Beruflich Dokumente
Kultur Dokumente
(Theory)
T. SAITO
1
1. INTRODUCTION
1-1. Section
bD 3 b3 D
D A bD Ix Iy
12 12
1) One-Dimensional Problem
L δ
N D
N
b
E
N
E
A L
EA
N
L
2
2) Two-Dimensional Problem
x x x x
x y
x
E
x x y
y x
E E E
y x
: Poisson Ratio y
E E
xy
xy
xy xy
xy
xy G xy
1
G E : Shear Modulus
2(1 )
1 1
0
x E E x 1 0 x
1 1 x E
y 0 y or y 1 0 y
1 1
2
E E
1 xy xy 0 0 xy
xy 0 0 2
G
3
1-3. Beam Theory
P
Q
M
Q : Shear Force
M : Moment
M
Q
P
M(x)
x d2y 1
2
M ( x)
dx EI
P
dM ( x)
Q ( x)
dx
y
Example )
d2y 1
2
M ( x), M ( x) Px
dx EI
d2y P
2 x
dx EI dy
at x L, 0 and y 0 :
dy 1 P 2 dx
x c1
dx 2 EI Therefore,
1 P 3 PL2 1 PL3
y x c1 x c 2 c1 , c2
6 EI 2 EI 3 EI
1 P 3 1 PL2 1 PL3
y x x
6 EI 2 EI 3 EI
4
1.4 Properties of Reinforced Concrete Structure
Unit Weight
Nominal Strength
Concrete Type Unit Weight (kN/m3)
(N/mm2 = MPa)
Material Parameters
Thermal
Young’s Modulus
Poisson’s Ratio Expansion
(N/mm2 = MPa)
Coefficient(1/℃)
22 000 ( Fc = 18 )
Concrete 25 000 ( Fc = 24 ) 1/6 1 x 10-5
28 000 ( Fc = 30 )
5
2. SIMPLE EXAMPLE FOR FEM FORMULATION
The problem is to obtain the deformation of a simple supported beam under various load
conditions.
If you change the load condition, you will get the different deformation pattern. Actually,
there are infinite variations for the deformation pattern.
etc.
We assume a particular function for the deformation pattern to fix the variation, such
as the following function:
v( x) a sin( x) (1)
L
x=0 x=L
x
v
v(0.5L) a (2)
A
δ
The relation between nodal displacement and element deformation is then expressed as,
v( x) sin( x) (3)
L
6
Step.4: Stiffness equation at the node
We obtain the relation between the nodal force and the nodal displacement, for example,
by using the “Principle of Virtual Work Method.”
P K (4)
A
δ
The above example tells the essence of the finite element analysis, which is:
“Assume the deformation pattern to reduce the degree of freedom of the element, then,
obtain the deformation from the limited number of nodal displacements.”
7
3. TRIANGULAR ELEMENT FOR PLANE ANALYSIS
etc.
To fix the variation for the deformation patterns, we assume a linear function for the
deformation pattern.
u ( x, y ) 1 2 x 3 y
(5)
v ( x, y ) 4 5 x 6 y
In a matrix form,
1
2
u 1 x y 0 0 0 3
(6)
v 0 0 0 1 x y 4
5
6
8
y
v3
3
y1 u3
v2
v1
y2 u2
2
y3 u1
1
x
x1 x2 x3
1
2
Node 1: u1 1 x1 y1 0 0 0 3 (7)
v 0
1 0 0 1 x1 y1 4
5
6
1
2
Node 2: u 2 1 x2 y2 0 0 0 3 (8)
v 0
2 0 0 1 x2 y 2 4
5
6
1
2
Node 3: u 3 1 x3 y3 0 0 0 3 (9)
v 0
3 0 0 1 x3 y 3 4
5
6
It is summarized as,
u1 1 x1 y1 0 0 0 1
u 2 1 x2 y2 0 0 0 2
u 1 x y3 0 0 0 3
3 3
(10)
v1 0 0 0 1 x1 y1 4
v 0 0 0 1 x2 y 2 5
2
v 0 0 y 3 6
3 0 1 x3
u = A α
9
We can obtain the coefficients 1 , 6 from the nodal displacements as,
α = A-1 u (11)
Substituting equation (11) into equation (6), the relation between nodal displacement
and element deformation is,
u1
u2
u 1 x y 0 0 0 1 u 3
A (12)
v 0 0 0 1 x y v1
v
2
v
3
u(x,y) = H(x,y) u
We obtain the relation between the nodal force and the nodal displacement, for example,
by using the “Principle of Virtual Work Method.”
