69×10-03. However, difference of
results will approach to zero, if few more digits after decimal points are taken in calculation.
Solution:
Using two point gauss Quadrature formula for the evaluation of three dimensional integration, we
have the following sampling points and weights.
w1 w 2 1
1 0.5773502692
2 0.5773502692
1 0.5773502692
2 0.5773502692
1 0.5773502692
2 0.5773502692
1 , 1 , 1 160.8886
1 , 1 , 2 0.8293
1 , 2 , 1 11.5513
1 , 2 , 2 0.0595
2 , 1 , 1 0.8293
2 , 1 , 2 0.0043
2 , 2 , 1 0.0595
2 , 2 , 2 0.0003
2 2 2
i , j, k
Now, I w i w j w k
i1 j1 k 1
1 , 1 , 1 w1w1w 2
Thus, I w1w1w1 1 , 1 , 2 w 2 w 2 w 2
2 , 2 , 2 = 174.222, where as
Iexact = 174.222.
1
4.1.1 Introduction
Analysis of frame structures can be carried out by the approach of stiffness method.
However, such types of structures can also be analyzed by finite element method. A unified
formulation will be demonstrated based on finite element concept in this module for the
analysis of frame like structures. A truss structure is composed of slender members pin
jointed together at their end points. Truss element can resist only axial forces (tension or
compression) and can deform only in its axial direction. Therefore, in case of a planar truss,
each node has components of displacements parallel to X and Y axis. Planar trusses lie in a
single plane and are used to support roofs and bridges. Such members will not be able to
carry transverse load or bending moment. The major benefits of use of truss structures are:
lightweight, reconstructable, reconfigurable and mobile. Configuration of few standard truss
structures are shown in Fig. 4.1.1.
α
u ( x ) =α 0 + α1 x =[1 x ] 0
α1 (4.1.1)
2
[N ] = 1 − x x
(4.1.3)
L L
So we get the element stiffness matrix as
=[k ] ∫∫∫ [ B ] [ D ][ B ] d Ω
T
(4.1.4)
Ω
d [N ] 1 1
Where, [B ] = = −
dx L L
So, the stiffness matrix will become:
1
− 1 1 AE 1 − 1
= ∫ [B ] E [B ]Adx = AE ∫ L −
L L
=
T
L − 1 1
dx
0 0
1 L
L
L
Thus, the stiffness matrix of the truss member along its member axis will be:
AE 1 − 1
[k ] =
L − 1 1
(4.1.5)
Now, let us find the stiffness matrix of a pin-jointed member of length L with respect to local
axis, having cross sectional areas A1 and A2 at the two ends of the member as shown in the
figure below.
From the above figure, the cross sectional area at a distance of x from left end can be
expressed as:
A − A1
Ax = A1 + 2 x (4.1.6)
L
As it is a pin-jointed member, the displacement at any point may be expressed in terms of
nodal displacement as= u N1u1 + N 2u2 .
Similarly the cross sectional area at any point may be represented in terms of the cross
sectional area of the two ends. Thus Ax = N 1 A1 + N 2 A2
x x
Where the shape functions are: N1 =1 − ; N2 =
L L
Now, the strain may be written as:
∂u ∂N1 ∂N 2 1 1 u1
εx = u 2 = − u1 + u 2 = 1 [− 1
= u1 + 1] = [B ]{u} (4.1.7)
∂x ∂x ∂x L L L
u 2
As the stress is proportional to strain according to Hook’s law, the stress-strain relationship
will be as follows:
u1
σ x = Eε x = [− 1 1] = E [B ]{u}
E
(4.1.8)
L
u 2
L L
U = ∫ ε x σ x dv = ∫ ε x Eε x Ax dx = ∫ {u} [B ] E [B ]{u}Ax dx
1 T 1 T 1 T T
(4.1.9)
2V 20 20
T u1
L
∂U
L
∫ [− 1 1] [− 1 1]Ax dx = [k ]{d }
E
{F } = = ∫ [B ] E [B ]{u}Ax dx = 2
T
∂{u} 0 L 0 u 2
(4.1.10)
1 − 1 L
E 1 − 1 A − A1 2
L
[k ] = E A1 + A2 − A1
∫
x dx = 2 A1 x + 2 x
L2 − 1 1 0 L L − 1 1 2L 0
E 1 − 1 A2 − A1 1 − 1
= + =
E
( + )
L − 1 1 − 1 1
A A A
2 2 L
1 1 2
(4.1.11)
Where, Fx1 and Fx 2 are the axial forces along the member axis X . Similarly, Fy1 and Fy 2 are
the forces perpendicular to the member axis X .
The relationship expressed in eq. (4.1.12) can be rewritten in matrix form as follows:
{F } = [T ]{F } (4.1.14)
Here, [T] is called transformation matrix. This relates between the global (𝑋, 𝑌 axis) and
member axis (𝑋�, 𝑌� axis). Similarly, the relations of nodal displacements between two
coordinate systems may be written as:
{d } = [T ]{d } (4.1.15)
Again, the equation stated in (4.1.5) can be generalized and expressed with respect to the
member axis including force and displacement vector as:
Fx1 1 0 −1 0 u1
Fy1 AE 0 0 0 0 v1
= (4.1.16)
Fx 2 L −1 0 1 0 u2
Fy 2
0 0 0 0 v2
Where, the nodal forces in Y direction are zero. The above equation may also be expressed in
short as:
{F } = [k ]{d } (4.1.17)
Where, the matrices in the above equation are written with respect to the member axis. Now,
eq. (4.1.17) can be rewritten with the use of eq. (4.1.14) and (4.1.15) as given below.
Here, the transformation matrix [T] is orthogonal, i.e., [T]-1 is equal to [T]T. Therefore, from
the above relationship, the generalized stiffness matrix can be expressed as:
[k ] = [T ]T [k ][T ] (4.1.20)
Thus,
(4.1.21)
Or,
The above stiffness matrix can be used for the analysis of two-dimensional truss problems.
7
Here, the displacement function using Pascal’s triangle can be expressed as:
α 0
u ( x ) =α 0 + α1 x + α 2 x = 1 x x α1
2 2
(4.2.1)
α
2
d [ N ] 3 4x 4 8x 1 4x
Where, [ B ] = =− + − − +
d x L L2 L L2 L L2
3 4𝑥
⎧− + 2 ⎫
𝐿 𝐿 ⎪
𝐿⎪
4 8𝑥 3 4𝑥 4 8𝑥 1 4𝑥
= 𝐴𝐸 � − 2 × �− + 2 − − + 2 � 𝑑𝑥
0 ⎨ 𝐿 𝐿 ⎬ 𝐿 𝐿 𝐿 𝐿2 𝐿 𝐿
⎪ 1 4𝑥 ⎪
⎩− 𝐿 + 𝐿2 ⎭
(4.2.5)
After integrating the above equation, the stiffness matrix of the 3-node truss member will
become:
7 −8 1
AE
[𝑘] = �−8 16 −8� (4.2.6)
3L
1 −8 7
Analyze the truss shown below by finite element method. Assume the cross sectional area of
the inclined member as 1.5 times the area (A) of the horizontal and vertical members. Assume
modulus of elasticity is constant for all the members and is E.
Solution
The analysis of truss starts with the numbering of members and joints as shown below:
The member information for the truss is shown in Table 4.2.1. The member and node
numbers, modulus of elasticity, cross sectional areas are the necessary input data. From the
coordinate of the nodes of the respective members, the length of each member is computed.
Here, the angle θ has been calculated considering anticlockwise direction. The signs of the
direction cosines depend on the choice of numbering the nodal connectivity.
Now, let assume the coordinate of node 1 as (0, 0). The coordinate and restraint joint
information are given in Table 4.2.2. The integer 1 in the restraint list indicates the restraint
exists and 0 indicates the restraint at that particular direction does not exist. Thus, in node no.
2, the integer 0 in x and y indicates that the joint is free in x and y directions.
The stiffness matrices of each individual member can be found out from the stiffness matrix
equation as shown below.
Thus the local stiffness matrices of each member are calculated based on their individual
member properties and orientations and written below.
1 2 3 4 3 4 5 6
0 0 0 0 1 1 −1 −1 1 3
0 1 0 − 1 −1 1 1 − 1 4
AE
2 3 AE
[k ]1 = [k ]2 =
L 0 0 0 0 3
4 2 L − 1 1 1 − 1 5
0 −1 0 1 4 1 −1 −1 1 6
and
5 6 1 2
1 0 −1 0 5
0 0 0 − 1 6
AE
[k ]3 =
L − 1 0 1 0 1
0 0 0 0 2
Global stiffness matrix can be formed by assembling the local stiffness matrices into globally.
Thus the global stiffness matrix are calculated from the above relations and obtained as
follows:
1 2 3 4 5 6
1
6
11
1 0 0 0 −1 0
0 1
0 −1 0 0
3 3 3 3
0 0 − −
4 2 4 2 4 2 4 2
3
[K ] = AE 0 − 1 − 3
1+
3 3
−
L 4 2 4 2 4 2 4 2
− 1 0 − 3 3 3 3
+1 −
4 2 4 2 4 2 4 2
3 3 3 3
0 0 − −
4 2 4 2 4 2 4 2
Fx1 0
F 0
y1
Fx 2 2 P
{F } =
The equivalent load vector for the given truss can be written as:=
Fy 2 P
Fx 3 0
Fy 3 0
Let us assume that u and v are the horizontal and vertical displacements respectively at joints.
Thus the displacement vector will be expressed as follows:
u1 0
v 0
1
u u
{d } = 2 = 2
v 2 v 2
u 3 0
v3 0
Therefore, the relationship between the force and the displacement will be:
12
1 0 0 0 −1 0
0 1 0 −1 0 0
Fx1 3 3 3 3 0
F 0 0 − − 0
y1 4 2 4 2 4 2 4 2
2 P AE 3 3 3 3 u2
= 0 −1 − 1+ −
P L
4 2 4 2 4 2 4 2 v2
Fx 3 3 3 3 3 0
−1 0 − +1 −
Fy 3 4 2 4 2 4 2 4 2 0
3 3 3 3
0 0 − −
4 2 4 2 4 2 4 2
From the above relation, the unknown displacements u2 and v2 can be found out through
computer programming. However, as numbers of unknown displacements in this case are
only two, the solution can be obtained by manual calculations. The above equation may be
rearranged with respect to unknown and known displacements in the following form:
Fα kαα kαβ dα
=
Fβ kβα kββ d β
Thus the developed matrices for the truss problem can be rearranged as:
2P 3 −3 −3
3
0 0
4 2
4 2 4 2 4 2 u2
P − 3 3 3 −3
1+ 0 −1 v2
4 2 4 2 4 2 4 2
Fx1 AE 0
= −1 0
L
0 0 1 0
Fy1 0 −1 0 1 0 0 0
−3 3 3 3
−1 0 +1 −
Fx3 4 2 4 2 4 2 4 2 0
3 −3 3 3
0 0 −
4 2 4 2
Fy3 0
4 2 4 2
.
The above relation may be condensed into following
3 −3
2 P AE 4 2 4 2 u
= 2
P L −3 1+
3 v2
4 2 4 2
13
The unknown displacements can be derived from the relationships expressed in the above
equation.
−1
3 −3 3 3
1+
u2 AE 4 2 4 2 2 P 4 2 L 4 2 4 2 2 P
=
v2 L −3 3 P 3 AE 3 3 P
4 2 1+
4 2 4 2
4 2
Thus the unknown displacement at node 2 of the truss structure will become:
8 2
u2 PL 3 +
= 3
v2 AE 3
Support Reactions:
The support reactions {Ps} can be determined from the following relation:
{Ps } =
− { Pcs } + K βα {dα }
Where, {Pcs} correspond to equivalent loadings at supports. Thus, the support reaction of the
present truss structure will be:
0 0
0 0
0 −1
0 8 2 − 3P
{Ps } = AE −3
− +
3 PL 3 + =
AE 3 − 2 P
0 L 4 2 4 2
3
0 3 −3 2P
4 2 4 2
Now, the member end actions can be obtained from the corresponding member stiffness and
the nodal displacements. The member end forces are derived as shown below.
Member –1
0
Fmx1 0 0 0 0 0
0 1 0 −1 0
Fmy1 AE PL −3P
= 8 2
Fmx 2 L 0 0 0 0 3 + AE 0
3
Fmy 2 0 −1 0 1
3
3P
14
Member – 2
8 2
Fmx 2 1 −1 −1 1 3 + 2P
F −1 1 1 −1 3 −2 P
my 2 3 AE 3 PL
=
Fmx 3 4 2 L −1 1 1 −1 AE −2 P
Fmy 3 0 2 P
1 −1 −1 1
0
Member –3
Fmx 3 1 0 − 1 0 0 0
F
my 3 AE 0 0 0 0 0 PL 0
= =
Fmx1 L − 1 0 1 0 0 AE 0
0
Fmy1 0 0 0 0 0
Thus the member forces in all members of the truss will be:
3P 3P
2
{ m} ( ) ( ) =
2
F = − 2 P + 2 P −2 2 P
0
0
The reaction forces at the supports of the truss structure will be:
0
−3P
{FR } =
−2 P
2 P
4.3.1 Introduction
A beam is a structural member which is capable of withstanding load primarily by resisting
bending. The primary tool for analysis of beam is the Euler–Bernoulli beam equation. Other
methods for determining the deflection of beams include "slope deflection method" and
"method of virtual work". For calculation of internal forces of beam include "moment
distribution method", force or flexibility method and stiffness method. However, all these
methods have limitations if either of geometry, loading, material properties or boundary
conditions becomes arbitrary in nature. Finite element techniques can well handle such cases
and relieve the analyzer of making simplifications to arrive approximate solutions.
Now, applying boundary conditions, the following expressions from the above relations can
be obtained:
At x=0:
α 0 α 0
α α
V1 = [1 0 0 0] 1 ; θ1 = [ 0 1 0 0] 1 ;
α 2 α 2
α 3
α 3
At x=L:
α 0 α 0
α α
3 1 2 1
V2 = 1 L L2 L θ 2 = 0 1 2 L 3L
;
α 2 α 2
α 3 α 3
So,
−1
α 0 1 0
0 0 V1 1 0 0 0
V1
α 0 θ 0 1 0 0
1 1 0
0 2 3 1 θ 1
= = − −
2 3
−
(4.3.4)
α 2 1 L
L2 L3 V2 L2 L L2 L V2
α 3 0 2
1 2 L 3L θ 2 2 1 2
− 3
1 θ
2
3
L L2 L L2
Therefore,
1 0 0 0
V1 V1
0
3 1 0 0 θ
1 θ1
[
v(x ) = 1 x x2 ]
x 3 − 2 −
2 3
− = [N 1 N2 N3
N 4 ] 1 (4.3.5)
l l l2 l V2 V2
2 1 2 1 θ θ 2
− 3 2
l 3 l2 l l 2
Where,
3 2 2 3 2 2 x3 3x 2 2 x 3 x2 x3
N1 = 1 − x + x ; N 2 = x − x + ; N 3 = − and N 4 = − +
L2 L3 L L2 L2 L3 L L2
(4.3.6)
17
N is called shape function which interpolates the beam displacement in terms of its nodal
displacements.
1 0 0 0 V1
0 0 θ
Where, [ B ] [ 0=
1 0 2
= 0 2 6 x ] ; [ A] = ; {d }
1 L L2 L3 V2
θ 2
0 1 2 L 3L2
Thus,
∂U
[ ] [B] [B][A ]{d }dx
L
{F } = = EI ∫ A −1
T T −1
(4.3.10)
∂{d } 0
0 0 0 0 0 0 0 0 0
L L
0
∫[ ]
0 0 0
L
0 0 0
B T [B ]dx = ∫ [0 0 2 6 x ]dx = ∫
0 0
Now, dx =
0 0
2 0 0 4 12 x 0 0 4L 6 L2
0
6 x 0 12 x 36 x 2
0 0 0 6 L2 12 L3
(4.3.12)
18
So,
0 0 0 0
0 0 −1
[ ]
0 0
[k ] = EI A −1 T [A]
0 0 4L 6 L2
0 3
0 6 L2 12 L
3 2
1 0 − 0 1 0 0 0
L2 L3 0 0 0
1
0 1 −
2
0 0 0 0 0 1 0 0
[k ] = EI L L2 3 2 3 1
2 − − −
0 − 3 0
3 2 0 4L 6 L L2
0 L L2 L
L2 L 2
1 2 1
1 1 0 0 6 L2 12 L3 3 − 3
0 0 − L L2 L L2
L L2
0 12 6 12 6
0 0 0 6 1 0 0 L3 L2
6
L2 L3
0 6
0 0 1 0 4 6
0 0 −2 2 − 2 − 2
= EI 3 2 3 1 = EI L L L L
0 0 0 − 6 − L2 −
L L2 L
− 12 6
− 2
12
− 2
6
2 L3 L3
1 2 1 6
L L
0 0 2 6l 3 − 3 2 6 4
L L2 L L2 − 2
L2 L L L
12 6 L −12 6 L
6 L 4 L2 −6 L 2 L2
EI
[ k ] = 3 −12 −6 L 12 −6 L (4.3.13)
L
2
6 L 2 L −6 L 4 L
2
Where, Fx1 and Fx 2 are the axial forces along the member axis X . Similarly, Fy1 and Fy 2 are
the forces perpendicular to the member axis X . M 1 and M 2 are the moment about its axis at
node 1 and 2 respectively.
The relationship expressed in eq. (4.3.14) can be rewritten in matrix form as follows:
{F } = [T ]{F } (4.3.16)
20
Similarly, the displacement vector in local coordinate system (𝑋�, 𝑌�) may be transformed to
global (𝑋, 𝑌) coordinate system by the following relation.
{d } = [T ]{d } (4.3.17)
0 0 0 0 0 0
6 EI
Fx1 0 12 EI 6 EI 12 EI
0 − 3 u1
L3 L2 L L2
Fy1 6 EI 4 EI 6 EI 2 EI v1
0 0 − 2
M1 L2 L L L θ1
= (4.3.18)
Fx2 0 0 0 0 0 0 u
2
12 EI 6 EI 12 EI 6 EI v
Fy2 0 − 3 − 0 − 2 2
M L L2 L3 L
θ
2 6 EI 2 EI 6 EI 4 EI 2
0 0 − 2
L2 L L L
The matrices in the above equation are written with respect to the member axis. Now, the eq.
(4.3.18) can be rewritten as follows with the use of eqs. (4.3.16) and (4.3.17).
Here, the transformation matrix [T] is orthogonal. Thus, from the above relationship, the
generalized stiffness matrix can be expressed as:
Considering λ = cos θ and µ = sin θ the above expression can be written as follows:
21
0 0 0 0 0 0
12 6 12 6
λ −µ 0 0 0 − 3 λ µ 0
L2
0 0 0 0 0 0
µ L3 L2 L
λ 0 0 0 0 6 4 6 2 −µ λ 0 0 0 0
0 0 − 2
0 0 1 0 0 0 L2 L 0 0 1 0 0 0
[ k ] = EI
L L
0 0 0 λ − µ 0 0 0 0 0 0 0 0 0 0 λ µ 0
0 0 0 µ λ 0 12 6 12 6 0 0 0 −µ λ 0
0 − 3 − 2 0 − 2
0 0 0 0 0 1 L L L3 L 0 0 0 0 0 1
6 2 6 4
0 0 − 2
L2 L L L
(4.3.22)
12 µ 2 12 µλ 6µ 12 µ 2 12 µλ 6µ
− − − −
L3
L3 L2 L3 L3 L2
12 µλ 12λ 2 6λ 12 µλ 12λ 2 6λ
− L3 L3 L2 L3
− 3
L L2
− 6µ 6λ 4 6µ 6λ
− 2
2
2
L2 L2 L
[ k ] = EI L 2 L L
(4.3.23)
− 12 µ 12 µλ 6µ 12 µ 2 12 µλ
− 3
6µ
L3 L3 L2 L3 L L2
12 µλ 12λ 2
− 3
6λ
− 2
12 µλ
− 3
12λ 2 6λ
− 2
L3 L L L L3 L
6µ 6λ 2 6µ 6λ 4
− 2 − 2
L L2 L L2 L L
22
(4.4.1)
The loading function for the present case can be written as:
w2 − w1
w ( x=
) w1 + x
L (4.4.2)
From eqs. (4.4.1) and (4.4.2), the equvalent nodal load will become
L 2 x 3 3x 2 7 w 3w
∫ 3 − 2 + 1 w( x )dx 1 + 2 L
F1 0L 3
L L 20 20
2
1 2 2
M 1 ∫ L2 − L + x w( x )dx 20 + 30 L
x 2 x w w
{Q} = = 0L
= 3w 7 w (4.4.3)
1 + 2 L
3 2
F2 − 2 x 3 x
( )
∫0 L3
+ w x dx
L2 20 20
M 2 L 3 w w 2
x x
2
∫ 2 − w( x )dx − 30 − 20 L
1 2
0 L L
23
{Q} = 12 (4.4.4)
wL
2
2
− wL
12
Here, [N*] is obtained by evaluating [N] at point where the concentrated load F
is applied. Thus,
2 x3 3x 2 2a 3 3a 2
3 − 2 + 1 3 − 2 + 1
L3 L L3 L
x 2 x2 a 2a 2
L2 − L + x L2 − L + a
[ N *] = 3 2
at dista nce a 3 2 (4.4.6)
− 2 x + 3x − 2a + 3a
L3 L2 L3 L2
3 2 3 2
x −x a −a
L2 L L2 L
24
2a 3 3a 2
3 − 2 + 1
F1 3
L L
a 2 a 2
M 1 L2 − L + a
Therefore, {Q} = =
(4.4.7)
F2 − 2 a 3
3 a 2
+ 2
L3 L
M 2
a − a
3 2
L2 L
Now, if load F is acting at midspan (i.e., a=L/2), then equivalent nodal load will be
F
2
FL
{Q} = 8 (4.4.8)
F
2
FL
−
8
With the above approach, the equivalent nodal load can be found for various loading function
acting on beam members.
