Rev Topic04 StructuralDynamicsofMDOFSystems

Structural Dynamics of Linear Elastic
Multiple-Degrees-of-Freedom (MDOF) Systems
u1
u2
u3
Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 1

Structural Dynamics of Elastic
MDOF Systems
• Equations of motion for MDOF systems

• Uncoupling of equations through use of natural
mode shapes
• Solution of uncoupled equations
• Recombination of computed response
• Modal response history analysis
• Modal response spectrum analysis
• Equivalent lateral force procedure

Symbol Styles Used in this Topic
M
U Matrix or vector (column matrix)
m Element of matrix or vector or set

u (often shown with subscripts)
W
Scalars
g

Relevance to ASCE 7-05
ASCE 7-05 provides guidance for three specific

analysis procedures:
• Equivalent lateral force (ELF) analysis
• Modal superposition analysis (MSA)
• Response history analysis (RHA)
ELF usually allowed ELF not allowed
Cs
See ASCE 7-05

Table 12.6-1
Ts 3.5Ts T

Planar Frame with 36 Degrees of Freedom
uy
Majority of mass
is in floors rz Typical nodal
DOF
9 10 11 12 ux
5 6 7 8
1 2 3 4
Motion is
predominantly
lateral
13 14 15 16

Planar Frame with 36 Static Degrees of Freedom
But with Only THREE Dynamic DOF
u1
u2 ⎧ u1 ⎫
⎪ ⎪
u3
U = ⎨u2 ⎬
⎪u ⎪
⎩ 3⎭

Development of Flexibility Matrix
f1 = 1 kip
d1,1
d2,1
d1,1
d2,1
d3,1 d3,1

(continued)
d1,2
f2=1 kip
d2,2 d1,1 d1,2
d2,1 d2,2
d3,2 d3,1 d3,2

(continued)
d1,3
d2,3 d1,1 d1,2 d1,3

d2,1 d2,2 d2,3
f3 = 1 kip
d3,1 d3,2 d3,3
d3,3

Concept of Linear Combination of Shapes
(Flexibility)
⎡ d1,1 d1,2 d1,3 ⎤ ⎧ f1 ⎫

⎢ ⎥⎪ ⎪
U = ⎢d2,1 d2,2 d2,3 ⎥ ⎨f2 ⎬
⎢d3,1 d3,2 d3,3 ⎥ ⎪⎩f3 ⎪⎭
⎣ ⎦
⎧ d1,1 ⎫ ⎧ d1,2 ⎫ ⎧ d1,3 ⎫
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
U = ⎨d2,1 ⎬ f1 + ⎨d2,2 ⎬ f2 + ⎨d2,3 ⎬ f3
⎪d ⎪ ⎪d ⎪ ⎪d ⎪
⎩ 3,1 ⎭ ⎩ 3,2 ⎭ ⎩ 3,3 ⎭
DF=U K = D-1 KU=F

Static Condensation
⎡K m,m K m,n ⎤ ⎧Um ⎫ ⎧ Fm ⎫ DOF with mass
⎢K ⎥ ⎨ ⎬=⎨ ⎬
⎣ n,m K n,n ⎦ ⎩ Un ⎭ ⎩{0}⎭ Massless DOF
1 K m,mUm + K m,nUn = Fm
2 K n,mUm + K n,nUn = {0}

Static Condensation
(continued)
−1
Rearrange 2 Un = −K K n,mUm n ,n
−1
Plug into 1 K m,mUm − K m,nK K n,mUm = Fm n,n
Simplify ⎡⎣K m,m − K m,nK K n,m ⎤⎦ Um = Fm −1

n,n
−1
K̂ = K m,m − K m,nK K n,m n,n
Condensed stiffness matrix

Idealized Structural Property Matrices
m1 f1(t), u1(t) ⎡ k1 -k1 0 ⎤
K = ⎢⎢-k1 k1+ k 2 -k 2 ⎥⎥
k1 ⎢⎣ 0 -k 2 k 2 + k 3 ⎥⎦
m2 f2(t), u2(t)
⎡m1 0 0 ⎤
M = ⎢⎢ 0 m2 0 ⎥⎥
k2 ⎢⎣ 0 0 m3⎥⎦
m3 f3(t), u3(t)
⎧ f1(t ) ⎫ ⎧u1(t ) ⎫
⎪ ⎪ ⎪ ⎪
k3 F(t ) = ⎨f2 (t )⎬ U(t ) = ⎨u2 (t )⎬
⎪ f (t )⎪ ⎪u (t )⎪
⎩3 ⎭ ⎩ 3 ⎭
Note: Damping to be shown later
Coupled Equations of Motion
for Undamped Forced Vibration
&&(t ) + KU(t ) = F(t )
MU
⎡m1 0 0 ⎤ ⎧u &&1 ( t ) ⎫ ⎡ k1 −k1 0 ⎤ ⎧ u 1 ( t ) ⎫ ⎧ f1 ( t ) ⎫
⎢ 0 ⎪ && ⎪ ⎪ ⎪ ⎪ ⎪
⎢ m2 0 ⎥⎥ ⎨ u 2 ( t ) ⎬+
⎢−k1
⎢ k1 + k 2 − k 2 ⎥⎥ ⎨ u 2 ( t ) ⎬ = ⎨ f 2 ( t ) ⎬
⎢⎣ 0 0 m 3 ⎥⎦ ⎩⎪ u
&& 3 ( t ) ⎭⎪ ⎣⎢ 0 −k 3 k 2 + k 3 ⎦⎥ ⎩⎪ u 3 ( t ) ⎭⎪ ⎩⎪ f3 ( t ) ⎭⎪
DOF 1 m1u&&1(t ) + k1u1(t ) − k1u2 (t ) = f1(t )

DOF 2 m 2u&&2 (t ) − k1u1(t ) + k1u2 (t ) + k 2u2 (t ) − k 2u3 (t ) = f2 (t )
DOF 3 &&3 (t ) − k 2u2 (t ) + k 2u3 (t ) + k 3u3 (t ) = f3 (t )
m3u

Developing a Way To Solve
the Equations of Motion
• This will be done by a transformation of coordinates

from normal coordinates (displacements at the nodes)
To modal coordinates (amplitudes of the natural
Mode shapes).
• Because of the orthogonality property of the natural mode

shapes, the equations of motion become uncoupled,
allowing them to be solved as SDOF equations.
• After solving, we can transform back to the normal

coordinates.

Solutions for System in Undamped Free Vibration
(Natural Mode Shapes and Frequencies)
&&(t ) + KU(t ) = {0}

MU
Assume U(t ) = φ sin ωt U(
&& t ) = −ω2φ sinωt
Then Kφ − ω 2
Mφ = {0} has three (n) solutions:
⎧φ1,1 ⎫ ⎧φ1,2 ⎫ ⎧φ1,3 ⎫

⎪ ⎪ ⎪ ⎪ ⎪ ⎪
φ1 = ⎨φ2,1 ⎬ , ω 1 φ2 = ⎨φ2,2 ⎬ , ω 2 φ3 = ⎨φ2,3 ⎬ , ω 3
⎪φ ⎪ ⎪φ ⎪ ⎪φ ⎪
⎩ 3,1 ⎭ ⎩ 3,2 ⎭ ⎩ 3,3 ⎭
Natural frequency
Natural mode shape

Solutions for System in Undamped Free Vibration
(continued)
For a SINGLE Mode
K Φ = MΦΩ2 For ALL Modes
Where: Φ = [φ1 φ2 φ3 ]
⎡ω12 ⎤
⎢ ⎥
Ω =⎢
2 2
ω2 ⎥ Kφ = ω2
Mφ
⎢ ω2⎥
⎣ 3⎦
φ1,i = 1.0
Note: Mode shape has arbitrary scale; usually or
ΦT MΦ = I
Mode Shapes for Idealized 3-Story Frame
φ1,2
φ1,1 φ1,3
Node
φ2,1 φ2,2 φ2,3
Node Node
φ3,1 φ3,2 φ3,3
MODE 1 MODE 2 MODE 3

