STRUCTURAL DYNAMICS

References:
1. ‘Dynamics of Structures’, Clough & Penzien, McGrawhill, 2nd Ed.,
1993.
2. ‘Dynamics of Structures’, Chopra, Prentice Hall, 2nd Ed., 2001.

1. OVERVIEW

Objectives: to present methods for analyzing the stresses and deflections in

any given structure subjected to an arbitrary dynamic loading.

Dynamic = Time-varying
Dynamic Load  Its magnitude, direction and/or position varies with time
 the resulting stresses and deflections are also time-varying.

2. DEGREE OF FREEDOM (DOF) AND LUMPED MASS

Degree of freedom (dof) : Σ independent displacement

components/coordinates required to specify the position of all significant
mass particles in the structure.
m1 m2

1 2

Beam Lumped-Mass Equivalent

m3

m3

m2
m2

m1
m1

m: mass of floor & column

Figure 1: Lumped-Mass Idealization

1
For distributed mass and distributed elastic properties  infinite dof. In
practice  replace the distributed mass of the structure by an equivalent
number of finite masses lumped (or concentrated) at nodal points.

3. VIBRATION THEORY: EQUATION OF MOTION

Newton 2nd law of motion: ‘the rate of change of momentum of any mass
particle m = the force acting on it’.

d dv
p(t) = (m ) (1)
dt dt

where:
p(t) : applied force vector
v(t) : position vector of particle mass m

If m does not vary with time

d 2v
p(t) = m = m v (t)
dt 2

p(t) - m v (t) = 0 (2)

m v (t) = inertia force

Equation (2) also known as D’Alembert Principle: ‘A mass develops an

inertial force proportional to its acceleration v (t) and opposing it. p(t) can be
expressed to include many types of forces applying on the mass.
Weightless spring of stiffness k: to represent the elastic resistance
Basic elements to displacement; damper c to represent the energy-loss
mechanism
v(t): dynamic response
C (displacement from the static
v(t) equilibrium position.
m
p(t) P(t): effective load acting on
k
the system

Figure 2: Idealized Single Degree of Freedom (SDOF) System

2
By d’Alembert equilibrium of forces

P inertia + P damping + P stiffness = P external

m v (t) + c v (t) + k v(t) = P(t) (3)

viscous damping mechanism elastic force

4. UN-DAMPED FREE VIBRATION

Suppose the motion of a structure is initiated by disturbing the system from

its static equilibrium position; e.g. v(0) and v (0) are non zero, then equation
(3) becomes:

m v(t)
k

Figure 3: Idealized Un-damped Single Degree of Freedom (SDOF) System

m v (t) + c v (t) + k v(t) = 0 (4)

The solution for equation (4) may be expressed in the following form:

v(t) = G e st (5)

G : arbitrary complex constant

e st : exponential function

To derive the vibration response, substitute (5) into (4), leading to:

( ms2 + cs + k ) G e st = 0 (6)

3
dividing by mGe st and introducing

k k
ω2 = ( ie., ω = ) (7)
m m
equation (6) becomes

c
s2 + s + ω2 = 0 (8)
m

The type of motion given by (8) depends on the amount of damping in the
system.

For un-damped vibration, c=0, the two values of s given by (8) are:

S 1,2 = ± i ω (9)

Thus the total response includes two terms of (5), i.e.:

v(t) = G 1 e iωt + G 2 e – iωt (10)

an alternative form of this motion expression in trigonometric form is given

by:

v(t) = A cos ωt + B sin ωt (11)

The values of constants A and B may be determined from the initial

conditions, i.e. v(0) and v (0) at t=0. It can be shown that

v(0)
A = v(0) and B= (12)
ω

Thus equation (11) becomes

v(0)
v(t) = v(0) cos ωt + sin ωt (13)
ω

2π
v(t) from equation (13) has the same value at time t and (t + n ) n=1,2,3 . So
ω
equation (13) is periodic, and the period of motion is

4
 2π
T= [seconds] (14)
ω

T : period of motion = time required to complete one cycle

The natural frequency

1 ω
f= = (15)
T 2π

ω [radian/second] = angular velocity f [cycles/second] or [Hertz]

The solution is simple

harmonic motion (SHM)
as can be seen in Figure 4

Figure 4: Un-damped free-vibration response

Alternatively, equation (13) may be expressed as

v(t) = ρ cos ( ωt + θ ) (16)

where:

2
 v(0) 
amplitude ρ = [v(0)] + 
2
 (17)
 ω 

 − v(0) 
phase angle θ = tan –1   (18)
 ωv(0) 

5
The natural vibration properties ω, T and f depend only on the mass and
stiffness of the structure. Therefore, these properties are the natural
properties of the structure.