Solution
Step 1: Numbering of Nodes and Members
The analysis of beam starts with the numbering of members and joints as shown below:
The member AB and BC are designated as (1) and (2). The points A,B,C are designated
by nodes 1, 2 and 4. The member information for beam is shown in tabulated form as shown
in Table 4.4.1. The coordinate of node 1 is assumed as (0, 0). The coordinate and restraint
joint information are shown in Table 4.4.2. The integer 1 in the restraint list indicates the
restraint exists and 0 indicates the restraint at that particular direction does not exist. Thus, in
node no. 2, the integer 0 in rotation indicates that the joint is free rotation.
1 1 2 EI
2 2 3 2EI
1 2 3 4
12 EI 6 EI 12 EI 6 EI
L3 − 1
L2 L3 L2
6 EI 4 EI 6 EI 2 EI
− 2 2
[k ]1 = 12
L2 L L L
− EI 6 EI
− 2
12 EI
− 2
6 EI 3
L3 L L3 L
6 EI 2 EI 6 EI 4 EI
− 2 4
L2 L L L
3 4 5 6
24 EI 12 EI 24 EI 12 EI
L3 − 3
L2 L3 L2
12 EI 8 EI 12 EI 4 EI
− 2 4
[k ]2 = L L
2
L L
− 24 EI 12 EI
− 2
24 EI
− 2
12 EI 5
L3 L L3 L
12 EI 4 EI 12 EI 8 EI
− 2 6
L2 L L L
1 2 3 4 5 6
12 EI 6 EI 12 EI 6 EI
L3 − 0 0 1
L2 L3 L2
6 EI 4 EI 6 EI 2 EI
2
− 2 0 0 2
L L L L
− 12 EI 6 EI 36 EI 6 EI 24 EI 12 EI
− 2 − 3
[K ]
= L
3
L L L2 L3 L2
6 EI 2 EI 6 EI 12 EI 12 EI 4 EI
2
− 2 4
L L L2 L L L
0 − 2
24 EI 12 EI 24 EI 12 EI
0 − 3 − 2 5
L L L3 L
12 EI 4 EI 12 EI 8EI
0 0 − 2 6
L2 L L L
27
0
0
0
{d } =
θ 2
0
θ 3
P
−
2
PL
−
8
P wL
− 2 + 2
{F } = PL wL2
−
8 12
wL
−
2
wL2
12
{ F } = [ K ] {d }
{d } = [ K ] { F }
−1
12 EI
−1 P
6 EI 12 EI
− 3
6 EI
0 0 −
L3 L2 L L2
2
−
PL
0 6 EI 4 EI
− 2
6 EI 2 EI
0 0
0 L2 L L L 8
12 EI 12 EI P wL
− + 2
6 EI 36 EI 6 EI 24 EI
0 − L3 − 2
L L L2
−
L3 L2 × 2
= 6 EI 4 EI PL wL2
θ 2
2 EI 6 EI 12 EI 12 EI
− 2 −
0 L 8 12
2
L L2 L L L
0 24 EI 12 EI 24 EI 12 EI
θ 3 0 − 3 − 2 − 2 −
wL
L L L3 L
12 EI 4 EI 12 EI 8EI 2
2
0 0 − 2 wL
L2 L L L 12
−1
12 EI 4 EI PL wL2 PL wL2
− −
θ 2 L L 8 12 L 2 −1 8
12
= × =
−1 3
θ3 4 EI 8 EI wL
2 2
wL 20 EI
L L 12 12
PL wL2
θ 2 L 4 − 4
= 2
θ 3 20 EI − PL + wL
8 3
PL2 wL3
θ2 = −
80 EI 80 EI
PL wL3
θ3 = − +
160 EI 60 EI
Member-(1)
29
12 EI 6 EI 3P 3wL
40 − 40
6 EI 12 EI
L3 −
L2 L3
2
PL wL2
L 0
F1 6 EI 2 EI
M
4 EI 6 EI
− 2 0 −
1 L L2 L L L × = 40 40
= 3P 3wL
6 EI 0
F2 20 EI − 3
12 EI 6 EI 12 EI
− 2 − 2 PL wL −
2 +
M 2 L L L3 L − 40 40
6 EI 2 EI 6 EI 4 EI 4 4 PL wL2
− 2 20 − 20
L2 L L L
Member-(2)
24 12 24 12 wL 3P
− +
L2 L L2 L 0 20 40
F2 12 12 PL wL2 3PL wL2
M − 4 − −
4 = 40
8
2 EI L
= L × 4 30
12 wL 3P
F3 L − 2
24 12 24 0
− − 2 − −
M 3 L L L2 L − PL + wL 20 40
12 8 3 wL2
8
12
4 −
L L 12
Support reactions:
R A 7.425
{FR } = RB = 17.5
R 6.075
C
31
4.5.1 Introduction
The plane frame is a combination of plane truss and two dimensional beam. All the members
lie in the same plane and are interconnected by rigid joints in case of plane frame. The
internal stress resultants at a cross-section of a plane frame member consist of axial force,
bending moment and shear force.
Therefore, the stiffness matrix of the frame in its local coordinate system will be the
combination of 2-d truss and 2-d beam matrices:
u1 v1 θ1 u2 v2 θ2
𝐴𝐸 𝐴𝐸
⎡ 𝐿 0 0 − 0 0 ⎤
𝐿
⎢ 0 12𝐸𝐼 6𝐸𝐼
0 −
12𝐸𝐼 6𝐸𝐼 ⎥
⎢ 𝐿3 𝐿2 𝐿3 𝐿2 ⎥
6𝐸𝐼 4𝐸𝐼 6𝐸𝐼 2𝐸𝐼
⎢ 0 0 − ⎥
���� 𝐿2 𝐿2
[𝑘]=⎢ 𝐴𝐸 𝐿
𝐴𝐸
𝐿 ⎥
(4.5.1)
⎢− 0 0 0 0 ⎥
⎢ 𝐿 12𝐸𝐼 6𝐸𝐼
𝐿
12𝐸𝐼 6𝐸𝐼 ⎥
⎢ 0 −
𝐿3
−
𝐿2
0
𝐿3
− 2⎥
𝐿
⎢ 6𝐸𝐼 2𝐸𝐼 6𝐸𝐼 4𝐸𝐼 ⎥
⎣ 0 𝐿2 𝐿
0 −
𝐿2 𝐿 ⎦
before formulating the global stiffness matrix by assembly. The generalized stiffness matrix
of a frame member can be obtained by transferring the matrix of local coordinate system into
its global coordinate system. The transformation matrix can be expressed as:
cos 𝜃 sin 𝜃 0 0 0 0
⎡− sin 𝜃 cos 𝜃 0 0 0 0⎤
⎢ ⎥
0 0 1 0 0 0⎥
[T]=⎢ (4.5.2)
⎢ 0 0 0 cos 𝜃 sin 𝜃 0⎥
⎢ 0 0 0 − sin 𝜃 cos 𝜃 0⎥
⎣ 0 0 0 0 0 1⎦
Now, the generalized stiffness matrix of the member can be obtained from the relation of
� ][𝑇] . Thus considering 𝜆 = cos 𝜃 and 𝜇 = sin 𝜃 the stiffness matrix in global
[𝐾] = [𝑇]𝑇 [𝐾
coordinate system can be written as follows:
AE AE
⎡ 0 0 − 0 0 ⎤
⎢ L L ⎥
⎢ 0 12EI 6EI 12EI 6EI ⎥
0 − 3
λ −µ 0 0 0 0 ⎢ L3 L2 L L2 ⎥
⎡µ λ 0 0 0 0⎤ ⎢ 6EI 4EI 6EI 2EI ⎥
⎢ ⎥ ⎢ 0 0 − 2
[K] = EI ⎢0 0 1 0 0 0⎥ × ⎢ L 2 L L L ⎥
⎢0 0 0 λ −µ 0⎥ − AE AE ⎥
⎢0 0 0 µ λ 0⎥ ⎢ 0 0 0 0 ⎥
⎢ L L ⎥
⎣0 0 0 0 0 1⎦ 12EI 6EI 12EI 6EI
⎢ 0 − 3 − 2 0 − 2⎥
⎢ L L L3 L ⎥
⎢ 6EI 2EI 6EI 4EI ⎥
⎣ 0 L 2 L
0 − 2
L L ⎦
λ µ 0 0 0 0
⎡−µ λ 0 0 0 0⎤
⎢ ⎥
⎢ 0 0 1 0 0 0⎥
×
⎢ 0 0 0 λ µ 0⎥
⎢ 0 0 0 −µ λ 0⎥
⎣ 0 0 0 0 0 1⎦
(4.5.3)
Solution
The members AB and BC are designated as (1) and (2). The points A, B and C are designated
by nodes 1, 2 and 3. The member information for the frame is shown in tabulated form as
shown in Table 1(a). The coordinate of node 1 is assumed as (0,0). The coordinate and
restraint joint information are shown in Table 1(b). The integer 1 in the restraint list indicates
34
the restraint exists and 0 indicates the restraint at that particular direction does not exist.
Thus, in node no. 2, the integer 0 all the restraint type indicates that the joint is in free all the
three directions.
1 2 3 4 5 6
12EI 6EI 12EI 6EI
⎡ (1.5L)3 0 − − 0 − 1
(1.5L)2 (1.5L)3 (1.5L)2 ⎤
⎢ AE AE ⎥
⎢ 0 0 0 − 0 2
(1.5L) (1.5L) ⎥
⎢ 6EI 4EI 6EI 2EI ⎥
−
⎢ (1.5L)2 0 0 ⎥ 3
(1.5L) (1.5L)2 1.5L
[𝑘]1 =⎢ 12EI 6EI 12EI 6EI ⎥
⎢− (1.5L)3 0
(1.5L)2 (1.5L)3
0
(1.5L)2 ⎥ 4
⎢ AE AE ⎥
⎢ 0 −
(1.5L)
0 0
(1.5L)
0 ⎥ 5
⎢ 6EI 2EI 6EI 4EI ⎥
⎣− (1.5L)2 0
(1.5L) (1.5L)2
0
(1.5L) ⎦ 6
4 5 6 7 8 9
35
A(1.5 E) A(1.5 E)
⎡ L
0 0 − L
0 0 ⎤ 4
⎢ 0
12(1.5 E)I 6(1.5 E)I
0 −
12(1.5 E)I 6(1.5 E)I ⎥
5
⎢ L3 L2 L3 L2 ⎥
⎢ 6(1.5 E)I 4(1.5 E)I 6(1.5 E)I 2(1.5 E)I ⎥
0 2
0 − L2
6
[𝑘]2 = ⎢ A(1.5 E) L L
A(1.5 E)
L ⎥
⎢− 0 0 0 0 ⎥ 7
L L
⎢ 12(1.5 E)I 6(1.5 E)I 12(1.5 E)I 6(1.5 E)I
⎥
⎢ 0 − L3 − 0 − L2 ⎥ 8
L2 L3
⎢ 6(1.5 E)I 2(1.5 E)I 6(1.5 E)I 4(1.5 E)I ⎥ 9
⎣ 0 L2 L
0 − L2 L ⎦
The distributed load on member (2) can be replaced by its equivalent joint load as shown in
the figure below.
Member – (1)
In case of member (1), the member forces will be:{𝐹𝑚 }1 = [𝐾](1) {𝑑}(1)
𝐹𝑥
⎡ 1⎤ 56.17 0 −126.4 −56.17 0 −126.4
𝐹𝑦 ⎡ 0 7110 0 0 −7110 0 ⎤
⎢ 1⎥ ⎢ ⎥
⎢ 𝑀1 ⎥ −126.4 0 379.2 126.4 0 189.6 ⎥
⎢𝐹𝑥 2 ⎥ = 10 ⎢⎢−56.17
6
0 379.2 56.17 0 126.4 ⎥
⎢ ⎥ ⎢
𝐹
⎢ 𝑦 2⎥ 0 −7110 0 0 7110 0 ⎥
⎣−126.4 0 189.6 126.4 0 379.2 ⎦
⎣ 𝑀2 ⎦
0
⎡ 0 ⎤
⎢ ⎥
0
×⎢ −12
⎥
⎢ 4.327 𝑋 10 𝑤 ⎥
⎢−1.7127 𝑋 10−10 𝑤 ⎥
⎣−5.4978 𝑋 10−10 𝑤 ⎦
0.0697𝑤
⎡ 1.2177𝑤 ⎤
⎢ ⎥
−0.10479𝑤 ⎥
=⎢
⎢−0.06925𝑤 ⎥
⎢−1.21661𝑤 ⎥
⎣−0.20793𝑤 ⎦
It is to be noted that {Fm} are the end actions due to joint loads. Hence it must be added to the
corresponding end actions in the restrained structure in order to obtain the end actions due to
the loads. Therefore, {Fm}actual are the true member end actions due to actual loading system
can be expressed as
{𝐹𝑚 }𝑎𝑐𝑡𝑢𝑎𝑙 = {Fm} + {Ffm}
Where, {Ffm} are the end actions in the restrained structure. Since there is no load acting on
member (1), the actual end actions will be:
0.0697𝑤 0 0.0697𝑤
⎡ 1.2177𝑤 ⎤ ⎡0⎤ ⎡ 1.2177𝑤 ⎤
⎢ ⎥ ⎢ ⎥ ⎢ ⎥
{Fm}actual = ⎢−0.10479𝑤 ⎥ + ⎢0⎥ = ⎢−0.10479𝑤 ⎥
⎢−0.06925𝑤 ⎥ ⎢0⎥ ⎢−0.06925𝑤 ⎥
⎢−1.21661𝑤 ⎥ ⎢0⎥ ⎢−1.21661𝑤 ⎥
⎣−0.20793𝑤 ⎦ ⎣0⎦ ⎣−0.20793𝑤 ⎦
Member (2)
In similar way, the member forces in member (2) will be {Fm}(2) = [K](2){d}(2)
38
𝐹𝑥
⎡ 2⎤ 16 0 0 −16 0 0 4.327 𝑋 10−12 𝑤
⎡ 0 ⎤ ⎡ ⎤
⎢ 𝐹𝑦2 ⎥ 0.284 0.426 0 −0.284 0.426 −1.7127 𝑋 10−10 𝑤
⎢ 𝑀2 ⎥ ⎢ ⎥ ⎢ ⎥
9⎢ 0 0.426 0.853 0 −0.426 0.426 ⎥ ⎢−5.4978 𝑋 10−10 𝑤 ⎥
⎢𝐹𝑥 3 ⎥ = 10 ×
⎢ −16 0 0 16 0 0 ⎥ ⎢ 0 ⎥
⎢ ⎥ ⎢ ⎥
⎢𝐹𝑦 3 ⎥ 0 −0.284 −0.426 0 0.284 −0.426 ⎢ 0 ⎥
⎣ 0 0.426 0.426 0 −0.426 0.853 ⎦ ⎣ 0 ⎦
⎣ 𝑀3 ⎦
0.069232𝑤
⎡−0.28325𝑤 ⎤
⎢ ⎥
−0.54215𝑤 ⎥
=⎢
⎢−0.06923𝑤 ⎥
⎢ 0.283245𝑤 ⎥
⎣ −0.3076𝑤 ⎦
0.069232𝑤 0 0.0692𝑤
⎡−0.28325𝑤 ⎤ ⎡ 1.5𝑤 ⎤ ⎡ 1.2167𝑤 ⎤
⎢ ⎥ ⎢ ⎥ ⎢ ⎥
−0.54215𝑤 ⎥ ⎢ 0.75𝑤 ⎥ ⎢ 0.2078𝑤 ⎥
{Fm}actual = ⎢ + =
⎢−0.06923𝑤 ⎥ ⎢ 0 ⎥ ⎢−0.0692𝑤 ⎥
⎢ 0.283245𝑤 ⎥ ⎢ 1.5𝑤 ⎥ ⎢ 1.7832𝑤 ⎥
⎣ −0.3076𝑤 ⎦ ⎣−0.75𝑤 ⎦ ⎣−1.0576𝑤 ⎦
4.6.1 Introduction
The property of a grid member is basically a combination of 2-d beam with torsional effect.
The plane frame is assumed to be loaded in its own plane where as loading in the grid is
normal to its plane. As a result torsional effects are included in the grid analysis. Thus the
grid member can withstand bending moment, shear force as well as torsional moment.
Therefore, the stiffness matrix of the grid in its local coordinate system will be:
GI x GI x
L 0 0 − 0 0
Mx1
L
4 EI y 6 EI y 2 EI y 6 EI y
0 − 0 Mz1
L L 2
L L2
6 EI y 12 EI y 6 EI y 12 EI y
0 − 0 − − Fy1
[k ] = GI L2 L3
GI x
L2 L3 (4.6.1)
− x 0 0 0 0 Mx2
L L
2 EI y 6 EI y 4 EI y 6 EI y
0 − 2
0 Mz2
L L L L2
0 6 EI y 12 EI y 6 EI y 12 EI y
− 0 Fy2
L2 L3 L2 L3
cos 𝜃 sin 𝜃 0 0 0 0
⎡− sin 𝜃 cos 𝜃 0 0 0 0⎤
⎢ ⎥
0 0 1 0 0 0⎥
[T]=⎢
⎢ 0 0 0 cos 𝜃 sin 𝜃 0⎥
⎢ 0 0 0 − sin 𝜃 cos 𝜃 0⎥
⎣ 0 0 0 0 0 1⎦
40
Now, the generalized stiffness matrix of the member can be obtained from the relation
of [𝑘] = [𝑇]𝑇 �𝑘��[𝑇] . Thus considering 𝜆 = cos 𝜃 and 𝜇 = sin 𝜃 the stiffness matrix in
global coordinate system can be written as follows:
xGI GIx
⎡ L 0 0 − 0 0 ⎤
L
𝜆 −𝜇 0 0 0 0 ⎢ 0 4EIy
−
6EIy
0
2EIy 6EIy ⎥
⎡𝜇 ⎢ L2 ⎥
0⎤ ⎢
L L2 L
𝜆 0 0 0 6EIy 12EIy 6EIy 12EIy
⎢
0 0 1 0 0
⎥
0⎥ ⎢ 0 − L2 0 − − L3 ⎥
L3 L2 ⎥
[𝑘] = ⎢ × GIx GIx
⎢0 0 0 𝜆 −𝜇 0 ⎥ ⎢− 0 0 0 0 ⎥
⎢0 0 0 𝜇 𝜆 0⎥ ⎢ L 2EIy 6EIy
L
4EIy 6EIy ⎥
⎣0 0 0 0 0 1⎦ ⎢ 0 − 0 − L2 ⎥
L L2 L
⎢ 6EIy 12EIy 6EIy 12EIy ⎥
⎣ 0 L2
− L3
0 L2 L3 ⎦
𝜆 𝜇 0 0 0 0
⎡−𝜇 𝜆 0 0 0 0⎤
⎢ ⎥
0 0 1 0 0 0⎥
×⎢
⎢ 0 0 0 𝜆 𝜇 0⎥
⎢0 0 0 −𝜇 𝜆 0⎥
⎣0 0 0 0 0 1⎦
GI x 2 4 EI y 2 GI x 4 EI y 6 EI y GI x 2 4 EI y 2 GI 2 EI y 6 EI y
− x + λµ
λ + µ L − L λµ L2
µ − λ + µ −
L2
µ
L L L L L L
GI x 4 EI y GI x 2 4 EI y 2 6 EI y GI 2 EI y GI x 2 2 EI y 2 6 EI y
− λµ µ + λ − λ − x + λµ − µ + λ λ
L L L L L2 L L
L L L2
6 EI y 6 EI y 12 EI y 6 EI y 6 EI y 12 EI y
2
µ − 2 λ µ − 2 λ −
= L L L3 L2 L L3
GI x 2 4 EI y 2 GI x 2 EI y 6 EI y GI x 2 4 EI y 2 GI x 4 EI y 6 EI y
− L λ + L µ − + λµ µ λ + µ
L − L λµ
− 2 µ
L L L2 L L L
GI x 2 EI y GI 2 EI y 2 6 EI y GI x 4 EI y GI x 2 4 EI 6 EI y
+ λµ − x µ2 + λ − λ
L − L λµ
µ + λ2 λ
y
L2 L2
L L L L L L
6 EI y 6 EI y 12 EI y 6 EI y 6 EI y 12 EI y
− 2 µ λ − − 2 µ λ
L L2 L3 L L2 L3
(4.6.2)
Solution
The member AB and BC are designated as (1) and (2). The points A, B and C are designated
by nodes 1, 2 and 3. The member information for the grid is shown in tabulated form as
shown in Table 4.6.1. The coordinate of node 1 is assumed as (0, 0). The coordinate and
restraint joint information are shown in Table 4.6.2. The integer 1 in the restraint list indicates
the restraint exists and 0 indicates the restraint at that particular direction does not exist.
Thus, in node no. 2, the integer 0 all the restraint type indicates that the joint is free in all the
three directions.
1 2 3 4 5 6
GJ GJ 1
L 0 0 − 0 0
L
12 EI 6 EI 12 EI 6 EI
0 0 − 2
L3 L2 L3 L2
0 6 EI 4 EI 6 EI 2 EI
0 − 2 3
[k ]AB
= L2 L L L
GJ GJ
− 0 0 0 0 4
L L
0 12 EI 6 EI 12 EI 6 EI
− − 0 − 2 5
L3 L2 L3 L
6 EI 2 EI 6 EI 4 EI
0 0 − 2 6
L2 L L L
1 2 3 4 5 6
K K
L 0 0 − 0 0
1
L
36 K 18 K 36 K 18 K
0 0 − 2
L3 L2 L3 L2
0 18 K 12 K 18 K 6K
0 − 2 3
[k ]AB
= L2 L L L
K K
− 0 0 0 0 4
L L
0 − 2
36 K 18 K 36 K 18 K
− − 0
L3 L2 L3 L 5
18 K 6K 18 K 12 K
0 0 − 2
L2 L L L 6
43
As the member BC member is also horizontal, the value of Cos θ = 1 and Sin θ = 0 and thus,
the stiffness matrix will be:
6 5 4 7 8 9
K K
L 0 0 − 0 0
6
L
36 K 18 K 36 K 18 K
0 0 − 5
L3 L2 L3 L2
0 18 K 12 K 18 K 6K
0 − 2 4
[k ]BC = K L 2
L L L
K
− 0 0 0 0 7
L L
0 36 K 18 K 36 K 18 K
− 3 − 2 0 − 2
L L L3 L 8
18 K 6K 18 K 12 K
0
Step 3: Formation of global stiffness 0
matrix: − 2
L 2
L L 9 L
The global stiffness matrix can be obtained by assembling the local stiffness matrix of
members (AB) and (BC). Now looking at the grid structure, the displacements at the fixed
supports, are known and all are equal to zero. Only the displacement at co-ordinates 4, 5, 6
are unknown. So the global system stiffness matrix, corresponding to the displacement at co-
ordinate 4, 5, 6 will be:
𝐾 12𝐾 18𝐾
⎡𝐿𝐴𝐵 𝐿𝐵𝐶 + 0 ⎤
𝐿𝐵𝐶 2
⎢ 18𝐾 36𝐾 36𝐾 18𝐾 ⎥
[𝐾] = ⎢ 𝐿 2 + − 2 ⎥
𝐵𝐶 𝐿𝐴𝐵 3 𝐿𝐵𝐶 3 𝐿𝐴𝐵
⎢ 18𝐾 𝐾 12𝐾 ⎥
⎣ 0 − +
𝐿𝐴𝐵 2 𝐿𝐵𝐶 𝐿𝐴𝐵 ⎦
2.65 0.72 0
= K �0.72 0.8505 −1.125�
0 −1.125 3.2
0
0
0
d 4
{d } = d5
d
6
0
0
0
Here, d4, d5, and d6 indicate the displacement vectors at point B.
