1. Preparedby
Prof. V.V.S.H. Prasad, Madhav, Professor, Dept of Mechanical Engineering
Mr . C. Labesh Kumar, Assistant Professor, Dept of Mechanical Engineering
Finite Element Methods
III B Tech II Semester
INSTITUTE OF AERONAUTICAL ENGINEERING
(Autonomous)
Dundigal, Hyderabad- 500043
1
2. 2
UNIT- 1
• Introduction to FEM:
– Stiffness equations for a axial bar element in local co-ordinates
bars
using Potential Energyapproach and Virtual energyprinciple
– Finite element analysis of uniform, stepped and tapered
subjected to mechanical and thermalloads
– Assembly of Global stiffness matrix and loadvector
– Quadratic shapefunctions
– properties of stiffnessmatrix
5. Axially Loaded Bar – GoverningEquations
and BoundaryConditions
• Differential Equation
• BoundaryCondition Types
• prescribeddisplacement(essential BC)
•prescribedforce/derivative of displacement
(natural BC)
dx dx
5
d EA(x)
du f (x) 0 0 x L
6. 6
Axially Loaded Bar –BoundaryConditions
• Examples
• fixed end
• simplesupport
• free end
7. Potential Energy
- Axially loadedbar
- Elasticbody
x
Stretchedbar
• Elastic Potential Energy(PE)
- Spring case
Unstretched spring
PE 0
2
1
2
kx
PE
undeformed:
deformed:
0
7
PE 0
1 L
PE Adx
2
2 V
PE
1
σT
εdv
8. Potential Energy
• Work Potential(WE)
B
L
WP u fdx Pu
0
P
f
f: distributed force over aline
P: point force
u: displacement
A B
• T
otalPotential Energy
L
0 0
1 L
Adx u fdx PuB
2
• Principle of Minimum PotentialEnergy
Forconservative systems, of all the kinematically admissible displacementfields,
those corresponding to equilibrium extremize the total potential energy. If the
extremum condition is aminimum, the equilibrium state isstable.
8
9. Potential Energy +Rayleigh-RitzApproach
P
f
A B
Example:
Step1: assumeadisplacement field
i
i 1 to n
u aii x
is shapefunction / basis function
n is the order ofapproximation
Step2: calculate total potential energy
9
11. Galerkin’sMethod
P
f
A
Example:
EA(x)
du
P
dx xL
d du
dx
EA(x)
dx
f (x) 0
ux 0 0 dx
w
V
i
P
d du
dx
dx xL
du
~
EA(x)
u~x 0 0
f ( x)dV 0
EA(x)
B
Seekan approximation u
~so
~
In the Galerkin’s method, the weight function is chosen to be the sameasthe shape
function.
11
12. Galerkin’sMethod
P
f
A B
Example:
f ( x)dV 0
dx
du
~
EA(x)
dx
d
V
wi
0 0
0
L L
dx 0
du~L
du~dw
EA(x) i
dx wi fdx wi EA(x)
dx dx
1 2 3
1
2
3
12
14. FEM Formulation of AxiallyLoaded
Bar – Governing Equations
• WeakForm
• Differential Equation
d du
dx
EA(x)
dx
f (x) 0 0 x L
• Weighted-Integral Formulation
0
d du
w EA(x) f (x)dx 0
dx dx
L
0
L
du
0
du
L
dw
0 dx
EA(x)
dx
wf (x)dx wEA(x)
dx
14
15. Approximation Methods –Finite
Element Method
Example:
Step1: Discretization
Step2: Weakform of oneelement P2
P1
x1 x2
1
1
0
x
x
x
du
2
du
dx
EA(x)
dx
w(x) f (x)dx w(x) EA(x)
dx
x2
dw
1 1
2 2
P w x P 0
dx
EA(x) w(x) f (x) dx w x
dx
x2
dw du
x1
15
16. Approximation Methods –Finite