Concept of Linear Combination of Mode Shapes
(Transformation of Coordinates)
U =ΦY
Mode shape
⎡φ1,1 φ1,2 φ1,3 ⎤ ⎧ y1 ⎫ ⎧φ1,1 ⎫ ⎧φ1,2 ⎫ ⎧φ1,3 ⎫

⎢ ⎥⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪
U = ⎢φ2,1 φ2,2 φ2,3 ⎥ ⎨y2 ⎬ U = ⎨φ2,1 ⎬ y1 + ⎨φ2,2 ⎬ y2 + ⎨φ2,3 ⎬ y3
⎢⎣φ3,1 φ3,2 φ3,3 ⎥⎦ ⎪⎩y3 ⎭⎪ ⎪φ ⎪
⎩ 3,1 ⎭
⎪φ ⎪
⎩ 3,2 ⎭
⎪φ ⎪
⎩ 3,3 ⎭
Modal coordinate =
amplitude of mode
shape

Orthogonality Conditions
Φ = [φ1 φ2 φ3 ]
Generalized mass Generalized stiffness
⎡m1* ⎤ ⎡k1* ⎤
⎢ ⎥ ⎢ ⎥
Φ MΦ = ⎢
T *
m2 ⎥ Φ KΦ = ⎢
T *
k2 ⎥
⎢ m*⎥
⎢ *⎥
⎣ 3⎦ ⎣ k3⎦
Generalized damping Generalized force

⎡c1* ⎤ ⎧f1* (t )⎫
⎢ ⎥ ⎪* ⎪
Φ CΦ = ⎢
T *
c2 ⎥ Φ F(t ) = ⎨f2 (t )⎬
T
⎢ c *⎥
⎪f * (t )⎪
⎣ 3⎦
⎩3 ⎭

Development of Uncoupled Equations of Motion
MDOF equation of motion: && &

MU+CU+KU = F(t )
Transformation of coordinates: U = ΦY
Substitution: && C Φ Y+K
M Φ Y+ & Φ Y = F(t )
&& + ΦTCΦY& + ΦT KΦY = ΦT F(t )
Premultiply by Φ T : ΦT MΦY
Using orthogonality conditions, uncoupled equations of motion are:

⎡m1* ⎤ ⎧ &&y1 ⎫ ⎡c 1* ⎤ ⎧ y& 1 ⎫ ⎡k 1* ⎤ ⎧ y1 ⎫ ⎧ f1* (t )⎫
⎢ ⎥ ⎪ && ⎪ ⎢ ⎥⎪& ⎪ ⎢ ⎥⎪ ⎪ ⎪ * ⎪
⎢ m *2 ⎥ ⎨y2 ⎬ + ⎢ c 2* ⎥ ⎨y2 ⎬ + ⎢ k 2* ⎥ ⎨ y 2 ⎬ = ⎨ f2 (t ) ⎬
⎢
⎣ m 3* ⎥⎦ ⎪⎩ &&
y 3 ⎪⎭ ⎢⎣ c 3* ⎥⎦ ⎪⎩ y& 3 ⎪⎭ ⎢⎣ k 3* ⎥⎦ ⎪⎩ y 3 ⎪⎭ ⎩⎪ f3* (t )⎭⎪

(Explicit Form)
* * * *
Mode 1 && &
m y +c1y1+k1y1 = f1 (t )
1 1
** * *
Mode 2 && &
m y +c2 y2 +k2 y2 = f2 (t )
2 2
* * * *
Mode 3 && &
m y +c3 y3 +k3 y3 = f3 (t )
3 3

(Explicit Form)
*
ci
Simplify by dividing through by m* and defining ξi =
2mi*ωi
Mode 1 y&&1 + 2ξ1ω1y&1 + ω12 y1 = f1* (t )/ m1*
Mode 2 y 2 + 2ξ2ω2 y 2 + ω2 y 2 = f2 (t ) / m2
&& & 2 * *
Mode 3 y&&3 + 2ξ3ω3 y& 3 + ω32 y 3 = f3* (t ) / m3*

Earthquake “Loading” for MDOF System
u&&g U&&r ,1
⎧u&&g (t ) + u
&&r ,1(t ) ⎫
⎪ &&r ,2 (t )⎬⎪ =
FI (t ) = M ⎨u&&g (t ) + u
⎪u&& (t ) + u&& (t ) ⎪
⎩ g r ,3 ⎭
⎧1.0⎫ ⎧u
&&r ,1(t ) ⎫
⎪ ⎪ && ⎪&& ⎪
M ⎨1.0⎬ ug (t ) + M ⎨ur ,2 (t )⎬
⎪1.0⎪ ⎪u
&& ( ) ⎪
⎩ ⎭ ⎩ r ,3 ⎭t
Move to RHS as FEFF (t ) = - M Ru&&g (t )

Modal Earthquake Loading F (t ) = −Φ MRu&&g (t )
* T
m1 = 2
u1
⎡2 ⎤ ⎧1⎫
⎪ ⎪
m2 = 3
u2 ⎢
M=⎢ 3 ⎥ ⎥ R = ⎨1⎬
⎪1⎪
m3 = 1
u3
⎢⎣ 1⎥⎦ ⎩ ⎭
u&&g (t )
m3 = 1
m1 = 2 u1
u3 ⎡2 + 1 ⎤ ⎧1⎫
⎢ ⎥ ⎪ ⎪
m2 = 3 u2 M=⎢ 3 ⎥ R= ⎨1⎬
⎢⎣ ⎥
1⎦ ⎪0 ⎪
⎩ ⎭
u&&g (t )
Definition of Modal Participation Factor
For earthquakes: fi (t ) = −φ MRu

*
&&g (t ) T
i
Typical modal equation:
fi (t ) φ MR * T
y&&i + 2ξ i ωi y& i + ω y i =
i
2
*
= − *
&&
u g (t )
i
mi mi
Modal participation factor pi

Caution Regarding Modal Participation Factor
φ MR T
pi = i
*
m i
φ Mφi
T
i
Its value is dependent on the (arbitrary) method

used to scale the mode shapes.

Variation of First Mode Participation Factor
with First Mode Shape
p1 = 1.0 p1 =1.4 p1 =1.6

1.0 1.0 1.0

Concept of Effective Modal Mass
For each Mode I, mi = pi mi 2 *
• The sum of the effective modal mass for all

modes is
equal to the total structural mass.
• The value of effective modal mass is

independent of mode shape scaling.
• Use enough modes in the analysis to provide

a total effective mass not less than 90% of the
total structural mass.

Variation of First Mode Effective Mass
with First Mode Shape
m1 / M = 1.0 m1 / M= 0.9 m1 / M= 0.7

1.0 1.0 1.0

Derivation of Effective Modal Mass
(continued)
For each mode:
y&&i + 2ξi ωi y& i + ω y i = − pi u&&g i

2
SDOF system:
qi + 2ξ i ωi qi + ωi qi = −u&&g
&& & 2
Modal response history, qi(t) is obtained by first

solving the SDOF system.