The natural frequency and period of various structures vary over a wide
range:

Figure 5: The Natural Frequency and Period of various structures

6
Figure 6: The Natural Frequency and Period of various building structure
(From Dynamics of Structures’, Chopra, Prentice Hall, 2nd Ed., 2001)

7
Example 1:
8W

EI
EI 2EI h
0.8W 1.2W Figure 7

Consider the frame shown. Assume beam weight + loading = 8W, and the
weight of the two columns are 0.8W and 1.2W respectively. Determine the
approximate value of the natural frequency and natural period of vibration !

Solution:

Total lumped weight at beam level = 8W + 0.5(0.8W+1.2W) = 9W

9W
Lumped mass, m =
g
1
k Total stiffness:

12 EI 12(2 EI ) 36 EI
k= 3
+ 3
= 3
L L L

k EIg
ω= =2
m WL3

Figure 8: Column stiffness

ω 1 EIg
The natural frequency f= =
2π π WL3

1 WL3
The natural period T= =
f EIg

8
Practice Problems:

Figure 9

Q2. The building frame shown in Figure 10 supports a rotating machine.

Neglect the axial and shear deformations and the weight of the columns, and
assume that the girder is rigid. Take E=200 GPa; I=12x106 mm4; m=2500 kg
and L=5 m. Calculate the natural frequency of vibration sidesway and the
natural period of vibration for the frame.

m
B C

I I L 2EI 2EI

A D

L
L

Figure 10 Figure 11
2L

Q3. Same as Q2 for frame shown in Figure 11. Assume E 200 GPa;
I=10x106 mm4; m=2700 kg; and L=4 metres.

9
5. DAMPED FREE VIBRATION

In this case, damping is present in the structure, a realistic case. The solution
of equation (8) becomes:

2
c  c 
  −ω
2
S 1,2 = - ± (19)
2m  2 m 

Three types of motions are represented by this expression, according to the

quantity under the square root. Sign is +ve, -ve or zero.

CRITICALLY DAMPED STRUCTURE

If the radical term of equation (19) set to zero

c
=ω (20)
2m

thus, the critical value of damping coefficient, C c , is

C c = 2mω (21)

The both values of S from equation (19) are the same

Cc
s1 = s2 = - = -ω (22)
2m

The general solution of equation (4) may be written as

S1t S 2t
v(t) = G 1 e + G2 t e (23a)

or

v(t) = (G 1 + G 2 t) e −ωt (23b)

Using the initial displacement and initial velocity, equation (23) becomes

v(t) = [v(0)(1 − ωt ) + v(0)t ]e −ωt (24)

10
 Graphical presentation of equation (24) is:

v(t) v(0)

v(0)

Figure 12: Free Vibration Response with Critical Damping

Critical damping represents the transition between the non-oscillatory and

oscillatory motions. The motion is not periodic, represents the smallest
amount of damping at which no oscillation occurs in the free-vibration
system/response.

UNDER CRITICALLY DAMPED SYSTEM ( c < c c )

In this case, c < c c or c < 2mω or the quantity under the radical sign is
negative. It is convenient to introduce the non-dimensional ratio, called
damping ratio ξ

c c
ξ= = (25)
cc 2mω

Most structures have ξ < 0.2

The equation (4)

mv(t ) + cv(t ) + kv(t ) = 0

becomes

v(t ) + 2ξωv(t ) + ω 2 v(t ) = 0 (26)

solution of equation (26) is given by:

11
 v(t)= e −ξωt ( A cos ω D t + B sin ω D t ) (27)

where:

A = v(0) (28)

v(0) + ξωv(0)
B= (29)
ωD
And
ωD = ω 1−ξ 2 (30)

Using equation (30), it can be shown that ω D ≅ ω for most structures, which
have ξ < 0.2.

Alternatively, equation (27) can be presented as:

v(t ) = e −ξωt ρ cos(ω D t − θ ) (31)

where

ρ = A2 + B 2 (32)

 B
θ = tan −1   (33)
 A

Figure 13: Effects of Damping on Free Vibration

12
 Figure 14: Vibration of Systems with Four Levels of Damping Ratio;
ξ = 0.02, 0.05, 0.1 and 0.2

DAMPING CHARACTERISTICS (APPROXIMATE METHOD)

Figure 15: Free vibration response of under-critically-damped system

Consider any two successive positive peaks such as v n and v n+1 respectively.
The ratio of these two successive values is given by (using equation 31)

 2πξω 
 
vn  
= e ωD 
(34)
v n +1

13
Taking the natural logarithm (ln) of both sides of this equation, and
substituting ω D = ω 1 − ξ 2 , the logarithmic decrement of damping, δ, is
obtained by:

vn 2πξ
δ = ln = (35)
v n +1 1−ξ 2

For low values of damping, equation (35) can be approximated by:

δ = 2πξ (36)

Thus,

vn δ 2πξ
= e ≅ e = 1 + 2πξ +
(2πξ )
2
+ ... (37)
v n +1 2!

