0% found this document useful (0 votes)
33 views46 pages

F Inite Element Method

The chapter discusses finite element analysis for 3D solid elements. It covers the tetrahedron element and hexahedron element. For the tetrahedron element, it describes the shape functions, strain matrix, and element matrices. Similarly, it does the same for the hexahedron element. It also briefly mentions higher order elements and elements with curved surfaces.
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
33 views46 pages

F Inite Element Method

The chapter discusses finite element analysis for 3D solid elements. It covers the tetrahedron element and hexahedron element. For the tetrahedron element, it describes the shape functions, strain matrix, and element matrices. Similarly, it does the same for the hexahedron element. It also briefly mentions higher order elements and elements with curved surfaces.
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 46

Finite Element Method

A Practical Course

CHAPTER 6:

FEM FOR 3D SOLIDS


CONTENTS
 INTRODUCTION
 TETRAHEDRON ELEMENT
– Shape functions
– Strain matrix
– Element matrices
 HEXAHEDRON ELEMENT
– Shape functions
– Strain matrix
– Element matrices
– Using tetrahedrons to form hexahedrons
 HIGHER ORDER ELEMENTS
 ELEMENTS WITH CURVED SURFACES
 CASE STUDY
INTRODUCTION
 For 3D solids, all the field variables are dependent
of x, y and z coordinates – most general element.
 The element is often known as a 3D solid element
or simply a solid element.
 A 3-D solid element can have a tetrahedron and
hexahedron shape with flat or curved surfaces.
 At any node there are three components in x, y and
z directions for the displacement as well as forces.
TETRAHEDRON ELEMENT
 3D solid meshed with tetrahedron elements
TETRAHEDRON ELEMENT
Consider a 4 node tetrahedron element
w4
 u1  
v  
 node 1 4= l
 1  v4
 w1  
  w3
u4
 u2   w1

 v2   node 2 v3
   i
 w2   1= u3 3= k
v1
de   
fsz
 u3   u1
 v3   z=Z fsy
 node 3 w2
  
fsx
 w3  
u  y=Y u2
 4  2= j
 v4   x=X
 node 4 u2
w  
 4 
z = Z


Shape functions
U h ( x, y, z )  N( x, y, z )d e
 node
  1   node
   2   node
  3    node
   4 

 N1 0 0 N2 0 0 N3 0 0 N4 0 0
 0 
where N  0 N1 0 0 N2 0 0 N3 0 0 N4
 0 0 N1 0 0 N2 0 0 N3 0 0 N 4 

Use volume coordinates (Recall Area coordinates for 4=l


2D triangular element)
1=i P 3=k
V P 234
L1 
V1234 z
y
x 2=j
Shape functions
VP134 VP124 VP123
Similarly, L2  , L3  , L4 
V1234 V1234 V1234

Can also be viewed as ratio of distances

d d d d 4=l
L1  P  234 , L2  P 134 , L3  P 124 , L4  P 123
d1 234 d1 234 d1 234 d1 234

L1  L2  L3  L4  1 (Partition of unity) 1=i P 3=k

since z

VP 234  VP134  V P124  VP123  V1234 x


y
2=j
Shape functions
1 at the home node i
Li   (Delta function property)
0 at the the remote nodes jkl

x  L1 x1  L2 x 2  L3 x3  L4 x 4
y  L1 y1  L2 y 2  L3 y 3  L4 y 4
z  L1 z1  L2 z 2  L3 z 3  L4 z 4 L1  L2  L3  L4  1

 1  1 1 1 1   L1 
x  x x x3 x 4   L2 
   1 2
    
 y   y1 y 2 y3 y 4   L3 
 z   z1 z 2 z3

z 4   L4 
Shape functions (Adjoint matrix)

 L1   a1 b1 c1 d1  1  i= 1,2
Therefore, L   d 2   x  i
 2  1 a 2 b2 c2
    l = 4,1
 L3  6V  a3 b3 c3 d 3  y l j
 L4    
a 4 b4 c4 d 4  z  j = 2,3
k
(Cofactors) k = 3,4
xj yj z j  1 yj zj
where ai  det  xk yk zk  , bi   det 1 yk z k 
 xl yl zl  1 yl z l 
yj 1 zj  yj zj 1
ci   det  yk 1 z k  , d i   det  y k zk 1 
 yl 1 zl   yl zl 1 
Shape functions
 1 xi y i zi 
 1 xj yj z j 
1
V   det  (Volume of tetrahedron)
6  1 xk y k zk 
 