{d } = [ K ] { F } .
−1
Now, eliminating the rows in the force matrix, corresponding to zero
element in displacement matrix, the reduced matrix will be as follows.
0
𝑑4 −1 ⎡ 4𝑤 ⎤
2.65 0.72 0
⎢− ⎥
�𝑑5 � = k −1 �0.72 0.8505 −1.125� ⎢ 2 ⎥
𝑑6 0 −1.125 3.2 ⎢ 16𝑤 ⎥
⎣ 12 ⎦
0
1 0.662 −1.047 −0.368 −2𝑤
= �−1.047 3.856 1.3556 � � 4 �
𝑘
−0.368 1.355 0.789 𝑤
3
Member - AB
In case of member (AB), the member forces will be: {Fm}(AB) = [K](AB) {d}(AB)
1 −1
⎡ 0 0 0 0 ⎤
⎢ 4 4 ⎥
⎢ 0 36 18 36 18 ⎥
0 − 3
𝑇1 ⎢ 43 42 4 42 ⎥ 0
⎡𝐹 ⎤ 18 12 18 6 ⎥ ⎡ ⎤
1 ⎢ 0
⎢ ⎥ 0 0 − ⎢ ⎥
⎢
⎢𝑀1 ⎥ = 𝐾 ⎢ 42 4 42 4 ⎥× 1⎢ 0
⎥ 𝐾 ⎢ 1.603𝑤 ⎥⎥
⎢ 𝑇2 ⎥ 1 1
⎢ 𝐹2 ⎥ ⎢− 0 0 0 0 ⎥ ⎢−5.905𝑤 ⎥
⎢ 4 4 ⎥ ⎣−1.658𝑤 ⎦
⎣𝑀2 ⎦ 36 18 36 18
⎢ 0 − − 0 − ⎥
⎢ 43 42 43 42 ⎥
⎢ 18 6 18 12 ⎥
⎣ 0 4 2 4
0 − 2
4 4 ⎦
Thus,
𝑇1 −0.4𝑤
⎡𝐹 ⎤ ⎡
1 1.456𝑤 ⎤
⎢ ⎥ ⎢ ⎥
⎢𝑀1 ⎥ = ⎢ 4.156𝑤 ⎥
⎢ 𝑇2 ⎥ ⎢ 0.4𝑤 ⎥
⎢ 𝐹2 ⎥ ⎢−1.456𝑤 ⎥
⎣𝑀2 ⎦ ⎣ 1.70𝑤 ⎦
46
It is to be noted that {Fm} are the end actions due to joint loads. Hence it must be added to the
corresponding end actions in the restrained structure in order to obtain the end actions due to
the loads. Therefore, {Fm} Actual are the true member end actions due to actual loading system
and can be expressed as
{Fm} Actual = {Fm} + {Ffm}
Where, {Ffm}are the end actions in the restrained structure. Since there is no load acting on
member (1), the actual end action will be:
0
⎡ 4𝑤 ⎤
⎢ ⎥
−0.4𝑤 2
⎡ 1.456𝑤 ⎤ ⎢ ⎥ ⎡−0.4𝑤 ⎤
42 𝑤
⎢ ⎥ ⎢ ⎥ ⎢ 3.46𝑤 ⎥
4.156𝑤 ⎥ ⎢ 12 ⎥ ⎢ 5.49𝑤 ⎥
{𝐹𝑚 }𝐴𝑐𝑡𝑢𝑎𝑙 = ⎢ + =
⎢ 0.4𝑤 ⎥ ⎢ 0 ⎥ ⎢ 0.40𝑤 ⎥
⎢−1.456𝑤 ⎥ ⎢ 4𝑤 ⎥ ⎢ 0.54𝑤 ⎥
⎣ 1.70𝑤 ⎦ ⎢ 2 ⎥ ⎣ 0.34𝑤 ⎦
⎢ 42 𝑤 ⎥
⎣− 12 ⎦
Member - BC
In similar way, the member forces in member (BC) will be: {Fm}(BC) = [K](BC) {d}(BC)
1 −1
⎡ 0 0 0 0 ⎤
⎢ 5 4 ⎥
⎢ 0 36 18 36 18 ⎥
0 − 3
𝑇2 ⎢ 53 52 5 52 ⎥ −1.658𝑤
⎡𝐹 ⎤ 18 12 18 6 ⎥ ⎡−5.905𝑤 ⎤
2 ⎢
⎢ ⎥ ⎢ 0 0 − 2 ⎢ ⎥
⎢𝑀2 ⎥ = 𝐾 52 5 5 5 ⎥ × 1 ⎢ 1.603 ⎥
⎢ 𝑇3 ⎥ ⎢ 1 1 ⎥ 𝐾⎢ 0 ⎥
𝐹 ⎢ − 0 0 0 0 ⎥ ⎢ 0 ⎥
⎢ 3⎥ 5 5
⎢ ⎥ ⎣ ⎦
⎣𝑀3 ⎦ 0
⎢ 0 − 36 − 18 0 36
−
18⎥
⎢ 53 52 53 52 ⎥
⎢ 18 6 18 12 ⎥
⎣ 0 5 2 5
0 − 2
5 5 ⎦
Thus,
47
𝑇2 −0.33𝑤
⎡𝐹 ⎤ ⎡ ⎤
⎢ ⎥ ⎢−0.55𝑤 ⎥
2
⎢𝑀2 ⎥ = ⎢−0.40𝑤 ⎥
⎢ 𝑇3 ⎥ ⎢ 0.33𝑤 ⎥
⎢ 𝐹3 ⎥ ⎢ 0.55𝑤 ⎥
⎣𝑀3 ⎦ ⎣−2.33𝑤 ⎦
Since there is no load acting on member (BC), the actual end action will be:
−0.33𝑤 0 −0.33𝑤
⎡−0.55𝑤 ⎤ ⎡0⎤ ⎡−0.55𝑤 ⎤
⎢−0.40𝑤 ⎥ ⎢ ⎥ ⎢−0.40𝑤 ⎥
{𝐹𝑚 }𝐴𝑐𝑡𝑢𝑎𝑙 = ⎢ ⎥ + ⎢0⎥ = ⎢ ⎥
⎢ 0.33𝑤 ⎥ ⎢0⎥ ⎢ 0.33𝑤 ⎥
⎢ 0.55𝑤 ⎥ ⎢0⎥ ⎢ 0.55𝑤 ⎥
⎣−2.33𝑤 ⎦ ⎣0⎦ ⎣−2.33𝑤 ⎦
Thus, the reaction forces at the support and load at the joints will be:
𝑇1 −0.4𝑤
⎡𝐹 ⎤ ⎡ 3.46𝑤 ⎤
⎢𝑀1 ⎥ ⎢ 5.49𝑤 ⎥
⎢ 1⎥ ⎢ ⎥
⎢ 𝑇2 ⎥ ⎢ 0 ⎥
⎢ 𝐹2 ⎥ = ⎢ 0 ⎥
⎢𝑀2 ⎥ ⎢ 0 ⎥
⎢ 𝑇3 ⎥ ⎢ 0.33𝑤 ⎥
⎢ 𝐹3 ⎥ ⎢ 0.55𝑤 ⎥
⎣𝑀3 ⎦ ⎣−2.33𝑤 ⎦
1 2 3 4 5 6 7 8 9 10 11 12
EAX EAX
1 L 0 0 0 0 0 − 0 0 0 0 0
L
0 12 EI Z
0 0 0
6 EI Z
0 −
12 EI Z
0 0 0
6 EI Z
2 L3 L2 L3 L2
12 EIY 6 EIY 12 EIY 6 EIY
0 0 0 − 0 0 0 − 0 − 0
3 L3 L2 L3 L2
GI X GI X
0 0 0 0 0 0 0 0 − 0 0
4 L L
6 EIY 4 EIY 6 EIY 2 EIY
5 0 0 − 0 0 0 0 0 0
L2 L L2 L
6 EI Z 4 EI Z 6 EI Z 2 EI Z
6 0 0 0 0 0 − 0 0 0
L2 L L2 L
− EAX 0 0 0 0 0
EAX
0 0 0 0 0
7 L L
12 EI Z 6 EI Z 12 EI Z 6 EI Z
8 0 − 0 0 0 − 0 0 0 0 − 2
L3 L2 L3 L
12 EIY 6 EIY 12 EIY 6 EIY
9 0 0 − 0 0 0 0 0 0
L3 L2 L3 L2
GI X GI X
10 0 0 0 − 0 0 0 0 0 0 0
L L
6 EIY 2 EIY 6 EIY 4 EIY
11 0 0 − 0 0 0 0 0 0
L2 L L2 L
0 6 EI Z 2 EI Z 6 EI Z 4 EI Z
12 0 0 0 0 − 0 0 0
L2 L L L
The generalized stiffness matrix of a rigid jointed space frame member can be obtained by
transferring the matrix of local coordinate system into its global coordinate system. The
transformation matrix will become a square matrix of size 12×12 in this case as the degrees
of freedom for each node/joint is six.
Lecture 1: Constant Strain Triangle
The triangular elements with different numbers of nodes are used for solving two dimensional
solid members. The linear triangular element was the first type of element developed for the
finite element analysis of 2D solids. However, it is observed that the linear triangular element is
less accurate compared to linear quadrilateral elements. But the triangular element is still a very
useful element for its adaptivity to complex geometry. These are used if the geometry of the 2D
model is complex in nature. Constant strain triangle (CST) is the simplest element to develop
mathematically. In CST, strain inside the element has no variation (Ref. module 3, lecture 2) and
hence element size should be small enough to obtain accurate results. As indicated earlier, the
displacement is expressed in two orthogonal directions in case of 2D solid elements. Thus the
displacement field can be written as
u
d (5.1.1)
v
Here, u and v are the displacements parallel to x and y directions respectively.
u 1 u v
2 2
x
x 2 x x
v 1 u v
2 2
y
y 2 y y
(5.1.3)
v u u u v v
xy
x y x y x y
In case of small amplitude of displacement, one can ignore the nonlinear term of the above
equation and will reach the following expression.
∂u
εx =
∂x
∂v
εy = (5.1.4)
∂y
∂v ∂u
γ= +
∂x ∂y
xy
∂N ∂N 2 ∂N 3 u1
1 0 0 0 u
ε x ∂x ∂x ∂x 2
∂N1 ∂N 2 ∂N 3 u3
=ε =εy 0 0 0 (5.1.5)
γ ∂y ∂y ∂y v1
xy ∂N ∂N 2 ∂N 3 ∂N1 ∂N 2 ∂N 3 v2
1
∂y ∂y ∂y ∂x ∂x ∂x v3
Or, ε = [ B ] {d } (5.1.6)
In the above equation [B] is called as strain displacement relationship matrix. The shape
functions for the 3 node triangular element in Cartesian coordinate is represented as,
1
x 2 y 3 x 3 y 2 y 2 y 3 x x 3 x 2 y
N 2A
1 1
N 2 x 3 y1 x1y3 y3 y1 x x1 x 3 y
2A
3
N 1
x1y 2 x 2 y1 y1 y 2 x x 2 x1 y
2A
Or,
1
x y
1 1 1
N1 2A
1
N 2 2 2 x 2 y (5.1.7)
2A
3
N
1
3 3 x 3 y
2A
Where,
1 x 2 y 3 x 3 y 2 , 2 x 3 y1 x1y3 , 3 x1y 2 x 2 y1 ,
1 y 2 y3 , 2 y3 y1 , 3 y1 y 2 , (5.1.8)
1 x 3 x 2 , 2 x 2 x1 , 3 x 2 x1 ,
Hence the required partial derivatives of shape functions are,
∂N1 β1 ∂N 2 β 2 ∂N 3 β3
= , = , = ,
∂x 2 A ∂x 2 A ∂x 2 A
∂N1 γ 1 ∂N 2 γ 2 ∂N 3 γ 3
= , = , = ,
∂y 2 A ∂x 2A ∂x 2 A
Hence the value of [B] becomes:
∂N ∂N 2 ∂N 3
1 0 0 0
∂x ∂x ∂x
∂N1 ∂N 2 ∂N 3
[ B] = 0 0 0
∂y ∂y ∂y
∂N1 ∂N 2 ∂N 3 ∂N1 ∂N 2 ∂N 3
∂y ∂y ∂y ∂x ∂x ∂x
β1 β2 β3 0 0 0
1
Or, [ B] = 0 0 0 γ1 γ 2 γ 3 (5.1.9)
2A
γ 1 γ2 γ3 β1 β 2 β3
According to Variational principle described in module 2, lecture 1, the stiffness matrix is
represented as,
=[k ] ∫∫∫ [ B ] [ D ][ B ] d Ω
T
(5.1.10)
Ω
Since, [B] and [D] are constant matrices; the above expression can be expressed as
=[k ] [=
B ] [ D ][ B ] ∫∫∫ d V [ B ] [ D ][ B ]V
T T
(5.1.11)
V
For a constant thickness (t), the volume of the element will become A.t . Hence the above
equation becomes,
[ k ] = [ B ] [ D ][ B ] At
T
(5.1.12)
For plane stress condition, [D] matrix will become:
1 µ 0
E
[ D] = 2
µ 1 0 (5.1.13)
1− µ
1− µ
0 0
2
Therefore, for a plane stress problem, the element stiffness matrix becomes,
β1 0 γ1
β 0 γ 2
2 1 µ 0 β1 β 2 β3 0 0 0
β3 γ3
0 0 0 γ 3
Et 0
[k ] = µ 1 0 γ1 γ 2 (5.1.14)
4 A (1 − µ 2 ) 0 γ1 β1
1 − µ γ 1 γ 2 γ3 β1 β 2 β3
0 γ2 β2 0 0
2
0 γ3 β3
Or,
2 (1 + µ ) β γ
β1 + Cγ 1 β1β 2 + Cγ 1γ 2 β1β3 + Cγ 1γ 3 µβ1γ 2 + C β 2γ 1 µβ1γ 3 + C β3γ 1
2
1 1
2
β + Cγ
2 2
β 2 β3 + Cγ 2γ 3 µβ 2γ 1 + C β1γ 2
(1 + µ ) β γ
µβ 2γ 3 + C β 3γ 2 (5.1.15)
2 2 2 2
2
Et
[k ] = (1 + µ ) β γ
4 A (1 − µ )
2
β32 + Cγ 32 µβ3γ 1 + C β1γ 3 µβ 3γ 2 + C β 2γ 3 3 3
2
γ + Cβ
2
1 1
2
γ 1γ 2 + C β1β 2 γ 1γ 3 + C β1β3
Sym. γ 22 + C β 22 γ 2γ 3 + C β 2 β 3
γ 32 + C β32
Where, C =
(1 − µ )
2
Similarly for plane strain condition, [D] matrix is equal to,
(1 − µ ) µ 0
E
= [ D] µ (1 − µ ) 0 (5.1.16)
(1 + µ )(1 − 2µ ) 1 − 2µ
0 0
2
Hence the element stiffness matrix will become:
M β12 + γ 12 M β1β 2 + γ 1γ 2 M β1β3 + γ 1γ 3 ( µ + 1) β1γ 1 µβ1γ 2 + β 2γ 1 µβ1γ 3 + β3γ 1
M β 22 + γ 22 M β 2 β 3 + γ 2γ 3 µβ 2γ 1 + β1γ 2 ( µ + 1) β 2γ 2 µβ 2γ 3 + β3γ 2
M β32 + γ 32 µβ3γ 1 + β1γ 3 µβ3γ 2 + C β 2γ 3 ( µ + 1) β3γ 3
(5.1.17)
Et
[k ] =
2 A (1 + µ ) M γ 12 + β12 M γ 1γ 2 + β1β 2 M γ 1γ 3 + β1β3
Sym. M γ 22 + β 22 M γ 2γ 3 + β 2 β 3
M γ 32 + β32
Where M= (1 − µ )
Where, FΓ, and FΩ are the surface and body forces respectively. Using the relationship between
stress-stain and strain displacement, one can derive the following expressions:
D ][ B ]{d } , δ {ε } [ B ]δ=
{σ } [ = {d } and δ {u} [ N ]δ {d }
(5.1.19)
Hence eq. (5.1.18) can be rewritten as,
NS F d dT N T F d (5.1.20)
d B DBdd d
T T T T
Or,
B D Bdd N F d N F d
S T (5.1.21)
T T
s
Here, [N ] is the shape function along the boundary where forces are prescribed. Eq.(5.1.21) is
equivalent to k d F , and thus, the nodal load vector becomes
(5.1.22)
For a constant thickness of the triangular element eq.(5.1.22) can be rewritten as
For the a three node triangular two dimensional element, one can represent F and F as,
Fx Fx
F and F
Fy Fy
Fx 0
For example, in case of gravity load on CST element, F
Fy g
For this case, the shape functions in terms of area coordinates are:
L1 L 2 L3 0 0 0
N (5.1.24)
0 0 0 L1 L 2 L3
As a result, the force vector on the element considering only gravity load, will become,
L 0 0 0
1
0
L 0 0
2
L 0 0 0
dA t
0
F t 3
dA gt dA (5.1.25)
0 L L g L
A
1
g A
1 A
1
0 L
3 L g 3 L 3
The, [D] matrix is the constitutive matrix which will be taken according to plane stress or plane
strain condition. The nodal strain and stress vectors are given by,
{ε n } = {ε x1 ε x 2 ε x3 ε y1 ε y 2 ε y 3 γ xy1 γ xy 2 γ xy 3}
T
(5.2.9a)
{σ n } = {σ x1 σ x 2 σ x3 σ y1 σ y 2 σ y 3 τ xy1 τ xy 2 τ xy 3}
T
(5.2.9b)
[ Bn1 ] [ 0]
{ε n } = [ 0] [ Bn 2 ] {d } (5.2.10)
[ Bn 2 ] [ Bn1 ]
Referring to section 3.3.1, using proper values of area coordinates in [B] matrix, one can find
3b1 −b2 −b3 4b2 0 4b3
1
[ Bn1 ] = −b1 3b2 −b3 4b1 4b3 0 (5.2.11a)
2A
−b1 −b 3b3 0 4b2 4b1
And,
3a1 −a2 −a3 4a2 0 4a3
[ Bn 2 ] = −a1 3a2 −a3 4a1 4a3 0
1
(5.2.11b)
2A
−a1 −a 3a3 0 4a2 4a1
Thus, the element stiffness can be evaluated by putting the values from eq. (5.2.11) in eq. (5.2.7).
The element 1 is connected with node 1, 3 and 4 and let assume its Cartesian coordinates are (x1,
y1), (x3, y3) and (x4, y4) respectively. If we consider nodes 1, 3 and 4 are similar to node 1, 2 and
3 in eq.(5.1.9) then the [B] can be written as
β1 β 2 β3 0 0 0
[ B ] = 0 0 0 γ 1 γ 2 γ 3
1
2A
γ 1 γ 2 γ 3 β1 β 2 β3
By introducing values of β & γ discussed in previous lecture note, we can get value of [B] as
0 1 −1 0 0 0
1
=[ B] 0 0 0 −3 0 3
1500
−3 0 3 0 1 −1
For plain stress problem, putting the values of E and µ one can find the following values.
1 µ 0 16 4 0
E 4 ×104
= [ D] = µ 1
1 − µ 2
0 4 16 0
3
1− µ 0 0 6
0 0
2
Therefore the stiffness matrix for the element 1 will be
[ k ]1 = tA [ B ] [ D ][ B ]
T
The rectangular elements are widely used for solving two dimensional continuums. The main
advantage of this type of element is the easy formulation and easy development of computer
code. The element stiffness of such elements is derived here using the concept of isoparametric
formulation.
u 1 u v
2 2
x
x 2 x x
v 1 u v
2 2
y
y 2 y y
(5.3.6)
v u u u v v
xy
x y x y x y
In case of small amplitude of displacement, one can ignore the nonlinear term of the above
equation and will reach the following expression.
∂u
εx =
∂x
∂v
εy = (5.3.7)
∂y
∂v ∂u
γ= +
∂x ∂y
xy
If the element thickness is t, then dΓ1 =t.dl. Thus the eq.(5.3.19) can be replaced as
0
(1 − η )
+1
1 −η 1 + η qx 2
{Fx } = t ∫ 2 dl (5.3.21)
−1 (1 + η )
2 2 qx 3
2
0
After integrating the above expression, the nodal load vector along x direction will become as
follows.
0
2q + q
{Fx } = x 2 x3
t
(5.3.22)
3 q x 2 + 2q x 3
0
Lecture 4: Numerical Evaluation of Element Stiffness
Derivation of element stiffness for a four node rectangle element has been demonstrated in last
lecture. The stiffness matrix of each element can be calculated easily by developing a suitable
computer algorithm. To help students for developing their own computer code, a numerical
example has been solved and demonstrated here.
