Professional Documents
Culture Documents
Chapter 3 - English
Chapter 3 - English
CONTENTS
3.1 FEM FOR 2D SOLIDS
INTRODUCTION
LINEAR TRIANGULAR ELEMENTS
– Field variable interpolation
– Shape functions construction
– Using area coordinates
– Strain matrix
– Element matrices
LINEAR RECTANGULAR ELEMENTS
– Shape functions construction
– Strain matrix
– Element matrices
– Gauss integration
– Evaluation of me
Assoc. Prof. Dr. Tich Thien TRUONG 2 HCMUT 2007
FINITE ELEMENT METHOD
Chapter 3. FEM FOR 2D AND AXISYMMETRIC SOLIDS
─ Variational formulation
─ Treating plane strain as a limit case
INTRODUCTION
For plane strain problems, the thickness of the
element is unit, but for plane stress problems,
the actual thickness must be used.
In this course, it is assumed that the element
has a uniform thickness h.
Formulating 2D elements with a given variation
of thickness is also straightforward, as the
procedure is the same as that for a uniform
element.
y fy
y fx
fx
x
Assume,
N1 a1 b1 x c1 y N i ai bi x ci y i= 1, 2, 3
N 2 a2 b2 x c2 y ai
or Ni
1 x y bi pT
N 3 a3 b3 x c3 y pT ci
x2 y3 x3 y2 y2 y3 x3 x2
Solving, a1 , b1 , c1
2 Ae 2 Ae 2 Ae
N i ai bi x ci y
1
where ai ( x j yk xk y j ) , i = 1, 2, 3
2 Ae
1 j, k determined from cyclic permutation
bi ( y j yk )
2 Ae
i = 1, 2, 3
1 i
ci ( xk x j )
2 Ae
k j
k = 3, 1, 2 j = 2, 3, 1
A2 A2 A3 A2 A2 A3
L1 L2 L3 1
Ae Ae Ae Ae
Therefore, N1 L1 , N 2 L2 , N 3 L3
U h ( x , y ) N ( x , y )d e
Strain matrix xx
u
0
x x
v
yy LU where L 0
y
y
u v
xy
y x
y x
LU LNd e Bd e
0 b1 0 b2 0 b3 0
x
B 0 c1 0 c2 0 c3
B LN 0
N c1 b1 c2 b2 c3 b3
y
(constant strain element)
y x
Assoc. Prof. Dr. Tich Thien TRUONG 17 HCMUT 2007
FINITE ELEMENT METHOD
Chapter 3. FEM FOR 2D AND AXISYMMETRIC SOLIDS
Element matrices
h
k e B cBdV ( dz )BT cBdA hBT cBdA
T
0
Ve Ae Ae
Constant matrix
k
e hAe BT cB
With c is the matrix of elastic constants.
For plane stress problem of isotropic material:
1 0
E 1 0
c
1
2
1
0 0 2
Assoc. Prof. Dr. Tich Thien TRUONG 18 HCMUT 2007
FINITE ELEMENT METHOD
Chapter 3. FEM FOR 2D AND AXISYMMETRIC SOLIDS
Element matrices
For plane strain problem of isotropic material:
1 1
0
c
1 E
1 0
1 1 2 1
0 1 2
0
2 1
Element matrices
After computing, we get:
Element matrices
Plane stress:
E
C1
1 2
Plane strain:
C1
1 E
1 1 2
and the components kij are computed as following:
Element matrices
Element matrices
With C2 and are also the elastic constants depending
on the stress state.