(continued)
From previous slide y i (t ) = pi qi (t )
Recall ui (t ) = φi y i (t )
Substitute ui (t ) = pi φi qi (t )

(continued)
Applied “static” forces required to produce ui(t):
Vi (t ) = Kui (t ) = Pi Kφi qi (t )
Recall: Kφi = ωi Mφi 2
Substitute:
Vi (t ) = Mφi Pi ωi qi (t ) 2

(continued)
Total shear in mode: Vi = Vi T R
Vi = (Mφi ) RPi ω qi (t ) = φi MRPi ω qi (t )

T 2 T 2
“Acceleration” in mode
Define effective modal mass:
Mi = φ MRPi
T
i
and
Vi = Mi ω qi (t )
2

(continued)
⎡ φi MR ⎤ T
T
Mi = φi MRPi = ⎢ T
T
⎥ φi Mφi Pi
⎣ φi Mφi ⎦ i
Mi = Pi mi
2 *

Development of a Modal Damping Matrix
In previous development, we have assumed:
⎡c1* ⎤
⎢ ⎥
Φ CΦ = ⎢
T *
c2 ⎥
⎢ *⎥
c3 ⎦
⎣
Two methods described herein:
• Rayleigh “proportional damping”
• Wilson “discrete modal damping”

Rayleigh Proportional Damping
(continued)
MASS
PROPORTIONAL
DAMPER
STIFFNESS
PROPORTIONAL
DAMPER
C = αM + β K

(continued)
C = αM + β K
For modal equations to be uncoupled:
Assumes
2ωnξn = φ Cφn T
n
ΦT MΦ = I
Using orthogonality conditions:
2ωnξn = α + βω 2
n
1 ωn
ξn = α+ β
2ωn 2
(continued)
Select damping value in two modes, ξm and ξn
Compute coefficients α and β:
⎧α ⎫ ωm ω n ⎡ ωn −ωm ⎤ ⎧ξ m ⎫
⎨ ⎬=2 2 ⎢ −1/ ω 1/ ω ⎥ ⎨ ξ ⎬
⎩β ⎭ ωn − ω m2
⎣ n m⎦⎩ n ⎭
Form damping matrix C = αM + β K

Rayleigh Proportional Damping (Example)
5% critical in Modes 1 and 3
0.15
TYPE
Structural frequencies MASS α = 0.41487
Modal Damping Ratio

β = 0.00324
STIFFNESS
ω
TOTAL
Mode
0.10
1 4.94
2 14.6
3 25.9 0.05
4 39.2
5 52.8
0.00
0 20 40 60
Frequency, Radians/Sec.

Rayleigh Proportional Damping (Example)
5% Damping in Modes 1 & 2, 1 & 3, 1 & 4, or 1 & 5
0.15
MODES
1,2
Proportionality factors 1,3
Modal Damping Ratio

(5% each indicated mode) 1,4
1,5
0.10
Modes α β
1&2 .36892 0.00513
1&3 .41487 0.00324
1&4 .43871 0.00227 0.05
1&5 .45174 0.00173
0.00
0 20 40 60
Fequency, Radians/sec

Wilson Damping
ξ
Directly specify modal damping values i
*
⎡c1* ⎤ ⎡2m1*ω1ξ1* ⎤
⎢ ⎥ ⎢ ⎥
Φ CΦ = ⎢
T *
c2 ⎥=⎢ 2m2ω2ξ2
* *
⎥
⎢ *⎥ ⎢
c3 ⎦ ⎣ * *⎥
2m3ω3ξ3 ⎦
⎣

Formation of Explicit Damping Matrix
From “Wilson” Modal Damping
(NOT Usually Required)
⎡2ξ1ω1 ⎤
⎢ 2ξ 2ω2 ⎥
⎢ ⎥
ΦT CΦ = ⎢ ••• ⎥=c
⎢ ⎥
⎢ 2ξ n −1ωn −1 ⎥
⎢⎣ 2ξ nωn ⎥⎦
(Φ ) T −1
cΦ−1 = C
⎡n T ⎤
C = M ⎢∑ 2ξi ωi φi φi ⎥M
⎣ i =1 ⎦
Wilson Damping (Example)
5% Damping in Modes 1 and 2, 3
10% in Mode 5, Zero in Mode 4
12
10
10
Modal Damping Ratio
6
5 5 5
0
0
4.94 14.57 25.9 39.2 52.8
Frequency, Radians per second

Wilson Damping (Example)
5% Damping in all Modes
12
10
Modal Damping Ratio
6
5 5 5 5 5
0
4.94 14.57 25.9 39.2 52.8
Frequency, Radians per second

Solution of MDOF Equations of Motion
• Explicit (step by step) integration of coupled equations
• Explicit integration of FULL SET of uncoupled equations
• Explicit integration of PARTIAL SET of uncoupled
Equations (approximate)
• Modal response spectrum analysis (approximate)

Computed Response for
Piecewise Linear Loading
Vt2
Vt 2 −Vt 2
δ=
t2 − t1
δ
Vt1
Force, V(t)
t0 t1 t2 t3 Time, t
Time

Example of MDOF Response of Structure
Responding to 1940 El Centro Earthquake
u1(t) Example 1
m1=1.0 k-s2/in
10 ft k1=60 k/in Assume Wilson
damping with 5%
u2(t) critical in each mode.
m2=1.5 k-s2/in
10 ft k2=120 k/in
u3(t)
m3=2.0 k-s2/in
10 ft k3=180 k/in
N-S component of 1940 El Centro earthquake

Maximum acceleration = 0.35 g

Example 1 (continued)
Form property matrices:

m1=1.0 k-s2/in
u1(t) ⎡1.0 ⎤
m2=1.5 k-s2/in
k1=60 k/in
M = ⎢⎢ 1.5 ⎥kip − s2/in
⎥
u2(t) ⎢⎣ 2.0⎥⎦
k2=120 k/in
m3=2.0 k-s2/in
u3(t) ⎡ 60 − 60 0 ⎤
k3=180 k/in K = ⎢⎢− 60 180 − 120⎥⎥kip/in
⎢⎣ 0 − 120 300 ⎥⎦

Solve eigenvalue problem:

m1=1.0 k-s2/in KΦ = MΦΩ2
u1(t)
k1=60 k/in ⎡21.0 ⎤
m2=1.5 k-s2/in ⎢
Ω =⎢
2
96.6 ⎥ sec−2
u2(t) ⎥
k2=120 k/in ⎢⎣ 212.4⎥⎦
m3=2.0 k-s2/in
u3(t)
⎡1.000 1.000 1.000 ⎤
k3=180 k/in
Φ = ⎢0.644 − 0.601 − 2.57⎥
⎢ ⎥
⎢⎣0.300 − 0.676 2.47 ⎥⎦

Normalization of Modes Using Φ MΦ = I
T
⎡0.749 0.638 0.208 ⎤ ⎡1.000 1.000 1.000⎤

⎢0.644 −0.601 −2.57⎥
Φ = ⎢⎢0.478 −0.384 −0.534 ⎥⎥
vs ⎢ ⎥
⎢⎣0.223 −0.431 0.514 ⎥⎦ ⎢⎣0.300 −0.676 2.47 ⎥⎦

Mode Shapes and Periods of Vibration
MODE 1 MODE 2 MODE 3

ω = 4.58 rad/sec ω = 9.83 rad/sec ω = 14.57 rad/sec
T = 1.37 sec T = 0.639 sec T = 0.431 sec

⎧ 4.58 ⎫ ⎧ 1.37 ⎫
⎪ ⎪ ⎪ ⎪
m1=1.0 k-s2/in ωn = ⎨ 9.83 ⎬rad/ sec Tn = ⎨0.639⎬sec
u1(t) ⎪14.57⎪ ⎪0.431⎪
⎩ ⎭ ⎩ ⎭
k1=60 k/in
m2=1.5 k-s2/in
u2(t)
Compute Generalized Mass:
k2=120 k/in
m3=2.0 k-s2/in ⎡1 . 801 ⎤
u3(t) M * = Φ T M Φ = ⎢⎢ 2 . 455 ⎥kip − sec 2 / in
⎥
⎢⎣ 23 . 10 ⎥⎦
k3=180 k/in

m1=1.0 k-s2/in
u1(t) Compute generalized loading:
k1=60 k/in
m2=1.5 k-s2/in
V * (t ) = −ΦT MR v&&g (t )
u2(t)
k2=120 k/in ⎧ 2.566 ⎫
m3=2.0 k-s2/in ⎪ ⎪ &&
u3(t) Vn = − ⎨−1.254⎬ vg (t )
*
⎪ 2.080 ⎪
k3=180 k/in ⎩ ⎭

Write uncoupled (modal) equations of motion:

m1=1.0 k-s2/in
u1(t) y&&1 + 2ξ1ω1y&1 + ω12 y1 = V1* (t )/ m1*
k1=60 k/in y&&2 + 2ξ2ω2 y& 2 + ω22 y 2 = V2* (t ) / m2*
m2=1.5 k-s2/in
u2(t) y&&3 + 2ξ3ω3 y& 3 + ω32 y 3 = V3* (t ) / m3*
k2=120 k/in
m3=2.0 k-s2/in y&&1 + 0.458y&1 + 21.0y1 = −1.425v&&g (t )
u3(t)
y&&2 + 0.983y&2 + 96.6y2 = 0.511v&&g (t )
k3=180 k/in
y&&3 + 1.457y&3 + 212.4y3 = −0.090v&&g (t )

Modal Participation Factors
Mode 1 1 .4 2 5 1 .9 1 1
Mode 2 − 0 .5 1 1 − 0 .7 9 9
Mode 3 0 .0 9 0 0 .4 3 5
Modal scaling φi,1 = 10

. φi Mφi =10
T
.