P1 u1
Q3
P2 u2
3 P3 P u
3 K 3 (13)
Q2
Q1 v1
Q1 Q v
P2 2 2
2 Q v
P1 3 3
1
F = Ku
10
4. STIFFNESS MATRIX FOR TRIANGULAR ELEMENT
Stiffness matrix in equation (13) can be obtained from the “Principle of Virtual Work
Method,” which is expressed in the following form:
dv u T F (14)
T
u u1
u
x x 0 1 0 0 0 0 2
v 1 u 3
y 0 0 0 0 0 1 A (16)
y v1
xy u v 0 0 1 0 1 0 v
y x 2
v
3
ε = B u
x
1 0 x
y
E 1 0 (17)
1
2 1 y
xy 0 0 xy
2
σ = C ε
11
Substituting equation (16) into equation (17),
σ= C B u (18)
Bu CBu dv u B
T
T T
CBdv u u T F (19)
V V
F Ku , K B T CBdv (20)
V
12
5. FROM ELEMENT STIFFNESS MATRIX TO GLOBAL STIFFNESS MATRIX
P
u1
u2
① ③ u u3
3
(2) u4 u4
v Boundary Condition v
(1) 1 ① fixed … u1=v1=0 3
v2 v
4
② fixed … u 2=v 2=0
② ④ v
3
v
4
P3 k 22
( 2) ( 2)
k 23 ( 2)
k 25 ( 2)
k 26 u 3
( 2)
P4 k 32 k (1)
k ( 2)
k ( 2)
k k u 4
( 2) ( 2)
Element (2) … Q k ( 2) (23)
33 33 35 36 36
k ( 2)
k ( 2)
k 56( 2) v3
3 52 53 55
Q k ( 2) k 63(1) k 63( 2 ) ( 2)
k 66(1) k 66( 2 ) v 4
4 62 k 65
F = K U
13
Set the load condition,
P3 0
P4 0
Q P (24)
3
Q 0
4
u3 0
u4 1 0
v K P (25)
3
v 0
4
14
6. HIGHER ORDER ELEMENT
Before After
deformation
Before
After
second order function of coordinate.
Before
u 1 2 x 3 y 4 x 2 5 xy 6 y 2
(26)
v 7 8 x 9 y 10 x 2 11 xy 12 y 2
In a matrix form,
1
u 1 x y x 2
xy y 2
0 0 0 0 0 0 2
(27)
v 0 0 0 0 0 0 1 x y x 2
xy y 2
12
5
In order to define the second order function, we need
4 an additional node in the middle of each side of the
v3
6 v2 triangle. At the result, the total number of nodes in
u3 one element is 6.
u2 3
1 2
15
The displacement of the element nodes are then expressed as,
u1 1 x1 y1 x12 x1 y1 y12 | 1
u 2 1 x2 y2 x 2
2 x2 y 2 y 2
2 | 0 2
|
u 6 1 x6 y6 x 2
6 x6 y 6 y 2
6 | 6
| (28)
v1 | 1 x1 y1 x12 x1 y1 y12 7
v2 | 1 x2 y2 x 22 x2 y 2 y 22 8
0 |
v6 | 1 x6 y6 x 2
6 x6 y 6 y 62 12
u = A α
u1
u2
u 1 x y x2 xy y2 0 0 0 0 0 0 1 u 6
A (29)
v 0 0 0 0 0 0 1 x y x 2
xy y 2 v1
v2
v
6
u(x,y) = H(x,y) u
As the same as the linear triangular element, the stiffness equation is obtained as
P1 u1
P2 u2
P6 u6
Q K v (30)
1 1
Q2 v2
Q v
6 6
F = Ku
16
The process is summarized as follows:
(1) Translate external forces into equivalent nodal force,
F = {P1, …, P6, Q1, …, Q6}T
(2) Calculate the nodal displacements from the stiffness equation,
u = K-1 F
(3) Obtain the element deformation from the nodal displacement.