 1 xl y l zl 

1
Therefore, N i  Li  (ai  bi x  ci y  d i z )
6V
Strain matrix
 x 0 0 
Since, U ( x, y, z )  N( x, y, z )d e
h
 0   y 0 
 
 0 0  z 
Therefore,   LU  LNd e  Bd e where B  LN   0  z  y  N
 
 z 0  x 
 
 b1 0 0 b2 0 0 b3 0 0 b4 0 0    y   x 0 
0 c1 0 0 c2 0 0 c3 0 0 c4 0 

1 0 0 d1 0 0 d 2 0 0 d3 0 0 d 4 
B  
2V  0 d1 c1 0 d 2 c2 0 d3 c3 0 d 4 c4 
 d1 0 b1 d2 0 b2 d3 0 b3 d3 0 b4 
 
 c1 b1 0 c2 b2 0 c3 b3 0 c4 b4 0 

(Constant strain element)


Element matrices

k e   BT cBdV Ve BT cB
Ve

 N11 N12 N13 N14 


N N 22 N 23 N 24 
m e    N NdV   
T  21
dV
Ve Ve
 N 31 N 32 N 33 N 34 
 
 N 41 N 42 N 43 N 44 

Ni N j 0 0 
 
where N ij   0 Ni N j 0 
 0 0 N i N j 

Element matrices
Eisenberg and Malvern, 1973 :
m !n ! p !q !

Ve
L1m Ln2 L3p Lq4 dV 
(m  n  p  q  3)!
6Ve

2 0 0 1 0 0 1 0 0 1 0 0
 2 0 0 1 0 0 1 0 0 1 0 

 2 0 0 1 0 0 1 0 0 1
 
 2 0 0 1 0 0 1 0 0
 2 0 0 1 0 0 1 0
 
V 2 0 0 1 0 0 1
m e  e 
20 2 0 0 1 0 0
 
 2 0 0 1 0
 2 0 0 1
 
 sy. 2 0 0
 
 2 0
 2
Element matrices

Alternative method for evaluating me: special natural


coordinate system
4= l
=constant

3= k
1= i Q =1
=0

z
P
2= j
y
=1
x
Element matrices

4= l
=constant

3= k
1= i =1
=0
P
z

2= j
y
=0
x
Element matrices

4= l =0
=constant

R Q
3= k
1= i =1
=1 P

2= j
y
=1
x
Element matrices
xP   ( x3  x2 )  x2 xB   ( xP  x1 )  x1   ( x3  x2 )   ( x2  x1 )  x1
yP   ( y3  y2 )  y2 yB   ( yP  y1 )  y1   ( y3  y2 )   ( y2  y1 )  y1
z P   ( z3  z 2 )  z 2 z B   ( zP  z1 )  z1   ( z3  z2 )   ( z2  z1 )  z1
x  x4   ( x4  xB )  x4   ( x4  x1 )   ( x2  x1 )   ( x2  x3 )
y  y4   ( y4  yB )  y4   ( y4  y1 )   ( y2  y1 )   ( y2  y3 )
z  z4   ( z4  z B )  z4   ( z4  z1 )   ( z2  z1 )   ( z2  z3 )
4= l =0
= constant
N1  (1   )
N 2   (1   ) O 3= k
=1
1= i
=1
N 3   =0
B =1
=0
P [xP(x3x2)+x2, yP(y3y2)+y2,zP(z3z2)+z2]
N 4  (1   ) z
=1
B [xB(xPx1)+x1, yB(yPy1)y1, zB(zPy1)z1]
=constant
y =1 2 = j = constant
=0
x =1
z=Z O [x=(1)(x4xB)xB, y=(1)(y4yB)yB, z=(1)(z4zB)zB]
Element matrices
 x x x 
    

y y y 
Jacobian: J
    
 z z z 
 
    

 x21   x31  x31  x41   x21   x31


det[J ]   y21   y31  y31  y41   y21   y31  6V  2
 z21   z31  z31  z41   z21   z31
1 1 1
m e   NT NdV    NT Ndet[J]d d d
0 0 0
Ve