(continued)
⎧1.000 ⎫ ⎧0.744 ⎫
⎪ ⎪ ⎪ ⎪
1.425 ⎨0.644 ⎬ = 1.911⎨0.480 ⎬
⎪0.300 ⎪ ⎪0.223 ⎪
⎩ ⎭ ⎩ ⎭
using φ1,1 =1 using φ T Mφ =1
1 1

Effective Modal Mass
Mn % Accum%
Mode 1 3.66 81 81
Mn = Pn m
2 *
n
Mode 2 0.64 14 95
Mode 3 0.20 5 100%
4.50 100%

Solving modal equation via NONLIN:
For Mode 1:
&y&1 + 2ξ1ω1y&1 + ω12 y1 = V1* (t ) / m1*
1.00y&&1 + 0.458y&1 + 21.0y1 = −1.425v&&g (t )
M = 1.00 kip-sec2/in
C = 0.458 kip-sec/in
K1 = 21.0 kips/inch
Scale ground acceleration by factor 1.425

Modal Displacement Response Histories (from NONLIN)
6.00
3.00
0.00
-3.00
T1 = 1.37 sec
-6.00
MODE 1 0 1 2 3 4 5 6 7 8 9 10 11 12
2.00
1.00
0.00
T2 = 0.64
-1.00
-2.00
MODE 2 0 1 2 3 4 5 6 7 8 9 10 11 12
0.20
0.10
0.00 T3 = 0.43
-0.10
-0.20
MODE 3 0 1 2 3 4 5 6 7 8 9 10 11 12
Time, Seconds
Maxima
Modal Response Histories:

5
4
Modal Displacement, Inches
3
2
1
0
-1
-2
-3 MODE 1
MODE 2
-4 MODE 3
-5
0 2 4 6 8 10 12
Time, Seconds

Compute story displacement response histories: u(t ) = Φy (t )
6
u1(t) 4
2
0
-2
-4
-6
0 1 2 3 4 5 6 7 8 9 10 11 12
6
4
u2(t) 2
0
-2
-4
-6
0 1 2 3 4 5 6 7 8 9 10 11 12
6
4
u3(t) 2
0
-2
-4
-6
0 1 2 3 4 5 6 7 8 9 10 11 12
Time, Seconds
= 0.300 x Mode 1 - 0.676 x Mode 2 + 2.47 x Mode 3

Compute story shear response histories:
400
u1(t) 200
0
-200
-400
0 1 2 3 4 5 6 7 8 9 10 11 12
400
200
u2(t) 0
-200
-400
0 1 2 3 4 5 6 7 8 9 10 11 12
400
u3(t) 200
0
-200
-400
0 1 2 3 4 5 6 7 8 9 10 11 12
Time, Seconds
= k2[u2(t) - u3(t)]
Displacements and forces at time of maximum displacements
(t = 6.04 sec)
134.8 k 0
5.11”
134.8
62.1 k 1348
2.86”
196.9
23.5 k 3317
1.22”
220.4
5521
Story Shear (k) Story OTM (ft-k)

Displacements and forces at time of maximum shear
(t = 3.18 sec)
38.2 k 0
3.91”
38.2
124.7 k 382
3.28”
162.9
182.7 k 2111
1.92”
345.6 5567

Modal Response Response Spectrum Method
• Instead of solving the time history problem for each
mode, use a response spectrum to compute the
maximum response in each mode.
• These maxima are generally nonconcurrent.
• Combine the maximum modal responses using some

statistical technique, such as square root of the sum of
the squares (SRSS) or complete quadratic combination
(CQC).
• The technique is approximate.
• It is the basis for the equivalent lateral force (ELF) method.

Example 1 (Response Spectrum Method)
Displacement Response Spectrum
1940 El Centro, 0.35g, 5% Damping
7.00
6.00 Modal response

Spectral Displacement, Inches
5.00
4.00 3.47”
3.00
3.04”
2.00
1.00
1.20”
0.00
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
Period, Seconds
Mode 3 Mode 1
T = 0.431 sec Mode 2 T = 1.37 sec
T = 0.639 sec
7.00
6.00

5.00
4.00
3.47”
3.00
3.04”
2.00
1.20”
1.00
0.00
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
Period, Seconds
Modal Equations of Motion Modal Maxima

y&&1 + 0.458 y&1 + 21.0 y1 = −1.425v&&g (t ) y1 = 1.425 * 3.47 = 4.94"
y&&2 + 0.983 y& 2 + 96.6 y 2 = 0.511v&&g (t ) y 2 = 0.511* 3.04 = 1.55"
y&&3 + 1.457 y& 3 + 212.4 y 3 = −0.090v&&g (t ) y 3 = 0.090 * 1.20 = 0.108"

7.00 Example 1 (continued)
6.00
5.00
The scaled response
4.00 3.47x 1.425”
spectrum values give
3.00
3.04 x 0.511” the same modal maxima
2.00
1.20 x 0.090” as the previous time
1.00
Histories.
0.00
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
Period, Seconds
6.00
3.00
Mode 1 0.00
T = 1.37 -3.00
-6.00
0 1 2 3 4 5 6 7 8 9 10 11 12
2.00
1.00
Mode 2 0.00
T = 0.64 -1.00
-2.00
0 1 2 3 4 5 6 7 8 9 10 11 12
0.20
0.10
Mode 3 0.00
T = 0.43 -0.10
-0.20
0 1 2 3 4 5 6 7 8 9 10 11 12
Time, Seconds

Computing Nonconcurrent Story Displacements
⎧1.000 ⎫ ⎧4.940 ⎫
⎪ ⎪ ⎪ ⎪
Mode 1 ⎨0.644 ⎬ 4.940 = ⎨ 3.181⎬
⎪0.300 ⎪ ⎪1.482 ⎪
⎩ ⎭ ⎩ ⎭
⎧ 1.000 ⎫ ⎧ 1.550 ⎫
⎪ ⎪ ⎪ ⎪
Mode 2 ⎨ − 0.601⎬ 1.550 = ⎨ − 0.931⎬
⎪−0.676 ⎪ ⎪−1.048 ⎪
⎩ ⎭ ⎩ ⎭
⎧ 1.000 ⎫ ⎧ 0.108 ⎫
Mode 3 ⎪ ⎪ ⎪ ⎪
⎨− 2.570⎬0.108 = ⎨− 0.278⎬
⎪ 2.470 ⎪ ⎪ 0.267 ⎪
⎩ ⎭ ⎩ ⎭

Modal Combination Techniques (for Displacement)
Sum of Absolute Values: At time of maximum displacement
“Exact”
⎧4.940 + 1.550 + 0.108⎫ ⎧6.60⎫
⎪ ⎪ ⎪ ⎪ ⎧5.15 ⎫
⎨ 3.181+ 0.931+ 0.278 ⎬ = ⎨4.39⎬ ⎪ ⎪
⎨2.86 ⎬
⎪1.482 + 1.048 + 0.267⎪ ⎪2.80⎪ ⎪⎩1.22 ⎭⎪
⎩ ⎭ ⎩ ⎭
Square Root of the Sum of the Squares:
Envelope of story displacement
⎧ 4.9402 + 1.5502 + 0.1082 ⎫
⎪ ⎪ ⎧5.18⎫ ⎧5.15⎫
⎪ 2 ⎪ ⎪ ⎪ ⎪ ⎪
⎨ 3.181 + 0.931 + 0.278 ⎬ = ⎨3.33⎬
2 2
⎨3.18 ⎬
⎪ 2 ⎪ ⎪1.84⎪ ⎩⎪1.93 ⎭⎪
⎪⎩ 1.482 + 1.048 + 0.267 ⎪⎭ ⎩ ⎭
2 2