u(x,y) = H(x,y)u
17
7. INTERPOLATION FUNCTION
u ( x) a 0 a1 x
u2 or
u1 1 2
x a
x1 x2 u ( x) 1 x 0 (31)
a1
l
The next step is to obtain the coefficients, a0, a1, from the nodal displacements. From
the relations:
u1 a 0 a1 x1
u 2 a 0 a1 x 2
or
u1 1 x1 a 0
(32)
u 2 1 x 2 a1
u = A α
The coefficients are obtained as, α = A-1 u. Then, the relation between the deformation
and the nodal displacements is,
u
u ( x) 1 x A 1 1 (33)
u2
1, x u1 1, x u 2
h1 ( x) , h2 ( x ) (35)
0, x u1 0, x u 2
18
From these characteristics, the functions are easily obtained as,
x2 x x x1
h1 ( x) , h2 ( x ) (36)
l l
In the same manner, if we assume a second order function for the deformation pattern,
the deformation can be directly expressed using interpolation functions as follows:
u2 u2
u1 1 u1 1 2
2 x
x
x1 3 x2
x1 x2
l l
h1(x)u1 h1(x)u1
u1 u1
x x
x1 x2 x1 x2
h2(x)u2 u2 u2
h2(x)u2
x x
x1 x2 x1 x2
u3 h3(x)u3
x
x1 x2
19
8. NATURAL COORDINATE
t = 7 + 0.5 x
x t
x=2(t–7)
5
W w( x)dx (38)
0
5 Global relationship:
4 x 2(t 7) (39)
3 dx
2
2
dt Local relationship:
1
dx 2dt (40)
0 t
6 7 8 9 10
Substituting equations (39) and (40) into (38), the total weight is expressed in t-axis as,
9.5
W 2 w x(t ) dt (41)
7.5
20
Next we consider a more complicated scale to measure the total weight of the pencil.
x
x = x(t)
x
0 1 2 3 4 5 6 5
t 4
α β 3 dx
2
w(x) : distribution of weight dt
1
x 0 t
α t β
w(x
x
x x+dx
Where dx(t)/dt represents the first derivative of x(x) by the variable t, which correspond
to the slope of x(t) at t. Substituting equations (42) and (43) into (38), the total weight
will be expressed in t-axis as,
W w x(t )
dx(t )
dt (44)
dt
f (t )dt
1
1 f (t1 ) 2 f (t 2 ) n f (t n ) (45)
21
f(tn)
f(t)
f(t2)
f(t1)
t
-1 t1 t2 tn +1
Going back to the problem to evaluate the total weight of the pencil, if we select the
scale which has the range [-1, 1], we can use the numerical integration method to
evaluate the integration.
x
0 1 2 3 4 5 6
t
-1 +1
1
f (t ) w x(t )
dx(t )
W f (t )dt ,
1
dt
(46)
22
9. ISOPARAMETRIC ELEMENT
We now introduce the natural coordinate for the example of one dimensional element.
x
u2 x2 x(t)
u1 1 2
x
x1 x2 x1
t
t
-1 +1
-1 +1
If we assume the linear transfer function x(t) between x-axis and t-axis, x(t) will be
expressed as
x(t ) h1 (t ) x1 h2 (t ) x 2 (47)
where
1 1
h1 (t ) (1 t ), h2 (t ) (1 t ) (48)
2 2
23
Advantages of using isplarametric elements are summarized below:
n
(1) The relation u (t ) h (t )u
i 1
i i does not require the calculation of inverse matrix.
n
(2) The relation x(t ) h (t ) x
i 1
i i enables to use the numerical integration method.
(3) Both functions u(t) and x(t) are expressed using the same interpolation functions.
24
10. SYSTEMATIC FORMULATION OF INTERPOLATION FUNCTION
As presented here, if you increase a node to define the second order function for the
deformation, the interpolation function changes in the following manners:
- Modify the existing interpolation functions, h1 and h2,
- Define a new interpolation function, h3.