∂N1 ∂N 2 ∂N 3 ∂N 4 x1 y1 x y1
−(1 − ξ ) +(1 − ξ ) +(1 + ξ ) −(1 + ξ ) 1
∂ξ
∂ξ ∂ξ ∂ξ x 2 y2 4 4 4 4 x2 y2
|J| =
∂N1 ∂N 2 ∂N 3 ∂N 4 x3 y3 −(1 − η ) −(1 + η ) +(1 + η ) +(1 − η ) x y3
∂η 3
∂η ∂η ∂η x4 y4 4 4 4 4 x y4
4
Considering the sampling point, (ξ=0 and η=0 ), the value of the Jacobian, [J] is
−1 +1 1 − 1 0 0
[J ] = 4 4 4 4 70 0
−1 −1 1 1 70 50
4 4 4 4 0 50
1
0
35 0 35 J11* J12*
Thus, [J ] = =
and [ J ]−1
= * *
; and |𝐽| = 875
0 25 0 1 J 21 J 22
25
Now, the strain vector for the element will become
∂u
∂ξ
J 11* J 12* 0 0 ∂u
* ∂η
[ε ] = 0 0 J 21 *
J 22 × ∂v
J 21
* *
J 22 J 11* J 12*
∂ξ
∂v
∂η
u1
∂N 1 ∂N 2 ∂N 3 ∂N 4
∂ξ 0 0 0 0 v1
∂ξ ∂ξ ∂ξ
u
0 ∂N 1 ∂N 2 ∂N 3 ∂N 4
J 11* 0
2
J 12* 0 0 0 0
* ∂η ∂η ∂η ∂η 2
v
[ε ] = 0 0 *
J 21 J 22 × ∂N 1 ∂N 2 ∂N 3 ∂N 4 u 3
J 21
* *
J 22 J 11* J 12* 0 0 0 0
∂ξ ∂ξ ∂ξ ∂ξ v3
∂N 1 ∂N 2 ∂N 3 ∂N 4
0 0 0 0 u4
∂η ∂η ∂η ∂η
v 4
J 11* J 12* 0 0
*
[ε ] = 0 × [B']{d } = [ B]{d }
*
0 J 21 J 22
J 21
* * *
J 22 J 11* J 12
1 1 1 1
− 4 0 4 0 4
0 −
4
0
1
1 1 1 1 35 0 0 0
− 0 − 0 1
0 0
[B ] =
/ 4 4 4 4 and [ B] = 0 0 0 [B']
0 −1 0 1 0 −
1 1 25
0
0 0
4 4 4 4 1 1
0 −
1
0 −
1
0
1
0
1
25 35
4 4 4 4
1− µ 2 1− µ
0 0 0 0 1 / 2
2
1 0 0
[C ][ B] = 2 × 10 0 1 0 ×
3
0 0 1 / 2
The natural co-ordinates of the sampling points for 2×2 Gauss Quadrature integration are
1 +0.57735 +0.57735
2 -0.57735 +0.57735
3 -0.57735 -0.57735
4 +0.57735 -0.57735
For a four node quadrilateral element, the shape functions and their derivatives are as follows.
1
0
35 0 35 J11* J12*
Thus, [J ] = 0 25 =
and [ J ]−1
= * *
; Thus |𝐽| = 875
0 1 J 21 J 22
25
∂u
∂ξ
J 11* J 12* 0 0 ∂u
* ∂η
[ε ] = 0 0 *
J 21 J 22 × ∂v
J 21
* *
J 22 J 11* J 12*
∂ξ
∂v
∂η
u1
∂N 1 ∂N 2 ∂N 3 ∂N 4
∂ξ 0 0 0 0 v1
∂ξ ∂ξ ∂ξ
u
0 ∂N 1 ∂N 2 ∂N 3 ∂N 4
J 11* 0
2
J 12* 0 0 0 0
* ∂η ∂η ∂η ∂η v2
[ε ] = 0 0 J *
J 22 ×
21
∂N 1 ∂N 2 ∂N 3 ∂N 4 u 3
J 21
*
J *
J *
J 12* 0 0 0 0
22 11
∂ξ ∂ξ ∂ξ ∂ξ v3
∂N 1 ∂N 2 ∂N 3 ∂N 4
0 0 0 0 u4
∂η ∂η ∂η ∂η
v 4
J 11* J 12* 0 0
*
[ε ] = 0 × [B']{d } = [ B]{d }
*
0 J 21 J 22
J 21
* *
J 22 J *
11 J 12*
1 − 0.57735 1 − 0.57735 1 + 0.57735 1 + 0.57735
− 4
0
4
0
4
0 −
4
0
− 1 − 0.57735 0 −
1 + 0.57735
0
1 + 0.57735
0
1 − 0.57735
0
4 4 4 4
′
[B ] =
1 − 0.57735 1 − 0.57735 1 + 0.57735 1 + 0.57735
0 − 0 0 0 −
4 4 4 4
1 − 0.57735 1 + 0.57735 1 + 0.57735 1 − 0.57735
0 − 0 − 0 0
4 4 4 4
Thus,
1
35 0 0 0 −0.1057 −0.3943
0 0.1057 0 0.3943 0 0
1 −0.1057 0 −0.3943 0 0.3943 0 0.1057 0
[ B] 0 0 0 ×
25 0 −0.1057 0 0.1057 0 0.3943 0 −0.3943
0 1 1
0 0 −0.1057 0 −0.3943 0 0.3943 0 0.1057
25 35
1− µ 2 1− µ
0 0 0 0 1 / 2
2
1 0 0
[C][B] = 2 × 10 0 1 0 ×
3
0 0 1 / 2
The values of gauss weights are wi=wj=1.0. Therefore, the stiffness matrix [k] at this sampling
point is [k ] = twi w j [ B ]ij [C ]ij [ B ]ij | J | , where t is thickness of the element. Thus at sampling point
T
1,
−0.003 0 −0.0042
0 −0.0042 −0.003
0.003 0 −0.0158
1 0 0
−0.0158 0.003
× 0 1 0 ×
0
[ k1 ]= 20 ×1×1× 875 × 2 ×103 ×
0.0113 0 0.0158
0 0 1/ 2
0 0.0158 0.0113
−0.0113 0 0.0042
0 0.0042 −0.0113
−0.003 0 0.003 0 0.0113 0 −0.0113 0
0 −0.0042 0 −0.0158 0 0.0158 0 0.0042
−0.0042 −0.003 −0.0158 0.003 0.0158 0.0113 0.0042 −0.0113
0.0632 0.0223 0.0848 −0.0223 −0.2358 −0.0834 0.0878 0.0834
0.0785 0.0834 0.2174 −0.0834 −0.2929 −0.0223 −0.0030
0.4672 −0.0834 −0.3162 −0.3109 −0.2358 0.3109
0.8866 0.0834 −0.8111 0.0223 −0.2929
[ k=
1 ] 10 ×
4
0.8795 0.3109 −0.3275 −0.3109
sym 1.0927 0.0834 0.0113
0.4755 −0.0834
0.2847
−0.0113 −0.0042
0
0 −0.0113
−0.0042
0.0113 −0.0158
0
0 −0.0158
0.0113
[ k2 ]= 20 ×1×1× 875 × 2 ×103 ×
0.003 0.0158
0
0 0.003
0.0158
−0.003 0.0042
0
0 −0.0030
0.0042
−0.0113 0 0.0113 0 0.0030 0 −0.0030 0
0 −0.0042 0 −0.0158 0 0.0158 0 0.0042
−0.0021 −0.0056 −0.0079 0.0056 0.0079 0.0015 0.0021 −0.0015
−0.0113 0 −0.0158
0 −0.0158 −0.0113
0.0113 0 −0.0042
0 −0.0042 0.0113
[ k3 ]= 20 ×1×1× 875 × 2 ×103 ×
0.0030 0 0.0042
0 0.0042 0.0030
−0.0030 0 0.0158
0 0.0158 −0.0030
−0.0030 0 0.0113 0 0.0030 0 −0.0030 0
0 −0.0158 0 −0.0042 0 0.0042 0 0.0158
−0.0079 −0.0056 −0.0021 0.0056 0.0021 0.0015 0.0079 −0.0015
The stiffness matrix of the element can be computed as the sum of the values at the four
sampling points: [k ] = [k1 ] + [k2 ] + [k3 ] + [k4 ] . Thus, the final value of the stiffness matrix will
become
′ = (1 ± r )(1 ± s )
1
N di (5.5.4)
4
′ are evaluated at r and s coordinates of point P. Let the point P coincides with the corner 1.
N di
To calculate stress σ x1 at corner 1 from σ x values at the four Gauss points, substitution of r and
s into the shape functions will give
σ xI =1.8666σ xI − 0.500σ xII + 0.134σ xIII − 0.500σ xIV (5.5.5)
Fig. 5.5.1 Natural coordinate systems used in extrapolation of stresses from Gauss points
σ 6 ( )
1− 3 4 ( )
1+ 3 4 (
1+ 3 4 ) ( )
1 − 3 4 σ IV
σ
7
(
)
1− 3 4 ( )
1− 3 4 (
1+ 3 4 ) ( )
1 + 3 4
σ 8 ( )
1 + 3 4 ( )
1− 3 4 (
1− 3 4 ) ( )
1 + 3 4
Here, σ 1 , σ 2 …….. σ 8 are the smoothened nodal values and σ I … σ IV are the stresses at the
Gauss points. Smoothened nodal stress values for four node rectangular element can be also be
evaluated in a similar fashion. The relation between the stresses at Gauss points and nodal point
for four nodel element will be
1
1 3
1
1
3
2 2 2 2
1 3 I
1 3 1
2 2 1 1
2 II
2 2
III
(5.5.7)
3 3 1 3 1
1 1
4 2 2 2 2 IV
3
1 1
3
1
1
2 2 2 2
The stress at particular node joining with more than one element will have different magnitude as
calculated from adjacent elements (Fig. 5.5.2(a)). The stress resultants are then modified by
finding the average of resultants of all elements meeting at a common node. A typical stress
distribution for adjacent elements is shown in Fig. 5.5.2(b) after stress smoothening.
Fig. 5.5.2 Stress smoothening at common node
∂u 1 ∂u ∂v
2 2
εx =+ +
∂x 2 ∂x ∂x
∂v 1 ∂u ∂v
2 2
εy = + +
∂y 2 ∂y ∂y
(5.5.8)
∂v ∂u ∂u ∂u ∂v ∂v
γ xy = + + +
∂x ∂y ∂x ∂y ∂x ∂y
The displacements at any point inside the node are expressed in terms of their nodla
displacements. Thus,
n n
=u N u [ N ]{u } =
∑= and v ∑
= N v [ N ]{v }
i =1
i i i i
i =1
i i i i (5.5.9)
Therefore,
∂u ∂ [ N i ]
= = {ui } [ B1 ]{ui }
∂x ∂x
∂v
= [ B1 ]{vi }
∂x
∂u
= [ B2 ]{ui }
∂y
∂v
= [ B2 ]{vi }
∂y
(5.5.10)
Here, [B1] and [B2] are the derivative of the shape function [Ni] with respect to x and y
respectively. The vectors {ui} and {vi} represent the nodal displacements vectors in x and y
directions respectively. The vector of strains at any point inside an element, {ε } may be
expressed in terms of nodal displacement as
{ε } = [ B ] {d }
(5.5.11)
where [B] is the strain displacement matrix. {d} is the nodal displacement vector and may be
expressed as
{u }
{d } = i
{vi }
(5.5.12)
The matrix [B] may be expressed with two components as
[=
B ] [ Bl ] + [ Bnl ]
(5.5.13)
where, [Bl] and [Bnl] are the linear and nonlinear part of the strain-displacement matrix
respectively and are expressed as follows:
[B1 ] [0]
[Bl ] = [0] [B2 ]
[B2 ] [B1 ]
(5.5.14)
and
1 T
2 {u } [B1 ] [B1 ] 2 {v } [B1 ] [B1 ]
T 1 T T
[Bnl ] = 1 {u }T [B2 ]T [B2 ] 1 {v }T [B2 ]T [B2 ]
2 T 2
{u } [B2 ]T [B1 ] {v }T [B2 ]T [B1 ]
(5.5.15)
5.5.2.1 Steps to include effect of geometrical nonlinearity
The nonlinear geometric effect of the structure at a particular instant of time can be obtained by
performing the following steps.
1. Calculation of displacement {d}1 considering linear part of strain matrix [Bl].
2. Evaluation of nonlinear part of the strain matrix [Bnl] (eq5.5.15) adopting {d}1 from
previous step.
3. Evaluation of total strain matrix [B] = [Bl] + [Bnl].
4. Calculation of displacement {d}2 considering both linear and nonlinear part of strain
matrix [B].
5. Repetition of steps 2 to 4 with {d}2, from which modified displacement, {d}3 are
obtained.
6. Step 5 is carried out until the displacements for two consecutive iteration converge
i.e.,
{d } j − {d } j −1
<ε
{d } j
Where ε is any pre-assigned small value and j is the number of iterations.
Thus, the displacements of these nodes are not required to formulate overall equilibrium
equations of the structure. This limits the usefulness of these elements. A technique known as
“static condensation” can be used to suppress the degrees of freedom associated with the internal
nodes in the final computation. The technique of static condensation is explained below. The
equilibrium equation for a system are expressed in the finite element form as
F = K d (5.5.16)
Where, {F}, [K] and {d} are the load vector, stiffness matrix and displacement vector for the
entire structure. The above equation can be rearranged by separating the relevant terms
corresponding to internal and external nodes of the elements.
Fi K ii K ie di
= (5.5.17)
Fe K ei K ee d e
Here, {di} and {de} are the displacement vectors corresponding to internal and external nodes
respectively. Similarly, {Fi} and {Fe} are force vectors corresponding to internal and external
nodes. Now, the above expression can be written in the following form separately.
Here, the equations are reduced to a form involving only the external nodes of the elements. The
above reduced substructure equations are assembled to achieve the overall equations involving
only the boundary unknowns. Thus the above equation can be rewritten as
F K
e ei Kii Fi K ee K ei Kii Kie de de
1 1
(5.5.22)
Or
Fc = K c de (5.5.23)
Fc = Fe - K ei Kii Fi and K c = K ee - K ei Kii Kie de . Here, [Kc]
-1 -1
Where,
is called condensed or reduced stiffness matrix and {Fc} is the condensed or effective nodal load
vector corresponding to external nodes of the elements. In this process, the size of the matrix for
inversion will be comparatively small. The unknown displacements of the exterior nodes, {de}
can be obtained by inverting the matrix [Kc] in eq.(5.5.23). Once, the values of {de} are obtained,
the displacements of internal nodes {di} can be found from eq.(5.5.20).
Lecture 6: Axisymmetric Element
5.6.1 Introduction
Many three-dimensional problems show symmetry about an axis of rotation. If the problem
geometry is symmetric about an axis and the loading and boundary conditions are symmetric
about the same axis, the problem is said to be axisymmetric. Such three-dimensional problems
can be solved using two-dimensional finite elements. The axisymmetric problem are most
conveniently defined by polar coordinate system with coordinates (r, θ, z) as shown in Fig. 5.6.1.
Thus, for axisymmetric analysis, following conditions are to be satisfied.
1. The domain should have an axis of symmetry and is considered as z axis.
2. The loadings on the domain has to be symmetric about the axis of revolution, thus they
are independent of circumferential coordinate θ.
3. The boundary condition and material properties are symmetric about the same axis and
will be independent of circumferential coordinate.
Axisymmetric solids are of total symmetry about the axis of revolution (i.e., z-axis), the field
variables, such as the stress and deformation is independent of rotational angle θ. Therefore, the
field variables can be defined as a function of (r, z) and hence the problem becomes a two
dimensional problem similar to those of plane stress/strain problems. Axisymmetric problems
includes, circular cylinder loaded with uniform external or internal pressure, circular water tank,
pressure vessels, chimney, boiler, circular footing resting on θsoil
• P( r,z, ) mass, etc.
If the loading consists of radial and axial components that are independent of θ and the material
is either isotropic or orthotropic and the material properties are independent of θ, the
displacement at any point will only have radial (𝑢𝑟 ) and axial (𝑢𝑧 ) components. The only stress
components that will be nonzero are 𝜎𝑟𝑟 , 𝜎𝑧𝑧 , 𝜎𝜃𝜃 𝑎𝑛𝑑 𝜏𝑟𝑧 .
Considering the original arc length versus the deformed arc length, the differential element
undergoes an expansion in the circumferential direction. Before deformation, let the arc length is
assumed as ds = rdθ. After deformation, the arc length will become ds = (r+u) dθ. Thus, the
tangential strain will be
r +u d - rd u
ε = (5.6.3)
rd r
Similarly, the shear strain will be
u w
rz
z r (5.6.4)
r 0 and z 0
Thus, there are four strain components present in this case and is given by
∂u ∂
∂r ∂r 0
εr
∂w 0 ∂
εz ∂z ∂z u
{ε } =
= = (5.6.5)
εθ u 1 0
w
γ rz r r
∂u ∂w ∂ ∂
+
∂z ∂r ∂z ∂r
5.6.3 Relation between Stress and Strain
The stress strain relation for axisymmetric case can be derived from the three dimensional
constitutive relations. We know the stress-strain relation for a three-dimensional solid is
1−𝜇 𝜇 𝜇 0 0 0
⎡ ⎤
𝜎𝑥 ⎢ 𝜇 1−𝜇 𝜇 0 0 0 ⎥ ϵ𝑥
⎧𝜎𝑦 ⎫ ⎢ ⎥ ⎧ϵ ⎫
𝑦
⎪𝜎 ⎪ ⎢ 𝜇 𝜇 1−𝜇 0 0 0 ⎥ ⎪ϵ ⎪
𝐸
𝑧
= (1+𝜇)(1−2𝜇) ⎢ ⎥ 𝑧
𝜏 0 ⎥ ⎨𝜈𝑥𝑦 ⎬
1−2𝜇
⎨ ⎬𝑥𝑦 ⎢ 0 0 0 2
0
⎪𝜏𝑦𝑧 ⎪ ⎢ 1−2𝜇 ⎥ ⎪𝜈𝑦𝑧 ⎪
⎩𝜏𝑧𝑥 ⎭ ⎢ 0 0 0 0 2
0 ⎥ ⎩𝜈𝑧𝑥 ⎭
⎢ 1−2𝜇 ⎥
⎣ 0 0 0 0 0 2 ⎦
(5.6.6)
Now, comparing the stress-strain components present in the axisymmetric case, the stress-strain
relation can be expressed from the above expression as follows
𝜎𝑟 1−𝜇 𝜇 𝜇 0 𝜀
⎧ ⎫ ⎡ ⎤⎧ 𝑟 ⎫
⎪𝜎𝑧 ⎪ ⎢ 𝜇 1−𝜇 𝜇 0 ⎥ ⎪𝜀𝑧 ⎪
𝐸
= ⎢ ⎥ (5.6.7)
⎨𝜎𝜃 ⎬ (1+𝜇)(1−2𝜇) ⎢ 𝜇 𝜇 1−𝜇 0 ⎥ ⎨𝜀𝜃 ⎬
⎪ ⎪ ⎢ 1−2𝜇 ⎥
⎪ ⎪
⎩𝜏𝑟𝑧 ⎭ ⎣ 0 0 0 2 ⎦
⎩𝜈𝑟𝑧 ⎭
Thus, the constitutive matrix [D] for the axisymmetric elastic solid will be
1−𝜇 𝜇 𝜇 0
⎡ ⎤
⎢ 𝜇 1−𝜇 𝜇 0 ⎥
𝐸
[D] = (1+𝜇)(1−2𝜇) ⎢ ⎥ (5.6.8)
⎢ 𝜇 𝜇 1−𝜇 0 ⎥
⎢ 1−2𝜇 ⎥
⎣ 0 0 0 2 ⎦
In which,
N y ε y 1 µ 0
Eh
{σ } = Nθ , {ε } = ε θ and [ D ] =µ 1 0 (5.6.11)
M χ 1− µ 2 2
y y 0 0 h
12
The generalized strain vector can be expressed in terms of the displacement vectors as follows.
{ε } = [ B ] {d } (5.6.12)
Where,
∂
0 ∂y
u 1
{d } = and [ B ] = 0 (5.6.13)
v
R
∂
2
− ∂y 2 0
Here, u and v are the displacement components in two perpendicular directions. With the use of
stress and strain vectors, the potential energy expression are written in terms of displacement
vectors as
∫ ({d } [ B ] [ D ][ B ]dy
{d } )
H
1
U = πR× 2
T T
(5.6.14)
2 0
Similarly, neglecting the rotary inertia, the kinetic energy can be expressed as
{d})
∫ ({d} [ N ] m [ N ]dy
H
1 T
T = πR× 2
T
(5.6.16)
2 0
{}
Where, m denotes the mass of the shell element per unit area and d represents the velocity
vector. Thus, the element mass matrix is given by
Le
[ M ] = 2πRm ∫ [ N ] [ N ] dy
T
(5.6.17)
0
Finite element formulation for the axisymmetric problem will be similar to that of the two
dimensional solid elements. As the field variables, such as the stress and strain is independent of
rotational angle θ, circumferential displacement will not appear. Thus, the displacement field
variables are expressed as
n
u r,z = N i r,z ui
i=1
n
(5.7.1)
wr,z = N i r,z wi
i=1
Here, ui and wi represent radial and axial displacements respectively at nodes. Ni (r, z) are the
shape functions. As the geometry and field variables are independent of rotational angle θ, the
interpolation function Ni (r, z) can be expressed similar to 2-dimensional problems by replacing
the x and y terms with r and z terms respectively.
5.7.1 Stiffness Matrix of a Triangular Element
Fig. 5.7.1 shows the cylindrical coordinates of a three node triangular element. Hence the
analysis of the axisymmetric element can be approached in a similar way as the CST element.
Thus the field variables of such an element can be expressed as
u =α 0 + α1r + α 2 z
(5.7.2)
w =α 3 + α 4 r + α 5 z
Or,
{d } = [φ ]{α } (5.7.3)
Where,
1 r z 0 0 0
u ,
[φ ] = {α } = {α 0 α1 α 2 α 3 α 4 α 5 }
T
{d } = and
w 0 0 0 1 r z
Using end conditions,
u1 1 ri zi 0 0 0 α 0
u 1 r zj 0 0 0 α1
2 j
u3 1 rk zj 0 0 0 α 2
= (5.7.4)
w1 0 0 0 1 ri zi α 3
w2 0 0 0 1 rj z j α 4
w3 0 0 0 1 rk zk α 5
Or,
{d } = [ A]{α }
(5.7.5)
[ A] {d }
⇒ {α } =
−1
{}
Here d are the nodal displacement vectors.