Computing Interstory Drifts
⎧4.940 − 3.181⎫ ⎧1.759 ⎫
⎪ ⎪ ⎪ ⎪
Mode 1 ⎨ 3.181 − 1.482 ⎬ = ⎨1.699 ⎬
⎪ ⎪ ⎪ ⎪
⎩ 1.482 − 0 ⎭ ⎩1.482 ⎭
⎧ 1.550−(−0.931) ⎫ ⎧ 2.481⎫
Mode 2 ⎪ ⎪ ⎪ ⎪
⎨− 0. −
931 (−1.048) =
⎬ ⎨ 0.117⎬
⎪ −1.048−0 ⎪ ⎪−1.048⎪
⎩ ⎭ ⎩ ⎭
⎧0.108− (−0.278)⎫ ⎧ 0.386 ⎫
⎪ ⎪ ⎪ ⎪
Mode 3 ⎨ − 0.278− 0.267⎬= ⎨− 0.545⎬
⎪ 0.267− 0 ⎪ ⎪ 0.267 ⎪
⎩ ⎭ ⎩ ⎭

Computing Interstory Shears (Using Drift)
⎧ 1.759(60) ⎫ ⎧105.5 ⎫
⎪ ⎪ ⎪ ⎪
Mode 1 ⎨1.699(120)⎬ = ⎨203.9 ⎬
⎪ ⎪ ⎪ ⎪
⎩1.482(180)⎭ ⎩266.8 ⎭
⎧ 2.481(60) ⎫ ⎧ 148.9 ⎫
Mode 2 ⎪ ⎪ ⎪ ⎪
⎨ 0.117(120) =
⎬ ⎨ 14.0 ⎬
⎪−1.048(180)⎪ ⎪−188.6⎪
⎩ ⎭ ⎩ ⎭
⎧ 0.386(60) ⎫ ⎧ 23.2 ⎫
⎪ ⎪ ⎪ ⎪
Mode 3 ⎨− 0.545(120) =
⎬ ⎨−65.4 ⎬
⎪ 0.267(180) ⎪ ⎪ 48.1 ⎪
⎩ ⎭ ⎩ ⎭

Computing Interstory Shears: SRSS Combination
⎧ 106 2 + 149 2 + 23.22 ⎫ ⎧220 ⎫
⎪⎪ ⎪
2 ⎪ ⎪ ⎪
⎨ 204 2
+ 14 2
+ 65 . 4 =
⎬ ⎨ 215 ⎬
⎪ 267 2 + 189 2 + 48.12 ⎪ ⎪ 331⎪
⎪⎩ ⎪⎭ ⎩ ⎭
“Exact” “Exact” “Exact”

⎧38.2⎫ ⎧135 ⎫ ⎧207 ⎫
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎨ 163 ⎬ ⎨197 ⎬ ⎨203 ⎬
⎪ 346 ⎪ ⎪220 ⎪ ⎪346 ⎪
⎩ ⎭ ⎩ ⎭ ⎩ ⎭
At time of At time of max. Envelope = maximum
max. shear displacement per story

Caution:
Do NOT compute story shears from the story

drifts derived from the SRSS of the story
displacements.
Calculate the story shears in each mode

(using modal drifts) and then SRSS
the results.

Using Less than Full (Possible)
Number of Natural Modes
Modal Response Histories:

5
4
3
2
1
0
-1
-2
-3 MODE 1
MODE 2
-4 MODE 3
-5
0 2 4 6 8 10 12
Time, Seconds

Using Less than Full Number of Natural Modes
Time-History for Mode 1
⎡ y1(t1) y1(t 2) y1(t 3) y1(t 4) y1(t 5) y1(t 6) y1(t 7) y1(t 8) .... y1(tn)⎤
y (t ) = ⎢⎢y 2 (t1) y 2 (t 2) y 2 (t 3) y 2 (t 4) y 2 (t 5) y 2 (t 6) y 2 (t 7) y 2 (t 8) .... y 2 (tn)⎥⎥
⎢⎣y3 (t1) y3 (t 2) y3 (t 3) y3 (t 4) y3 (t 5) y3 (t 6) y3 (t 7) y3 (t 8) .... y3 (tn)⎥⎦
u(t ) = [φ1 φ2 φ3 ]y (t )
Transformation: 3 x nt 3x3 3 x nt
3 x nt
Time History for DOF 1

⎡u1(t1) u1(t 2) u1(t 3) u1(t 4) u1(t 5) u1(t 6) u1(t 7) u1(t 8) .... u1(tn) ⎤
u(t ) = ⎢⎢u2 (t1) u2 (t 2) u2 (t 3) u2 (t 4) u2 (t 5) u2 (t 6) u2 (t 7) u2 (t 8) .... u2 (tn)⎥⎥
⎢⎣u3 (t1) u3 (t 2) u3 (t 3) u3 (t 4) u3 (t 5) u3 (t 6) u3 (t 7) u3 (t 8) .... u3 (tn)⎥⎦

Time History for Mode 1
⎡ y (t1) y1(t 2) y1(t 3) y1(t 4) y1(t 5) y1(t 6) y1(t 7) y1(t 8) .... y1(tn) ⎤
y (t ) = ⎢ 1 ⎥
⎣ 2
y (t 1) y 2 (t 2) y 2 (t 3) y 2 (t 4) y 2 (t 5) y 2 (t 6) y 2 (t 7) y 2 (t 8) .... y 2 (tn )⎦
NOTE: Mode 3 NOT Analyzed
u(t ) = ⎡⎣φ1 φ2 ⎤⎦ y (t )
Transformation: 3 x nt 3x2 2 x nt
3 x nt
Time history for DOF 1

⎡u1(t1) u1(t 2) u1(t 3) u1(t 4) u1(t 5) u1(t 6) u1(t 7) u1(t 8) .... u1(tn) ⎤
u(t ) = ⎢⎢u2 (t1) u2 (t 2) u2 (t 3) u2 (t 4) u2 (t 5) u2 (t 6) u2 (t 7) u2 (t 8) .... u2 (tn)⎥⎥
⎢⎣u3 (t1) u3 (t 2) u3 (t 3) u3 (t 4) u3 (t 5) u3 (t 6) u3 (t 7) u3 (t 8) .... u3 (tn)⎥⎦

(Modal Response Spectrum Technique)
Sum of absolute values:
⎧4.940 + 1.550 + 0.108 ⎫ ⎧6.60 ⎫ ⎧ 6.49 ⎫
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎨ 3.181 + 0.931 + 0.278 ⎬ = ⎨ 4.39 ⎬ ⎨4.112⎬ At time of maximum
⎪1.482 + 1.048 + 0.267 ⎪ ⎪2.80 ⎪ ⎪ 2.53 ⎪
⎩ ⎭ ⎩ ⎭ ⎩ ⎭ displacement
Square root of the sum of the squares: “Exact”:

⎧5.15⎫
⎧ 4.9402 + 1.5502 + 0.1082 ⎫ ⎪ ⎪
⎪ ⎪ ⎧5.18⎫ ⎧5.18⎫ ⎨ 2.86 ⎬
⎪ 2 ⎪ ⎪ ⎪ ⎪ ⎪ ⎪⎩1.22 ⎭⎪
⎨ 3.181 + 0.931 + 0.278 ⎬ = ⎨3.33⎬
2 2
⎨3.31⎬
⎪ 2 ⎪ ⎪1.84⎪ ⎪1.82⎪
⎪⎩ 1.482 + 1.048 + 0.267 ⎪⎭ ⎩ ⎭ ⎩ ⎭
2 2
3 modes 2 modes
Example of MDOF Response of Structure
Responding to 1940 El Centro Earthquake
u1(t) Example 2
m1=2.5 k-s2/in
10 ft k1=150 k/in Assume Wilson
damping with 5%
u2(t) critical in each mode.
m2=2.5 k-s2/in
10 ft k2=150 k/in
u3(t)
m3=2.5 k-s2/in
10 ft k3=150 k/in
N-S component of 1940 El Centro earthquake