25
(2) Two dimensional element
h9 = (1 r 2 )(1 s 2 )
26
(3) Three dimensional element
1
( , i ) (1 i ) for i 1
2 r , s, t
( , i ) (1 2 ) for i 0
27
11. STIFFNESS MATRIX FOR ISOPARAMETRIC ELEMENT
Using a two dimensional isoparametric element, we will see the procedure to derive the
stiffness matrix.
The coordinate transfer function {x, y} is expressed using the interpolation functions as
follows:
4
1 1 1 1
x(r , s ) hi ( r , s ) xi (1 r )(1 s ) x1 (1 r )(1 s ) x 2 (1 r )(1 s ) x3 (1 r )(1 s ) x 4
i 1 4 4 4 4
4
1 1 1 1
y (r , s ) hi (r , s ) y i (1 r )(1 s) y1 (1 r )(1 s) y 2 (1 r )(1 s ) y 3 (1 r )(1 s ) y 4
i 1 4 4 4 4
(51)
The deformation function {u, v} is also expressed using the same interpolation functions.
4
1 1 1 1
u (r , s ) hi ( r , s )u i (1 r )(1 s )u1 (1 r )(1 s )u 2 (1 r )(1 s )u 3 (1 r )(1 s )u 4
i 1 4 4 4 4
4
1 1 1 1
v( r , s ) hi (r , s )v i (1 r )(1 s )v1 (1 r )(1 s )v 2 (1 r )(1 s )v 3 (1 r )(1 s )v 4
i 1 4 4 4 4
(52)
Stiffness matrix can be obtained from the “Principle of Virtual Work Method,” which is
expressed in the following form:
28
dv u T F (53)
T
u 4
hi
i 1 x
ui
x x
v 4
hi
y vi
y i 1 y
xy u v 4 hi 4
hi
y x y u i x vi
i 1 i 1
u1
h1 h2 h3 h4 v1
0 0 0 0 u
x x x x 2
h1 h2 h3 h4 v 2
0 0 0 0 u
y y y y 3
h1 h1 h2 h2 h3 h3 h4 h4 v3
y x y x y x y x
u4
v
4
ε = B u (55)
x
1 0 x
y
E 1 0 (56)
1 1 y
xy 0 0 xy
2
σ = C ε
29
Substituting equation (55) into equation (56),
σ= C B u (57)
Bu CBu dv u B T CBdxdy u u T F
T
(58)
T
V V ( x, y )
F Ku , K B T CBdv (59)
V
If we assume the constant thickness of the plate (= t), using the relation dv tdxdy ,
K t B (60)
T
CBdxdy
V ( x, y )
Since this integration is defined in x-y coordinate, we must transfer the coordinate into
r-s coordinate to use the numerical integration method. Introducing the Jacobian
matrix,
x y
J r r ; Jacobian Matrix (61)
x y
s s
the above integration is expressed in r-s coordinate as,
( x, y )
1 1
K t B xr , s , y r , s CB xr , s , y r , s
T
drds (62)
1 1
(r , s )
where
30
x y
( x, y )
det J r r (63)
(r , s) x y
s s
1) Evaluation of Jacobian Matrix
x y hi x hi
4 4
r i 1 r
i r yi
J r i 1 (64)
x y 4 hi hi
4
s
xi yi
s i 1 s i 1 s
h1 h h h
The derivatives , , 4 , 1 ,, 4 are calculated as,
x x y y
In a matrix form,
31
h1 h2 h3 h4
J 1 r r r r (66)
h1 h2 h3 h4
s s s s
h1 1 h1 1
(1 s ) (1 r )
r 4 s 4
h2 1 h2 1
(1 s ) (1 r )
r 4 s 4
, (67)
h3 1 h3 1
(1 s ) (1 r )
r 4 s 4
h4 1 h4 1
(1 s ) (1 s )
r 4 s 4
32
12. NUMERICAL INTEGRATION METHOD
f(0.7746)
f(t)
f(-0.7746)
f(0)
t
-1 -0.7746 0 +0.7746 +1
( x, y )
1 1
K t B xr , s , y r , s CB xr , s , y r , s
T
drds
1 1
(r , s )
1 1
t F (r , s )drds (69)
1 1
3 3
t i j F (ri , s j )
i 1 j 1
where
( x, y )
F (r , s ) B xr , s , y r , s CB xr , s , y r , s
T
(r , s)
33