Dynamics of Structures

I. Introduction
1.1 Dynamic and Dynamics
Dynamic load--- any load of which its magnitude, direction, and/or point of application varies with
the time. If it is applied to an engineering structure, the resulting deflections and stresses
(response) are also time-varying (dynamic).
If the loading is a known function of time, the corresponding analysis of a specified structural
system is called a deterministic analysis. Otherwise, if the time history of the loading is not
completely known (or only known in a statistical sense---random), the structural response analysis
is nondeterministic.
Static, statics

1.2 Investigation of the Dynamical Behaviour of Structures

Design, analysis and testing
Existing Structure or
Design, drawings and data

Design

Analytical model
(assumption, sketches, etc.)
Mathematical Model
(Differential equations)

Analysis

Dynamic Bahviour
(Solutions of DE's)

Testing
Physical Model

Dynamic Testing

No
Meet
requirement

Yes

Stop

Fig.1.1 Steps in a dynamic investigation

1.3 Degrees of Freedom

The number of independent coordinates necessary to specify the configuration or position of a
system at any time is referred to as the number of degrees of freedom (DOF).
The analytical models in analysis stage fall into two basic categories:
discrete-parameter models
continuous models
Figure 1.2 shows these models of a cantilever beam (small deformation). The discrete-parameter
models here are called lumped-mass models because the mass of the system is assumed to be
represented by a small number of point masses, or particles. They are described by finite DOF systems,
while a continuous model represents an infinite DOF system (Fig.1.2c).

x
u1(t)

u2(t)

u(t)

(a) One DOF model

u(x, t)

u3(t)

(b) Three DOF model

(c) Distributed mass model

Fig.1.2 Discrete-parameter and continuous analytical models of a cantilever beam.

However, in cases where the mass of the system is quite uniformly distributed throughout, an
alternative approach to limiting the number of degrees of freedom may be preferable. This procedure is
based on the assumption that the deflected shape of the structure can be expressed as the sum of a
series of specified displacement patterns. The amplitudes of these patterns may be considered to be the
displacement coordinates of the system, and the infinite number of DOF of the actual beam is
represented by the infinite number of terms included in the series. Generally, the expression for the
displacement of any one-dimensional structure might be

u ( x)

Zn

( x)

where n(x) represents a geometric shape function which is compatible with the prescribed geometricsupport conditions and maintains the necessary continuity. Its contribution to the resulting shape of the
structure depends upon the amplitude Zn, which will be referred to as generalised coordinates.

1.4 Formulation of the Equations of Motion

Various types of force
Elastic force

fs

ku

, (u=u2-u1)

f D c u
Viscous force
c is called coefficient of viscous damping
(damping constant) (N s/m)
d
du
(m )
dt
dt

fI

mu

u2

u1

u2

fs

fs

k is called spring constant (N/m)

Inertial force

u1

fD

fD

ma

A force developed by the moving mass is proportional to its acceleration and opposing it---d'Alembert force. Typical inertia forces are shown in Fig.1.3.
may

may

max

max

b
c

I o

m 2
(b
12

c )
2

I o

1
m r 2
2

Fig.1.3 Inertia forces

Direct equilibrium (using d'Alembert's principle)
From the mathematical expressions of Newton's second law of motion

p(t )

d
du
(m )
dt
dt

mu

----suitable for many simple problems

Other methods to form the equation of motion include principle of virtual displacement and
variational approach based on Hamilton's principle (well-known principle of minimum potential
energy in static analysis).

II.