Maximum acceleration = 0.35 g

Form property matrices:

m1=2.5 k-s2/in
u1(t) ⎡2.5 ⎤
m2=2.5 k-s2/in
k1=150 k/in
M = ⎢⎢ 2.5 ⎥ kip − s2/in
⎥
u2(t) ⎢⎣ 2.5⎥⎦
k2=150 k/in
m3=2.5 k-s2/in
u3(t) ⎡ 150 −150 0 ⎤
k3=150 k/in K = ⎢⎢−150 300 −150⎥⎥ kip/in
⎢⎣ 0 −150 300 ⎥⎦

Solve = eigenvalue problem:

m1=2.5 k-s2/in KΦ = MΦΩ2
u1(t)
k1=150 k/in ⎡11.9 ⎤
m2=2.5 k-s2/in ⎢
Ω =⎢
2
93.3 ⎥ sec −2
u2(t) ⎥
k2=150 k/in ⎢⎣ 194.8⎥⎦
m3=2.5 k-s2/in
u3(t)
⎡1.000 1.000 1.000 ⎤
k3=150 k/in
Φ = ⎢⎢0.802 −0.555 −2.247⎥⎥
⎢⎣0.445 −1.247 1.802 ⎥⎦

Normalization of Modes Using Φ MΦ = I
T
⎡0.466 0.373 0.207 ⎤ ⎡1.000 1.000 1.000 ⎤

Φ = ⎢⎢0.373 −0.207 −0.465⎥⎥ vs ⎢0.802 −0.555 −2.247⎥
⎢ ⎥
⎢⎣0.207 −0.465 0.373 ⎥⎦ ⎢⎣0.445 −1.247 1.802 ⎥⎦

Mode Shapes and Periods of Vibration
Mode 1 Mode 2 Mode 3

ω = 3.44 rad/sec ω = 9.66 rad/sec ω = 13.96 rad/sec
T = 1.82 sec T = 0.65 sec T = 0.45 sec

⎧ 3.44 ⎫ ⎧1.82⎫
⎪ ⎪ ⎪ ⎪
m1=2.5 k-s2/in ωn = ⎨ 9.66 ⎬rad/ sec Tn = ⎨0.65⎬sec
u1(t) ⎪13.96⎪ ⎪0.45⎪
⎩ ⎭ ⎩ ⎭
k1=150 k/in
m2=2.5 k-s2/in
u2(t)
Compute generalized mass:
k2=150 k/in
m3=2.5 k-s2/in ⎡ 4 .6 0 3 ⎤
u3(t) M * = Φ T M Φ = ⎢⎢ 7 .1 5 8 ⎥ k ip − se c 2 / in
⎥
⎢⎣ 2 3 .2 4 1⎥⎦
k3=150 k/in

m1=2.5 k-s2/in
u1(t) Compute generalized loading:
k1=150 k/in
m2=2.5 k-s2/in
V * (t ) = −ΦT MR v&&g (t )
u2(t)
k2=150 k/in ⎧−5.617⎫
m3=2.5 k-s2/in ⎪ ⎪ &&
u3(t) Vn = ⎨ 2.005 ⎬ vg (t )
*
⎪−1.388⎪
k3=150 k/in ⎩ ⎭

Write uncoupled (modal) equations of motion:

m1=2.5 k-s2/in
u1(t) y&&1 + 2ξ1ω1y&1 + ω12 y1 = V1* (t )/ m1*
k1=150 k/in y&&2 + 2ξ2ω2 y& 2 + ω22 y 2 = V2* (t ) / m2*

m2=2.5 k-s2/in
u2(t) y&&3 + 2ξ3ω3 y& 3 + ω32 y 3 = V3* (t ) / m3*
k2=150 k/in
m3=2.5 k-s2/in y&&1 + 0.345y&1 + 11.88y1 = −1.22v&&g (t )
u3(t)
y&&2 + 0.966y&2 + 93.29y2 = 0.280v&&g (t )
k3=150 k/in
y&&3 + 1.395y&3 + 194.83y3 = −0.06v&&g (t )

Mode 1 − 1 .2 2 − 2 .6 1 5
Mode 2 0 .2 8 0 .7 4 8
Mode 3 − 0 .0 6 0 − 0 .2 8 7
Modal scaling φi,1 = 10

. φi Mφi =10
T
.

Effective Modal Mass
% Accum% Mn
Mode 1 6.856 91.40 91.40
Mn = Pn mn
2
Mode 2 0.562 7.50 98.90
Mode 3 0.083 1.10 100.0
7.50 100%

Solving modal equation via NONLIN:
For Mode 1:
&y&1 + 2ξ1ω1y&1 + ω12 y1 = V1* (t ) / m1*
1.00y&&1 + 0.345y&1 + 11.88 y1 = −1.22v&&g (t )
M = 1.00 kip-sec2/in
C = 0.345 kip-sec/in
K1 = 11.88 kips/inch
Scale ground acceleration by factor 1.22

Modal Displacement Response Histories (from NONLIN)
10.00
5.00
0.00 T=1.82
-5.00
-10.00
Mode 1 0 1 2 3 4 5 6 7 8 9 10 11 12
1.00
0.50
0.00
-0.50
T=0.65
-1.00
Mode 2 0 1 2 3 4 5 6 7 8 9 10 11 12
0.15
0.10
0.05
0.00 C
T=0.45
-0.05
-0.10
Mode 3 0 1 2 3 4 5 6 7 8 9 10 11 12
Maxima
Modal Response Histories
8
MODE 1
6
MODE 2
4 MODE 3
-2
-4
-6
-8
0 2 4 6 8 10 12
Time, Seconds

Compute story displacement response histories: u(t ) = Φy (t )
8.00
6.00
u1(t) 4.00
2.00
0.00
-2.00
-4.00
-6.00
-8.00
0 1 2 3 4 5 6 7 8 9 10 11 12
8.00
6.00
u2(t) 4.00
2.00
0.00
-2.00
-4.00
-6.00
-8.00
0 1 2 3 4 5 6 7 8 9 10 11 12
8.00
6.00
u3(t) 4.00
2.00
0.00
-2.00
-4.00
-6.00
-8.00
0 1 2 3 4 5 6 7 8 9 10 11 12
= 0.445 x Mode 1 – 1.247 x Mode 2 + 1.802 x Mode 3

Compute story shear response histories:
600.00
400.00
u1(t) 200.00
0.00
-200.00
-400.00
-600.00
0 1 2 3 4 5 6 7 8 9 10 11 12
600.00
400.00
u2(t) 200.00
0.00
-200.00
-400.00
-600.00
0 1 2 3 4 5 6 7 8 9 10 11 12
600.00
u3(t) 400.00
200.00
0.00
-200.00
-400.00
-600.00
0 1 2 3 4 5 6 7 8 9 10 11 12
Time, Seconds
=k2[u2(t)-u3(t)]
Displacements and Forces at time of Maximum Displacements
(t = 8.96 seconds)
222.2 k 0
6.935”
222.2
175.9 k 2222
5.454”
398.1
21.9 k 6203
2.800”
420.0 10403

Displacements and Forces at Time of Maximum Shear
(t = 6.26 sec)
130.10 k 0
6.44”
130.10
180.45 k 1301
5.57”
310.55
215.10 k 4406
3.50”
525.65 9663

Modal Response Response Spectrum Method
• Instead of solving the time history problem for each
mode, use a response spectrum to compute the
maximum response in each mode.
• These maxima are generally nonconcurrent.

• Combine the maximum modal responses using some
statistical technique, such as square root of the sum of
the squares (SRSS) or complete quadratic combination
(CQC).
• The technique is approximate.