Fig. 5.7.1 Axisymmetric three node triangle in cylindrical coordinates
1
N1 (=
r, z ) ( r2 z3 − r3 z2 ) + ( z2 − z3 ) r + ( r3 − r2 ) z
2A
1
N 2 (=
r, z ) ( r3 z1 − r1 z3 ) + ( z3 − z1 ) r + ( r1 − r3 ) z
2A
1
N 3 (=
r, z ) ( r1 z2 − r2 z1 ) + ( z1 − z2 ) r + ( r2 − r1 ) z
2A
Or,
1
N i ( r , z=
) (α i + r βi + zγ i )
2A
N j ( r , z=
)
1
2A
(α j + r β j + zγ j ) (5.7.8)
1
N k ( r , z=
) (α k + r β k + zγ k )
2A
Where,
αi =
rj zk − rk z j αj =
rk zi − ri zk αk =
ri z j − rj zi
βi =
z j − zk βj =
zk − zi βk =
zi − z j
γi =
rk − rj γi =
ri − rk γi =
rj − ri (5.7.9)
2 A=
1
2
( ri z j + rj zk + rk zi − ri zk − rj zi − rk z j )
Putting the value of {u,w} in eq. (5.7.7) from eq. (5.6.5),
∂N i ∂N j ∂N k
0 0 0 r
∂r ∂r ∂r 1
Ni Nj Nk r2
r 0 0 0 r3
{ε } =
r r
[ B ]{d } (5.7.10)
∂N i ∂N j ∂N k z1
0 0 0
∂z ∂z ∂z z2
∂N i ∂N j ∂N k ∂N i ∂N j ∂N k z3
∂z ∂z ∂z ∂r ∂r ∂r
Thus, the strain displacement matrix can be expressed as,
βi βj βk 0 0 0
Nj
N Nk
1 i 0 0 0
[ B] = r r r (5.7.11)
2A
0 0 0 γi γj γk
γ i γj γk βi βj β k
ri + rj + rk
Where, r = . Thus the stiffness matrix will become
3
=[k ] ∫∫∫ [ B ] [ D ][ B ] d Ω
T
2π
= [k ] ∫ ∫ [ B ] [ D ][ B ] r dθ dA 2π ∫ ∫ [ B ] [ D ][ B ] r dr dz
∫=
T T
Or, (5.7.12)
0
Since, the term [B] is dependent of ‘r’ terms; the term [ B ] [ D ][ B ] cannot be taken out of
T
integration. Yet, a reasonably accurate solution can be obtained by evaluating the [B] (denoted as
[B]) matrix at the centroid.
Hence, [ k ] = 2π r [ B ] [ D ][ B ] ∫ ∫ dr dz
T
Or,
[ k ] [ B ] [ D ][ B ] 2π rA
T
(5.7.13)
The strain-displacement relation for axisymmetric problem derived earlier (eq.(5.6.5)) can be
rewritten as
∂u
∂u ∂r
∂r
1 0 0 0 0 ∂u
εr
ε ∂w 0 0 0 1 0 ∂z
∂z
{ε } =
= z
=
1 ∂w (5.7.14)
εθ u 0 0 0 0
r ∂r
γ rz r
∂u ∂w 0 1 1 0 0 ∂w
+ ∂z
∂z ∂r u
∂u ∂u
∂r ∂ξ
J*
*
0 ∂u
∂u 11
J12 0 0
∂z J*21 J*22 0 0 0 ∂η
∂w = 0 0 *
J11 *
J12 0 ∂w (5.7.15)
∂rξ 0 J*21 J*22
0 ∂
0
∂w 0 0 0 0 1 ∂w
∂z ∂η
u
u
Hence, the strain components are calculated as
∂u
∂ξ
0 0 0 0 J11 0 ∂u
* *
1 J12 0 0
ε r *
ε 0 0 0 1 0 J 21 J*22 0 0 0 ∂η
z
= 1 0 0 ∂w
* *
0 J11 J12
ε θ 0 0 0 0
r 0 0 J*21 J*22
0 ∂ξ
γ rz
0 1 1 0 0 0 0 0 0 1 ∂w
∂η
u
Or,
51
u
J11
* *
J12 0 0 0
u
r
0
0
*
0 J J*22
z
21
1
w (5.7.16)
0 0 0 0
r
rz
J* J*22 * *
0 w
21 J11 J12
u
u u w w
With the use of interpolation function and nodal displacements, , , , can be expressed
x h x h
for a four node quadrilateral element as
u
N1 N 2 N 3 N 4 u1
0
x
0 0 0
x x x x u2
u N1 N 2 N 3 N 4 u3
0 0 0 0
h h h h h u4
(5.7.17)
w N1 N 2 N 3 N 4
w1
0 0 0 0
x x x x x w2
w N1 N 2 N 3
N 4
0 0 0 0
w3
h h h h
h
w4
N1 N 2 N 3 N 4
0 0 0 0
* *
0 0 N1 N 2 N 3 N 4
J11 J12 0 0 0 0 0
0
0 J 21 J*22 0
*
B
1 0 N1 N 2 N 3 N 4 (5.7.19)
0 0 0 0 0 0 0
* r
J 21 J*22 J11
* *
J12 0 N1 N 2 N 3 N 4
0 0 0 0
N
1 N2 N3 N4 N1 N2 N3 N 4
For a four node quadrilateral element,
(1ξ− )1( η− ) and
N1 = ⇒
∂N1
=
−
( 1η− ) ∂N1 1( ξ− )
=−
4ξ 4 ∂ η 4 ∂
(1ξ+ 1) ( η− ) and
N2 = ⇒
∂N 2 ( 1η− )
=
∂N1 1( ξ+ )
=− (5.7.20)
4ξ 4 ∂ η 4 ∂
=N3
(1ξ+ 1) ( η+ )
⇒=
∂N 2 ( 1η+ )
and
∂N1 1( ξ+ )
=
4ξ 4 ∂ η 4 ∂
(1ξ− )1( η+ ) and
N4 = ⇒
∂N 2
=
−
( 1η+ ) ∂N1 1( ξ− )
=
4ξ 4 ∂ η 4 ∂
Thus, the [B] matrix will become
J11
* *
J12 0 0 0
0 0 J 21 J*22 0
*
B 1
0 0 0 0
r
*
J 21 J 22 J11 J12 0
* * *
1 1 1 1
0 0 0 0
4 4 4 4
1 1 1 1
0 0 0 0
4 4 4 4
1 1 1 1
0 0 0 0
4 4 4 4
1 1 1 1
0 0 0 0
4 4 4 4
11 1 1 1 1 11 11 1 1 1 1 11
4 4 4 4 4 4 4 4
(5.7.21)
The stiffness matrix for the axisymmetric element finally can be found from the following
expression after numerical integration.
+1+1
=[k ] ∫ [ B=
] [ D ][ B ] d Ω ∫ ∫ [ B] [ D][πr.B].2.dξdηJ
T T
(5.7.22)
Ω −1−1
53
5.8.1 Introduction
Solid elements can easily be formulated by the extension of the procedure followed for two
dimensional solid elements. A domain in 3D can be discritized using tetrahedral or hexahedral
elements. For example, the eight node solid brick element is analogous to the four node rectangular
element. Regardless of the possible curvature of edges or number of nodes, the solid element can be
mapped into the space of natural co-ordinates, i.e the ξ = ±1,η = ±1, ζ = ±1 just like a plane element.
For three dimensional cases, each node has three degrees of freedom having u, v, and w as
displacement field in three perpendicular directions (X, Y and Z). In this case, one additional
dimension increases the computational expense manifolds.
∂u
∂x
∂u
∂y
∂u
∂z
ε x 1 0 0 0 0 0 0 0 0 ∂v
ε 0
y 0 0 0 1 0 0 0 0 ∂x
ε z 0 0 0 0 0 0 0 0 1 ∂v
=
γ xy 0 1 0 1 0 0 0 0 0 ∂y
γ yz 0 0 0 0 0 1 0 1 0 ∂v
γ zx 0 0 1 0 0 0 1 0 0 ∂z
∂w
∂x
∂w
∂y
∂w (5.8.1)
∂z
The following relation exists between the derivative operators in the global co-ordinates and the
natural co-ordinate system by the use of chain rule of partial differentiation.
∂ ∂x ∂y ∂z ∂
∂ξ ∂ξ
∂ξ ∂ξ ∂x
∂ ∂x ∂y ∂z ∂
= (5.8.2)
∂η ∂η ∂η ∂η ∂y
∂ ∂x ∂y ∂z ∂
∂ζ ∂ζ ∂ζ ∂ζ ∂z
Where the Jacobian Matrix will be
∂x ∂y ∂z
∂ξ
∂ξ ∂ξ
[J ] = ∂x ∂y ∂z
(5.8.3)
∂η ∂η ∂η
∂x ∂y ∂z
∂ζ ∂ζ ∂ζ
For an isoparamatric element the coordinates at a point inside the element can be expressed by its
nodal coordinate.
n n n
=x i i
=i 1 =i 1
∑=
N x ; y ∑ N=
y and z ∑ N z i i
=i 1
i i (5.8.4)
55
Substituting the above equations into the Jacobian matrix for an eight node brick element, we get
∂N i ∂N i ∂N i
∂ξ xi ∂ξ
yi zi
∂ξ
∂N i x ∂N i ∂N i
8
[J ] = ∑ yi zi (5.8.5)
i =1
∂η i ∂η ∂η
∂N ∂N i ∂N i
i xi yi zi
∂ζ ∂ζ ∂ζ
u
The strain displacement relation is given by {ε } = [B ]{d }, where, {d } = v .
w
The displacements in the x, y and z direction are u, v, and w respectively. Let consider the inverse of
Jacobian matrix as
J 11* J 13*
J 12*
* *
[J ]−1= J 21 *
J 22
J 23 (5.8.6)
J 31
* **
J 32
J 33
Thus, the relation between two coordinate systems can be rewritten as
−1
∂ ∂x ∂y ∂z ∂ ∂
∂x ∂ξ ∂ξ ∂ξ ∂ξ ∂ξ
J11
* *
J12 *
J13
∂ ∂x ∂y ∂z ∂ * * ∂
= = ∂η J 21
*
J 22 J 23 (5.8.7)
∂y ∂η ∂η ∂η * ∂η
J 31
* *
J 32 J 33
∂ ∂x ∂y ∂z ∂ ∂
∂ζ ∂ζ ∂ζ ∂ζ
∂z ∂ζ
∂u * ∂u * ∂u * ∂u
8
* ∂N i * ∂N i * ∂N i
8
= J 21 + J 22 + J 23 = ∑ J 21 + J 22 + J 23 i ∑ bi ui
u =
∂y
= ∂ξ ∂η ∂ζ i 1 =
∂ξ ∂η ∂ζ i 1
∂u * ∂u * ∂u * ∂u
8
* ∂N i * ∂N i * ∂N i
8
= J 31 + J 32 + J 33 = ∑ J 31 + J 32 + J 33 i ∑ ci ui
u =
∂z
= ∂ξ ∂η ∂ζ i 1 =
∂ξ ∂η ∂ζ i 1
Similarly,
∂v 8 ∂v 8 ∂v 8 ∂w 8 ∂w 8 ∂w 8
= ∑ = ai vi ; ∑ i i ∂z ∑
= b v ; = c v ; ∑ i i ∂y ∑
=
a w ; b w= and ∑ ci wi
∂x i 1 = ∂y i 1 = ∂x i 1 = ∂z i 1
i i i i
= i 1 = i 1 =
∂u
∂x
∂v
εx ∂y ai 0 0
0
ε y ∂w 0 bi
u
ε z ∂z 8 0 ci i
{ε } =
0
= ∂u = ∑ v (5.8.8)
γ xy + ∂v i =1 bi ai 0 i
bi i
w
γ yz ∂y ∂x 0 ci
γ zx ∂v + ∂w ci 0 ai
∂z ∂y
∂u + ∂w
∂z ∂x
Now, the strain displacement relationship matrix [B] can be identified from the above equation by
comparing it to {ε } = [B ]{d }.
[ k ] = ∫ ∫ ∫ [ B ] T [ D ][ B ]dξ dη dζ J (5.8.9)
−1 −1 −1
The size of the constitutive matrix [D] for solid element will be 6 × 6 and is already discussed in
module 1, lectures 3. For eight node brick element, the size of stiffness matrix will become 24 × 24
as number of nodes in one element is 8 and the degrees of freedom at each node is 3. It is well
established that 2 × 2 × 2 Gauss integration points are the optimal sampling points for eight node
isoparametric brick elements.
where {X } is the body forces per unit volume. The nodal load vector at any node i may be expressed
as
={Qi } ∫[N ] i
T
{ X }d Ω (5.8.11)
Ω
In case of gravity load, the force will act in the global negative Z direction. Therefore,
Ni 0 0 0
[ N i ] = 0 N i 0 and { X } = 0 (5.8.12)
0 0 N i
− ρ g
Here, the mass density of the material is ρ and the acceleration due to gravity is g. Thus, eq.(5.8.11)
will become
0
= {Qi } ∫ 0 d Ω (5.8.13)
Ω
− Ni ρ g
For isoparametric element the, the above expression will become
0
+1 +1 +1
{Qi } = ∫ ∫ ∫ 0 J dξdηdζ (5.8.14)
−1 −1 −1
− N ρ
i
Using Gauss Quadrature integration rule, the above expression may be evaluated as
0
n n n
{ Qi } = ∑∑∑ wi w j wk J ( ξi ,η j ,ζ k ) 0 (5.8.15)
=i 1 =j 1 =
k 1 − N ρ g
i ( ξi ,η j ,ζ k )
Where, n is the number of nodes in an element. For eight node linear brick element the value of n
will be 8 and the integration order suggested is 2×2×2. Similarly, for twenty node quadratic brick
element, the value of n will be 20 and the integration order suggested is 3×3×3.
In case of surface load, the value of [ N is ] in the above equation will become
58
N is 0 0
[ N is ] = 0 N i
s
0 (5.8.18)
0
0 N is
Here N is is the interpolation function for the node i. For example, the value of N is can be obtained
by substituting ξ =1 in N is for face 1. Thus, the surface pressure is expressed as,
ql3
{ p} = qm3 (5.8.19)
qn
3
Where, l3 , m3 , n3 are the direction cosines. Thus, eq.(5.8.17) can be expressed using eq.(5.8.19) in the
following form.
N is ql3
{Qi } = ∫−1 ∫−1 N is qm3 dA
+1 +1
(5.8.20)
N s qn
i 3
The value of dA is can be evaluated considering the cross product of vectors along the natural
coordinates parallel to the loaded faces of the element. Thus,
dA = e2 × e3 dηdζ (5.8.21)
σ x −σ τ xy τ xz
τ xy σ y −σ τ yz =
0 (5.8.22)
τ xz τ yz σ z −σ
The characteristic equation has three real roots σi, due to the symmetry of the stress tensor. The
principal stresses are arranged so that σ1 > σ2 > σ3. The maximum shear stress can be computed
from the following relations.
59
σ1 − σ 2 σ2 −σ3 σ 3 − σ1
τ max = largest of , and (5.8.23)
2 2 2
These three shear stress components will occur on planes oriented at 450 from the principal planes.
The distortion energy theory suggests that the total strain energy can be divided into two
components. They are (i) volumetric strain energy and (ii) distortion or shear strain energy. It is
anticipated that yield develops if the distortion component exceeds that at the yield point for a simple
tensile test. From the concept of distortion energy theory, the equivalent stress which is historically
known as Von Mises stress are defined as
(σ 1 − σ 2 ) 2 + (σ 2 − σ 3 ) 2 + (σ 3 − σ 1 ) 2
σe = (5.8.24)
2
The Von Mises stresses offer a measure of the shear or distortional stress in the material. In general,
this type of stress tends to cause yielding in metals.
60
From the above figure, the length of sides 1, 2 and 3 are calculated and will be 10, 8 and 6 cm
respectively. First, let consider side 1:
l1
{F }1 = ∫ [ N ] {FΓ } ds
T
Or,
0 0 0 0 0
L 0 L 0 L
2 2 2
l1
L 0 Fx1 l1
L3 0 10 l1
L3
{F }=1 ∫ 3 ds= ∫0 0 ds= 10 × ∫ ds
0
0 0 Fy1 0 0 0
0
0 L2 0 L2 0
0 L3 0 L3 0
p p !q !
Putting, ∫ L
1 L2 q ds =
( p + q + 1)!
l , we will get,
61
0 0 0
1 1 0.5
l1 1 0.1 1 0.5
{F }1 =10 × × =10 × × =
2 0 2 0 0
0 0 0
0 0 0
Similarly for side 2,
L1 0 L1 0 0 0
0 0 0 0 0 0
l2
L 0 Fx 2 l2
L3 0 0 l2 0 0
{F }2 = ∫ 3 ds = ∫0 0 ds = 20 × × = kN
0
0 L1 Fy 2 L1 20 2 1 0.8
0 0 0 0 0 0
0 L3 0 L3 1 0.8
Since no force is acting on side 3,
0
0
0
{F }3 =
0
0
0
Hence, the nodal load vector in all the nodes in x and y directions will become,
0
0.5
0.5
{F } = kN.
0.8
0
0.8
1
6.1.1 Introduction
A plate is a planer structure with a very small thickness in comparison to the planer dimensions. The
forces applied on a plate are perpendicular to the plane of the plate. Therefore, plate resists the applied
load by means of bending in two directions and twisting moment. A plate theory takes advantage of this
disparity in length scale to reduce the full three-dimensional solid mechanics problem to a two-
dimensional problem. The aim of plate theory is to calculate the deformation and stresses in a plate
subjected to loads. A flat plate, like a straight beam carries lateral load by bending. The analyses of
plates are categorized into two types based on thickness to breadth ratio: thick plate and thin plate
analysis. If the thickness to width ratio of the plate is less than 0.1 and the maximum deflection is less
than one tenth of thickness, then the plate is classified as thin plate. The well known as Kirchhoff plate
theory is used for the analysis of such thin plates. On the other hand, Mindlin plate theory is used for
thick plate where the effect of shear deformation is included.
Classical thin plate theory is based upon assumptions initiated for beams by Bernoulli but first applied to
plates and shells by Love and Kirchhoff. This theory is known as Kirchhoff’s plate theory. Basically,
three assumptions are used to reduce the equations of three dimensional theory of elasticity to two
dimensions.
1. The line normal to the neutral axis before bending remains straight after bending.
2. The normal stress in thickness direction is neglected. i.e., σ z = 0 . This assumption converts
the 3D problem into a 2D problem.
3. The transverse shearing strains are assumed to be zero. i.e., shear strains γxz and γyz will be
zero. Thus, thickness of the plate does not change during bending.
Et 3 ∂2w ∂2w
My =
− 2 +υ 2 =DP ( χ y + υχ x ) (6.1.9)
12 (1 − υ 2 ) ∂y ∂x
Et 3 ∂2w D (1 − υ )
M xy = M yx = = − P χ xy
12 (1 + υ ) ∂x∂y 2
Where, DP is known as flexural rigidity of the plate and is given by,
Et 3
DP = (6.1.10)
12 (1 − υ 2 )
Let consider the bending moments vary along the length and breadth of the plate as a function of x and
∂M x
y. Thus, if Mx acts on one side of the element, M='
Mx + dx acts on the opposite side. Considering
∂x
x
equilibrium of the plate element, the equations for forces can be obtained as
∂Qx ∂Qy
+ +p= 0 (6.1.11)
∂x ∂y
∂M x ∂M xy
+ =
Qx (6.1.12)
∂x ∂y
∂M xy ∂M y
+ =
Qy (6.1.13)
∂x ∂y
Using eq.6.1.9 in eqs.6.1.12 & 6.1.13, the following relations will be obtained.
∂ ∂2w ∂2w
Qx =
− DP 2 + 2 (6.1.14)
∂x ∂x ∂y
∂ ∂2w ∂2w
Qy = − DP 2 + 2 (6.1.15)
∂y ∂x ∂y
Using eqs. (6.1.14) and (6.1.15) in eq. (6.1.11) following relations will be obtained.
∂4w ∂4w ∂4w p
+ 2 + =
− (6.1.16)
∂x 4
∂x ∂y
2 2
∂y 4
DP
applicable for analysis of plates with smaller deformations, as higher order terms of strain-displacement
relationship cannot be neglected for large deformations. Moreover, as plate deflects its transverse
stiffness changes. Hence only for small deformations the transverse stiffness can be assumed to be
constant.
Contrary, Reissner–Mindlin plate theory (Fig. 6.1.5) is applied for analysis of thick plates, where
the shear deformations are considered, rotation and lateral deflections are decoupled. It does not require
the cross-sections to be perpendicular to the axial forces after deformation. It basically depends on
following assumptions,
1. The deflections of the plate are small.
2. Normal to the plate mid-surface before deformation remains straight but is not necessarily
normal to it after deformation.
3. Stresses normal to the mid-surface are negligible.
Thus, according to Mindlin plate theory, the deformation parallel to the undeformed mid surface, u and
v, at a distance z from the centroidal axis are expressed by,
8
𝑢 = 𝑧𝜃𝑦 (6.1.17)
v = −zθ𝑥 (6.1.18)
Where θx and θy are the rotations of the line normal to the neutral axis of the plate with respect to the x
and y axes respectively before deformation. The curvatures are expressed by
∂θ
χx = y (6.1.19)
∂x
∂θ
χy = − x (6.1.20)
∂y
Similarly the twist for the plate is given by,
∂θ ∂θ
=χ xy y − x (6.1.21)
∂y ∂x
Using eqs.6.1.9-6.1.10, the bending stresses for the plate is given by
Mx 1 υ 0 χx
Et 3
My = 2
υ 1 0 χy (6.1.22)
M 12 (1 − υ ) 1 − υ χ xy
xy 0 0
2
Or
{M } = [ D ]{ χ } (6.1.23)
Here "𝛼" is the numerical correction factor used to characterize the restraint of cross section against
warping. If there is no warping i.e., the section is having complete restraint against warping then α = 1
and if it is having no restraint against warping then α = 2/3. The value of α is usually taken to be π2/12 or
5/6. Now, the stress resultant can be combined as follows.
