Numerical
Evaluation of
Dynamic
Response
Chapter 5
January 26, 2015
CC Chang / HKUST
Time-Domain Analysis
given
mx i cx i kx i pi
mx i 1 cx i 1 kx i 1 pi 1
seek
Displacement x
mx (t) cx (t) kx(t) p(t)
Loading p
n
e
v
i
g
pi
p1 p
pi+1
Time stepping k
e
x1
x2
t1 t 2
time
se
xi
xi+1
t i tj
time
Time-Domain Analysis
Interpolation of Excitation Method
Central Difference Method
Newmarks Method
Wilson- Method
Commercial Software
Interpolation of Excitation
mx i cx i kx i pi
t i t t i 1
pi
linear pi
t i pi+1
pi
t i
pi
mx t cx t kx t pi
t i
mx i 1 cx i 1 kx i 1 pi 1
ti
pi
ti 1
Interpolation of Excitation
pi
mx cx kx pi
t i
constant
linear ramp
Solution of x(t) consists of 3 parts
- Step response due to pi
- Ramp response due to Pi
- Free vibration response
t i
Interpolation of Excitation
Displacement
x i 1 A x i B x i C p i D p i 1
where
A e t (
sin D t cos D t )
1
1
B e t (
sin D t )
D
2
1 2
t 1 2
C
e
[(
) sin D t (1
) cos D t ]
k t
D t
t
1 2
1
2
2
t 2 1
D 1
e
(
sin D t
cos D t )
k
t
D t
t
Interpolation of Excitation
Velocity
x i 1 A x i B x i C p i D p i 1
where
A e t (
sin D t )
B e t (cos D t
sin D t )
1
1
t
C e
[(
) sin D t cos D t ]
2
2
k t
t
t
1
1
1 e t ( sin D t cos D t )
D
kt
1 2
Interpolation of Excitation
Interpolation of Excitation
Note:
Linear system
SDOF system
Accuracy depends on the linear
interpolation of loading p
Central Difference Method
Finite difference
mx i cx i kx i p i
x x i 1 x i 1
i
2 t
x 2x i x i 1
x i i 1
t 2
xi
x i 1
x i 1
x i
x i 1 x i 1
2t
x i 1 2 x i x i 1
x i 1 x i 1
m
c
kx i pi
2
2 t
( t )
m
c
m
c
2m
[
]x i 1 pi [
]x i 1 [ k
]x i
( t ) 2 2t
( t ) 2 2t
( t ) 2
unknown
Pi
Central Difference Method
[
m
( t ) 2
c
m
c
2m
]x i 1 pi [
]x i 1 [ k
]x i
2
2
2t
2 t
( t )
( t )
At t=0, given x 0 and x 0
p cx 0 kx 0
x 0 0
m
x x1 x 1
0
2 t
x 2x 0 x 1
x 0 1
t 2
find
need
To find x1 we need x0 and x-1
mx i cx i kx i p i
x x i 1 x i 1
i
2t
x 2x i x i 1
x i i 1
t 2
( t ) 2
x 1 x 0 tx 0
x 0
2
Central Difference Method
Summary
Given m, c, k, p, x 0 and x 0
p0 cx 0 kx 0
( t ) 2
x 1 x 0 tx 0
x 0
x 0
m
2
m
c
2m
m
c
a
bk
k
( t ) 2 2 t
( t ) 2
( t ) 2 2 t
For step i
pi pi a x i 1 b x i
p
x i 1 i
k
x
x i 1
x i i 1
2 t
x 2 x i x i 1
x i i 1
t 2
Time-Domain Analysis
given
mx i cx i kx i pi
mx i 1 cx i 1 kx i 1 pi 1
seek
Displacement x
mx (t) cx (t) kx(t) p(t)
Loading p
n
e
v
i
g
pi
p1 p
pi+1
Time stepping k
e
x1
x2
t1 t 2
time
se
xi
xi+1
t i tj
time
Newmarks Method
Given
mx i cx i kx i pi
Seek
mx i1 cx i1 kx i1 pi1
Too many unknowns ! Need assumptions.