Element Method
Example(cont):
Step3: Choosingshape functions
- linear shapefunctions u 1u12u2
l
x1 x2
x
1
l
2
2
l
1
x x
x x
;
2
16
2
1
; 2
1
1
1
1
2
l
x
1l
x
2
x x 1;
17. Approximation Methods –Finite
Element Method
Example(cont):
Step4: Forming element equation
weak formbecomes
Letw 1
, 1 1 2 1 1
1
1
l
1 u u
l
x2
x
x2
x
EA 2 1
dx f dx P P 0
2
1
l
EA EA
l
x
x
u1 u2 1 f dx P
1
weak formbecomes
Letw 2
, 2 2 2 2 1
1
1
l
u u
l
x2
x
x2
1
x
EA 2 1
dx f dx P P 0
1
l
EA
l
EA
x2
x
u1 u2 2 f dx P2
1 1 1
2
P f P
x2
EA 1
1u1
x1
P f P
1 u x2
l 1
2 2 2 2
fdx
x1
1 fdx
E,Aare constant
17
18. Approximation Methods –Finite
Element Method
Example(cont):
Step5: Assembling to form systemequation
Approach 1:
Element1: 0 I I I
I I
u f P
E A
lI
f I
PI
1
1
1
1
2
2
2
0 0
0
0 0 0
0 0
0 0
Element2:
1
2 2
f
uII
f II
PII
lII
0 0
0 0
E II
AII 0 0 uII
II
PII
1
1
2
0
0
0
0
Element3:
1 1
III III III
III
III III III
f P
1 u f P
0 0
0 0
E III
AIII 0 0 0 0 0
1
0
1 1 0 0 uI
1 0
0 0
0 0
0 0 0
1 1
0 1 1 0
0 0 0
0 0 0
0 0
l 0 0 1 1 u
0 1
2 2 2
18
19. Approximation Methods –Finite
Element Method
Example(cont):
Step5: Assembling to form systemequation
Assembled System:
4 4
19
0 0
0
0
I I II II
lI
l
0
l l
l
EII
AII
EII
AII
EIII
AIII
EIII
AIII
lII
0
lIII
EIII
AIII
EIII
AIII
lIII
lIII
E I
AI
I I
E A
lI
f1
u1
EI
AI
E I
AI
E II
AII
II II
E A
3
lII
lIII
2 1 2 1
2 1 2 1
I I
II
II III II III
f1 P
1
f III
PIII
P1
f I PI
u f P f II
P
2 2 2
u f P
f f P P
3 3
u
f
P
4 2 2
20. Approximation Methods –Finite
Element Method
Example(cont):
Step5: Assembling to form systemequation
Approach 2: Element connectivitytable Element 1 Element 2 Element 3
1 1 2 3
2 2 3 4
global node index
(I,J)
local node
(i,j)
ij
20
IJ
k e
K
21. Approximation Methods –Finite
Element Method
Example(cont):
Step6: Imposing boundary conditions and forming condensesystem
Condensedsystem:
0
lI
lII
lII
lII
0
lIII
lIII
EIII
AIII
lIII
lIII
E I
AI
E II
AII
II II
E A
u2 f2 0
E II
AII
E III
AIII
E A
III III
u3 f3 0
f P
u
4
EII
AII
E III
AIII
4
lII
21
22. Approximation Methods –Finite
Element Method
Example(cont):
Step7: solution
Step8: post calculation
2
dx 1
dx
d2
dx
du
u
d1
u
1 1 2 2
u u u 2
1
dx
d2
dx
E Eu
d1
Eu
22
23. Summary - Major Steps inFEM
• Discretization
• Derivation of element equation
•weak form
•constructform of approximation solution
overone element
•derive finite element model
• Assembling– putting elements together
• Imposingboundary conditions
• Solvingequations
• Postcomputation
23
25. Linear Formulation for BarElement
2 2 12 22 u2
K
K
K12 u1
P f
P1
f1
K11
1
1
x2
x
i
x2
x
ji i
ij dx dx
EA i j
dx K , f
where K f dx
d d
x=x1 x=x2
2
1
x
x=x1 x=x2
u1 u2
P1 P2
f(x)
L= x2-x1
u
x
25
26. Higher Order Formulation for BarElement
u(x) u11(x) u22 (x) u33 (x)
1 3
u1 u3
u
x
u2
2
u1 u4
u
x
u2 u3
1 2 3 4
u(x) u11( x ) u22 ( x ) u33 ( x ) u44 ( x )