Single-Degree-of-Freedom Systems (SDOF)

2.1 Basic Dynamic System and its Equation of Motion

The essential physical properties of any linearly elastic structural or mechanical system subjected to an
external source of excitation or dynamic loading are its mass, elastic properties and energy-loss
mechanism or damping. In the simplest model of a SDOF system, each of these properties is assumed
to be represented by a single physical element (mass block, spring and dashpot). A sketch of such a
system is shown in Fig. 2.1

u(t)

k
m

u(t)

fs(t)

fI(t)

fD

p(t)

(a) basic components

p(t)

(b) free body diagram

Fig.2.1 SDOF model

(1)
(2)
(3)
(4)

Derivation of the equation of motion by d'Alembert's principle

select free body
apply external and internal loads
add inertia force
consider the equilibrium of the body

fI

fD

fS

p(t )

(2.1)

Substituting expressions of the relevant forces, the final equation of motion is

mu(t ) cu (t ) k u (t )

p(t )

(2.2)

Derivation of the equation of motion by virtual displacement principle

If the mass is given a virtual displacement u compatible with the system's constraints, the total virtual
work done by the equilibrium system of forces must be zero

fI u

fD u

fS u

p(t ) u

(2.3)

Considering u is arbitrary and nonzero, the same equation of motion as Eq.(2.2) can be obtained.
Example 1: Influence of weight (gravitational forces)
Rotating the model of Fig2.1(a) clockwise to a vertical position (90 degree) as shown in Fig.2.2,
the influence of weight of the mass must be considered in equilibrium equation of the system.
ut(t)
u(t)
unstretched
length

p(t)

k/2
Fixed
line

st

m
k/2
c

ur(t)
u(t)

Fig.2.2 Influence of weight

ug(t)

Fig.2.3 Influence of support excitation

mu(t ) cu (t ) k u (t )
Because

st

ur and

st

p(t )

(2.4)

W / k is the static displacement of the mass on the linear spring and

ur is the displacement of the mass measured relative to the static equilibrium position, the equation of
motion after substituting these relation into Eq.(2.4) is simplified to
mur (t ) cu r (t ) k u r (t ) p (t )
(2.5)
Example 2: Influence of support movement
Dynamic stresses and deflections can be induced in a structure not only by a time-varying applied
load, but also by motions of its support points such as the motions of a building caused be
earthquake etc. (Fig.2.3).
Since the inertial force in this case is given by

mut (t ) ,

f I (t )

(2.6)

the equilibrium equation leads

mut (t ) cu(t ) k u(t )

(2.4)
p(t )
Where total displacement ut , the displacement of the structure's base relative to a fixed reference axis
ug and the structure's displacement relative to the support u has a relation of

u t (t ) u (t ) u g (t )

(2.5)

The equation of motion of Eq.(2.4) can more conveniently be written

mu(t ) cu(t ) k u(t )

p(t ) mug (t )

p(t )

peff (t )

(2.6)

2.2 Analysis of SDOF System

The solution of the general equation (2.2) of motion is
u(t)=up(t)+uc(t)
(2.7)
where up(t) is a particular solution related directly to p(t) and uc(t) a complementary solution. To find
up(t) and uc(t), it is convenient to rewrite Eq.(2.2) as
2

u 2

p(t )

(2.8)

where 2=k/m and

is called the (undamped) circular frequency and
called the viscous damping factor and

c
2m
cc is critical damping coefficient. Both
of a SDOF system.

and

c
cc

is a dimensionless quantity

(2.9)

are very important parameters determining the response

2.2.1 Free vibration of undamped SDOF system

The equation of motion

The solution

0
u (0)

u (0) cos t

cos( t

(2.10)

sin

sin( t

(2.11)

where

u 2 (0)

u (0)

, tan

The natural frequency is defined by

u (0) /
u (0)

and

tan

u (0)
u (0) /

(2.12)

, which has a unit of Hz (=1cycle/s). Its reciprocal is the

undamped natural period T=1/f =2 / . The rotating vector representation in complex plane and general
response curve in time domain are shown in Fig.2.4 and Fig.2.5 respectively.

u(t)

I
u(0)

T=2 /

(0)

u(0)

t
t
(0)/

Fig.2.4 The vector representation of displacement

in complex plane

Fig.2.5 General response curve in time domain

2.2.2 Free vibration of damped SDOF system

2
u 2 u
u 0
rt
Assume a solution of the form u Ce , we obtain the characteristic equation
2
r2 2 r
0
r1
2
1
and
r2

The equation of motion

(2.13)
(2.14)
(2.15)

The solution depends on the value of which can be used to distinguish three cases:

>1, i.e., c>cc overdamped case

The solution

e
t

[C1ch

[u(0)ch

1 t

1 t

C 2 sh
u (0)

1 t]

u(0)
2

sh

1 t ] (2.16)

=1, i.e., c=cc, critically-damped case

The solution

(C1 C2 t )

[u(0) (u(0) u(0) )t ]

(2.17)

Example 3: An example for both nonoscillatory responses is seen in Fig. 2.6

u(t) (cm)
u (0) 0
u (0) 20cm / s

1.6

5rad / s

=1
0.8
=2

t (sec)
0.8

1.6

2.4

Fig.2.6 Nonoscillatory response systems

<1, i.e., c<cc underdamped case

r1
r2

(2.18)

and

is called damped circular frequency. The solution of Eq.(2.13) is

[u (0) cos

u (0)

u (0)

sin

t]

cos(

(2.19)

where
2

u (0)

u (0)

u (0)

and

u (0)

tan

u (0)
D u (0)

(2.20)

The free vibration response of undercritically damped (underdamped) system with zero initial velocity
is shown in Fig. 2.7.
u(t)
u (0) 0

e
u ( 0)

u1

u2

t
/

2 /

3 /

4 /

Fig.2.7 Free vibration response of underdamped system

Logarithmic decrement of damping which is defined by

ln

un
un 1

2
1

(2.21)

2.2.3 Response of SDOF system to harmonic loading

Assume a SDOF system is subjected to a load varying harmonically with time, saying sine- (or/and
cosine-) wave form. Without loss generality, p(t) in Eq.(2.2) can be assumed a force having an
amplitude p0 and circular frequency
and

mu(t ) cu(t ) k u(t )

p0 sin t

(2.22)

A complementary solution uc(t) and the particular solution up(t) of above equation are respectively

uc

p0
k (1

up

( A cos

1
2 2

(2

B sin

[(1

t )e

) sin t 2

(2.23)

cos t ]

(2.24)

where = /
is defined as the ratio of the applied loading frequency to the natural vibration
frequency (frequency ratio). The unknown constant A and B should be determined from the initial
condition of the system. For example, if u (0) u (0) 0 , then

p0
k (1

2
2 2
)

(2

and B

p0
k

2
D

(1

(1

2 2

(2

)
)2

(a) Response of undamped system ( =0). From Eqs.(2.23) and (2.24), we have

A cos t

p0 1
k 1

B sin t

sin t

(2.25)

Considering an initial condition of zero displacement and velocity, the above solution becomes

p0 1
k 1

(sin t

sin t )

(2.26)

sin t )

(2.27)

Response ratio

u (t )
p0 / k

R(t )

1
2

(sin t

(b) Steady-state harmonic response of damped system

The response is given by Eq.(2.24) because complementary will decay soon (with no interest) in
damped system. Eq.(2.24) can be written as
(2.28)
u p (t )
sin( t
) .
Where

p0
k

1
2 2

(1

(2

and

tan

2
1

(2.29)

Magnification factor

p0 / k

(1

2 2

(2.30)

)2

(2

Magnification factor D

4
3

2
1

0
0

1
Frequency ratio

Fig.2.8 Variation of magnification factor with frequency ratio

180
150

Phase angle

120
90
60

30
0
0

0.5

1.5

2.5

Frequency ratio

Fig.2.9 Variation of phase angle with frequency ratio

(c) Resonant response

It is apparent that from Fig.2.8 the steady-state response amplitude of an undamped system tends
towards infinity as the frequency ratio approaches unity. For some other lower value of damping, it is
seen that the maximum steady-state response amplitude occurs at a frequency ratio slightly less than
unity. To find the maximum or peak value of dynamic magnification factor, the differentiating result of
Eq.(2.30) with respect to will give

1 2 2
1

peak

Dmax

and

(2.31)
(2.32)
2

Even so, the response resulting from the unity frequency ratio, i.e., the frequency of the applied loading
equals to the undamped natural vibration frequency, is called resonance. Under this condition,