x i 1
x
x i
( x i x i 1 )
2
x i 1
x x i
x i
x i 1
x i
x i
xi
t
( x i x i 1 )
2
( x i 1 x i )
t
x i 1 x i
x i
t
( x i x i 1 )
2
x i 1 x i x i t
x i 1 x i x i t
t 2
( x i x i 1 )
4
xi
t 2
t 2
x i
x i 1
3
6
Newmarks Method
Given
mx i cx i kx i pi
Seek
mx i1 cx i1 kx i1 pi1
x i 1 x i [(1 ) t ]x i ( t )x i 1
1
2
2
x
t
x
[(
t
]
x
t
)x i 1
i
i
i
i 1
2
, are Newmarks parameters
Newmarks Method
Average acceleration
x
( x i x i 1 )
2
x i 1 x i
1
1
,
2
4
t
( x i x i 1 )
2
t 2
x i 1 x i x i t
( x i x i 1 )
4
x i
x i 1
x
( x i x i 1 )
2
x i 1
x i
x i
t
( x i x i 1 )
2
x i 1 x i x i t
xi
t 2
( x i x i 1 )
4
Newmarks Method
1
1
,
2
6
Linear acceleration
x x i
( x i 1 x i )
t
x i 1 x i
x i
t
( x i x i 1 )
2
2
x i 1
x x i
( x i 1 x i )
t
x i 1 x i
t
t
x i 1
x
i 1
i
i
i
3
6
x i
t
( x i x i 1 )
2
x i 1 x i x i t
xi
t 2
t 2
x i
x i 1
3
6
Newmarks Method
Incremental form
mx i cx i kx i pi
mx cx
i 1
i 1 kx i 1 pi 1
x i 1 x i [(1 ) t ]x i ( t ) x i 1
(1)
(2)
(3)
x i 1 x i tx i [( ) t 2 ]x i ( t 2 ) x i 1 (4)
2
(2) - (1) gives mx i cx i kx i pi
(5)
(3) gives x i x i 1 x i tx i ( t )x i
(6)
1 2
2
t
x
t
x
t
)x i (7)
(4) gives
i
i 1
i
i
i
2
Newmarks Method
From (6) and (7)
(5)
x i tx i ( t ) x i
(6)
x i tx i
1
1
x i
x i
xi
2
t
2
t
x i
x i x i t (1 ) x i
t
2
x i
mx i cx i kx i pi
1 2
t x i (t 2 ) x i (7)
2
(8)
(9)
(8) & (9) into (5)
kx i pi
1
k k
c
m
2
t
t
m
c
m
pi pi (
) x i [ t ( 1)c]x i
t
2
2
Newmarks Method
- Summary
Given m, c, k, p, x 0 and x 0
1
Assume and ( 2 , 6 4 )
1) Initialize
2) For each time step i ( i = 0,1,2,)
p cx 0 kx 0
x 0 0
m
1
k k
c
m
2
t
t
m
c
a
t
m
b
t ( 1)c
2
2
pi pi 1 pi ax i bx i
x i pi / k
x i
x i
1
1
x i
x i
t
2
t 2
x i
x i x i t (1 ) x i
t
Newmarks Method
- Example
Case study:
Given:
Physical properties of an SDF system:
2
m 0.253 kip sec / in, k 10 kips / in,
T 1 sec ( 6.28 rad / sec), 0.05
Half-cycle sine pulse force (shown right)
p, kips
10
10 sin(t / 0.6)
t, sec
0.6
Newmarks Method
- Example
Average acceleration using
t 0.1 sec
1.0 Initial calculations:
m 0.253 k 10 c 0.159
u 0 0 u 0 0 p0 0
1.1
1.2
1.3
1.4
p cx 0 kx 0
x 0 0
0
m
t 0.1
1
1
,
2
4
p cx 0 kx 0
x 0 0
m
1
k k
c
m
2
t
t
m
c
a
t
m
b
t ( 1)c
2
2
2
4
k k c
m 114 .5
2
t
( t )
4
a
m 2c 10.45; and b 2 m 0.507.
t
Newmarks Method
- Example
Average acceleration using
t 0.1 sec
2.0 calculations for each time step
2.1
pi pi ax i bx i pi 10.45x i 0.507 x i
2.2
x i
2.3
x i
2.4
x i
2.5
x i 1 x i x i , x i 1 x i x i , x i 1 x i x i
pi
pi
114 .5
k
2
x i 2 x i 20x i 2 x i
t
4
( x i tx i ) 2x i 400( x i 0.1x i ) 2x i
2
( t )
Newmarks Method
- Example
Average acceleration using
Results
t 0.1 sec
Newmarks Method
- Example
Linear acceleration using
t 0.1 sec
1.0 Initial calculations:
m 0.253 k 10 c 0.159
x 0 0 x 0 0 p0 0
1.1
p cx 0 kx 0
x 0 0
0
m
1.2
t 0.1
1.3
3
6
k k c
m 166.8
2
t
( t )
1.4
1
1
,
2
6
p cx 0 kx 0
x 0 0
m
1
k k
c
m
2
t
t
m
c
a
t
m
b
t ( 1)c
2
2
6
t
m 3c 15.68; and b 3m c 0.768.