 N11 N12 N13 N14 


1 1 1
N N 22 N 23 N 24 
m e  6Ve      2  21 d d d
0 0 0  N 31 N 32 N 33 N 34 
 
 N 41 N 42 N 43 N 44 
Element matrices
 f sx 
   0 31 
fe   [N]T
 f sy  dl
l 2 3  
f   sx 
f
 sz   f sy 
 
1  f sz 
For uniformly distributed load: fe  l23  
2  sx 
f
w4  f sy 
 
4= l
v4  f sz 
 0 
  31 
w3
u4
w1
v3
1= i u3 3= k
v1
f sz
z= Z u1
w2 f sy
f sx
y= Y
u2
x= X 2= j
z= Z u2
HEXAHEDRON ELEMENT
 3D solid meshed with hexahedron elements

P P’

P’’’ P’’
Shape functions 5 8

U  Nd e
6 fsz 4
 d e1  displacement components at node 1 7
d  displacement components at node 2 1
 e2  0
d e3  displacement components at node 3 z fsy
  2
fsx
d  displacement components at node 4
de   e4  0

d e 5  displacement components at node 5 y 3


d e 6  displacement components at node 6
  x
d e 7  displacement components at node 7
d  displacement components at node 8
 e8 
Ni 0 0
N i   0 0 
 u1 
 
d ei   v1  (i  1, 2, ,8)
Ni (i  1,2, ,8)
w 
 1
 0 0 N i 

N   N1 N2 N3 N4 N5 N6 N7 N8 
Shape functions

(-1, -1, 1)5 8(-1, 1, 1)
5 8
(1, -1, 1)6

6 fsz 4 7 (1, 1, 1)
7 
1 (-1, -1, -1)1 4(-1, 1, -1)
0
z fsy
2 fsx
(1, -1, -1)2
0
y 3 3(1, 1, -1)

x

8
x   N i ( , ,  ) xi 1
i 1 N i  (1   i )(1   i )(1   i )
8
y   N i ( , ,  ) y i
8
i 1
8
z   N i ( , ,  ) z i
(Tri-linear functions)
i 1
Strain matrix
  LU  LNd e  Bd e
B   B1 B2 B3 B4 B5 B6 B7 B8 
whereby
N i x 0 0 
 0 N y 0 
 i 
 0 0 N i z 
B i  LN i   
 0  N i  z N i y 
 N i z 0 N i x 
 
N i y N i x 0 

Note: Shape functions are expressed in natural


coordinates – chain rule of differentiation
Strain matrix
N i N i x N i y N i z
  
 x  y  z 
N i N i x N i y N i z Chain rule of
  
 x  y  z 
differentiation
N i N i x N i y N i z
  
 x  y  z 

 N i   N i   x y z 
    
 
 x 
   
 N 
 N i   i x y z 
    J   where J
    y      
 N i   N i   x y z 
    z   
      
Strain matrix
8 8 8

Since, x   N i ( , ,  ) xi , y   N i ( , ,  ) yi , z   N i ( , ,  ) z i
i 1 i 1 i 1

 x1 y1 z1 
 z2 
 N1 N 2 N 3 N 4 N5 N6 N 7 N8   x2 y2
           x3 y3 z3 
  
 N N 2 N 3 N 4 N5 N6 N 7 N8   x4 y4 z4 
J 1 
           x5 y5 z5 
 
 N1 N 2 N 3 N 4 N5 N6 N7 
N8  x6 y6 z6 
 
           x7 y7 z7 
 
 x8 y8 z8 
 8 N i 8
N i 8
N i 
 xi 
 y i

 zi  
 i 81 i 1 i 1

N i 8
N i 8
N i
or J   xi  y  z 
 i 1  i 1
i
 i 1
i
 
 8 N i 8
N i 8
N i 
 xi  y i  z i

 i 1  i 1  i 1  
Strain matrix
 N i   N i 
 x    
 N   
 i 1  N i  Used to replace derivatives
 J  
 y     w.r.t. x, y, z with
 N i   N i  derivatives w.r.t. , , 
 z    
 