• It is the basis for the equivalent lateral force (ELF) method.
Example 2 (Response Spectrum Method)
Displacement Response Spectrum
1940 El Centro, 0.35g, 5% Damping
7.00
MODAL RESPONSE
6.00
5.00 5.71”
4.00
3.02”
3.00
2.00
1.00 1.57”
0.00
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
Period, Seconds
Mode 3 Mode 2 Mode 1

T = 0.45 sec T = 0.65 sec T = 1.82 sec

7.00

6.00
5.00 5.71”
4.00
3.02”
3.00
2.00
1.00
1.57”
0.00
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
Period, Seconds
Modal Equations of Motion Modal Maxima

y&&1 + 0.345 y&1 + 11.88 y1 = −1.22v&&g (t ) y1 = 1.22 * 5.71 = 6.966"
y&&2 + 0.966 y& 2 + 93.29 y 2 = 0.280v&&g (t ) y 2 = 0.28 * 3.02 = 0.845"
y&&3 + 1.395 y& 3 + 194.83 y 3 = −0.060v&&g (t ) y3 = 0.060 *1.57 = 0.094"

7.00
5.71 x 1.22 Example 2 (continued)
S p e c tra l D is p la c e m e n t, In c h e s
6.00 in.
5.00
4.00 3.02 x 0.28 in. The scaled response

3.00 spectrum values give
2.00 1.57 x 0.06 in. the same modal maxima
1.00
as the previous time
0.00
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
histories.
Period, Seconds
10.00
5.00
0.00
Mode 1 -5.00
T = 1.82 -10.00
0 1 2 3 4 5 6 7 8 9 10 11 12
1.00
0.50
Mode 2 0.00
T = 0.65
-0.50
-1.00
0 1 2 3 4 5 6 7 8 9 10 11 12
0.15
0.10
Mode 3 0.05
T = 0.45 0.00
-0.05
C
-0.10
0 1 2 3 4 5 6 7 8 9 10 11 12
Computing Nonconcurrent Story Displacements
⎧1.000 ⎫ ⎧6.966 ⎫
⎪ ⎪ ⎪ ⎪
Mode 1 ⎨ 0.802 ⎬ 6.966 = ⎨ 5.586 ⎬
⎪0.445 ⎪ ⎪3.100 ⎪
⎩ ⎭ ⎩ ⎭
⎧ 1.000 ⎫ ⎧ 0.845 ⎫
⎪ ⎪ ⎪ ⎪
Mode 2 ⎨ − 0.555 ⎬ 0.845 = ⎨ − 0.469 ⎬
⎪−1.247 ⎪ ⎪−1.053 ⎪
⎩ ⎭ ⎩ ⎭
⎧ 1.000 ⎫ ⎧ 0.094 ⎫
Mode 3 ⎪ ⎪ ⎪ ⎪
⎨−2.247 ⎬ 0.094 = ⎨−0.211⎬
⎪ 1.802 ⎪ ⎪ 0.169 ⎪
⎩ ⎭ ⎩ ⎭
Modal Combination Techniques (For Displacement)
At time of maximum displacement
Sum of absolute values:
“Exact”
⎧6.966 + 0.845 + 0.108⎫ ⎧7.919⎫
⎪ ⎪ ⎪ ⎪ ⎧6.935⎫
⎨5.586 + 0.469 + 0.211⎬ = ⎨6.266⎬ ⎪ ⎪
⎪3.100 + 1.053 + 0.169⎪ ⎪4.322⎪ ⎨5.454⎬
⎩ ⎭ ⎩ ⎭
⎩⎪2.800⎭⎪
Square root of the sum of the squares Envelope of story displacement
⎧ 6.9662 + 0.8452 + 0.1082 ⎫
⎪ ⎪ ⎧7.02⎫ ⎧6.935⎫
⎪ 2⎪ ⎪ ⎪ ⎪ ⎪
⎨ 5.586 + 0.469 + 0.211 ⎬ = ⎨5.61⎬
2 2
⎨5.675⎬
⎪ 2 ⎪ ⎪3.28⎪
⎪⎩ 3.100 + 1.053 + 0.169 ⎪⎭ ⎩
2 2
⎭ ⎩⎪2.965⎭⎪
Computing Interstory Drifts
⎧ 6.966 − 5.586 ⎫ ⎧ 1.380 ⎫
⎪ ⎪ ⎪ ⎪
Mode 1 ⎨ 5.586 − 3.100 ⎬ = ⎨ 2.486 ⎬
⎪ ⎪ ⎪ ⎪
⎩ 3.100 − 0 ⎭ ⎩ 3.100 ⎭
⎧ 0.845 − (−0.469) ⎫ ⎧ 1.314 ⎫

Mode 2 ⎪ ⎪ ⎪ ⎪
⎨− 0.469 − ( −1.053) =
⎬ ⎨ 0.584 ⎬
⎪ −1.053 − 0 ⎪ ⎪−1.053⎪
⎩ ⎭ ⎩ ⎭
⎧0.108 − ( −0.211)⎫ ⎧ 0.319 ⎫
⎪ ⎪ ⎪ ⎪
Mode 3 ⎨ −0.211 − 0.169 ⎬ = ⎨−0.380⎬
⎪ 0.169 − 0 ⎪ ⎪ 0.169 ⎪
⎩ ⎭ ⎩ ⎭
Computing Interstory Shears (Using Drift)
⎧1.380(150) ⎫ ⎧ 207.0 ⎫
⎪ ⎪ ⎪ ⎪
Mode 1 ⎨ 2.486(150) ⎬ = ⎨ 372.9 ⎬
⎪ 3.100(150) ⎪ ⎪ 465.0 ⎪
⎩ ⎭ ⎩ ⎭
⎧ 1.314(150) ⎫ ⎧ 197.1 ⎫
Mode 2 ⎪ ⎪ ⎪ ⎪
⎨ 0.584(150) =
⎬ ⎨ 87.6 ⎬
⎪−1.053(150)⎪ ⎪−157.9⎪
⎩ ⎭ ⎩ ⎭
⎧ 0.319(150) ⎫ ⎧ 47.9 ⎫
⎪ ⎪ ⎪ ⎪
Mode 3 ⎨ − 0.380(150) =
⎬ ⎨ − 57.0 ⎬
⎪ 0.169(150) ⎪ ⎪ 25.4 ⎪
⎩ ⎭ ⎩ ⎭
Computing Interstory Shears: SRSS Combination
⎧ 2072 + 197.12 + 47.92 ⎫
⎪ ⎪ ⎧ 289.81⎫
⎪ 2 ⎪ ⎪ ⎪
⎨ 372.9 2
+ 87.6 2
+ 57 =
⎬ ⎨ 387.27 ⎬
⎪ 2⎪ ⎪491.73 ⎪
⎪⎩ 465 + 157.9 + 25.4 ⎪⎭ ⎩ ⎭
2 2
“Exact” “Exact” “Exact”

⎧ 130.1⎫ ⎧222.2 ⎫ ⎧304.0 ⎫
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎨310.5 ⎬ ⎨ 398.1⎬ ⎨398.5 ⎬
⎪525.7 ⎪ ⎪420.0 ⎪ ⎪525.7 ⎪
⎩ ⎭ ⎩ ⎭ ⎩ ⎭
At time of At time of max. Envelope = maximum
max. shear displacement per story
ASCE 7 Allows an Approximate
Modal Analysis Technique Called the
Equivalent Lateral Force Procedure
• Empirical period of vibration

• Smoothed response spectrum
• Compute total base shear, V, as if SDOF
• Distribute V along height assuming “regular”
geometry
• Compute displacements and member forces using
standard procedures
• Method is based on first mode response.