{M } [ D ] 0 { χ }
=
[ Ds ] {γ }
(6.1.30)
{Q} 0
Or,
𝑀𝑥 1 𝜇 0 0 0 χ
⎧ 𝑀 ⎫ ⎡ 𝐸𝑡 3 ⎤ ⎧ χ𝑥 ⎫
⎪ 𝑦 ⎪ ⎢12(1−𝜇2 ) �𝜇 1 0 � 0 0 ⎥⎪ 𝑦 ⎪
𝑀𝑥𝑦 = ⎢ 0 0
1−𝜇
0 0 χ (6.1.31)
⎥ 𝑥𝑦
⎨𝑄 ⎬ 2 ⎨ 𝛾𝑥 ⎬
⎪ 𝑥⎪ ⎢ 0 0 0 𝐸𝑡
�
𝛼 0 ⎥⎪ ⎪
� 𝛾
⎩ 𝑄𝑦 ⎭ ⎣ 0 0 0 2(1+𝜇) 0 𝛼 ⎦ ⎩ 𝑦 ⎭
Or,
𝜕𝜃𝑦
⎧ 𝜕𝑥 ⎫
𝑀𝑥 1 𝜇 0 0 0 ⎪ 𝜕𝜃𝑥 ⎪
⎧ 𝑀 ⎫ ⎡ 𝐸𝑡 3 𝜇 1 0 ⎤ ⎪ − 𝜕𝑦 ⎪
⎪ 𝑦 ⎪ ⎢12(1−𝜇2 ) � � 0 0 ⎪ ⎪
1−𝜇 ⎥ 𝜕𝜃𝑦 𝜕𝜃𝑥
𝑀𝑥𝑦 = ⎢ 0 0 2 0 0 ⎥ 𝜕𝑦 − 𝜕𝑥 (6.1.32)
⎨𝑄 ⎬ ⎨ ⎬
⎪ 𝑥⎪ ⎢ 0 0 0 𝐸𝑡
�
𝛼 0 ⎥ ⎪ 𝜃 + 𝜕𝑤 ⎪
�
⎩ 𝑄𝑦 ⎭ ⎣ 0 0 0 2(1+𝜇) 0 𝛼 ⎦ ⎪
𝑦 𝜕𝑥 ⎪
⎪ 𝜕𝑤⎪
⎩−𝜃𝑥 + 𝜕𝑦 ⎭
The above relation may be compared with usual stress-strain relation. Thus, the stress resultants and
their corresponding curvature and shear deformations may be considered analogous to stresses and
strains.
Similar to the above, the boundary conditions along x direction can also be obtained. Once the
displacements w(x,y) of the plate at various positions are found, the strains, stresses and moments
developed in the plate can be determined by using corresponding equations.
11
As nine displacement degrees of freedom present in the element, we need a polynomial with nine
independent terms for defining, w(x,y) . The displacement function is obtained from Pascal’s triangle by
choosing terms from lower order polynomials and gradually moving towards next higher order and so
on.
1
x y
x2 xy y2
x3 x2y xy2 y3
Thus, considering Pascal triangle, and in order to maintain geometric isotropy, we may consider the
displacement model in terms of the complete cubic polynomial as,
w ( x, y ) =α 0 + α1 x + α 2 y + α 3 x 2 + α 4 xy + α 5 y 2 + α 6 x 3 + α 7 ( x 2 y + xy 2 ) + α 8 y 3 (6.2.1)
12
∂2w
=2α 3 + 6α 6 x + 2α 7 y
∂x 2
∂2w
=2α 5 + 2α 7 x + 6α 8 y
∂y 2 (6.2.8)
∂ w
2
α 4 2α 7 ( x + y )
=+
∂x∂y
(6.2.10)
Further for isotropic material,
Mx 1 υ 0 χx
Et 3
My = 2
υ 1 0 χy (6.2.11)
M 12 (1 − υ ) 1 − υ χ xy
xy 0 0
2
Or
{M } [=
D ]{ χ } [ D ][ B ][ Φ ] {d }
−1
= (6.2.12)
Now the strain energy stored due to bending is
1 𝑎 𝑏 1 𝑎 𝑏
𝑈 = 2 ∫0 ∫0 [𝜒]𝑇 {𝑀}𝑑𝑥𝑑𝑦 = 2 ∫0 ∫0 {𝑑}𝑇 [𝜙 −1 ]𝑇 [𝐵]𝑇 [𝐷][𝐵][𝜙 −1 ]{𝑑}𝑑𝑥𝑑𝑦 (6.2.13)
Hence the force vector is written as
𝜕𝑈 𝑎 𝑏
{𝐹} =
𝜕{𝑑}
= [𝜙 −1 ]𝑇 ∫0 ∫0 [𝐵]𝑇 [𝐷][𝐵]𝑑𝑥𝑑𝑦 [𝜙 −1 ]{𝑑} = [𝑘]{𝑑} (6.2.14)
Thus, [k] is the stiffness matrix of the plate element and is given by
14
𝑎 𝑏
[𝑘] = [𝜙 −1 ]𝑇 ∫0 ∫0 [𝐵]𝑇 [𝐷][𝐵]𝑑𝑥𝑑𝑦 [𝜙 −1 ] (6.2.15)
For a triangular plate element with orientation as shown in Fig. 6.2.2, the stiffness matrix defined in
local coordinate system [k] can be transformed into global coordinate system.
[𝐾] = [𝑇]𝑇 [𝑘 ] [𝑇] (6.2.16)
Where, [K] is the elemental stiffness matrix in global coordinate system and [T] is the transformation
matrix given by
1 0 0 0 0 0 0 0 0
0 l m 0 0 0 0 0 0
x x
0 l y my 0 0 0 0 0 0
0 0 0 1 0 0 0 0 0
[T ] = 0 0 0 0 lx mx 0 0 0 (6.2.17)
0 0 0 0 l y my 0 0 0
0 0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 l x mx
0 0 0 0 0 0 0 l m
y y
Here, (lx , mx) and (ly, my) are the direction cosines for the lines OX and OY respectively as shown in Fig.
6.2.2.
Considering Pascal triangle, and in order to maintain geometric isotropy the following displacement
function is chosen for finite element formulation.
w(x, y) 0 1x 2 y 3 x 2 4 xy 5 y 2 6 x 3 7 x 2 y
(6.2.18)
8 xy 2 9 y3 10 x 3 y 11xy3
Hence Corresponding values for �𝜃𝑥 , 𝜃𝑦 �are,
w
x 2 4 x 2 5 y 7 x 2 2 8 xy 3 9 y 2 10 x 3 311xy 2 (6.2.19)
y
w
y 1 2 3 x 4 y 3 6 x 2 2 7 xy 8 y 2 310 x 2 y 11y3 (6.2.20)
x
The above can be expressed in matrix form
3
0
w 1 x y x 2 2 3 2 2 3 3
xy y x x y xy y x y xy
0 0 1 0 x 2y 0 x2 2xy 3y 2 x3 3xy 2 1
x
y 0 1 0 2x y 0 3x 2 2xy y 2 0 3x 2 y y3
11
(6.2.21)
In a similar procedure to three node plate bending element the values of {α} can be found from the
following relatios.
16
w1 1 0 0 0 0 0 0 0 0 0 0 0 0
x1 0 0 1 0 0 0 0 0 0 0 0 0 1
y1 0 1 0 0 0 0 0 0 0 0 0 0 2
w 2 1 0 b 0 0 b2 0 0 0 b3 0 0 3
0
0 1 0 0 2b 0 0 0 3b 2 0 0 4
x 2
y2 0 1 0 0 b 0 0 0 b 2 0 0 b3 5
w 3 1 a b a2 ab b 2 a3 a 2 b ab 2 b3 a 3b ab3 6
x3 0 0 1 0 a 2b 0 a2 2ab 3b 2 a3 3ab3 7
0
y3 1 0 2a b 0 3a 2 2ab b 2 0 3a 2 b b3 8
w 4 1 a 0 a2 0 0 a3 0 0 0 0 0 9
x 4 0 0 1 0 a 0 0 a2 0 0 a3 0 10
y4 0 1 0 2a 0 0 3a 2 0 0 0 0 0 11
(6.2.22)
Thus,
{α } = [Φ ] {d }
−1
(6.2.23)
Further,
𝜕2 𝑤
𝜒𝑥 ⎧− 𝜕𝑥 2 ⎫
⎪ 𝜕2 𝑤⎪
� 𝜒𝑦 � = − 𝜕𝑦 2 (6.2.24)
𝜒𝑥𝑦 ⎨ 2 ⎬
⎪ 2𝜕 𝑤 ⎪
⎩ 𝜕𝑥𝜕𝑦 ⎭
Where,
2w
2 3 6 6 x 2 7 y 610 xy
x 2
2w (6.2.25)
2 5 2 8 x 6 9 y 611xy
y 2
2w
4 2 7 x 2 8 y 310 x 2 311y 2
xy
Thus, putting values of eq. (6.2.25) in eq. (6.2.24), the following relation is obtained.
0
0 0 0 2 0 0 6x 2y 6xy
x
0 0 0
y 0 0 0 0 0 2 2x 6y 6xy
1 (6.2.26)
0 0 0
0 0 0 0 2 0 4y 6y 2
2
xy 0 4x 0 6x
11
Or,
B -1 d (6.2.27)
17
Further in a similar method to triangular plate bending element we can estimate the stiffness matrix for
rectangular plate bending element as,
Mx 1 υ 0 χx
Et 3
My = 2
υ 1 0 χy (6.2.28)
M 12 (1 − υ ) 1 − υ χ xy
xy 0 0
2
Or
{M } [=
D ]{ χ } [ D ][ B ][ Φ ] {d }
−1
= (6.2.29)
The bending strain energy stored is
1 𝑎 𝑏 1 𝑎 𝑏
𝑈 = 2 ∫0 ∫0 [𝜒]𝑇 {𝑀}𝑑𝑥𝑑𝑦 = 2 ∫0 ∫0 {𝑑}𝑇 [𝜙 −1 ]𝑇 [𝐵]𝑇 [𝐷][𝐵][𝜙 −1 ]{𝑑}𝑑𝑥𝑑𝑦 (6.2.30)
Hence, the force vector will become
𝜕𝑈 𝑎 𝑏
{𝐹} =
𝜕{𝑑}
= [𝜙 −1 ]𝑇 ∫0 ∫0 [𝐵]𝑇 [𝐷][𝐵]𝑑𝑥𝑑𝑦 [𝜙 −1 ]{𝑑} = [𝑘]{𝑑} (6.2.31)
Where, [k] is the stiffness matrix given by
𝑎 𝑏
[𝑘] = [𝜙 −1 ]𝑇 ∫0 ∫0 [𝐵]𝑇 [𝐷][𝐵]𝑑𝑥𝑑𝑦 [𝜙 −1 ] (6.2.30)
The stiffness matrix can be evaluated from the above expression. However, the stiffness matrix also can
be formulated in terms of natural coordinate system using interpolation functions. In such case, the
numerical integration needs to be carried out using Gauss Quadrature rule. Thus, after finding nodal
displacement, the stresses will be obtained at the Gauss points which need to extrapolate to their
corresponding nodes of the elements. By the use of stress smoothening technique, the various nodal
stresses in the plate structure can be determined.
18
6.3.1 Introduction
Finite element formulation of the thick plate will be similar to that of thin plate. The difference will be
the additional inclusion of energy due to shear deformation. Therefore, the moment curvature relation
derived in first lecture of this module for thick plate theory will be the basis of finite element
formulation. The relation is rewritten in the below for easy reference to follow the finite element
implementation.
𝑀𝑥 1 𝜇 0 0 0 χ𝑥
⎧ 𝑀 ⎫ ⎡ 𝐸𝑡 3 ⎤ ⎧ χ ⎫
⎪ 𝑦 ⎪ ⎢12(1−𝜇2 ) �𝜇 1 0 � 0 0 ⎥ ⎪ 𝑦⎪
𝑀𝑥𝑦 = ⎢ 0 0 2
1−𝜇
0 0 χ (6.3.1)
⎥ 𝑥𝑦
⎨𝑄 ⎬ ⎨ 𝛾𝑥 ⎬
⎪ 𝑥⎪ ⎢ 0 0 0 𝐸𝑡
�
𝛼 0 ⎥⎪ ⎪
� 𝛾
𝑄
⎩ 𝑦⎭ ⎣ 0 0 0 2(1+𝜇) 0 𝛼 ⎦ ⎩ 𝑦 ⎭
Or,
{M } [ D ] [ 0] { χ }
= (6.3.2)
{Q} [ 0] [ Ds ] {γ }
The above relation is comparable to stress-strain relations.
{σ }P = [C ]P {ε }P (6.3.3)
Where,
χ𝑥
⎧χ ⎫
⎪ 𝑦⎪
{𝜀}𝑃 = χ𝑥𝑦 = [𝐵]{𝑑𝑖 } (6.3.4)
⎨ 𝛾𝑥 ⎬
⎪ ⎪
⎩ 𝛾𝑦 ⎭
Where [B] is the strain displacement matrix and {di} is the nodal displacement vector. Thus, combining
eqs. (6.3.3) and (6.3.4), the following expression is obtained.
{σ }P = [C ]P [ B ]{di } (6.3.5)
Where, the shape function for the four node element is expressed as,
1
N i =(1 + ξ ξi )(1 + ηηi ) (6.3.7)
4
Here, ξi and ηi are the local coordinates ξ and η of the ith node. Using eq. (6.3.6), eq. (6.3.4) can be
rewritten as
4
∂N i
χ x = ∑ θ yi
i =1 ∂x
4
∂N i
χ=
y ∑
i =1
−θ xi
∂y
∂N i 4 4
∂N i
=χ xy yi
=i 1 =
∑θ
− ∑ θ xi
∂y i 1 ∂x
(6.3.8)
∂N4 4
=γx i
i
∂x
=i 1 =i 1
∑w
+ ∑ θ yi N i
4
∂N 4
=γ y ∑ wi i + ∑ θ xi N i
=i 1 = ∂y i 1
The above can be expressed in matrix form as follows:
∂N i
0 0
∂x
∂N i
χx 0 − 0
χ ∂y
y ∂N ∂N i
{ε }P =
= χ xy 0 − i {d } (6.3.9)
∂x ∂y i
γ
x ∂N i
γ y 0 Ni
∂x
∂N i − Ni 0
∂y
Here, for a four node quadrilateral element, the nodal displacement vector {di} will become
𝑇
{𝑑𝑖 } = �𝑤1 , 𝜃𝑥1 , 𝜃𝑦1 , 𝑤2 , 𝜃𝑥2 , 𝜃𝑦2 , 𝑤3 , 𝜃𝑥3 , 𝜃𝑦3 , 𝑤4 , 𝜃𝑥4 , 𝜃𝑦4 � (6.3.10)
Thus, the strain-displacement relationship matrix will be
20
∂N1 ∂N 2 ∂N 3 ∂N 4
0 0
∂x
0 0
∂x
0 0
∂x
0 0
∂x
0 ∂N1 ∂N 2 ∂N 3 ∂N 4
− 0 0 − 0 0 − 0 0 − 0
∂y ∂y ∂y ∂y
0 ∂N ∂N1 ∂N ∂N 2 ∂N ∂N 3 ∂N ∂N 4
[ B] = − 1
∂x ∂y
0 − 2
∂x ∂y
0 − 3
∂x ∂y
0 − 4
∂x ∂y
∂N1 ∂N 2 ∂N 3 ∂N 4
∂x 0 N1 0 N2 0 N3 0 N4
∂x ∂x ∂x
∂N1 − N1 0
∂N 2
− N2 0
∂N 3
− N3 0
∂N 4
− N4 0
∂y ∂y ∂y ∂y
(6.3.11)
Or,
[ B ] = [ B1 ](5×3) [ B2 ](5×3) [ B3 ](5×3) [ B4 ](5×3) (6.3.12)
Or,
[C ]P [ B ] = [CP B1 ](5×3) [CP B2 ](5×3) [CP B3 ](5×3) [CP B4 ](5×3) (6.3.14)
− µ t 2 ∂N i t 2 ∂N i
0 1 − µ ∂y 1 − µ ∂x
−t 2 ∂N i υ t ∂N i
2
0 1 − µ ∂x 1 − µ ∂y
Et −t 2 ∂N i t 2 ∂N i
[CBi ] =
(6.3.15)
12 (1 + µ ) 0 2 ∂x 2 ∂y
∂N i
6α ∂x
6α N i
0
∂N i
6α
∂y −6α N i 0
The bending and shear terms form above equation are separated and written as
21
t 2 N t 2 N i
i
1 y 1 x 0 0 0
0
0
0
t 2 N i t N i 0
2
0 0
Et 0 1 x 1 y 0
CBi 6 N
121 2
t 2
N i
t N x Ni
i
i
0
0
2 x
2 y
N i
0 0 0 y N i 0
0 0 0
(6.3.16)
The above expression can be written in compact form as
CBi = CBi b + CBi s (6.3.17)
Here, the contributions of bending and shear terms to stress displacement matrix is denoted as [CBi]b and
[CBi]s respectively. Generally, the contribution due to bending, [CBi]b in eq. (6.3.17) is evaluated
considering 2×2 Gauss points where as the shear contribution [CBi]s is evaluated considering 1×1 Gauss
point.
Using value of [C]P form eq. 6.3.1 and [Bi] from 6.3.11, the product of B Cp Bis evaluated as
T
22
N i N i
0 0 0
x y
N i N i
k B C B 0
T
0 N i
p y x
N N i
i 0 Ni 0
x y
i1,2,3,4
N i
0 0
x
0 0 N
1 0 0 i 0
Et 3
0 0 y
2 1 0
l 1 0 0 N N i
1
0 i
0 x y
0
2
N
0 0 0 Et 1 0 i 0 Ni
x
0 0 0 21 0 1
N i N 0
y i
i1,2,3,4
Or in short, (6.2.20)
k k12 k13 k14
11
k k 22 k 23 k 24
k BT C B 21 (6.2.21)
p k 32 k 33 k 34
k 33
k k 42 k 43 k 44
41
Where,
N i N i N i N i N i N i
6 6 Nj 6 Nj
x x y y y y
t 2 N N j t N N j
2
i i
N i 1 y y 1 y x
k ij Et
121 6N i
y t 2 N N
t 2 N N
j
i j
6N i N j i
2 x x 2 x y
t 2 N N j t 2 N N j
i i
N i 1 x y 1 x x
6N i
x t 2 N N t 2 N N
j
i i
j
2 y x 2 y y
23
(6.3.22)
By separating the bending and shear terms from above equation,
0
0 0
t N N j t N N j
t N i N j t N i N j
2 2 2 2
k ij Et i i
1 1 2 0 1 y y 2 x x
1 y x 2 x y
t N N j t N N j t 2 N N j t 2 N N j
2 2
i i i i
0 1 x y 2 y x 1 x x 2 y y
N N N i N i N N i
i i
i Nj Nj
x x y y y x
6 0
N i
Ni Ni N j
y
N i
N 0 N N
x
j i j
(6.3.23)
Thus, the matrix k can now be written as the sum of bending and shear contributions
k k k
b s (6.2.24)
Or,
k k12 k13 k14 k11 k12 k13 k14
11 b b b b s s s s
k 21 k 22 k 23 k 24 k 21 k 22 k 23 k 24
k b b b b s
s s s (6.2.25)
k
33 b k 32 k 33 k 34 k 33 k 32 k 33 k 34
b b b s s s s
k
41 b k 42 k 43 k 44 k 41 k 42 k 43 k 44
b b b s s s s
The stiffness matrix k can be evaluated from the following expression by substituting
k for BT C B in eq. (6.3.19) and is given as
p
1 1
k 1
1 k J d d (6.2.26)
Here, J is the determinate of the Jacobian matrix. The Gauss Quadrature integration rule is used to
compute the stiffness matrix [k].
The nodal load vectors from each element are assembled to find the global load vector at all the nodes.
25
6.4.1 Introduction
Skew plates often find its application in civil, aerospace, naval, mechanical engineering structures.
Particularly in civil engineering fields they are mostly used in construction of bridges for dealing
complex alignment requirements. Analytical solutions are available for few simple problems. However,
several alternatives are also available for analyzing such complex problems by finite element methods.
Commonly used three discretization methods for skew plates are shown in Fig 6.4.1.
If the skew plate is discretized using only rectangular plate elements, the area of continuum excluded
from the finite element model may be adequate to provide incorrect results. Another method is to use
combination of rectangular and triangular elements. However, such analysis will be complex and it may
not provide best solution in terms of accuracy as, different order of polynomials is used to represent the
field variables for different types of elements. Another alternative exist using skew element in place of
rectangular element.