1
Plane stress: C2 ;
2
1 2
Plane strain: C ;
1 2 1
2
xij xi x j
y jk y j yk
Assoc. Prof. Dr. Tich Thien TRUONG HCMUT 2007
FINITE ELEMENT METHOD
Chapter 3. FEM FOR 2D AND AXISYMMETRIC SOLIDS
Element matrices
h
m e N NdV
T
dx N T NdA h N T NdA
0
Ve Ae Ae
For elements with uniform density and thickness,
N1 N1 0 N1 N 2 0 N1 N 3 0
0 N1 N1 0 N1 N 2 0 N1 N 3
N 2 N1 0 N2 N2 0 N 2 N3 0
me h dA
Ae 0 N 2 N1 0 N2 N2 0 N 2 N3
N 3 N1 0 N3 N 2 0 N3 N3 0
0 N 3 N1 0 N3 N 2 0 N 3 N 3
Element matrices
2 0 1 0 1 0 y, v 3 (x3, y3)
2 0 1 0 1 (u3, v3)
fsy
fsx
hA 2 0 1 0 A
me
12 2 0 1
2 (x2, y2)
(u2, v2)
sy. 2 0 1 (x1, y1)
(u1, v1) x, u
2
f sx 0
fe [N]T
dl 0
f sy
2 3
l
1 fx
Uniform distributed load on side 23: fe l23
2 fy
fx
fy
Assoc. Prof. Dr. Tich Thien TRUONG 25 HCMUT 2007
FINITE ELEMENT METHOD
Chapter 3. FEM FOR 2D AND AXISYMMETRIC SOLIDS
Element matrices
If uniform distributed load are applied on three sides, we
use superposition principle:
31
f sy .l31
23
f sy .l23
N1 g x gx
N g g
1 y y
T x
g N 2 g x Ae h g x
fe [N] dV hdA
gy N g 3 g y
Ve
A 2 y
N3 g x gx
N3 g y g y
y jk 0 yik 0 yij 0
1
B 0 x jk 0 xik 0 xij
2A
x jk y jk xik yik xij yij
1 C2 0 y jk 0 yik 0 yij 0
C1
S e C2 1 0 0 x jk 0 xik 0 xij
2A
0 0 x jk y jk xik yik xij yij
LINEAR RECTANGULAR
ELEMENTS
Non-constant strain matrix
More accurate representation of stress and strain
Regular shape makes formulation easy
u1 y, v
v displacements at node 1
1 4 (x4, y4) 3 (x3, y3)
u2 (u4, v4) (u3, v3)
displacements at node 2
v2 fsy
de 2b fsx
u3
v3 displacements at node 3 2a
2 (x2, y2)
1 (x1, y1)
u4 (u1, v1) (u2, v2)
v displacements at node 4
4
x, u
x y
a 0
J ab
x y 0 b
y y
0 0
x
1 x x
B LN 0 N 0 N
y J
x x y y
y x
Strain matrix
1a 0 1
a 0 1
a 0 1
a 0
1 1 1 1 1
B LN 0 b 0 b 0 b 0 b
4 1
b 1
a
1
b
1 1 1 1 1
a
a b a b
Note: No longer a constant matrix!
Element matrices
x ( x1 a) y ( y1 b)
, dxdy =Jdd = abdd
a b 1 1
Therefore, k e hB cBdA
T
hB cB J d d
T
1 1
A
h 1 1
m e N NdV dx N NdA h N NdA
T T T
abh NT Nd d
0 1 1
V A A
With c is the matrix of elastic constants.
For plane stress problem of isotropic material:
1 0
E 1 0
c
1
2
1
0 0 2
Assoc. Prof. Dr. Tich Thien TRUONG 39 HCMUT 2007
FINITE ELEMENT METHOD
Chapter 3. FEM FOR 2D AND AXISYMMETRIC SOLIDS
Element matrices
For plane strain problem of isotropic material:
1 1
0
c
1 E
1 0
1 1 2 1
0 1 2
0
2 1
Element matrices
With C2 and are also the elastic constants depending
on the stress state.
1 C2 1
Plane stress: C2 ;
2 2
1 C2 1 2
Plane strain: C2 ;
1 2 2 1
Gauss integration
1 m
In 1 direction: I f ( )d w j f ( j )
1
j 1
nx ny
1 1
In 2 directions: I f ( , )d d wi w j f (i , j )
1 1
i 1 j 1
Gauss integration
m j , j wj Accuracy n
1 0 2 1
2 -1/3, 1/3 1, 1 3
w1 = 1, w2 = 1
Assoc. Prof. Dr. Tich Thien TRUONG HCMUT 2007
FINITE ELEMENT METHOD
Chapter 3. FEM FOR 2D AND AXISYMMETRIC SOLIDS
k e1,1 abh w1w1B T 1,1cB1,1 abh w1w1 f (1 ,1 ) abhf (1 ,1 )
ke(1,1)
1
C2
2,488.b2 a2 2,488. ab .
2 2,488
3b2 a2
0,5285
ab
C2 33 33
2 1
3a2 b2
2C
2
3 ab
2
3a2
2,488
b2
0,5285
ab
C2 33 33
2,488.
1 0,5285
a2 b2 ab
C2 33 33
2
3b2
2,488
a2
2C
2
3 ab
2 1
3a2 b2
0,5285
ab
C2 33 33
2
3a2
2,488
1
b2
2,488.