N i x 0 0 
 0 N y 0 
 i 
 0 0 N i z 
B i  LN i   
 0  N i  z N i y 
N i z 0 N i x 
 

 i
N y  N i x 0 
Element matrices
1 1 1
ke   B cBdV     B TcB det[J ]d dd
T
1 1 1
Ve

1 1 1 n m l
Gauss integration: I  1 1 1 f ( , )dd   wi w j wk f ( i , j ,  j )
i 1 j 1 k 1

1 1 1
m e   N NdV  T
  NT Ndet[J]d d d
1 1 1
Ve
Element matrices
For rectangular hexahedron:

det[J]  abc  Ve / 8

m11 m12 m13 m14 m15 m16 m17 m18 


 m 22 m 23 m 24 m 25 m 26 m 27 m 28 

 m 33 m 34 m 35 m 36 m 37 m 38 
 
m 44 m 45 m 46 m 47 m 48 
me  
 m55 m 56 m 57 m 58 
 
 m 66 m 67 m 68 
 sy. m 77 m 78 
 
 m88 
Element matrices
(Cont’d)
1 1 1
where m ij  1 1 1 abcN i N j ddd
Ni 0 0 N j 0 0 
0  
0   0
1 1 1
 abc    Ni Nj 0  ddd
1 1 1
 0 0 N i   0 0 N j 
Ni N j 0 0 
1 1 1  
 abc    0 Ni N j 0  ddd
1 1 1
 0 0 N i N j 

Element matrices
(Cont’d)
mij 0 0
 
or m ij   0 mij 0
0 0 mij 

1 1
where mij  abc  
1 1
N i N j ddd
abc 1 1 1

64 1
(1   i  )(1   j  ) d 1
(1   i )(1   j ) d 1
(1   i )(1   j  )d

hab
 (1  13  i j )(1  13  i j )(1  13  i j )
8
Element matrices
(Cont’d)
abc abc
E.g. m33  (1  3  1  1)(1  3  1  1)(1  3  1  1)  8 
1 1 1
8 216
8 abc
m11  m22  m33  m44  m55  m66  m77  m88 
216
4 abc
m12  m23  m34  m56  m67  m78  m14  m58  m15  m26  m37  m48 
216
2 abc
m13  m24  m16  m25  m36  m47  m57  m68  m27  m38  m45  m18 
216
1abc
m17  m28  m35  m46 
216
Element matrices
(Cont’d)
8 4 2 4 4 2 1 2
 8 4 2 2 4 2 1 
 
 8 4 1 2 4 2
  Note: For x direction only
abc  8 2 1 2 4 
m ex 
216  8 4 2 4
 
 8 4 2 
 sy. 8 4
 
 8 

(Rectangular hexahedron)
Element matrices
 f sx 
 
f e   [N]T  f sy  dl 5 8
l 3 4
f 
 sz   0 31 
 0  fsz
 31  6 4

For  f sx  1
7

uniformly   f  0

 sy  z fsy
 f sz 
fsx
distributed 2

  f  0
load: 1  sx 
f e  l 3  4    y 3
2  f sy 
 f sz 
x

 0 
 31 
 0 31 
 
 0
 31 
 0 31 
Using tetrahedrons to form hexahedrons
 Hexahedrons can be made up of several
tetrahedrons 8
5
4
8 1
1

Hexahedron 5 8 3
6 8
made up of 5
4
tetrahedrons: 1
6 7
8 6 7
2 3
3
6
6
1
1 3
3 2
Using tetrahedrons to form hexahedrons
 Element matrices can 5 8

be obtained by
4
assembly of 1
6 7
tetrahedron elements 8
5 8 2 3

6 4 7
1 6
4
2 3
2
Break into three

Hexahedron 5 8
5
made up of 6 1 6 4
tetrahedrons: 1 4 4
6
2 6
HIGHER ORDER ELEMENTS
 Tetrahedron elements

10 nodes, quadratic:
4
N i  (2 Li -1)Li for corner nodes i  1,2,3,4
9
N 5  4 L2 L3 
8
N 6  4 L1 L3 
 7
2
N 7  4 L1 L2  10
 for mid-edge nodes
N8  4 L1 L4 
1 5