u(x) u11( x ) u22 ( x ) u33 ( x ) u44 ( x ) unn ( x )
n
u1 un
u
x
u2 u3 u4
26
…
…
…
…
…
1 2 3 4 …
…
…
…
…
38. Arbitrarily Oriented 1-D Bar Element in 3-DSpace
2
2
1
y
y
y
x
x
x
y
y
y
2
2
1
0
0
0 w1
0 v
0 u1
w2 0 0
0
v
u
w1 0
0
v
u1
z w2
v
u
0 0
0 0
0 0 z z
z z
z
x x 0 0
0 0
0 0
x
-x,-x,-xare the Direction
Cosinesof the bar in thex-
y-zcoordinatesystem
P1,u1
2 2
P , u
x
x
y
z
2
1
-
x
-
x
-
38
2
1
2
1
1
0
0
0 R1
0 Q
0 P1
R2 0
Q2 0
P
R 0
Q 0
P
1
x
x
x
z
y
y
y
z R2
y
Q2
P
0 0
0 0 y y
0 0 0 z z
z z
x x x 0 0
0 0
0 0
39. Stiffness Matrix of 1-D Bar Element in 3-D Space
2
2
1
2
1
w
v
2
u
w1
v1
u
L
AE
R2
Q2
P2
R
Q1
P1
x
x x
x x
x
x x
x x
x x
x
x x
x x
x
x x
x x
x x
x
x x
x x
x
x xx xxx x x x x
x x
x
x x x x
x
x x
x x
x x
x x
x
2
2
2
2
2
2
2
2
2
2
x x x
2
1 1
P , u
2 2
P , u
x
x
x
y
z
2
1
-
-
x
-
39
0
0
0
0 0 0 0 0w 0
0v
0
0
0 0 1 0
0 0 0 0
1
L 1 0 0 1 0
AE 0
R 0
Q 0
0 0 0 0 0 v2 0
0 0 0 0 0 0w2 0
u2
1
1
u1
R2 0
P2
Q2 0
1
1
P1
47. BendingBeam
Normal strain:
Purebending problems:
Normal stress:
Normal stress with bendingmoment:
Moment-curvature relationship:
Flexure formula:
y
Review
x
y
Ey
x
M M
x
x ydA M
EI
1 M
I
x
My
I y2
dA
1 d2
y
M EI
EI
dx2
47
49. Governing Equation and BoundaryCondition
• Governing Equation
• BoundaryConditions -----
2
?,
at x 0
dx
d2
v
EI
dx
d
?& EI ?&
dv d2
v
v ? &
EssentialBCs– if vor
Natural BCs–if or
isspecdifviedat theboundary.
dx
{
dx
dx
dv
dx dx
d
at x L
?,
?& EI
v ? & ?& EI 2
d2
v
dx2
dx2
d2
v
d2
v
EI
dx2
2
dx
EI
dx
is
d
sp
ecifi
d
ed2
v
a
tthe boundary.
2
q(x) 0,
dx2
d2
v(x)
EI
dx
d2
0<x<L
49
50. Weak Formulation for BeamElement
• Governing Equation
2
q(x) 0,
dx2
d2
v(x)
EI
dx
d2
• Weighted-Integral Formulation for one element
2
x2
dx2
d2
v(x)
EI q(x)dx
dx
d2
0 w(x)
x1
• WeakFormfrom Integration-by-Parts ----- (1st time)
2
1
50
2
dx
dx
dx
dx dx
x
x
d2
v
2
EI
d
x1
x2
EI wq dx w
dw d d2
v
0
2
1
x x x
51. Weak Formulation
• WeakFormfrom Integration-by-Parts ----- (2nd time)
1
1
2
2
2
x
x
dx
dx
dx
dx
dx
x x
d2
v
2
dw d2
v
2
EI EI
d
x1
wq dx w
x2
d2
w d2
v
0 dx2 EI
V(x2)
x=x1
M(x2)
q(x)
y
x
x=x2
V(x1)
M(x1)
L= x2-x1
2
51
x1
dx
dw
dx
EI
x
2
wV M
x1
wq dx
d 2
v
x2
d2
w
0 dx2
79. Finiteelement formulation for2D:
x
y
S
u
S
T
u
v
x
px
Step 1:Divide the body into finite elements connected toeach
other through special points(“nodes”)
py
3
2
1
4
x
2
u1
u2
u3
u4
v4
v3
3
v2
Element‘e’
4
v1
y
1
v
u
79
4
3
2
2
1
u4
v
u3
v
v
v1
u
u
d
80. N3(x,y)u3 N4(x,y)u4
N3(x,y)v3 N4(x,y)v4
u(x,y) N1(x,y)u1 N2(x,y)u2
v(x,y) N1(x,y)v1 N2(x,y)v2
4
2
1
4 3
3
2
1
4
3
2
1
v
v3
u4
u2
v
u1
N u
N 0 N 0 N 0
0
0
v
0 N 0 N 0 N
v (x,y)
u (x, y) N
u
u N d
80
83. Strain approximation in terms of strain-displacement matrix
Stressapproximation
Summary: Foreachelement
Displacement approximation in terms ofshape functions
Element stiffnessmatrix
Element nodal loadvector
u N d
DB d
ε B d
e
V
k B DB dV
T
S
e
b f
ST