t
2
Newmarks Method
- Example
Linear acceleration using
t 0.1 sec
2.0 calculations for each time step
2.1
pi pi ax i bx i pi 15.68x i 0.7679x i
2.2
x i
2.3
x i
2.4
x i
2.5
x i 1 x i x i , x i 1 x i x i , x i 1 x i x i
pi
pi
166.8
k
3
t
x i 3x i x i 30x i 3x i 0.05x i .
t
2
6
( x i tx i ) 3x i 600( x i 0.1x i ) 3x i
2
( t )
Newmarks Method
- Example
Linear acceleration using
Results
t 0.1 sec
Newmarks Method - Stability
To keep the algorithm stable
t
1
T 2
Average acceleration
1
2
1
1
,
2
4
Unconditional stable
Linear acceleration
1
1
,
2
6
Conditional stable
Usually for accuracy t 0.1T or less!
t
0.55
T
Newmarks Method
for nonlinear systems
EOM of nonlinear system
mx ( t ) cx ( t ) f ( x ) p( t )
mx i ( t ) cx i ( t ) f ( x i ) p i
mx i 1 ( t ) cx i 1 ( t ) f ( x i 1 ) p i 1
mx i cx i f i pi
f i f ( x i 1 ) f ( x i )
f i
x i ( k i ) T x i
x i
Tangential stiffness
function of displacement
Linear
( k i )T
f i 1
fi
x i
f i
Nonlinear
xi
x i 1 x
Newmarks Method
for nonlinear systems
mx i ( t ) cx i ( t ) (k i ) T x i p i
Newmarks Scheme
x i
x i 1, x i 1, x i 1
Correction (minimize the unbalanced force)
mx i ( t ) cx i f i pi
Original
Approximate mx i ( t ) cx i ( k i ) T x i pi
Unbalanced
x i
( j)
R ( j)
( k i )T
f i ( k i )T x i R
Newton-Raphson Iteration
x i t x i x i (1) x i ( 2)
Newmarks Method
for nonlinear systems
Given m, c, k, p, x 0 and x 0
1
Assume and ( 2 , 6 4 )
1) Initialize
p cx 0 (fs )0
x 0 0
m
m
c
a
t
m
b
t ( 1)c
2
2
Newmarks Method
for nonlinear systems
2) For each time step i ( i = 0,1,2,)
pi pi ax i bx i
x
( j)
R ( j)
k
T
R (i ) pi
x i 1(0) x i
k k c 1 m
i
i
t
t 2
x i 1( j) x i 1( j1) x ( j)
j=1, 2, ...
f ( j) f s( j) fs( j1) ( k T k T ) x ( j)
R ( j1) R ( j) f ( j)
1
k T k T
c
m
(Newton-Raphson iteration)
2
t
t
x i
x i
1
1
x i
x i
x i
2
t
2
t
x i x i t (1 ) x i
t
Newmarks Method
for nonlinear systems
Case study
Given:
Physical properties of an SDOF system:
m 0.253 kip sec 2 / in, 0.05
Force-deformation relation is elastoplastic with x y 0.75 in
Half-cycle sine pulse force
f, kips
7.5
p, kips
10
e
o
0.75
d
10 sin( t / 0.6)
x, in.
10
c
10
1
t, sec
0.6
Newmarks Method
for nonlinear systems
Average acceleration without iteration using
t 0.1 sec
1.0 Initial calculations:
m 0.253 k 10 c 0.159
x 0 0 x 0 0 p0 0
1.1
1.2
1.3
p0 cx 0 kx 0
x 0
0
m
t 0.1
4
a
m 2c 10.45; and b 2 m 0.507.
t
Newmarks Method
for nonlinear systems
Average acceleration without iteration using
t 0.1 sec
2.0 calculations for each time step
2.1
pi pi ax i bx i pi 10.45x i 0.507x i
2.2
ki k
2.3
2.4
2.5
for branches oa, bc and de; and
2
4
k i k i c
m k i 104.5
2
t
( t )
pi
x i
k
2
x i x i 2 x i 20x i 2x i
t
2.6
x i 1 x i x i , x i 1 x i x i
2.7
(fs )i 1 (fs )i k i x i .