1
b2 a2
2C
2
3 ab
2
3a2
2,488
b2
0
ab
,5285
C2 33 33
2 1
3a2 b2
2,488. 2
C
.
ab
1
2,488.
a2 b2
0,5285
ab
C2 33 33
2
3a2
2,488
1
b2
2,488.
C2
.
ab
2 1
3a2 b2
Ke(1,1)
1
2,488.
b2 a2
2,488. 2
C
.
ab
2 2,488
3b2 a2
0,5285
ab
C2 33 33
1
2,488.
a2 b2
0,5285
ab
C2 33 33
2
3b2
2,488
a2
1 2C2
2,488.
a2 b2 3 ab
1
2,488.
a b
2 2
ke(2,1)
0,1786 2,488
2 2
b a
2C2
3 ab
2 2,488
2
3b a
2
0,5285
ab
C2 33 33
2 1
2 2
3a b
2,488C20,1786
ab
2
3a2
0,1786 2
b
0,5285
ab
C2 33 33
a
1 0,5285
2,488. 2 0,1786 2
b ab
C2 33 33
2
3b2
2,488 2
a
2C
2
3 ab
2 1
2 2
3a b ab
0,5285
C2 33 33
2
3a 2
0,1786 2
b
1
0,1786 2,488
b 2
2
a
2,488C20,1786
ab
2
3a2
b
0,1786 2
0,5285
ab
C 2 33 33
2 1
2 2
3a
b
2C2
3 ab
ke(2,1)
a
b
1 0,14
2,488. 2 0,1786 2 C2 33 33
ab
2
3a2
b
1
0,1786 2
2C2
3 ab
2 1
2 2
3a b
0,1786 2,488
b2
2
a
2C2
3 ab
2 2,488
3b2 a2
0,5285
ab
C 2 3 3 33
a
1
2,488. 2 0,1786 2
b
0,14
C2 33 33
ab
2
3b2
2,488
a2
0,1786 2,488 2,488C20,1786
b2 a2 ab
1
sym 2,488. 2 0,1786 2
a b
ke(2,2)
1
C2
0,1786.b2 a2 0,1786. a.b
2 0,1786
3b 2
2
a
0,14
ab
C2 33 33
2 1
2 2
3a b
2C
2
3 ab
2
3a2
0,1786 2
b
0,14
ab
C2 33 33
1 0,14
0,1786. 2 2
a b ab
C2 33 33
2
3b2
0,1786 2
a
2C
2
3 ab
2 1
3a2 b2 ab
0,14
C2 33 33
2
3a 2
0,1786 2
b
1
1
0,1786. 2 2
b a
2C
2
3 ab
2
3a2
0,1786 2
b ab
0,14
C2 33 33
2 1
2 2
3a b
0,1786. 2
C
a.b
1
0,1786. 2 2
a b
0,14
ab
C2 33 33
3
2
a2
0,1786 2
b
1
0,1786. 2
C
a.b
2 1
3a2 b2
ke(2,2)
1
0,1786. 2 2
b a
0,1786. 2
C
a.b
2 0,1786
3b2
a2
0,14
ab
C
2 33 33
1
0,1786. 2 2
a b ab
0,14
C2 33 33
2
3b2
0,1786 2
a
1 2C2
0,1786. 2 2
b a 3 ab
1
sym 0,1786. 2 2
a b
ke(1,2)
2,488 0,1786
b2 a2
2C2
3 ab
2 0,1786
2
3b a
2
0,14
C2 33 33
ab
2 1
2 2
3a b
1,786C22,488
ab
2
3a 2
2,488 2
b
0,5285
ab
C2 33 33
0,1786. 2 2,488 2
a b ab
1 0,5285
C2 33 33
3b
2
2
0,1786 2
a
2C2
3 ab
2 1
2 2
3a b
0,5285
ab
C2 33 33
2
3a 2
2,488 2
b
1
0,1786 2,488
b2
a2
1,786C22,488
ab
2
3a2
2,488
b2
0,5285
ab 2
C
3 3 33
2 1
3a2 b2
2C2
3 ab
ke(1,2)
2,488. 2 0,1786 2
a
1
b
0
,5285
ab
C2 33 33
2
3a2
2,488
1
b2
2 C