Taking only the first two terms of the Taylor’s Series expansion on the right
hand side

1 v n − v n +1
ξ≅ (38)
2π v n +1

The damping ratio obtained by the approximate method (equation 38) may
be corrected by the following graph

Figure 16:
Damping ratio correction
factor to be applied to result
obtained from equation 38

For lightly damped systems, greater accuracy in evaluating the damping

ratio can be obtained by considering response peaks, which are several
cycles apart, say N cycles, then

14
 vn − vn+ N
ξ≅ (39)
2mπv n + N

OVER-CRITICAL DAMPING

c
In this case ξ = >1
cc

Equation (19) can be written in the form

s 1,2 = − ξω ± ω ξ 2 − 1 = − ξω ± ωˆ (40)

in which

ωˆ = ω ξ 2 − 1 (41)

substituting the two values of s given by equation (40) into equation (5), i.e.
v(t)=G e st, and simplifying, yield:

v(t) = (A sinh ω̂ t + B cosh ω̂ t) e ( −ξωt ) (42)

in which

A and B can be evaluated using the initial conditions v(0) and v(0) .

The response of an over-critically damped system is similar to the motion of

a critically damped system. The motion is not periodic, and the damping is
large enough to prevent oscillation.

15
EXAMPLE 2:

v(t)
Weight W=mg
p = jacking force

k/2 c k/2

Figure 17: Vibration test of simple building

A one-story building is idealized as a rigid girder supported by weightless

columns. In order to evaluate the dynamic properties of this structure, a
vibration test is made, in which the rigid girder is displaced laterally by a
hydraulic jack and then suddenly released. During jacking, it was observed
that a force of 9,072 N was required to displace the girder 0.508 cm. After
the instantaneous release of this initial displacement, the maximum
displacement on the first return swing is only 0.406 cm, and the period of
this displacement is T=1.40 sec.

From these data, the following dynamic properties are determined:

a. Effective weight of the girder

2π m
T= = 2π = 1.40 sec
ω k

2 2

m = 
1.40   1.40  9072
 k =  × = 868 × 103 kg
 2π   2π  0.00508

16
b. Un-damped frequency of vibration

1 1
f= = = 0.714 cycles/seconds = 0.714 Hertz
T 1.40

ω = 2πf = 4.48 rad/sec

c. Damping Properties

Logarithmic decrement:

vn 0.508
δ = ln =ln = 0.224
v n +1 0.406

δ
Damping ratio ξ ≅ = 3.55%
2π

Damping coefficient

c = ξcc = ξ × 2mω = 0.0355 × 2 × 868 × 10 × 4.48=281.74 N-

3

sec/m

Damped frequency

ω D = ω 1 − ξ 2 = ω 1 − 0.0355 2 = ω 0.999 ≅ ω

d. Amplitudes after 6 cycles

6
v 
6

 v0 = 
0.406 
v6 =  1  × 0.508 = 0.1324 cm
 v0   0.508 

17
6. FORCED VIBRATION - RESPONSE TO HARMONIC LOADING

v(t)
m
p(t)=P o sin ϖt

k/2 c k/2
Po=amplitude of harmonic
load of sine wave form
ϖ =circular frequency of load

Figure 18

Equation of motion is:

mv(t ) + cv(t ) + kv(t ) = Po sin vt (43)

FOR UN-DAMPED SYSTEM  c=0

mv(t ) + kv(t ) = Po sin vt (44)

which has a complimentary solution of free vibration form of equation (11)

v c (t)=A cosωt + B sinωt (45)

Particular solution:

The general solution must also include the particular solution, which
depends on the form of the dynamic loading. For the case of harmonic
loading, the particular solution is:

v p (t)= G sin ϖt (46)

the amplitude G is to be evaluated.