S
f
V
T T
N T dS
f e
N X dV
83
84. Constant Strain Triangle (CST) :Simplest 2Dfinite element
y
v
v1
u1
3
1
(x,y)
84
v
u
1 1
v2
(x ,y)
2 (x ,y )
u2
2 2
x
• 3 nodes per element
• 2 dofs per node (each node canmove in x- and y-directions)
• Hence6 dofs per element
3
(x3,y3)
u3
85. u 21 N26 d61
1
1
3 2
2
1
v3
u3
v
u
N v
0
0 u 2
0 N2 0 N3
N 0 N 0
v (x,y)
u
u (x, y)
N1
Thedisplacement approximation in terms of shape functionsis
85
3
2
1
N
0
0 N2 0 N3 0
N 0 N 0
N
N1
u (x,y) N1u1 N2u2 N3u3
v(x,y) N1v1 N2 v2 N3v3
86. N
2A
3
2
1 1 1
1
N
a2 b2x c2 y
2A
a3 b3x c3 y
2A
a b x c y
N
2
x3
2
b1 y2
b2 y3
b3 y1
x3 y2
x1 y3
x2 y1
a1 x2 y3
a2 x3 y1
a3 x1 y2
y
2
y3
1
y3 c1 x3 x2
y1 c2 x1 x3
y2 c3 x2 x1
1 x1 y1
A area of triangle
1
det1 x
where
x
y
u3
v3
Formula for the shape functionsare
v1
u1
u2
3
1
(x,y)
86
v
u
1 1
v2
(x ,y)
2 2
2 (x ,y)
3 3
(x ,y)
87. Properties of the shapefunctions:
1.Theshape functions N1, N2 and N3 are linear functions of xand
y
x
y
2
3
1
1
N1
2
3
1
N2
1
2
3
1
1
N3
at node 'i'
0 at other nodes
N
1
i
87
88. 2.At every point in thedomain
y
x
3
i1
3
Ni yi
i1
Ni xi
i1
3
1
Ni
88
89. 3.Geometric interpretation of the shapefunctions
At any point P(x,y) that theshape functions are evaluated,
y
2
x
3
1
P (x,y)
A3
A2
A1
3
2
1
N
N
N
A1
A
A2
A
A3
A
89
91. 91
Element stresses (constant inside eachelement)
DB d
Inside each element, all components ofstrain are constant: hence
the name Constant StrainTriangle
93. Element stiffnessmatrix
e
V
k B DB dV
T
e
V
T
dVBT
DBAt
k B DB t=thickness of theelement
A=surfacearea of the element
SinceBisconstant
t
93
A
95. Element nodal load vector due to bodyforces
e
Ae
V
T T
b
f N X dV t N X dA
e
e
e
A
A
A
b
N X dA
t
N X dA
t
N X dA
t
f
f
f
f
f
tAe
N3 Xb
tAe
N2 Xb dA
tAe
N3 X a dA
dA
2 a
1 b
1 a
b3 y
fb3x
b2 y
b2x
fb1y
b1x
x
y
fb3x
fb3y
fb1y
fb1x
fb2x
fb2y
2
3
1
(x,y)
95
Xb
Xa
96. EXAMPLE:
If Xa=1andXb=0
96
0
3
3
0
0
3
0
0
3
2
1
tA
tA
tA
N dA
t
N dA
0
t
N dA
t
t
t
N X dA
N X dA
t
N X dA
t
N X dA
t
N X dA
t
f
f
f
e
e
e
e
e
e
e
e
A
A
A
A
N3 X b dA
A
3 a
A
2 b
A
2 a
A
1 b
A
1 a
fb3 y
b3x
fb2y
b2x e
fb1x
fb1y
b
97. Element nodal load vector due totraction
e
ST
S
T
S
dS
f N T
EXAMPLE:
y
fS3x
fS3y
fS1y
fS1x
2
x
3
1
97
e
l
S
S
T dS
f t
13 along13
NT
98. Element nodal load vector due totraction
EXAMPLE:
y
fS3x
2
3
S
T S dS
f tl23 along 23
NT
e
fS3y
fS2x
fS2y
(2,0)
(2,2)
1
(0,0)
0
x
1
S
T
t 21 t
1
23
2 x l
fS t N (1)dy
e 2 along 23
fS 0
2 y
fS t
3x
fS 0
3 y
2
Similarly, compute
1
98
2
99. 99
Recommendations for use ofCST
1.Use in areas where strain gradients aresmall
2.Use in mesh transition areas(fine mesh to coarse mesh)
3.Avoid CSTin critical areas of structures (e.g.,stress
concentrations, edges of holes,corners)
4.In general CST
sare not recommended for general analysis
purposes as a very large number of these elements are required
for reasonableaccuracy.