It is important to note that the terms (ξ, η) in the above equations represent the absolute coordinate of the
point “P” in skew coordinate system, not the natural coordinates. Since the above element has four
corner nodes and each node have three degrees of freedom present, a polynomial with minimum 12
independent terms are necessary for defining the displacement function w(x,y). Considering Pascal
triangle, and in order to maintain geometric isotropy, the displacement function may be considered as
follows:
w 0 1 2 3 2 4 5 2 63 7 2
(6.4.3)
82 + 93 103 113
Hence corresponding values for �𝜃𝑥 , 𝜃𝑦 �are,
w
2 4 2 5 7 2 2 8 3 9 2 103 311 2 (6.4.4)
w
1 2 3 4 3 6 2 2 7 8 2 310 2 113 (6.4.5)
Or, in matrix form,
w 1 2 2
3 2 2 3 3 3 0
0 0 1 0 2 0 2 3
2 2
3 3 2 1
0 1 0 2 0 3 2 2 2 0 3 2 3
11
(6.4.6)
Or,
28
{d } = [ Φ ]{α } (6.4.7)
The value of [α] can be determined using value of �𝑤, 𝜃𝜉 , 𝜃𝜂 � at four nodes as
w1 1 0 0 0 0 0 0 0 0 0 0 0
0
x1 0 0 1 0 0 0 0 0 0 0 0 1
0
y1 0 1 0 0 0 0 0 0 0 0 0 2
0
w 2 1 0 b 0 0 b2 0 0 0 b3 0 0 3
0
0 1 0 0 2b 0 0 0 3b 2 0 0 4
x 2
y2 0 1 0 0 b 0 0 0 b 2 0 0 b3 5
w 3 1 a b a2 ab b 2 a3 a 2 b ab 2 b3 a 3b ab3 6
x3 0 0 1 0 a 2b 0 a2 2ab 3b 2 a3 3ab3 7
0
y3 1 0 2a b 0 3a 2 2ab b 2 0 3a 2 b b3 8
w 4 1 a 0 a2 0 0 a3 0 0 0 0 0 9
x 4 0 0 1 0 a 0 0 a2 0 0 a3 0 10
y4 0 1 0 2a 0 0 3a 2 0 0 0 0 0 11
(6.4.8)
Or,
{α } = [ Φ ] {di }
−1
(6.4.9)
Further, considering eq. (6.4.2)
0, cos ec
x y
(6.4.10)
1, cot
x y
Again,
𝜕2 𝑤
𝜒𝑥 ⎡− 𝜕𝑥 2 ⎤
⎢ 𝜕2 𝑤⎥
� 𝜒𝑦 � = ⎢− 𝜕𝑦 2 ⎥ (6.4.11)
𝜒𝑥𝑦 ⎢ 2𝜕2 𝑤 ⎥
⎣ 𝜕𝑥𝜕𝑦 ⎦
The values in right hand side of the eq. (6.4.11) can be calculated by using chain rule as,
w w w w
. (6.4.12)
x x x
Therefore,
2w 2w
2 (6.4.13)
x 2
29
Similarly,
w w w w w
cos cosec
y y y
Thus, further derivation provides
2w 2 w
2
2w 2w
cosec 2 cot 2 2cos cosec
2
(6.4.14)
y 2
And,
2w w 2w 2w
cot 2 cosec
2
xy x y
(6.4.15)
Hence eq. (6.4.11) is converted to
2 w 2 w
2 2
x 1
2 w 0 0
2 w
cot 2 cosec 2 -cotcosec
2
(6.4.16)
y 2
2 -2cot 0 cosec 2 2
2 w 2 w
xy
Or,
x,y H , (6.4.17)
Further, by partial differentiation of eq. (6.4.3),
2w
2 3 6 6 2 7 610
2
2w
2 2 5 2 8 6 9 611 (6.4.18)
2w
2 4 7 4 8 610 2 6112
Or, in matrix form,
30
2 w
2
0 0 0 2 0 0 6 2 0 0 6 0 0
2 w
2
0 0 0 0 0 2 0 0 2 6 0 6
2 0 0 0 0 0 0 4 4 6 2 11
62
0 0
2 w
(6.4.19)
Or,
, B B1 di
Or,
x,y H B1 di (6.4.20)
Again,
M x,y D x,y (6.4.21)
Where [D] for plane stress condition is
1 0
Eh 3
D 1 0 (6.4.22)
l21 2
1
0 0
2
Using eq. (6.4.20) in eq. (6.4.21),
M x,y D H B A1 d (6.4.23)
The expression for bending strain energy stored,
a b
0 0
Where,
31
x y
x, y 1 0
J
x
sin (6.4.26)
, y cos sin
k sin 1 B, H D H B, dd 1
T T T
0 0
(6.4.28)
Thus, the element stiffness matrix of a skew element for plate bending analysis can be evaluated from
the above expression using Gauss Quadrature numerical integration.
32
6.5.1 Introduction
The finite strip method (FSM) was first developed by Y. K. Cheung in 1968. This is an efficient tool for
analyzing structures with regular geometric platform and simple boundary conditions. If the structure is
regular, the whole structure can be idealized as an assembly of 2D strips or 3D prisms. Thus the
geometry of the structure needs to be constant along one or two coordinate directions so that the width
of the strips or the cross-section of the prisms does not change. Therefore, the finite strip method can
reduce three and two-dimensional problems to two and one-dimensional problems respectively. The
major advantages of this method are (i) reduction of computation time, (ii) small amount of input (iii)
easy to develop the computer code etc. However, this method will not be suitable for irregular geometry,
material properties and boundary conditions.
The strips are assumed to be connected to each other along a discreet number of nodal lines that coincide
with the longitudinal boundaries of the strip. The general form of the displacement function in two
dimensions for a typical strip is given by
𝑤 = 𝑤(𝑥, 𝑦) = ∑𝑛𝑚=1 𝑓𝑚 (𝑦)𝑋𝑚 (𝑥) (6.5.1)
Here, the functions ƒm(y) are polynomials and the functions Xm(x) are trigonometric terms that satisfy the
end conditions in the x direction. The functions Xm(x) can be taken as basic functions (mode shapes) of
the beam vibration equation.
33
𝑑4 𝑋 𝜇4
𝑑𝑦 4
− 𝐿4 𝑌 = 0 (6.5.2)
Here L is the length of beam strip and µ is a parameter related to material, frequency and geometric
properties. The general solution of the above equation will become
𝑛𝜋𝑥 𝑛𝜋𝑥 𝑛𝜋𝑥 𝑛𝜋𝑥
𝑋𝑚 (𝑥) = 𝐶1 sin � 𝐿
� + 𝐶2 cos � 𝐿
� + 𝐶3 sinh � 𝐿
� + 𝐶4 cosh � 𝐿
� (6.5.3)
Four conditions at the boundaries are necessary to determine the coefficients C1 to C4 in the above
expression.
The formulation of the finite strip method is similar to that of the finite element method. For example,
for a strip subjected to bending, the moment curvature relation will become
𝜕2 𝑤
− 𝜕𝑥 2 ⎫
𝑀𝑥 𝐷𝑥 𝐷1 0 ⎧ ⎪ 𝜕2 𝑤 ⎪
� 𝑀𝑦 � = �𝐷1 𝐷𝑦 0 � − 2
𝜕𝑦
(6.5.15)
𝑀𝑥𝑦 0 0 2𝐷𝑥𝑦 ⎨ 𝜕2 𝑤 ⎬
⎪2 ⎪
⎩ 𝜕𝑥𝜕𝑦⎭
Where Mx , My and Mxy are moments per unit length and [D] is the elasticity matrix. From eq. (6.5.7)
the following expressions are evaluated to incorporate in the above equation.
𝜕2 𝑤 𝑛𝜋 2 𝑛𝜋𝑥
− 𝜕𝑥 2 = ∑𝑚
𝑛=1 𝑤(𝑦) � 𝐿 � 𝑠𝑖𝑛 𝐿
𝜕2 𝑤 𝑛𝜋𝑥
− 𝜕𝑦 2 = − ∑𝑚
𝑛=1 −(2𝛼2 + 6𝛼3 𝑦)𝑠𝑖𝑛 𝐿
(6.5.16)
𝜕2 𝑤 𝑛𝜋 𝑛𝜋𝑥
2 𝜕𝑥𝜕𝑦 = ∑𝑚 2
𝑛=1 2(𝛼1 + 2𝛼2 𝑦 + 3𝛼3 𝑦 ) � 𝐿 � 𝑐𝑜𝑠 𝐿
The above expression is written in matrix form in the below
𝜕2 𝑤
⎧ − 𝜕𝑥 2 ⎫
⎪ 𝜕2 𝑤 ⎪ ∑ 𝑤(𝑦)𝑁 2 𝑠𝑖𝑛𝑁𝑥
− 2 = �− ∑ −(2𝛼2 + 6𝛼3 𝑦)𝑠𝑖𝑛𝑁𝑥 � (6.5.17)
⎨ 𝜕𝑦 2
⎬ 2
∑ 2(𝛼1 + 2𝛼2 𝑦 + 3𝛼3 𝑦 )𝑁𝑐𝑜𝑠𝑁𝑥
⎪2 𝜕 𝑤 ⎪
⎩ 𝜕𝑥𝜕𝑦⎭
𝑛𝜋
Here, 𝐿 is denoted as N. Rearranging the above expression, one can find the following.
𝜕2 𝑤
⎧ − 𝜕𝑥 2 ⎫
⎪ 𝜕2 𝑤 ⎪ sin 𝑁𝑥 0 0 𝑤(𝑦)𝑁 2
− 𝜕𝑦 2 = � 0 sin 𝑁𝑥 0 �� −(2𝛼2 + 6𝛼3 𝑦) �
⎨ 2 ⎬ 0 0 cos 𝑁𝑥 2(𝛼1 + 2𝛼2 𝑦 + 3𝛼3 𝑦 )𝑁
2
⎪2 𝜕 𝑤 ⎪
⎩ 𝜕𝑥𝜕𝑦⎭
𝛼0
sin 𝑁𝑥 0 0 𝑁2 𝑁2𝑦 𝑁2𝑦2 𝑁2𝑦3
𝛼
=� 0 sin 𝑁𝑥 0 ��0 0 −2 −6𝑦 � �𝛼1 � (6.5.18)
2
0 0 cos 𝑁𝑥 0 2𝑁 4𝑁𝑦 6𝑁𝑦 2 𝛼3
Thus, in short, the curvature and moment equation will become
{𝜒} = [𝐻(𝑥)][𝐵(𝑦)]{𝛼} = [𝐻(𝑥)][𝐵(𝑦)][𝐴]−1 {𝑑} (6.5.19)
{𝑀} = [𝐷]{𝜒} = [𝐷][𝐻(𝑥)][𝐵(𝑦)][𝐴] {𝑑} −1
(6.5.20)
Now, the strain energy for the bending element can be written similar to plate bending formulation.
𝐿 𝑏 𝐿 𝑏
1 1
𝑈 = � �{𝜒}𝑇 {𝑀} 𝑑𝑥 𝑑𝑦 = � �{𝛼}𝑇 [𝐴−1 ]𝑇 [𝐵(𝑦)]𝑇 [𝐻(𝑥)]𝑇 [𝐷][𝐻(𝑥)][𝐵(𝑦)][𝐴]−1 {𝑑} 𝑑𝑥 𝑑𝑦
2 2
0 0 0 0
(6.5.21)
Thus the force vector can be derived as
36
𝜕𝑈 𝐿 𝑏
{𝐹} =
𝜕{𝑑}
= ∫0 ∫0 [𝐴−1 ]𝑇 [𝐵(𝑦)]𝑇 [𝐻(𝑥)]𝑇 [𝐷][𝐻(𝑥)][𝐵(𝑦)][𝐴]−1 𝑑𝑥 𝑑𝑦 {𝑑} = [𝑘]{𝑑}
(6.5.22)
Thus, the stiffness matrix of a strip element can be obtained from the following expression.
𝐿 𝑏
[𝑘] = ∫0 ∫0 [𝐴−1 ]𝑇 [𝐵(𝑦)]𝑇 [𝐻(𝑥)]𝑇 [𝐷][𝐻(𝑥)][𝐵(𝑦)][𝐴]−1 𝑑𝑥 𝑑𝑦
𝐿 𝑏
= [𝐴−1 ]𝑇 ∫0 ∫0 [𝐵(𝑦)]𝑇 [𝐻(𝑥)]𝑇 [𝐷][𝐻(𝑥)][𝐵(𝑦)] 𝑑𝑥 𝑑𝑦 [𝐴]−1 (6.5.23)
𝐿
The stiffness matrix [k] can be simplified by integrating the term∫0 [𝐻(𝑥)]𝑇 [𝐷][𝐻(𝑥)] 𝑑𝑥 as follows.
𝐿
∫0 [𝐻(𝑥)]𝑇 [𝐷][𝐻(𝑥)] 𝑑𝑥
sin 𝑁𝑥 0 0 𝐷𝑥 𝐷1 0 sin 𝑁𝑥 0 0
𝐿
= ∫0 �
0 sin 𝑁𝑥 0 � �𝐷1 𝐷𝑦 0 �� 0 sin 𝑁𝑥 0 � 𝑑𝑥
0 0 cos 𝑁𝑥 0 0 2𝐷𝑥𝑦 0 0 cos 𝑁𝑥
𝐷𝑥 sin 𝑁𝑥 𝐷1 sin 𝑁𝑥 0 sin 𝑁𝑥 0 0
𝐿
= ∫0 � 𝐷1 sin 𝑁𝑥 𝐷𝑦 sin 𝑁𝑥 0 �� 0 sin 𝑁𝑥 0 � 𝑑𝑥
0 0 2𝐷𝑥𝑦 cos 𝑁𝑥 0 0 cos 𝑁𝑥
𝐷𝑥 sin2 𝑁𝑥
𝐷1 sin2 𝑁𝑥 0
𝐿 2 2
= ∫0 � 𝐷1 sin 𝑁𝑥
𝐷𝑦 sin 𝑁𝑥 0 � 𝑑𝑥 (6.5.24)
2
0 0 2𝐷𝑥𝑦 cos 𝑁𝑥
Here, the terms Dx, Dy and Dxy are constant and not varied with x or y. Following integrations are carried
out to simplify the above expression further.
𝐿 𝐿 1−𝑐𝑜𝑠2𝑁𝑥 1 𝑠𝑖𝑛2𝑁𝑥 𝐿 1 𝑠𝑖𝑛2𝑁𝐿
Now ∫0 𝑠𝑖𝑛2 𝑁𝑥 𝑑𝑥 = ∫0 � 2
� 𝑑𝑥 = 2 �𝑥 − 2𝑁
� = 2 �𝐿 − 2𝑁
�
0
𝑛𝜋
𝑛𝜋 𝐿 1 𝑠𝑖𝑛2 𝐿 1
Putting, 𝑁 = 𝐿
finally ∫0 𝑠𝑖𝑛2 𝑁𝑥 𝑑𝑥 will become 2 �𝐿 − 𝑛𝜋
𝐿
� = 2𝐿
2
𝐿
𝐿 𝐿 1+𝑐𝑜𝑠2𝑁𝑥 1 𝑠𝑖𝑛2𝑁𝑥 𝐿 1
Similarly; ∫0 𝑐𝑜𝑠 2 𝑁𝑥 𝑑𝑥 = ∫0 � 2
� 𝑑𝑥 = 2 �𝑥 + 2𝑁
� =2𝐿
0
Thus,
𝐷𝑥 𝐷1 0
𝐿 𝐿 𝐿
∫0 [𝐻(𝑥)]𝑇 [𝐷][𝐻(𝑥)] 𝑑𝑥 = 2 �𝐷1 𝐷𝑦 0 �=
2
[𝐷] (6.5.25)
0 0 2𝐷𝑥𝑦
Using eq. (6.5.25), the expression for stiffness matrix [k] in eq. (6.5.23) is simplified as follows.
𝑏 𝐿
[𝑘] = [𝐴−1 ]𝑇 ∫0 [𝐵(𝑦)]𝑇 [𝐷][𝐵(𝑦)] 𝑑𝑦 [𝐴−1 ]
2
2
𝑁 0 0
⎡ 2 ⎤ 𝐷𝑥 𝐷1 0 𝑁2 𝑁2𝑦 𝑁2𝑦2 𝑁2𝑦3
𝐿 𝑏 𝑁 𝑦 0 2𝑁
= 2 [𝐴−1 ]𝑇 ∫0 ⎢ 2 2 ⎥ �𝐷1 𝐷𝑦 0 �� 0 0 −2 −6𝑦 � 𝑑𝑦[𝐴]−1
⎢𝑁 𝑦 −2 4𝑁𝑦 ⎥
0 0 2𝐷𝑥𝑦 0 2𝑁 4𝑁𝑦 6𝑁𝑦 2
⎣𝑁 2 𝑦 3 −6𝑦 6𝑁𝑦 2 ⎦
37
𝑁 2 𝐷𝑥 𝑁 2 𝐷1 0
𝑏⎡ ⎤ 2 𝑁2𝑦
𝐿 𝑁 2 𝑦𝐷𝑥 𝑁 2 𝑦𝐷1 4𝑁𝐷𝑥𝑦 ⎥ 𝑁 𝑁2𝑦2 𝑁2𝑦3
⎢
= [𝐴−1 ]𝑇 � ⎢ 2 2 �0 0 −2 −6𝑦 � 𝑑𝑦[𝐴]−1
2 𝑁 𝑦 𝐷𝑥 − 2𝐷1 𝑁 2 𝑦 2 𝐷1 − 2𝐷𝑦 8𝐷𝑥𝑦 𝑁𝑦 ⎥
𝑜 ⎢ 2 3 ⎥ 0 2𝑁 4𝑁𝑦 6𝑁𝑦 2
⎣𝑁 𝑦 𝐷𝑥 − 6𝑦𝐷1 𝑁 2 𝑦 3 𝐷1 − 6𝑦𝐷𝑦 12𝑁𝑦 2 𝐷𝑥𝑦 ⎦
(𝑁 4 𝐷𝑥 ) (𝑁 4 𝑦𝐷𝑥 ) (𝑁 4 𝑦 2 𝐷𝑥 − 2𝑁 2 𝐷1 ) (𝑁 4 𝑦 3 𝐷𝑥 − 6𝑁 2 𝑦𝐷1 )
⎡ ⎤
⎢ (𝑁 4 𝑦𝐷𝑥 ) �𝑁 4 𝑦 2 𝐷𝑥 + 8𝑁 2 𝐷𝑥𝑦 � �𝑁 4 𝑦 3 𝐷𝑥 − 2𝑁 2 𝑦𝐷1 + 16𝑁 2 𝑦𝐷𝑥𝑦 � �𝑁 4 𝑦 4 𝐷𝑥 − 6𝑁 2 𝑦 2 𝐷1 + 24𝑁 2 𝑦 2 𝐷𝑥𝑦 �⎥
𝑏⎢ ⎥
𝐿 𝑁 4 𝑦 4 𝐷𝑥 − 4𝑁 2 𝑦 2 𝐷1 𝑁 4 𝑦 5 𝐷𝑥 − 2𝐷1 𝑁 2 𝑦 3 − 6𝑁 2 𝑦 3 𝐷1
⎢ � � � � ⎥
= [𝐴−1 ]𝑇 � ⎢ +4𝐷 𝑦 + 32𝑁 2 2
𝑦 𝐷 𝑥𝑦 −12𝑦𝐷 𝑦 + 48𝑁 2 3
𝑦 𝐷 𝑥𝑦 ⎥ 𝑑𝑦[𝐴]
−1
2 (𝑁 4 2
𝑦 𝐷 − 2𝑁 2
𝐷 ) �𝑁 4 𝑦 3 𝐷𝑥 − 2𝑁 2 𝑦𝐷1 + 8𝑁 2 𝑦𝐷𝑥𝑦 �
0 ⎢ 𝑥 1 4 5
𝑁 𝑦 𝐷𝑥 − 6𝑁 𝑦 𝐷1 2 3 4 6
𝑁 𝑦 𝐷𝑥 − 6𝑁 𝑦 𝐷1 2 4 ⎥
⎢(𝑁 4 𝑦 3 𝐷𝑥 − 6𝑁 2 𝑦𝐷1 ) �𝑁 4 𝑦 4 𝐷𝑥 − 6𝑁 2 𝑦 2 𝐷1 + 24𝑁 2 𝑦 2 𝐷𝑥𝑦 � 2 3 2 4 2 ⎥
⎢ � −2𝑁 𝑦 𝐷 1 + 12𝑦𝐷 𝑦 � � −6𝑁 𝑦 𝐷1 + 36𝑦 𝐷𝑦 � ⎥
2 3 2 4
⎣ +48𝑁 𝑦 𝐷𝑥𝑦 +72𝑁 𝑦 𝐷𝑥𝑦 ⎦
(6.5.26)
Thus, by putting the assumed shape function, the stiffness matrix of a strip element can be evaluated
numerically using Gaussian Quadrature or other numerical integration methods.
38
6.6.1 Introduction
A shell is a curved surface, which by virtue of their shape can withstand both membrane and bending
forces. A shell structure can take higher loads if, membrane stresses are predominant, which is primarily
caused due to in-plane forces (plane stress condition). However, localized bending stresses will appear
near load concentrations or geometric discontinuities. The shells are analogous to cable or arch structure
depending on whether the shell resists tensile or, compressive stresses respectively. Few advantages
using shell elements are given below.
1. Higher load carrying capacity
2. Lesser thickness and hence lesser dead load
3. Lesser support requirement
4. Larger useful space
5. Higher aesthetic value.
The example of shell structures includes large-span roof, cooling towers, piping system, pressure vessel,
aircraft fuselage, rockets, water tank, arch dams, and many more. Even in the field of biomechanics,
shell elements are used for analysis of skull, Crustaceans shape, red blood cells, etc.
Depending upon deflection in transverse direction due to transverse shear force per unit length, the shell
can be classified into structurally thin or thick shell. Further, depending upon the thickness of the shell
in comparison to the radii of curvature of the mid surface, the shell is referred to as geometrically thin or
thick shell. Typically, if thickness to radii of curvature is less than 0.05, then the shell can be assumed as
a thin shell. For most of the engineering application the thickness of shell remains within 0.001 to 0.05
and treated as thin shell.
This approach has the advantage of being independent of any particular shell theory. This approach can
be used to formulate a general shell element for geometric and material nonlinear analysis. Such element
has been employed very successfully when used with 9 or, in particular, 16 nodes. However, the 16-
node element is quite expensive in computation. In a degenerated shell model, the numbers of unknowns
present are five per node (three mid-surface displacements and two director rotations). Moderately thick
shells can be analysed using such elements. However, selective and reduced integration techniques are
necessary to use due to shear locking effects in case of thin shells. The assumptions for degenerated
shell are similar to the Reissner-Mindlin assumptions.
x
xi
V
8
i
y N xh
, i
y V3i
(6.6.3)
i 1
2
z
zi
Where,
xi x xi xi xi
1 i
yi yi yi and, V3i yi yi (6.6.4)
2
zi
zi
top
zi
bottom
zi
top
zi bottom
For small thickness, the vector V3i can be represented as a unit vector tiv3i:
x xi
y
8
V
, yi ti v3i
N i xh (6.6.5)
i 1
2
z
zi
th
Where, ti is the thickness of shell at i node. In a similar way, the displacement at any point of the shell
element can be expressed in terms of three displacements and two rotation components about two
orthogonal directions normal to nodal load vector V3i as,
u ui
8
Vti ai
v N i xh
, vi v1i v2i (6.6.6)
i 1
2 b
i
w
wi
Where, (𝛼𝑖 , 𝛽𝑖 ) are the rotations of two unit vectors v1i & v2i about two orthogonal directions normal to
nodal load vector V3i.The values of v1i and v2i can be calculated in following way:
The coordinate vector of the point to which a normal direction is to be constructed may be defined as
x xiˆ yjˆ zkˆ (6.6.7)
43
In which, iˆ, ˆj , kˆ are three (orthogonal) base vectors. Then, V1i is the cross product of iˆ & V3i as shown
below.