Substitute (46) into (44) gives:

− mϖ 2 G sin ϖt + kG sin ϖt = Po sin ϖt (47)

18
 k
Dividing both sides by sin ϖt and k, and noting that =ω2,
m

Po  1 
G=  2 
(48)
k 1 − β 

In which

β = ratio of the applied loading frequency to the natural free-vibration

frequency

ϖ
β = (49)
ω

General solution

General solution= Complimentary solution + particular solution

v(t) = v c (t) + v p (t)

P  1 
=  ω + ωt + o  sin ϖt (50)
k 1 − β 2 
A cos
 t B sin
 
  

complimentary solution particular solution

A and B depend on the initial conditions with which the response was
initiated, i.e. v(0) and v(0) .

For v(0)=0 and v(0) =0

Po β  1 
A=0 ; B= − (51)
k 1 − β 2 

And equation (50) becomes

Po  1 
v(t ) =  2 
(sin vt − β sin ωt ) (52)
k 1 − β 

in which:

19
Po
= the maximum value of static response of the harmonic force
k
p o sin ϖt , denoted by v st

Po
vs t =
k

1
= the amplification factor or magnification factor, MF, of the
1− β 2
harmonically applied loading

Response Ratio:
Response ratio is ratio of the dynamic displacement response to the
displacement produced by static application of load P o , i.e:

v(t )  1 
R(t) = =  (sin ϖt − β sin ωt ) (53)
v st 1 − β 2 

FOR DAMPED SYSTEM: C≠ 0

The general solution is:

P
ϖt  = e −ξωt ρ cosω D t − θ  + o Rd sinϖt − φ  (54)
  k

transient steady state

where:

Po
= maximum value of the static response of harmonic force po sin ϖt
k
Rd = Dynamic response factor

1
= 2
(55)
 ϖ    ϖ 
2 2

1 −    +  2ξ 
  ω    ω 

20
  ϖ 
 2ξ 
φ = tan − 1 ω  (56)
 ϖ 2 
1 −   
  ω  

= phase angle by which response lags behind load

The constants ρ and θ are determined by using the initial conditions v(0) and
v(0) .

Figure 19: (a) Harmonic force; (b) response of un-damped system to

ϖ P
harmonic force; = 0.2 , v(0)=0 and v(o) = ω o
ω k

21
 ϖ
Figure 20: Response of damped system to harmonic force; = 0.2 , ξ = 0.05
ω
Po
, v(0) = 0 , v(o) = ω
k

The transient response decays exponentially with time. The total response,
as time goes on, approaches the steady-state response. However, the largest
deformation peak may occur before has reached the steady-state.

RESONANT RESPONSE

It apparent from equation (50) that the steady-state response amplitude of an

ϖ
un-damped system tends towards infinity as the frequency ratio β =
ω
approaches unity.

22
 ϖ
ω

Quasi static Resonant Inertia Range

- Essentially the same as Rd >> 1 is sensitive to -The effects of inertia of

static deformation damping mass are dominated
- Controlled by ‘k’ - Deformation is in the
opposite direction of the
force

Figure 21: Deformation response factor and phase angle for a damped
system excited by harmonic force

23
Though the dynamic response factor at ϖ ≅ ω is very large for system with
low damping, a very long time period is needed before the deformation
amplitude gradually reaches the steady state. The amplitude has to build-up
gradually cycle by cycle as shown by the figure below.

t/T

Figure 22: Response of Damped system with ξ = 0.05 to sinusoidal force of

frequency ϖ = ω ; v(0) = v(0) = 0

The number of cycles required to reach 95% of steady state amplitude is:

10 cycles for Rd=10 ( ξ = 0.05 )

24 cycles for Rd=25 ( ξ = 0.02 )
48 cycles for Rd=50 ( ξ = 0.01 )

PRACTICE PROBLEMS:

Q 4. An un-damped SDOF system having mass m and stiffness k is initially

stationary at its equilibrium position. If a constant load P is suddenly applied
to the mass, describe the subsequent motion.

Q 5. An SDOF system has mass m = 10 kg and stiffness k = 25 kN/m.

When the mass is given an initial disturbance and then let to oscillate, it is

24
found that the amplitude of vibrations reduces by 50% each cycle. Calculate
the damping ratio n, the viscous damping coefficient c and damped natural
period T d . How long will it take for the amplitude of vibration to reduce to
1% of its initial value?

Q6. The frame shown in the Figure 23 supports a rotating machine which
exerts a horizontal force at the girder level, P(t) = 1000 sin 4.8t N. The
columns are made of steel (E = 200 GPa ) and 110 mm square sections.
Assuming 5% of critical damping, determine:

(a) steady-state amplitude of vibration and

(b) the maximum dynamic stress in the columns.

Assume the girder is rigid.