2.8
p cx i 1 (fs )i 1
x i 1 i 1
.
m
ki 0
for ab and cd
Newmarks Method
for nonlinear systems
Average acceleration without iteration using
Results
t 0.1 sec
Newmarks Method
for nonlinear systems
Newton-Raphson iteration using
t 0.1 sec
1.0 Initial calculations:
m 0.253 k 10 c 0.159
x 0 0 x 0 0 p0 0
1.1
p cx 0 kx 0
x 0 0
0
m
1.2
t 0.1
1.3
4
m 2c 10.45; and b 2 m 0.507.
t
Newmarks Method
for nonlinear systems
Newton-Raphson iteration using t 0.1 sec
2.0 calculations for each time step
2.1
2.2
pi pi ax i bx i pi 10.45x i 0.507 x i
2.3
2
4
k i k i c
m k i 104.5
2
t
( t )
( j)
R
x ( j)
x i 1( j) x i 1( j1) x ( j)
k
T
2.4
2.5
2.6
2.7
ki k
for branches oa, bc and de; and k i 0 for ab and cd
f ( j) fs( j) fs( j1) ( k T k T ) x ( j)
R ( j1) R ( j) f ( j)
1
k T k T
c
m
2
t
t
2
x i
x i 2 x i 20x i 2 x i
t
4
4
x i
x i
x i 2x i 400x i 40 x i 2 x i
2
t
( t )
x i 1 x i x i ,
for nonlinear systems
Linear acceleration using
Results
t 0.1 sec
No iteration
Newmarks Method
Wilson- Method
Modification of the Newmarks linear acceleration meth
1
1
,
2
6
Incremental form
mx i cx i kx i pi
x i tx i ( t ) x i
1 2
t x i (t 2 ) x i
x i 1
2
Linear acceleration and
force
x i
between
t
and
t
mx i ci x i i+
kx i pi
x i tx i
1
tx i
2
1
1
x i tx i t 2 x i t 2 x i
2
6
x i tx i
x i
x i
t
ti
t i 1
t t
ti
time
Wilson- Method
mx i cx i kx i pi
where
pi
x i
k
3
6
k k
c
m
2
t
t
6
t
m 3c x i 3m
c x i
2
t
3
t
x i 3x i x i
t
2
1
6
6
x i
x i x i 3x i
i
2
t
t
pi pi 1 pi
then
x i
x i
Wilson- Method
- Summary
Given m, c, k, p, x 0 and x 0
x i pi / k
Assume t and 1 normally 1.4
1) Initialize
p cx 0 kx 0
x 0 0
m
3
6
k k
c
m
2
t
t
6
m 3c
t
1
b 3m tc
2
2) For each time step i ( i = 0,1,2,)
pi pi 1 pi ax i bx i
pi
k
3
t
x i x i 3x i x i
t
2
6
6
1
x i x i
x i
x i x i 3x i
t 2
1
x i tx i tx i
2
1
1
x i tx i t 2 x i t 2 x i
2
6
x i
Commercial Software - MATLAB
Demonstration
- SDOF structure under earthquake excitation
1940 El Centro EQ
ground acc. m/s
4
2
-2
-4
10
time s
15
20
75-story building
m 4.61 107 kg
c 1.04 106 N - s/m (1%)
k 5.83 107 N/m
x g
Commercial Software - MATLAB
load elcen;
ground acc. m/s
clear all;
Response of a tall building under earthquake
4
2
0
-2
m=4.61*10^7; c=1.04*10^6; k=5.83*10^7;
-4
A=[0 1; -k/m -c/m]; B=[0;-1]; C=[1 0]; D=[0];
10
15
20
10
time s
15
20
0.4
[x temp]=lsim(A,B,C,D,acc,time);
subplot(2,1,1);
plot(time,acc);
ylabel('ground acc. m/s^2');
title('Response of a tall building under earthquake');
subplot(2,1,2);
plot(time,x);
xlabel('time');
ylabel('disp. m');
disp. m
0.2
0
-0.2
-0.4
Commercial Software - MATLAB
The lsim command operates in the STATE-SPACE domai
cx kx p
Original 2nd ordermx
State-Space 1st order mx cx kx p
1
0
x
k c
x
m
m
output
x x
x 0
1 p
x m
x
y 1 0 0 p
x
Commercial Software - MATLAB
AX BU
X
Y CX DU
1
0
c
A k
m
m
x
X
x
0
B 1
m
C and D depend on the output desired.
[x temp]=lsim(A,B,C,D,-acc,time);
Demonstration !
Summary
mx (t) cx (t) kx(t) p(t)
5 methods are introduced
Interpolation of Excitation Method
Central Difference Method
Newmarks Method
Wilson- Method
Commercial Software
Accuracy
Extension to multiple degrees and/or nonlinear
structures
Newmarks Method
Wilson- Method
Commercial Software