105. 105
Step 4:Assemblethe global stiffness matrix correspondingto the nonzerodegrees of
freedom
Notice that
u3 v3 u4 v4 v1 0
Hencewe need to calculate only asmall (3x3) stiffnessmatrix
K
1
0.983 0.45 0.2 u
0.45 0.983 0 107 1
u
2
0.2
u
0
u2
1.4
v2
v2
106. Step 5:Compute consistent nodalloads
f f 2x
f f
2 y 2 y
0
f1x 0
f2 y 1000 fS
2 y
Theconsistent nodal load due to traction on the edge3-2
N dx
3
3
3
S2 y
9
0
x2
3
50 2 50
2
225 lb
(300)(0.5)
N (300)tdx
f
150
x
dx
3
x0 3
x0 32
x0 3 32
3 2
N
106
x
3
2 32
107. Hence
f2 y 1000 fS
2 y
1225 lb
Step 6:Solve the system equations to obtain the unknownnodal loads
K d f
107
0
0
0.2
0
0.983
2
107
0.45 u2
0.45 0.2u1
0.983
0 1.4v 1225
Solveto get
v
u
4
2
2
1
0.9084104
in
u 0.106910 in
0.2337 104
in
108. 108
0.9084104
0 0
0.1069104
0.2337104
0
1 1 2 2 4 v4
v u v u
u
d(1)T
76.1
Calculate
114.1
(1)
1391.1 psi
Step 7:Compute the stresses in theelements
In Element#1
(1)
DB(1)
d(1)
With
109. 109
3
0.9084104
3
0 0 0 0 0.1069104
2 v2
u v u4 v4 u
d(2)T
Calculate
114.1
(2)
28.52 psi
In Element#2
(2)
DB(2)
d(2)
With
363.35
Notice that the stressesare constant ineach element
114. Formulation for 1-D LinearElement
Let T(x) T11(x)T22(x)
f2
1
x1
2
x2
x
T1 T2
f1
2 1
1 2
l l
x x
,
x x
( x) ( x)
x2
x1
1T1
2T2
2
2
1
1
x
x
, f (x) A
T
f (x) A
T
114
115. Formulation for 1-D LinearElement
Letw(x)= i (x), i =1, 2
2 x2
i j
j i i 2 2 i 1 1
j1
x2
d d
( x )f
dx AQdx (x ) f
dx dx
x
1 1
0 T A
x
2
Qi i (x2 ) f2 i (x1) f1
KijTj
j1
11
1
1
K22 T2
K12
K12 T1
K
2 Q2
f
f Q
1
115
1
2
x2 x2
i j
ij i i 1
x1 x2
x
dT dT
dx dx
d d
where K A dx, Q AQdx, f A , f A
dx dx
x
116. Element Equations of 1-D LinearElement
1
1
1 T2
1 1T1
L 1
2 Q2
Q
f
f A
i i 1
dT dT
dx xx1 xx2
where Q AQ dx, f A , f2 A
dx
x2
x1
f2
1
x1
2
x2
T1
x
T2
116
f1
117. 1-D Heat Conduction -Example
t1 t2 t3
0
T 200o
C T 50o
C
x
1 2 3
Acomposite wall consists of three materials, asshown in the figure below.
Theinside wall temperature is 200oCand the outside air temperature is50oC
with aconvection coefficient of h =10 W(m2.K). Find the temperature along
the composite wall.