3 ab
2
2 1
3a2 b2
0,1786 2,488
b2
a2
2C2
3 ab
2 0,1786
3b2 a2
0,5285
ab
C 2 3 3 33
2,488. 2 0,1786 2
a
1
b
0,5285
ab
C2 33 33
2
3b2
2,488
a2
0,1786 2,488 1,786C22,488
2
b2 a ab
1
sym 2,488. 2 0,1786 2
a b
ab
0,5285
C2 3
3 3 3
0,5285
ab
C2 3 3
3 3
C
k14 abh 1
ab
2 ab
2
0,14 C 3 3 3 3 0,14 C 3 3 3 3
16
2 1 2 1 2 1 2 1 C
k15 abh 2 2 2 2 2 2 2 2 1
3 a b 3 a b 3 a b 3 a b 16
2 2 2 2 C
k17 abh 2 2,488 2 2 0,1786 2 2 0,1786 2 2 2,488 2 1
3a b 3a b 3a b 3a b 16
0,5285
ab
C2 3
3 3 3
0,14
ab
C 2 3 3
3 3
C
k18 abh 1
ab
2 ab
2
0,14 C 3 3 3 3 0,5285 C 3 3 3 3
16
1 1 1 1 C
k22 abh 2,488. 2 2 2,488. 2 0,1786 2 0,1786. 2 2 0,1786. 2 2,488 2 1
a b a b a b a b 16
2 2 2 2 C
k24 abh 2 2,488 2 2 2,488 2 2 0,1786 2 2 0,1786 2 1
3b a 3b a 3b a 3b a 16
2 1 2 1 2 1 2 1 C
k26 abh 2 2 2 2 2 2 2 2 1
3 a b 3 a b 3 a b 3 a b 16
2 1 2 1 2 1 2 1 C
k28 abh 2 2,488 2 2 0,1786 2 2 0,1786 2 2 2,488 2 1
3a b 3a b 3a b 3a b 16
G11 G17
16a2 16b2 16b2 16a2
C1 C2 1 1 C1 1 C2 C2
G12 G18 G23
16ab 16ab
C1 1 C1 1
2 2
C1 1 C1 1 1
2
G13 G22
16a 2
16b 2 16b2 16a2
C1 1 C2 C2 C1 1 1 C1 1
2
G14 G24 2
16ab 16b 16a2
C 1 1 C1 1 1 C1 1 1 C1 1 1
G15 1 G 26 2
16a 2
16b 2
16b 16a2
C1C2 1 1 C1 1 1 C1 1 C1 1 1
2
G16 G28
16ab 16ab 16b2 16a2
We calculate 12 components: G11 G12 G13 G14 G15 G16 G17 G18
G22 G24 G26 G28. Other components are calculated from
above relations.
f (,)
1 1
f (,).d.d
c1 1
2
C1 1
2
1
G11 C1 2 2
16a2 16b2 3a 3b
C1 C2 1 1 C1 C2
G12
16ab 4ab
C1 1
2
C1 1 1
1
G13
C1 2 2
16a 2
16b2 3a 6b
k38 k47 k56 k12 k33 k55 k77 k11 k35 k17
f (,)
1 1
f (,).d.d
f (,)
1 1
f (,).d.d
C1 1
2
C1 1 1 1
G17 C1 2 2
6a 3b
2
16b 16a2
C1 1
2
C1 1
2 1
G22
C1 2 2
16b2 16a2 3b 3a
f (,)
1 1
f (,).d.d
C1 1
2
C1 1 1
1
G28
C1 2 2
16b2 16a2 6a 3b
Element matrices
After computing, we get:
k11 k12 k13 k14 k15 k16 k17 k18
k22 k23 k24 k25 k26 k27 k28
k33 k34 k35 k36 k37 k38
C1h k44 k45 k46 k47 k48
ke
4ab k55 k56 k57 k58
k66 k67 k68
k77 k78
sym k88
With C1 is the elastic constant depending on the stress
state.