• Higher modes can be included empirically.
• Has been calibrated to provide a reasonable
estimate of the envelope
of story shear, NOT to provide accurate
estimates of story force.
• May result in overestimate of overturning
moment.
• Assume first mode effective mass = total Mass = M = W/g
• Use response spectrum to obtain total acceleration @ T1
Acceleration, g
Sa1
T1 Period, sec
W
VB = (Sa1g )M = (Sa1g ) = Sa1W
g
dr
Assume linear first mode response
hx 2 Wx
Wx fx (t ) = d r (t )ω1
fx h g
nstories
d r (t )ω12 nstories
h hx
VB (t ) = ∑i =1
fi (t ) =
hg
∑i =1
hiWi
Portion of base shear applied to story i

fx (t ) hxWx
=
VB (t ) nstories
∑i =1
hiWi
ELF Procedure Example
m1=1.0 k-s2/in
k1=60 k/in
m2=1.5
Recall
3h k2=120 k/in T1 = 1.37 sec
m3=2.0
k3=180 k/in
ELF Procedure Example
Total weight = M x g = (1.0 + 1.5 + 2.0) 386.4 = 1738 kips
Spectral acceleration = w2SD = (2p/1.37)2 x 3.47 = 72.7
in/sec2 = 0.188g
7.00
6.00
5.00
4.00 3.47 in.
3.00
2.00
1.00
1.37sec
0.00
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
Period, Seconds
Base shear = SaW = 0.188 x 1738 = 327 kips
ELF Procedure Example (Story Forces)
386(3h )
f3 = = 0.381VB = 0.375(327) = 125 kips
386(3h ) + 579(2h ) + 722(h )
125 k
W=386 k
125 k
(220 k)
125 k
W=579 k
250 k
3h (215 k)
77 k
W=722 k
327 k
327 k (331 k)
Story Shear (k)
ELF Procedure Example (Story Displacements)
Units = inches
Time History Modal Response

(Envelope) Spectrum
ELF
⎧5.15 ⎫ ⎧5.18 ⎫ ⎧5.98 ⎫

⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎨3.18 ⎬ ⎨3.33 ⎬ ⎨3.89 ⎬
⎪⎩1.93 ⎪⎭ ⎪1.84 ⎪ ⎪1.82 ⎪
⎩ ⎭ ⎩ ⎭
ELF Procedure Example (Summary)
• ELF procedure gives good correlation with

base shear (327 kips ELF vs 331 kips modal
response spectrum).
• ELF story force distribution is not as good.

ELF underestimates shears in upper stories.
• ELF gives reasonable correlation with displacements.
Higher Mode Effects
+ =
1st Mode 2nd Mode Combined
ASCE 7-05 ELF Approach
• Uses empirical period of vibration

• Uses smoothed response spectrum
• Has correction for higher modes
• Has correction for overturning moment
• Has limitations on use
Approximate Periods of Vibration
Ta = Ct h x
n
Ct = 0.028, x = 0.8 for steel moment frames
Ct = 0.016, x = 0.9 for concrete moment frames
Ct = 0.030, x = 0.75 for eccentrically braced frames
Ct = 0.020, x = 0.75 for all other systems
Note: For building structures only!
Ta = 0.1N
For moment frames < 12 stories in height, minimum
story height of 10 feet. N = number of stories.
Adjustment Factor on Approximate Period
T = TaCu ≤ Tcomputed
SD1 Cu
> 0.40g 1.4
0.30g 1.4
0.20g 1.5
0.15g 1.6
< 0.1g 1.7
Applicable only if Tcomputed comes from a “properly

substantiated analysis.”
ASCE 7 Smoothed Design Acceleration Spectrum
(for Use with ELF Procedure)
SDS
V = CSW CS =
1
⎛R⎞
⎜I ⎟
“Short ⎝ ⎠
period”
SDS acceleration
SD1
Spectral Response
CS =
Acceleration, Sa
1 2
⎛R⎞
“Long period” T⎜ ⎟
acceleration ⎝I ⎠
2
SD1
3 Varies
3
T = 0.2 T = 1.0 Period, T
R is the response modification factor, a
function of system inelastic behavior. This is covered
in the topic on inelastic behavior. For now, use
R = 1, which implies linear elastic behavior.
I is the importance factor which depends on the

Seismic Use Group. I = 1.5 for essential facilities,
1.25 for important high occupancy structures,
and 1.0 for normal structures. For now, use I = 1.
Distribution of Forces Along Height
Fx = CvxV
k
wxh
Cvx = n
x
∑w h
i =1
i i
k
k Accounts for Higher Mode Effects
k = 0.5T + 0.75
(sloped portion only)
k
2.0
1.0
0.5 2.5
Period, sec
k=1 k=2
ELF Procedure Example (Story Forces)
V = 327 kips T = 1.37 sec k = 0.5(1.37) + 0.75 = 1.435
146 k
W = 386k (125 k)
146 k
120 k
W = 579k (125 k)
266 k
3h
61 k
W = 722k (77 k)
327 k
327 k
Story Shear (k)
ASCE 7 ELF Procedure Limitations
• Applicable only to “regular” structures with T
less than 3.5Ts. Note that Ts = SD1/SDS.
Adjacent story stiffness does not vary more than 30%.

Adjacent story strength does not vary more than 20%.
Adjacent story masses does not vary more than 50%.
If violated, must use more advanced analysis (typically

modal response spectrum analysis).
ASCE 7 ELF Procedure
Other Considerations Affecting Loading
• Orthogonal loading effects

• Redundancy
• Accidential torsion
• Torsional amplification
• P-delta effects
• Importance factor
• Ductility and overstrength

Rev Topic04 StructuralDynamicsofMDOFSystems

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Rev Topic04 StructuralDynamicsofMDOFSystems

Uploaded by

Copyright:

Available Formats

Structural Dynamics of Linear Elastic

Multiple-Degrees-of-Freedom (MDOF) Systems

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 1

• Equations of motion for MDOF systems

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 2

m Element of matrix or vector or set

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 3

ASCE 7-05 provides guidance for three specific

See ASCE 7-05

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 4

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 5

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 6

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 7

d3,2 d3,1 d3,2

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 8

d2,3 d1,1 d1,2 d1,3

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 9

⎡ d1,1 d1,2 d1,3 ⎤ ⎧ f1 ⎫

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 10

⎡K m,m K m,n ⎤ ⎧Um ⎫ ⎧ Fm ⎫ DOF with mass

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 11

Simplify ⎡⎣K m,m − K m,nK K n,m ⎤⎦ Um = Fm −1

Condensed stiffness matrix

DOF 1 m1u&&1(t ) + k1u1(t ) − k1u2 (t ) = f1(t )

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 14

• This will be done by a transformation of coordinates

• Because of the orthogonality property of the natural mode

• After solving, we can transform back to the normal

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 15

&&(t ) + KU(t ) = {0}

⎧φ1,1 ⎫ ⎧φ1,2 ⎫ ⎧φ1,3 ⎫

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 16

For a SINGLE Mode

K Φ = MΦΩ2 For ALL Modes

φ3,1 φ3,2 φ3,3

MODE 1 MODE 2 MODE 3

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 18

⎡φ1,1 φ1,2 φ1,3 ⎤ ⎧ y1 ⎫ ⎧φ1,1 ⎫ ⎧φ1,2 ⎫ ⎧φ1,3 ⎫

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 19

Generalized damping Generalized force

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 20

MDOF equation of motion: && &

Using orthogonality conditions, uncoupled equations of motion are:

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 21

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 22

Mode 1 y&&1 + 2ξ1ω1y&1 + ω12 y1 = f1* (t )/ m1*

Mode 3 y&&3 + 2ξ3ω3 y& 3 + ω32 y 3 = f3* (t ) / m3*

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 23

Move to RHS as FEFF (t ) = - M Ru&&g (t )

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 24

For earthquakes: fi (t ) = −φ MRu

Typical modal equation:

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 26

Its value is dependent on the (arbitrary) method

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 27

p1 = 1.0 p1 =1.4 p1 =1.6

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 28

For each Mode I, mi = pi mi 2 *

• The sum of the effective modal mass for all

• The value of effective modal mass is

• Use enough modes in the analysis to provide

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 29

m1 / M = 1.0 m1 / M= 0.9 m1 / M= 0.7

Instructional Material Complementing FEMA 451, Design Examples MDOF Dynamics 4 - 30