P (t)
C
B

5m m = 3000 kg

A D

Figure 23

25
 MULTI DEGREE OF FREEDOM (MDOF) SYSTEM

UN-DAMPED FREE VIBRATION OF MDOF SYSTEM

Infinitely stiff beams

m2 -k 2 1 k2

k2
m1 1 k 1 +k 2 -k 2
m1 m2
k1 k1 k2

Figure 24. Idealization of 2- DOF building frame

Ignore axial deformation, infinitely stiff beams  2 degree of freedom (dof)

system (2 independent lateral translations at floor levels).

Equation of motions

m1v1 + (k1 + k 2 )v1 − k 2 v 2 = 0 (1a)

m2 v2 − k 2 v1 + k 2 v 2 = 0 (1b)

or in matrix form

m1 0   v1  k1 + k 2 − k 2   v1  0

0  +  = (2)
 m2  v2   − k 2 k 2  v 2  0

For n degree of freedom (dof) system

m1 0 k11 k12 ... k1n 

 m2   k 22 ... k 2 n 
[m] =   [k ] = 
 ...   ... ... 
   
0 mn   k nn 

26
 [m]{v} + [k ]{v} = [0] (3)

Pre multiplying by [m]-1

{v} + [m]−1 [k ]{v} = [0]

or

{v} + [D]{v} = [0] (4)

where [D ] = [m]−1 [k ] = Dynamic matrix

Solution for this equation:

vi = Ai sin(ωt + ψ ) , where i = 1,2,….n (5)

A i = amplitude of the deflected shape at n coordinate

vi = ωAi sin(ωt + ψ )

vi = −ω 2 Ai sin(ωt + ψ ) = −ω 2 vi

Thus,

− ω 2 {vi } + [D ]{vi } = [0] (6)

[[D] − ω [I ]]{v } = [0]

2
i (7)

Equation 7 is called EIGEN VALUE or CHARACTERISTIC VALUE

problem. ω 2 is the Eigen values, indicating the square of free-vibration
frequencies. [v i ] express the corresponding shapes of the vibrating system.

Non-trivial solution for equation (7)

[D] − ω 2 [I ] = 0 (8)

27
This determinant gives n degree polynomial in ω 2 . The mode having the
lowest frequency is called the first mode, the next higher frequency is the
second mode etc.

For each ω 2 , an Eigen vector can be found which is relative amplitudes of

the displacements at nodes, called NATURAL MODES. If one amplitude is
set to unity, the amplitude vector is normalized to give the NORMAL
MODE CHARACTERISTIC SHAPE.

Normal Coordinates

In general, any deflection represented by coordinate d can be described by

arbitrary set of compatible mode shape Φ with undetermined amplitude u

D = Φu

u is known as generalized coordinates.

Φ 31 Φ 32 Φ 33

Φ 21 Φ 22 Φ
23

Φ 11 Φ 12 Φ 13

3-dof system Mode 1 Mode 2 Mode 3

Figure 25. Modal shapes and normal coordinates

Φ is formed by Eigen values for n natural mode.

28
EXAMPLE 3:
m 1 =m2
m2 EI constant
L
Determine:
m1 - Natural frequencies
L - Corresponding normal mode characteristic
shape

Figure 26.
m 0  m 0
Mass matrix [m] =  1 = 
 0 m2   0 m

k + k − k 2  EI  48 − 24
Stiffness [k] =  1 2 = 3
 − k2 k 2  
L − 24 24 

EI  48 − 24
Dynamic matrix [D] = [m]-1 [k] =  
mL3 − 24 24 

And so,

EI  48 − 24 2 1 0
− 24 24  − ω 0 1  = 0
mL3    

EI
for simplicity of solution, let = 10
mL3

480 − ω 2 − 240
=0
− 240 240 − ω 2

for λ = ω 2 , gives λ2 − 720λ + 57600 = 0

roots are λ1 = 91.7  ω1 = 9.6 rad/sec

λ 2 = 628.3  ω 2 = 25.1 rad/sec

29
using ω1 2 = 91.7 , substitute to equation (7)

480 − 91.7 − 240   A1  0

=
 − 240
 240 − 91.7   A2  0

Letting A 2 =1, gives solution

 A1  0.618
A  =  1 
 2  

Similarly, for ω 2 2 = 628.3 , substitute to equation (7)

480 − 628.3 − 240   A1  0

=
 − 240
 240 − 628.3  A2  0

Letting A 2 =1, gives solution

 A1  − 1.162
A  =  1 
 2  

This can be shown diagrammatically as follows:

1 1

0.618 -1.162

Mode 1 Mode 2

Figure 27: Normal mode characteristic shapes

30