Assoc. Prof. Dr. Tich Thien TRUONG HCMUT 2007
FINITE ELEMENT METHOD
Chapter 3. FEM FOR 2D AND AXISYMMETRIC SOLIDS
b2 a 2 C2 2 b 2 a 2
k11 k16 ab k24
3 4 6
C2 2 a 2
b 2
C2
k12 k
ab 17 k25 ab
4 6 4
a 2 2b 2 C2 a 2 b 2
k13 k18 ab k26
6 4 6
C2 a 2 b 2 C2
k14 ab k22 k27 ab
4 3 4
a 2 b2 C2 b 2 2a 2
k15 k23 ab k28
6 4 6
4 b2 a 2 2 a 2 2b 2
k11 ; k12 ab C2 ; k13
3 3
2 b2 a 2
k14 C2 ab; k15 ; k16 ab C2
3
2 b 2 2 a 2 4 a 2 b 2
k17 ; k18 ab C2 ; k22
3 3
2 a 2 2 b 2 2 a 2 b 2
k23 k18 ; k24 ; k25 k16 ; k26
3 3
2 b 2 2a 2
k27 k14 ; k28 ; k33 k11 ; k34 k16 ; k35 k17 ;
3
k36 k14 ; k37 k15 ; k38 k12 ; k44 k22 ; k45 k18 ; k46 k28
k47 k12 ; k48 k26 ; k55 k11 ; k56 k12 ; k57 k13 ; k58 k14
k66 k22 ; k67 k18 ; k68 k24 ; k77 k11 ; k78 k16 ; k88 k22
Evaluation of me
1 1
mij hab N i N j d d
1 1
hab 1 1
16
1
(1 i )(1 j )d (1 i )(1 j )d
1
hab
(1 13 i j )(1 13 i j )
4
hab hab
E.g. m33 (1 3 1 1)(1 3 1 1) 4
1 1
4 9
Note: In practice, Gauss integration is often used
Evaluation of me
4 0 2 0 1 0 2 0
4 0 2 0 1 0 2
4 0 2 0 1 0
hab 4 0 2 0 1
me
9 4 0 2 0
4 0 2
sy. 4 0
4
Element matrices
Load vector of element inferred by uniform volume load {g} =
{gx gy}T: N g g
1 x x
N g g
1 y y
N2 g x gx
1 1
T x
g N2 g y gy
fe [N] dV habd d hab
g y Ng gx
Ve
1 1 3 x
N3 g y gy
N4 g x gx
N4 g y gy
Assoc. Prof. Dr. Tich Thien TRUONG 68 HCMUT 2007
FINITE ELEMENT METHOD
Chapter 3. FEM FOR 2D AND AXISYMMETRIC SOLIDS
Element matrices
For uniform distributed load on the side 23:
f sx
fe [N]T
dl
l 2 3 f
sy
y, v
0 0
4 (x4, y4) 3 (x3, y3)
0 0 (u4, v4) (u3, v3)
fsy
f sx f sx 2b fsx
1 f sy f sy 2a
l23 b 1 (x1, y1) 2 (x2, y2)
2 f sx f sx (u2, v2)
(u1, v1)
f sy f sy
x, u
0 0
0 0
Assoc. Prof. Dr. Tich Thien TRUONG HCMUT 2007
FINITE ELEMENT METHOD
Chapter 3. FEM FOR 2D AND AXISYMMETRIC SOLIDS
Element matrices
If uniform distributed load are applied on four sides, we
use superposition principle:
f sx12 .l12 f sx41.l41 f sx12 .a f sx41.b
12 12 41
f sy .l12 f sy .l41 f sy .a
41
f sy .b
f sx12 .l12 f sx23 .l23 f sx12 .a f sx23 .b
12 12 23
1 f sy .l12 f sy .l23 f sy .a
23
f sy .b
fe 23 23 34
2 f sx .l23 f sx .l34 f sx .b
34
f sx .a
f sy23 .l23 f sy34 .l34 f sy23 .b f sy34 .a
34 34 41
f sx .l34 f sx .l41 f sx .a
41
f sx .b
f 34 .l f sy41.l41 f sy34 .a f sy41.b
sy 34
Assoc. Prof. Dr. Tich Thien TRUONG HCMUT 2007
FINITE ELEMENT METHOD
Chapter 3. FEM FOR 2D AND AXISYMMETRIC SOLIDS
y4 C2 x2 y4 C2 x1 y1 C2 x1 y1 C2 x2
C1
e
S C 2 y 4 x2 C2 y 4 x1 C2 y1 x1 C2 y1 x2
4ab
x2 y4 x1 y4 x1 y1 x2 y1
with : xi xi x; yi yi y;
LINEAR QUADRILATERAL
ELEMENTS
Rectangular elements have limited application
Quadrilateral elements with unparallel edges are
more useful
Irregular shape requires coordinate mapping
before using Gauss integration
Coordinate mapping
y 4 (x4, y4)
3 (x3, y3)
4 (1, +1) 3 (1, +1)
2 (x2, y2)
1 (x1, y1) 1 (1, 1) 2 (1, 1)
x
Physical coordinates Natural coordinates
U h ( , ) N ( , )d e (Interpolation of displacements)
Coordinate mapping
x1
y coordinate at node 1
X( , ) N( , )x e 1
x2
x coordinate at node 2
where X , y2
xe
y x3