1 70W m K , 2 40W m K , 3 20W m K
t1 2cm, t2 2.5cm, t3 4cm
117
118. Thermal Conduction andConvection-
Fin
Objective: to enhanceheattransfer
dx
t
x
w
loss
Q
2h(T T)dx w 2h(T T)dx t
2h(T T)w t
Ac dx Ac
Governing equation for 1-D heat transfer in thin fin
c c
d A
dT A Q 0
dx dx
c
118
c
d A
dT Ph T T A Q 0
dx dx
P 2w t
where
125. let:
)
1 1 2 2
1
a
x2
wf dx Q w(x ) Q w(x
u
t
cw
w u
x x
0
x
Transient HeatConduction
n
ux,t u j tj x
j1
and w i x
KuMu
F
x2
x1
ij
x x
K a i j
dx
Mij
x2
cijdx
x1
x2
Fi i fdx Qi
x1
ODE!
125
126. TimeApproximation – First OrderODE
dt
126
0 t T 0
a
du
bu f t u0u
Forward difference approximation - explicit
Backward difference approximation - implicit
k k k
a
k1
u u
t
f bu
k k k
k1
u u
t
f bu
a bt
138. FiniteElement Model – Single-VariableProblem
u u jj
j
Ritzmethod:
e
n
e e
K u f Qe
ij j i i
j1
where j (r, z) j (x, y)
w i
Weakform
22
138
e
j j
a
i i
K e
a a
rdrdz
00 i j
r r z z
ij 11
where
i i
e
f e
frdrdz
i i n
e
Qe
q ds
140. 3-Node Axi-symmetricElement
T (r, z) T11 T22 T33
1
2
3
140
1 2 3
e
1 r z
2A
r2z3 r3z2
z z
r r
3 2
3 1 1 3
2 3 1
e
1 r z
2A
r z r z
z
z
r r
1 3
1 2 2 1
3 1 2
e
1 r z
2A
r z r z
z z
r r
2 1
141. 4-Node Axi-symmetricElement
T (r, z) T11 T22 T33 T44
1 2
3
4
a
b
r
141
z
2
3
a b
1
1
1
1
a b a b
1
4 a b
142. Single-Variable Problem –Example
Step1: Discretization
Step2: Element equation
e
142
j
i
ij
K e
i
j
rdrdz
r r z z
i i
e
f e
frdrdz
i i n
e
Qe
q ds
143. Reviewof CSTElement
• Constant Strain Triangle (CST)- easiest and simplest
finite element
– Displacement field in terms ofgeneralized coordinates
– Resulting strain field is
– Strains do not vary within the element. Hence, the name
constant strain triangle(CST)
• Other elements are not solucky.
• Canalso be called linear triangle becausedisplacement field is
linear in xand y - sidesremainstraight.
143
144. Constant Strain Triangle
• Thestrain field from the shapefunctions lookslike:
– Where, xi and yi are nodal coordinates (i=1, 2,3)
– xij =xi - xj and yij=yi - yj
– 2Ais twice the area of the triangle, 2A=x21y31-x31y21
• Node numbering is arbitrary except that the sequence123
must go clockwise around the element if Ais tobe positive.
144
145. 145
Constant Strain Triangle
• Stiffness matrix for element k =BTEBtA
• TheCSTgivesgood results in regions of theFE
model where there is little straingradient
– Otherwise it does notwork well.
146. Linear Strain Triangle
• Changesthe shape functions and results in
quadratic displacement distributions and
linear strain distributions within theelement.
146
148. ExampleProblem
• Consider the problem we were lookingat:
5in.
1in.
0.1in.
I 0.113
/12 0.008333 in4
M c
1 0.5
I 0.008333
60 ksi
E
0.00207
2
ML
25
2EI 2 29000 0.008333
0.0517in.
148
1k
1k
149. Bilinear Quadratic
• TheQ4element is aquadrilateral element
that hasfour nodes. In terms ofgeneralized
coordinates, its displacement field is:
149
154. 154
Quadratic Quadrilateral Element
• Should we try to use this element to solve our
problem?
• Or try fixing the Q4element for our purposes.
– Hmm…toughchoice.
155. 155
Isoparametric Elements and Solution
• Biggestbreakthrough in the implementation of the
finite element method is the development of an
isoparametric element with capabilities to model
structure (problem) geometries of any shapeandsize.
• Thewhole idea works onmapping.
– Theelement in the real structure is mapped to an
‘imaginary’ element in an ideal coordinatesystem
– Thesolution to the stress analysis problem is easyand
known for the ‘imaginary’element
– Thesesolutions are mapped back to the element inthe
real structure.