V1i iˆV3i & V2i V3i V1i (6.6.8)
and,
V1i & v2i V2i
v1i (6.6.9)
V1i V2i
u v w
N i
N
N
V i
V i
xxx
xx
x
b1 a1
T T
u v w N N i t v N i
i ui vi wi i 2i V
8 8 8
tv
b 2 i 1i V a
hhhhh
2
i 1 2 h
2
i1 i 1
u v w
0
N b 3 i
N
a3 i
i
i
V V V
(6.6.14)
It may be important to note that the constitutive relation expressed in eq. (6.6.19) is same as for the case
of plane stress formulation. Also, eq. (6.6.20) with a multiplication of thickness h is similar to the terms
corresponds to shear force in case of plate bending problem.
7.1.1 Introduction
There are two types of structural failure associated: namely (i) material failure and (ii) geometry or
configuration failure. In case of material failure, the stresses exceed the permissible values which may
lead to formation of cracks In configuration failure, the structure is unable to maintain its designed
configuration under the external disturbance even though the stresses are in permissible range. The
stability loss due to tensile forces falls in broad category of material instability. The loss of stability
under compressive load is termed as structural or geometric instability which is commonly known as
buckling. Thus, buckling is considered by a sudden failure of a structural element subjected to large
compressive stress, where the actual compressive stress at the point of failure is less than the ultimate
allowable compressive stresses. Buckling is a wide term that describes a range of mechanical
behaviours. Generally, it refers to an incident where a structural element in compression deviates from a
behaviour of elastic shortening within the original geometry and undergoes large deformations involving
a change in member shape for a small increase in force. Various types of buckling may occur such as
flexural buckling, torsional buckling, torsional flexural buckling etc. Here, the use of finite element
technique for elastic stability analysis of various structural members will be briefly discussed.
x + (7.1.1)
x 2 x x
Here, u and v are the displacements in local X and Y directions respectively. Here, the strain-
displacement relation is nonlinear. The total potential energy in the member with uniform cross sectional
area and subjected to external forces can be written as
L
EA
=Π
2 0
ε x 2 dx - P1v1 + P2 v2 (7.1.2)
Where, P1 and P2 are the external forces in nodes 1 and 2, v1 and v2 are the vertical displacements in
nodes 1 and 2, E and A are the modulous of elasticity and cross sectional area respectively (Fig. 7.1.1).
The length of the member is considered to be L.
2
EA u 1 u v
2
L 2 2
L u 2 v 2
EA u u
3 2
2 2
+ + dx - P1v1 + P2 v2
v u 1
2 0 x x
=
x x 4 x x
(7.1.3)
Neglecting higher order terms and considering EA u = P the above expression can simplify to the
x
following form.
EA u P v
L 2 L 2
2 0 x 2 0 x
=Π dx + dx - P1v1 + P2 v2 (7.1.4)
Considering nodal displacements at nodes 1 and 2 as u1, v1 and u2, v2, the displacement at any point
inside the element can be represented in terms of its interpolation functions and nodal displacements.
The shape function for a two node truss element is
[N ] = 1 − x x
(7.1.5)
L L
Equation expressed in eq. (7.1.5) is expressed in terms of the nodal displacement vectors {ui} and {vi}
with the help of interpolation function.
L L
1 P
=
Π ui N ' EA N 'ui dx + vi N ' N 'vi dx - P1v1 + P2 v2
T T T T
(7.1.6)
2 0 2 0
Where,
u v
= N ui = N' ui and = N vi = N' vi (7.1.7)
x x x x
Making potential energy (eq. 7.1.6) stationary one can find the equilibrium equation as
3
L L
Or,
F = k A ui P kG vi (7.1.9)
Where, [kA] is the axial stiffness of the member and [kG] is the geometric stiffness of the member in its
local coordinate system and can be derived as follows.
1
-
L 1 1 AE 1 -1
L
1
-
L 1 1 1 1 -1
L
kG = N ' N ' dx = 0 L -
dx =
T
(7.1.11)
-1 1
0 1 L L L
L
However, in a generalised form to accommodate both the direction, these stiffness matrices can be
written as
1 0 1 0 0 0 0 0
AE 0 0 0 0 1 0 1 0 1
k A = and kG = (7.1.12)
L 1 0 1 0 L 0 0 0 0
0 0 0 0 0 1 0 1
The generalised stiffness matrix of a plane truss member in global coordinate system can be derived
using the transformation matrix as derived module 4, lecture 1. The transformation matrix can be
recalled and written here as follows.
cos θ sinθ 0 0
− sinθ cos θ 0 0
[T ] =
0 0 cos θ sinθ
0 0 − sinθ cos θ
(7.1.13)
Thus, the stiffness matrices with respect to global coordinate system will become
K A = T k A T and KG = T kG T
T T
(7.1.14)
Here, [KA] and [KG] are the axial stiffness and geometric stiffness of the member in global coordinate
system. The force-displacement relationship in global coordinate system can be written from eq. (7.1.9)
as
4
Here, w is the transverse displacement and I is the moment of inertia of the member.
Considering nodal displacements at nodes 1 and 2 as w1, θ1 and w2, θ2, the displacement at any point
inside the element can be represented in terms of its interpolation functions and nodal displacements.
𝑤1
𝜃
𝑤 = [𝑁1 𝑁2 𝑁3 𝑁4 ] �𝑤1 � = [𝑁]{𝑑} (7.1.18)
2
𝜃2
Thus, the above energy equation can be rewritten as
1 L N " EI N " d + d P N ' T N ' d dx
2 0
T T
=Π d
(7.1.19)
Where,
5
d 2w d 2 d wd
2
= 2 N d = N" d and = N d = N' d (7.1.20)
dx dx dx dx
Applying the variational principle one can express
=
∂Π
∂ {d }
T
(
{F } = ∫0 N " EI [ N "] + P [ N '] [ N '] dx {d }
l T
) (7.1.21)
Thus, the stiffness matrix will be obtained as follows which have two terms.
[k ] = (
∫0 N EI [ N "]
l
" T
) dx +P∫ ([ N '] [ N ']) dx =
l
0
T
[k ] + P [k ] F G (7.1.22)
The first term resembles ordinary stiffness matrix for the bending of a beam. So this matrix is called
flexural stiffness matrix. The second matrix is known as geometric stiffness matrix as it only depends on
the geometrical parameters. Thus, the flexural stiffness matrix [kF] and geometric stiffness matrix [kG]
can be derived from the following expressions.
[ kF ] ∫=
0 ( T
0 )
N " EI [ N "] dx and [ kG ] ∫ ([ N '] [ N ']) dx
l l T
(7.1.23)
The above matrices can be derived from the assumed interpolation function. For example, the
interpolation functions for a two node beam element are expressed by the following equation.
𝑥2 𝑥3 𝑥2 𝑥3 𝑥2 𝑥3 𝑥2 𝑥3
[𝑁] = ��1 − 3 2 + 2 3 � , �𝑥 − 2 + 𝐿2 � , �3 𝐿2 − 2 𝐿3 � , �− + 𝐿2 �� (7.1.24)
𝐿 𝐿 𝐿 𝐿
Now, the first and second order derivative of the above function will become
𝑥 𝑥2 𝑥 𝑥2 𝑥 𝑥2 𝑥 𝑥2
[𝑁 ′ ] = ��−6 2 + 6 3 � , �1 − 4 + 3 2 � , �6 2 − 6 3 � , �−2 + 3 2 �� (7.1.25)
𝐿 𝐿 𝐿 𝐿 𝐿 𝐿 𝐿 𝐿
6 𝑥 4 𝑥 6 𝑥 2 𝑥
�𝑁 " � = ��− 𝐿2 + 12 𝐿3 � , �− 𝐿 + 6 𝐿2 � , �𝐿2 − 12 𝐿3 � , �− 𝐿 + 6 𝐿2 �� (7.1.26)
Using the above expressions in eq. (7.1.23), flexural and geometric stiffness matrices can be derived and
obtained as follows.
12 6𝐿 −12 6𝐿
2
𝐸𝐼
[𝐾𝐹 ] = 3 � 6𝐿 4𝐿 −6𝐿 2𝐿2 � (7.1.27)
𝐿 −12 −6𝐿 12 −6𝐿
6𝐿 2𝐿2 −6𝐿 4𝐿2
36 3𝐿 −36 3𝐿
2
[𝐾𝐺 ] =
1
� 3𝐿 4𝐿 −3𝐿 −𝐿2 � (7.1.28)
30𝐿 36 −3𝐿 36 −3𝐿
3𝐿 −𝐿2 −3𝐿 4𝐿2
Due to external forces {F}, one can find the displacement vectors from the following equation.
{F }
= [=
k ] {d } [ k F ] {d } + P [ k G ] { d } (7.1.29)
The above is the beam-column equation in finite element form. In the absence of transverse load, the
member will become column and P will be considered to be undetermined.
[ k F ] {d } + P [ kG ] { d } =
0 (7.1.30)
Denoting P as -λ we can reach to the familiar eigenvalue problem as given below.
6
[ k F ] {d } = λ [ kG ] {d } (7.1.31)
Solving above equation, values of λ and associated nodal displacement vectors can be obtained.
Mathematically this can be expressed as
( [ k ] − λ [ k ] ) {d } =
F G 0 (7.1.32)
Thus,
[ k F ] − λ [ kG ] =
0 (7.1.33)
For the matrix of size n, one can find n+1 degree polynomial in λ. The smallest root of the above
equation will become the first approximate buckling load. From this value λ, one can find a set of ratios
for the nodal displacement components. From this, the first buckling mode shape can be calculated.
Higher mode approximations can also be found in a similar process. The procedure to determine the
critical load by the above method is illustrated in the following example.
Example 7.1.1
Consider a column with one end clamped and other end free as shown in Fig. 7.1.2.
Fig. 7.1.3 Column with one end fixed and other end free
Now, the finite element equation of the column considering single element will be
12 6𝐿 −12 6𝐿 36 3𝐿 −36 3𝐿
𝐸𝐼 6𝐿 4𝐿2 −6𝐿 2𝐿2 � − λ 1 � 3𝐿 4𝐿2 −3𝐿 −𝐿2 � = 0
�
𝐿3 −12 −6𝐿 12 −6𝐿 30𝐿 36 −3𝐿 36 −3𝐿
6𝐿 2𝐿2 −6𝐿 4𝐿2
3𝐿 −𝐿2 −3𝐿 4𝐿2
7
Here,
= kinematic viscosity
= mass density of fluid
vi = Velocity components
f i = Body forces
p = fluid pressure
The suffix ,j and ,ji are the derivatives along j and j & i direction respectively. The dot represents the
derivative with respect to time. Neglecting non linear convective terms, viscosity and body forces, eq.
(7.2.1) can be simplified as:
p,i v i 0 (7.2.2)
Now, the continuity equation of the fluid is expressed by
p c 2 v k,k 0 (7.2.3)
Here, c is the acoustic wave speed in fluid. In the above expression, two sets of variables, the velocity
and the pressure are used to describe the behaviour of fluid. Now it is possible to combine equation
(7.2.2) and (7.2.3) to obtain a single variable formulation. For the small amplitude of fluid motion, one
can assume
vi u i (7.2.4)
Where u i is the displacement component of fluid. To obtain single variable formulation, eq. (7.2.4) may
be substituted into eq. (7.2.3) and one can get
p c 2 u k,k 0 (7.2.5)
Integrating eq. (5) w. r. t. time we have
10
p c 2 u k,k (7.2.6)
Now differentiating the above expression w. r. t. xi following expression will be arrived:
Similarly, it is possible to obtain the fluid equation in terms of pressure variable only. Differentiating eq.
(7.2.3) w. r. t. Time, the following expression can be obtained.
p c 2 p,ii 0 (7.2.11)
The above expression is basically the Helmholtz wave equation for a compressible fluid having acoustic
speed c.
1
2 p 2
p0 (7.2.12)
c
Thus, the general form of fluid equations of 2D linear steady state problems can be expressed by the
Helmholtz equation. For incompressible fluid c becomes infinitely large. Hence for incompressible
fluid, eq. (7.2.12) can be written as
2 p 0 (7.2.13)
For the ideal, irrotational fluid flow problems, the field variables are the streamline, ψ and potential φ
functions which are governed by Laplace’s equations
2 2
0
x 2 y 2
2 2
0
x 2 y 2 (7.2.14)
11
Derivation of the above expression and many other related equations can be found in details in fluid
mechanics related text books.
w u c u d 0
2
i i k,ki (7.2.16)
Performing integration by parts on the second terms, one can arrive at the following expression:
Now from earlier relation (eq. 7.2.6) we have, p c 2 u k,k . Thus, the above equation may be
written as:
In case of fluid filled rigid tank, the weighting function wi must satisfy the condition w i n i 0 on its
rigid boundary. Therefore, the above eq. will become
For finite element implementation of the above expression, let consider the interpolation function as N
and u as the nodal displacement vector. Thus,
12
u Nu and w Nw (7.2.20)
Now the divergence of the displacement vector can be expressed as:
u i,i Lu LNu Bu (7.2.21)
wT T
N Nu B c Bu d w N np d
T 2 T T
(7.2.22)
p
Or,
K c 2 B B d
T
(7.2.24)
F N n p d
T
p
Using eq. (7.2.23), the displacements in fluid domain can be determined under external forces applying
proper boundary conditions.
Now, performing integration by parts on the first term, the following expression can be obtained.
1
w i p,i d w i,i p,i d
c2
0
w i pd (7.2.27)
Thus,
1
c2 w p d w
i p d w i p,i d
i,i ,i (7.2.28)
13
Assuming interpolation function as N and p as the nodal pressure vector, the pressure (p) at any point
can be written as: p Np and similarly, w Nw . The divergence of the pressure can be expressed
. Again,
as: p, i Lp LNp Bp , where, L
x y
w i,i Lw LNw Bw
p Nw Np
w T N T Np
(7.2.29)
w i
T
Thus, eq. (7.2.28) will become:
1 w BT Bpd w N T p d
T T T
c2 w N T Npd n
(7.2.30)
Or,
E p G p B (7.2.31)
Where,
1 T
E
c2 N N d
T
G B B d (7.2.32)
T p
B N d
n
Applying boundary conditions, eq. (7.2.31) can be solved to calculate the dynamic pressure developed in
the fluid under applied accelerations on the domain.
where aeiωt is the horizontal component of the ground acceleration in which, ω is the circular frequency
of vibration and i = − 1 , n is the outwardly directed normal to the elemental surface along the interface.
In case of vertical dam-reservoir interface ∂p ∂n may be written as ∂p ∂x as both will represent normal to the
element surface. For an inclined dam-reservoir interface ∂p ∂x cannot represent the normal to the element
surface. Therefore, to generalize the expressions ∂p ∂n is used in eq. (7.2.38). If {a} is the vector of nodal
accelerations of generalized coordinates, {B2} may be expressed as
B2 R 2 a (7.2.39)
where,
R 2 = N r T N d d
T
(7.2.40)
2
Here, [T] is the transformation matrix for generalized accelerations of a point on the dam reservoir
interface and [Nd] is the matrix of shape functions of the dam used to interpolate the generalized
acceleration at any point on their interface in terms of generalized nodal accelerations of an element.
δ ∫ Ldt = 0 (7.3.1)
t1
The Langrangian functional, L, is defined as follows using a set of permissible time histories of
displacement.
L T (7.3.2)
Where T is the kinetic energy and is the potential energy which is basically elastic strain energy. The
kinetic energy, T is defined in the integral form as
1
T x T x d (7.3.3)
2
Here Ω represents the whole volume of the domain, ρ is the mass density of structure and x is the set of
time histories of velocities. The strain energy in the entire domain of elastic structure can be expressed
as
1 1
e T s d ee T
D d (7.3.4)
2 2
Here ε are the strains obtained from the set of time histories of displacements. Here, L can be expressed
in terms of the generalized variables x1 , x 2 x n an d x 1 , x 2 x n . Here, xi is the displacement and
the velocity is expressed by
17
d
x i xi (7.3.5)
dt
Considering xi as generalised displacement, the Lagrange’s equation of motion are expressed as
d L L
0 where i 1 to n
dt x x
(7.3.6)
i i
2 2
Now, substituting the values in Lagrangean L, we can obtain
1 1 1 1 1
L T m1x 12 m 2 x 22 k1x12 k 2 x 22 k 2 x1x 2 k 2 x12 (7.3.8)
2 2 2 2 2
Thus,
d L L
m1x1 k1x1 k 2 x 2 k 2 x1
dt x 1 x1
m1x1 k1x1 k 2 x 2 x1 0
(7.3.9)
d L L
m 2 x 2 k 2 x 2 k 2 x1
dt x 2 x 2
m 2 x 2 k 2 x 2 x1 0
Above equations can be written in the matrix form as shown below:
18
m1 0
x1 k1 k 2 k 2 x1 0
0 m (7.3.10)
x
2
k k x 2
0
2 2 2
Thus, for a multi degree freedom system, the above expression can be written as
M x K x 0 (7.3.11)
For a steady state condition, one can consider x = a sin t . Thus, x = -2 a sin t = -2 x . As a result,
the above equation can be re-written as
K x 2 M x 0
K 2 M x 0
2 M 1 K
(7.3.12)
Thus, for a single degree of freedom system, the fundamental frequency can be found as
K
(7.3.13)
M
l
M N T Nd A 0 N T Ndx AL 0 N T Nd
0 1 2 d
l
AL 1 d AL (7.3.16)
0
13 2 3
1
1 1
AL 3 6 AL 2 1
AL 3 2 3
2 3 3 1 1 6 1 2
3 0 6 3
2 3
Thus, the consistent mass matrix will be
AL 2 1
M (7.3.17)
6 1 2
Whereas the lumped mass matrix can be expressed as follows.
AL 1 0
M (7.3.18)
2 0 1
Now, to obtain a generalised mass matrix of a two dimensional truss member, consider Fig. 7.3.2.
The displacement vector and the shape function for such case are expressed as
u1
v1
N N2 0
d and N 1
0
0
(7.3.19)
u2 N1 0 N 2
2
v
1 0
l 0 1 1 0 0
M N T Nd AL 0
0 1 0
d (7.3.20)
0
0
After performing integration, the mass matrix will be obtained as
2 0 1 0
pAL 0 2 0 1
M (7.3.21)
6 1 0 2 0
0 1 0 2
The lumped mass matrix of two dimensional truss member will become
1 0 0 0
pAL 0 1 0 0
M= (7.3.22)
2 0 0 1 0
0 1
0 0
Substitution for the interpolation functions and performing the required integrations, one can arrive the
following mass matrix for the beam element.
156 22𝐿 54 −13𝐿
𝜌𝐴𝐿 22𝐿 4𝐿2 13𝐿 −3𝐿2
[M] = 420 � � (7.3.24)
54 13𝐿 156 −22𝐿
−13𝐿 −3𝐿2 −22𝐿 4𝐿2
1 0 0 0
𝜌𝐴𝐿 0 1 0 0
[M] = � � (7.3.25)
2 0 0 1 0
0 0 0 1
1/3 0 0 1/6 0 0
⎡ ⎤
⎢ 13/35 11𝐿/210 0 9/70 −13𝐿/420⎥
⎢ 𝐿2 /105 0 13𝐿/420 −𝐿2 /140 ⎥
[M] =𝜌𝐴𝐿 ⎢ ⎥ (7.3.27)
1/3 0 0
⎢ ⎥
⎢𝑆𝑦𝑚. 13/35 −11𝐿/210⎥
⎣ 𝐿2 /105 ⎦
The lumped mass matrix of such two dimensional frame element will be
1 0 0 0 0 0
⎡0 1 0 0 0 0⎤
⎢ ⎥
𝜌𝐴𝐿 0 0 1 0 0 0
[M] = 2 ⎢ ⎥ (7.3.28)
⎢0 0 0 1 0 0⎥
⎢0 0 0 0 1 0⎥
⎣0 0 0 0 0 1⎦
1
3
13
0
35
0 13
0
35
0 0 0
J
3A
11 L2
0 0 L 0
210 105
11 L2
0 L 0 0 0
210 105
M AL
1 0 0 0 0 0
1
6 3
9 13 13
0 0 0 0 L 0
70 420 35
9 13 13
0 0 0 L 0 0 0
70 420 35
0 J J
0 0 0 0 0 0 0
6A 3A
L2 L2
0 13
11
0 L 0 0 0 0 L 0
420 140 210 105
L2 2
0 13 L 0
11 L
0 0 0 L 0 0 0
105
420 140 210
(7.3.29)
In a similar process, one can derive the mass matrix of other types of structures such as two dimensional
plane stress/strain element, three dimensional solid element, plate bending element, shell element etc.
1
M N N d
T
2
1
C N N d
T
(7.3.31)
2
1
K B D Bd
T
2
In a linear dynamic system, these values remain constant throughout the time history analysis. Damping
is a mechanism which dissipates energy, causing the amplitude of free vibration to decay with time. The
various types of damping which influences structural dynamical behaviour are viscous damping,
hysteresis damping, radiation damping etc. Viscous damping exerts force proportional to velocity, as
exhibited by the term. A formulation for this kind of problem was developed by Rayleigh. In this case,
the energy dissipated is proportional to frequency and to the square of amplitude. Viscous damping is
provided by surrounding gas or liquid or by the viscous damper attached to the structure. Radiation
damping refers to energy dissipation to a practically unbounded medium, such as soil that supports
structure. Among the various types of physical damping, viscous damping is easy represent
computationally in dynamic equations. Fortunately, damping in structural problems is usually small
enough regardless of its actual source. Its effect on structural response is modelled well by regarding it
as viscous. The Rayleigh damping defines the global damping matrix [C ] as a linear combination of the
global mass and stiffness matrices.
[C ] = α ′[M ] + β ′[K ] (7.3.32)
Here, α ′ and β ′ are the stiffness and mass proportional damping constants respectively. The
relationship between the fraction of critical damping ratio, ξ , α and β at frequency ω is given by
1 β
ξ = αω + (7.3.33)
2 ω
Damping constants α ′ and β ′ are determined by choosing the fractions of critical damping ( ξ1′ and ξ 2′ )
at two different frequencies (ω 1 and ω2) and solving simultaneously as follows
2 1 2 2 2 11
22 12
(7.3.34)
2 1 2 2 1
22 12
There are several numerical integration schemes are available to obtain the time history response of the
structural system. Amongst, the implicit type of methods such as the Newmark’s method is one of the
most popular one.