coordinate at node 3
y3
x4
N1 (1 )(1 )
1
y
4
coordinate at node 4
4
N 2 14 (1 )(1 ) 4
N 3 14 (1 )(1 ) x N i ( , ) xi
i 1
N 4 14 (1 )(1 )
4
y N i ( , ) yi
i 1
Coordinate mapping
4
On the side 23, substitute = 1 into x N i ( , ) xi
i 1
x 12 (1 ) x2 12 (1 ) x3 x 12 ( x2 x3 ) 12 ( x3 x2 )
or
y 12 (1 ) y2 12 (1 ) y3 y 12 ( y2 y3 ) 12 ( y3 y2 )
( x3 x2 )
Eliminating , y {x 12 ( x2 x3 )} 12 ( y2 y3 )
( y3 y2 )
y 4 (x4, y4)
3 (x3, y3)
4 (1, +1)
3 (1, +1)
2 (x2, y2)
1 (x1, y1) 1 (1, 1) 2 (1, 1)
x
Strain matrix
N i N i x N i y N i N i
x
x y
or J
N i N i x N i y N i N i
y
x y
x y
where J (Jacobian matrix)
x y
N 4 x1 y1
N1 N 2 N3
y2
x2
Since X ( , ) N ( , ) x e
, J
N1 N 2 N3 N 4 x3 y3
x4 y4
Strain matrix
N i N i (Relationship between differentials of
x
Therefore, N J
1 shape functions w.r.t. physical
i N i coordinates and differentials w.r.t.
y natural coordinates)
Element matrices
1 1
ke hB cB J d d
T
1 1
h
m e N NdV dx NT NdA hNT NdA
T
0
V A A
1 1
h NT N J d d
1 1
Remarks
Shape functions used for interpolating the coordinates
are the same as the shape functions used for
interpolation of the displacement field. Therefore, the
element is called an isoparametric element.
Note that the shape functions for coordinate
interpolation and displacement interpolation do not have
to be the same.
Using the different shape functions for coordinate
interpolation and displacement interpolation,
respectively, will lead to the development of so-called
subparametric or superparametric elements.
nd = (p+1)(p+2)/2
(1,0,p1) (0,1,p1)
Node i, IJ K p
(2,0,p2)
Argyris, 1968 :
L1
L2
i (I,J,K)
N i lII ( L1 )lJJ ( L2 )lKK ( L3 )
L3 (0,p1,1) ( L L 0 )( L L 1 ) ( L L ( 1) )
l ( L )
( L I L 0 )( L I L 1 ) ( L I L ( 1) )
(p,0,0) (p1,1,0) (0,p,0)
y, v 3 N1 N 2 N 2 (2 L1 1) L1
N 4 N 5 N 6 4 L1 L2
5 1
6 y, v 3 N1 N 2 N 3 (3L1 1)(3L1 2) L1
2
7 9
2 8 N 4 N9 L1 L2 (3L1 1)
4 6 2
1 9 N10 27 L1 L2 L3
10
x, u 2
5
1 4
Cubic element x, u
Quadratic element
( k 0 )( k 1 ) ( k k 1 )( k k 1 ) ( k n )
0
(0,0) (n,0)
6
(nine node quadratic element)
8 9
1
N 5 N 31D ( ) N11D ( ) (1 )(1 )(1 )
2
1
1 5 2 N 6 N 21D ( ) N 31D ( ) (1 )(1 )(1 )
1 2
N1 N11D ( ) N11D ( ) (1 ) (1 )
4 1
1 N 7 N 31D ( ) N 21D ( ) (1 )(1 )(1 )
N 2 N 21D ( ) N11D ( ) (1 ) (1 ) 2
4
1
1
N3 N 21D ( ) N 21D ( ) (1 )(1 )
N8 N11D ( ) N11D ( ) (1 )(1 )
4 2
1 N 9 N 31D ( ) N 31D ( ) (1 2 )(1 2 )
N 4 N11D ( ) N 21D ( ) (1 )(1 )
4
Assoc. Prof. Dr. Tich Thien TRUONG 83 HCMUT 2007
FINITE ELEMENT METHOD
Chapter 3. FEM FOR 2D AND AXISYMMETRIC SOLIDS
N j 14 (1 j )(1 j )( j j 1), j 1, 2, 3, 4
6
8 0 N j 12 (1 2 )(1 j ) , j 5, 7
N j 12 (1 j )(1 2 ) , j 6,8
1 5 2
=1
(eight node quadratic element)
Assoc. Prof. Dr. Tich Thien TRUONG 84 HCMUT 2007
FINITE ELEMENT METHOD
Chapter 3. FEM FOR 2D AND AXISYMMETRIC SOLIDS
4 7 3
3
4 7
8 6 6
8
1 2
5
1 5 2
CASE STUDY
Side drive micro-motor
CASE STUDY
10N/m
CASE STUDY
CASE STUDY
CASE STUDY
CASE STUDY
CASE STUDY
24 bilinear,
1 quadrilateral
41 0.0139
CASE 2
96 bilinear,
quadrilateral
129 0.0180
24 quadratic,
4 quadrilateral
105 0.0191
3.2.1.