– All the loads and boundary conditions are also mapped
from the real to the ‘imaginary’ element in this approach
157. Isoparametric element
Y
1 2 3 4
0 0 0 N1
X N
0
• The mappingfunctions areq
x1u
ite simple:
4
x
2
x3
N N N 0 0 0 0 x
N2 N3 N4 y1
y2
1
4
N 1 (1)(1)
2
4
N
1
(1 )(1)
3
4
N 1 (1 )(1 )
4
4
N
1
(1)(1 )
y3
y4
Basically, thex and y coordinatesof any point in the
elementare interpolations of the nodal (corner)
coordinates.
Fromthe Q4 element,thebilinear shape functions
are borrowed to be used as the interpolation
functions. They readily satisfy the boundaryvalues
too. 157
158. Isoparametric element
v
1 2 3 4
0 0 0 N1
u N
0
• Nodal shape functions for d
u
i1s
placements
4
u
2
u3
N N N 0 0 0 0 u
N2 N3 N4 v1
v2
v3
v4
1
4
N 1 (1)(1)
2
4
N
1
(1 )(1)
3
4
N 1 (1 )(1 )
4
4
N
1
(1)(1 ) 158
160. Isoparametric Element
X Y
u
u
X Y u
X Y u
X
Y
It is easier to obtain
X
and
Y
X Y
J X Y Jacobian
defines coordinate transformation
Hencewe will do itanother way
u
u
X
u
Y
X Y
u
u
X
u
Y
X
Ni
X
i
Y N
i
Yi
X N
i X i
Y N
i Yi
X
u
Y
u
1
u
J
u
169
161. 161
GaussQuadrature
• Themapping approach requires usto be able to
evaluate the integrations within the domain (-1…1)of
the functionsshown.
• Integration canbe done analytically by usingclosed-
form formulas from atable ofintegrals (Nah..)
– Or numerical integration can be performed
• Gaussquadrature is the more common form of
numerical integration - better suited for numerical
analysis and finite elementmethod.
• Itevaluated the integral of afunction asasum of a
becomes
n
I Wii
i1
finite numb1er of terms
I d
1
163. Numerical Integration
Calculate:
b
I f xdx
a
• Newton – Cotesintegration
• Trapezoidalrule – 1storderNewton-Cotesintegration
• Trapezoidalrule – multipleapplication
1
b a
f (x) f (x) f (a)
f (b) f (a)
(x a)
2
f (a) f (b)
b b
I f (x)dx f1(x)dx (b a)
a a
b
163
xn1
xn x1 x2 xn
a x0 x0 x1
I f (x)dx fn (x)dx f (x)dx f (x)dx f
(x)dx
2
n1
i1
i
f (x ) f (b)
f (a) 2
h
I
164. Numerical Integration
Calculate:
b
I f xdx
a
• Newton – Cotesintegration
• Simpson1/3 rule – 2ndorderNewton-Cotesintegration
2
2
0 1
1
0 2
0 1
2 f (x )
(x2 x0 )(x2 x1)
(x x0 )(x x1)
f (x ) f (x )
(x0 x1)(x0 x2 ) (x x )(x x )
(x x1)(x x2 ) (x x )(x x )
f (x) f (x)
b b
a a
6
f (x ) 4 f (x ) f (x )
I f (x)dx f2 (x)dx (x2 x0 ) 0 1 2
164
165. Numerical Integration
Calculate:
b
I f xdx
2
(b a)
f (a)
(b a)
f(b)
2 2
165
I (b a)
f (a) f (b)
I c0 f (x0 ) c1 f (x1)
a
• GaussianQuadrature
Trapezoidal Rule: GaussianQuadrature:
Choose according to certaincriteria
c0,c1,x0 , x1
166. Numerical Integration
Calculate:
b
I f xdx
a
• GaussianQuadrature
3
1
1
1
1
I f xdx f
3
f
1
• 3pt GaussianQuadrature
I f xdx 0.55 f 0.77 0.89 f 0 0.55 f0.77
1
1
1
• 2pt GaussianQuadrature
I f xdx c0 f x0 c1 f x1 cn1 f xn1
b a
~
x 1
2(x a)
Let:
b 1
1
a
166
~ ~
1 1 1
f (x)dx
2
(b a) f 2
(a b)
2
(b a)xdx