Mode Superposition Example 9.
1. Change of Basis to Modal Generalized Displacements
Equation of equilibrium :
MU&& + CU& + KU = R
Using the following transformation:
P is a n X n square matrix.
U (t ) = PX (t ) X(t) time-dependent of order n.
Substituting,
P T
MP && (t ) + P T CP X& (t ) + P T KP X (t ) = P T R
X
12 3 123 12 3 { ~
~ ~ ~ R
K C P
Mode Superposition Example 9.7
Equilibrium equations with damping neglected,
MU&& + KU = 0
The solution of this equation is:
ɸ vector of order n
U = Φ sin ω (t − t0 ) t time variable
t0 time constant
ω constant identified (frequency of vibration)
The second derivated:
U&& = −ω 2 Φ sin ω (t − t0 )
Substituing :
KΦ = ω 2 MΦ
Mode Superposition Example 9.7
We can write the n solutions like:
KΦ = MΦΩ 2
Eigenvectors are M-orthonormal:
Φ T
142 K Φ =
4 ~ 43Ω
4
2
K
⇒ P = Φ ⇒ U (t ) = ΦX (t )
Φ
14
T
Φ4
M2 =3I
~
M
Equilibrium equation:
X&& (t ) + Φ T CΦX& (t ) + Ω 2 X (t ) = Φ T R(t )
Mode Superposition Example 9.7
2. Analysis with Damping Neglected
The velocity- dependent damping effects are not included:
X&& (t ) + Ω 2 X (t ) = Φ T R(t )
Individual equations:
&x&i (t ) + ωi2 xi (t ) = ri (t )⎫ xi t =0 = Φ Ti M 0U
⎬i = 1,2,..., n. ⇒
ri (t ) = Φ Ti R(t ) ⎭ x&i t =0 = Φ Ti M 0U&
xi t =0 = Φ Ti M 0U
x&i t =0 = Φ Ti M 0U&
Mode Superposition Example 9.7
EXAMPLE 9.7: Use mode superposition to calculate the displacement
response of the system considered in Exemples 9.1 to 9.4 and 9.6.
Consider a simple system: MU&& + CU& + KU = R
⎡ 2 0⎤
M =⎢ ⎥
⎣0 1 ⎦
⎡2 0⎤ ⎡ U&&1 ⎤ ⎡ 6 − 2⎤ ⎡U1 ⎤ ⎡ 0 ⎤ ⎡ 6 − 2⎤
⎢0 1⎥ ⎢ && ⎥ + ⎢− 2 4 ⎥ ⎢U ⎥ = ⎢10⎥ ⇒ K =⎢ ⎥
⎣ ⎦ ⎣U 2 ⎦ ⎣ ⎦⎣ 2 ⎦ ⎣ ⎦ ⎣− 2 4 ⎦
⎡0⎤
R=⎢ ⎥
⎣10⎦
The generalized eigenproblem:
⎡ 6 − 2⎤ 2 ⎡2 0⎤
⎢ − 2 4 ⎥ Φ = ϖ ⎢0 1 ⎥ Φ
⎣ ⎦ ⎣ ⎦
Mode Superposition Example 9.7
⎡ 1 ⎤
⎢ ⎥
ω12 = 2; Φ1 = ⎢ 3 ⎥
⎢ 1 ⎥
⎢⎣ 3 ⎥⎦
Two solutions without derivation;
⎡1 2 ⎤
⎢ ⎥
ω2 = 5; Φ 2 =
2
⎢ 2 3 ⎥
⎢ 2⎥
⎢⎣ − 3 ⎥⎦
Equilibrium equation: X&& (t ) + Φ T CΦX& (t ) + Ω 2 X (t ) = Φ T R(t )
⎡ 10 ⎤
⎢ ⎥
&X& (t ) + ⎡2 0⎤ X (t ) = ⎢ 3
⎥
⎢0 5⎥
⎣ ⎦ ⎢− 10 2⎥
⎢⎣ 3 ⎥⎦
(1) Calculate the exact response by integrating each of two decoupled
equilibrium equations exactly.
(2) Use the Newmark method with time step Δt=0.28 for the time
integration.
Two equilibrium equations:
10 2
&x&1 + 2 x1 = and &x&2 + 5 x2 = −10
3 3
Initial conditions: U t =0 = 0 ; U& t =0 = 0
Using the equation: X&& (t ) + Ω X (t ) = Φ R(t )
2 T
&x&i (t ) + ωi2 xi (t ) = ri (t )⎫ xi t =0 = Φ Ti M 0U
⎬i = 1,2,..., n. ⇒
ri (t ) = Φ Ti R (t ) ⎭ x&i t =0 = Φ Ti M 0U&
x1 t =0 = 0 x&1 t =0 = 0
x2 t = 0 = 0 x&2 t =0 = 0
Mode Superposition Example 9.7
5 2
The exact solution is: x1 = (1 − cos 2t ); x2 = 2 (−1 + cos 5t )
3 3
Using the relation: U (t ) = ΦX (t )
⎡ 1 1 2⎤
⎢ ⎥
U (t ) = ⎢ 3 2 3 ⎥ X (t )
⎢ 1 2⎥
⎢ −
⎣ 3 3 ⎥⎦
⎡ 1 1 2 ⎤⎡ 5 ⎤
⎢ ⎥ (1 − cos 2t ) ⎥
U (t ) = ⎢ 3 2 3 ⎥⎢ 3
⎢ ⎥
⎢ 2 ⎥⎢
(−1 + cos 5t )⎥
1 2
⎢ − 2
⎣ 3 3 ⎥⎦ ⎢⎣ 3 ⎥⎦
Mode Superposition Example 9.7
Employing Newmark method we obtain:
Time Δt 2Δt 3Δt 4Δt 5Δt 6Δt 7Δt 8Δt 9Δt 10Δt 11Δt 12Δt
tU 0.003 0.038 0.176 0.486 0.996 1.66 2.338 2.861 3.052 2.806 2.131 1.157
0.382 1.41 2.78 4.09 5.00 5.29 4.986 4.277 3.457 2.806 2.484 2.489
Evaluating the desplacements, where Δt=0.28 we obtain:
Time Δt 2Δt 3Δt 4Δt 5Δt 6Δt 7Δt 8Δt 9Δt 10Δt 11Δt 12Δt
X(t) 0.2258 0.8199 1.807 2.379 4.123 5.064 5.579 5.774 5.521 4.855 3.866 2.773
-0.304 -0.792 -2.123 -2.939 -3.258 -2.632 4.986 -1.156 -0.330 -0.004 -0.248 -1.088
Displacem ent of the system
3,500
Exact
Newmark
3,000 Central Difference method
Houbolt
Wilson θ
2,500
2,000
1,500
1,000
0,500
0,000
Δt 2Δt 3Δt 4Δt 5Δt 6Δt 7Δt 8Δt 9Δt 10Δt 11Δt 12Δt
t
Displacem ent of the system
6,000
Exact
Newmark
Central Diff erence method
5,000
Houbolt
Wilson θ
4,000
3,000
2,000
1,000
0,000
Δt 2Δt 3Δt 4Δt 5Δt 6Δt 7Δt 8Δt 9Δt 10Δt 11Δt 12Δt
t