3.2.2.
Remark 3.2.1.1
Remark 3.2.1.2
Note that (r , z) form a Cartesian coordinate system on the θ =
const planes, whereas (r , θ) form a polar coordinate system
on the z = const planes.
3.2.1.2.2. Displacement, Strains, Stresses
The displacement field is a function of r and z only, defined by
two components:
u r (r, z)
u(r, z) (3.2.1)
u
z (r, z)
e rr e rz e r
[e] e zr e zz e z (3.2.2)
er ez e
Because of the assumed axisymmetric state, er and ez
vanish, leaving only four distinct components:
e rr e rz 0
[e] e zr e zz 0 (3.2.3)
0 0 e
Each of these non-vanishing components is a function of r and
z only. As usual in preparation for finite element work, the non-
vanishing components are arranged as a 4 × 1 strain vector:
e rr
e
e zz (3.2.4)
e
rz
in which rz = erz + ezr = 2erz. This differs from the plane stress
case considered in the introductory course in the appearance
of e , the “hoop” or circumferential strain.
rr rz r
[] zr zz z (3.2.5)
r z
rr rz 0
[] zr zz 0 (3.2.6)
0 0
Each of the non-vanishing components is a function of r and z.
Collecting these four components into a stress vector:
rr
zz (3.2.7)
rz
where rz ≡ σrz. The difference with respect to the plane stress
problem is again the appearance of the “hoop” or
circumferential stress σ .
In matrix form:
e rr e rr r 0
e e 0 u
zz zz z r
e Du (3.2.9)
e e 1 r 0 u z
rz 2e rz z r
Remark 3.2.1.3
(3.2.12)
(3.2.13)
(3.2.15)
In (3.2.15) the strains are a slave field which are derived from
displacements.
The external work potential is the sum of contributions due to
body force and prescribed surface tractions:
(3.2.16)
dV 2rdA (3.2.17)
where dA is the element of area in the generating cross
section. Insertion in (3.2.15) and the second of (3.2.16)
reduces U and Wb to area integrals:
1
U 2 reT EedA (3.2.18)
2 A
dS 2rds (3.2.20)
Wt 2 rt uds
T
(3.2.21)
St
where (ur , uz) are the displacements of the “ring” point. Thus
the ubiquitous 2π term can be suppressed
A concentrated or point load F, however, can only act along
the z direction at points on the axis of revolution as illustrated
in Figure 3.2.2. The corresponding work term is
WF Fu z (3.2.23)
so the factor 2π is missing. To render this compatible with the
other energy terms the load is divided by 2π, so the
contribution to the external loads potential is
F
WF 2 u z (3.2.24)
2
1
(3.2.25)
2
to the strain energy density. This in turn introduces some
additional terms in the structure of the strain-displacement
matrix B, which translate into additional contributions in the
formation of the element stiffness matrix and consistent node
force vector.
(3.2.26)
Remark 3.2.2.2
Generating cross-section
B DN (3.2.27)
(3.2.28)
(3.2.29)
(3.2.30)
(3.2.31)
(3.2.32)
(3.2.33)
Premultiplying by BT gives
(3.2.34)
in which
(3.2.35)
(3.2.36)
(3.2.40)