N 9  4 L2 L4  6
 3
N10  4 L3 L4 
HIGHER ORDER ELEMENTS
 Tetrahedron elements (Cont’d) 4

20 nodes, cubic: 14
12
N i  12 (3Li  1)(3Li  2)Li for corner nodes i  1,2,3,4 20 13
N 5  92 (3L1  1)L1L3 N11  92 (3L1  1)L1L4  11 16 2
8
N 6  92 (3L3  1)L1L3 N12  92 (3L4  1)L1L4  7 17
 19 9
N 7  92 (3L1  1)L1L2 N13  92 (3L2  1)L2 L4  1 5 18 15
 for edge nodes
N8  92 (3L2  1)L1 L2 N14  92 (3L4  1)L2 L4  5 10
N 9  92 (3L2  1)L2 L3 N15  92 (3L3  1)L3 L4  6
 3
N10  92 (3L3  1)L2 L3 N16  92 (3L4  1)L3 L4 
N17  27 L2 L3 L4 
N18  27 L1 L2 L3 

 for center surface nodes
N19  27 L1L3 L4 
N 20  27 L1L2 L4 
HIGHER ORDER ELEMENTS


 Brick elements nd=(n+1)(m+1)(p+1) nodes 


(n,m,p)

Lagrange type:
(n,m,0)
Ni  N N N 1D
I
1D
J
1D
K  l ( )l ( )l ( )
n
I
m
J
p
K i(I,J,K)

where (0,0,0) (n,0,0)

(   0 )(  1 )  (   k 1 )(   k 1 ) (   n )
lkn ( ) 
( k   0 )( k  1 )  ( k   k 1 )( k   k 1 )  ( k   n )
HIGHER ORDER
ELEMENTS


(-1, -1, 1)5 15 8(-1, 1, 1)


 Brick elements (Cont’d) 16
14
(1, -1, 1)6 13 3
19
Serendipity type elements: 20
7 (1, 1, 1)


17 18
(-1,-1,-1)1
11(-1,0,-1) 4(-1, 1, -1)
12(0-1,-1)
20 nodes, tri-quadratic: 10(0,1,-1)
(1, -1, -1)2
9(1,0,-1) 3(1, 1, -1)
N j  18 (1   j )(1   j )(1   j )( j   j   i  2) 

for corner nodes j  1,  , 8


N j  14 (1   2 )(1   j )(1   j ) for mid-side nodes j  10,12,14,16
N j  14 (1   2 )(1   j )(1   j ) for mid-side nodes j  9,11,13,15
N j  14 (1   2 )(1   j )(1   j ) for mid-side nodes j  17,18,19, 20
HIGHER ORDER ELEMENTS 

 Brick elements (Cont’d)



32 nodes, tri-cubic:
Nj  1
64 (1   j )(1   j )(1   j )(9 2  9 2  9 2  19)
for corner nodes j  1,  , 8
Nj  9
64 (1   2 )(1  9 j )(1   j )(1   j ) 

for side nodes with  j   13 , j  1 and  j  1


Nj  9
64 (1   2 )(1  9 j )(1   j )(1   j )
for side nodes with  j   13 ,  j  1 and  j  1
Nj  9
64 (1   2 )(1  9 j )(1   j )(1   j )
for side nodes with  j   13 ,  j  1 and  j  1
ELEMENTS WITH CURVED
SURFACES 4
4
9
9 8
8
10 2
7
7 2 10
5 1
1 5
6 3 6
3

2 10 3
2
9 10 9 11
14
1 14 11 3 1 15
12 12 4
4 15 
13 6 18 13 6 16
7 7
18
17 16 17
5 19 19
20 5
8 8
 20
CASE STUDY
 Stress and strain analysis of a quantum dot
heterostructure
Material E (Gpa) 
GaAs cap layer
GaAs 86.96 0.31

InAs 51.42 0.35

InAs wetting
InAs quantum dot
layer

GaAs substrate
CASE STUDY
CASE STUDY
30 nm

30 nm
CASE STUDY
CASE STUDY

You might also like