1
2

Dmax

(2.33)

(the difference between Eqs.(2.32) and (2.33) is only 2 percent for =0.2). At the resonant exciting
frequency ( =1), the displacement response becomes

( A cos

B sin

t )e

p0 cos t
k 2

(2.34)

If the system starts from rest (zero velocity and displacement initial conditions), then

1 p0
[(cos
2 k

1
1
u
The response ratio
has an envelope of
.
p0 / k
2

sin

t )e

cos t ]

(2.35)

(d) Vibration isolation

Two types of isolation problems:
Prevent oscillatory forces of the system from support
A rotating unbalance SDOF system is shown in left figure and its
p(t ) p0 sin t
steady-state displacement response is

p0 D
sin( t
k

u p (t )

u(t)
k

The force transfer to the foundation can be estimated on

fS

ku(t )

f=fS+fD

fD

cu (t )

p0 D sin( t
)
cp0 D
cos( t
k

p0 D cos( t

Fig. 2.10
The total maximum force transfer to the foundation is

f max

( f s2

f D2 )1 / 2 | max

) 2 , and

p0 D 1 (2

f max / p 0

D 1 (2

)2

Transmissibility
Isolate harmful support movement from a structural system
The governing equation of relative displacement is
mu(t ) cu (t ) k u(t )
mug (t ) peff (t ) (2.36)
A harmonic support movement leads a steady-state relative
displacement according to Eqs.(2.24) or (2.28) for the Eq.(2.36)

u p (t )

sin( t

p0
D sin( t
k

ut(t)=ug(t)+u(t)
m

ug

u0 sin t

)
Fig. 2.11

(1

u0
2 2
)

(2

The total displacement is

[(1

u t (t )

) sin t

cos t ]

u g (t ) u(t )
(Note: the main contribution to u(t) is only from up(t)).

Similarly

t
max

/ u0

D 1 (2

(2.37)

Transmissibility

4
3
2
1
0
0

0.5

1
1.5
Frequency ratio

2.5

Fig.2.12 Variation of transmissibilitywith frequency ratio

2.3 Response to Impulsive Loading

Property
The load is of short duration
Damping has much less importance in controlling the maximum response of a structure because
the maximum response to a particular impulsive load will be reached in a very short time, before the
damping forces can absorb much energy from the structure.
Two phases: forced and free vibration during and after loading respectively.
Example 4: Response of sine wave impulse
Phase 1
During this phase, the structure is subjected to a single half sine wave loading as shown in the right figure. Assuming
that the system starts from rest, the undamped response
ratio, including the transient as well as the steady-state term,
is given by Eq.(2.27)

u (t )
1
(sin t
2
p0 / k 1
2
where
and
2t1 t1
R(t )

sin t ) (t1 t) (2.38)

p(t)

p0

t
t1
Phase I

t1

T
2t1

t-t1
Phase II

Fig.2.13

Phase 2
The free vibration which occurs during this phase (t t1) depends on the displacement u(t1) and velocity
u (t1 ) existing at the end of Phase I. In terms of the response ratio, it depends on the values of R(t1) and

R (t1 ) . Considering the displacement response of free vibration of undamped system Eq.(2.11), we
have

R(t )

R (t1 )

R(t1 ) cos ( t t1 )

Note that from Eq.(2.38) R(t1 )

R(t )

1
1

[sin

sin
2

cos( t

(t t1 )

(t t1)

(2.39)

(t )
and R
1

sin

) (1 cos ) sin( t

[sin t sin( t

) , then

)]
(2.40)

)]

(1 cos

cos(

) sin( t

) (t

t1 )

The maximum (displacement) response can occur in either Phase I or Phase II depending on the ratio of
load duration t1 to the period of vibration of the structure.
However, if =1, both Eq.(2.38) and Eq.(2.40) are not suitable. Please derive the similar expressions
for the displacement response (in terms of response ratio).

Exercise: Response for rectangular and/or triangular impulses

p(t)

Response of undamped system to a very short duration impulse

Let the undamped system be at rest for 0 t. The equation of
motion and initial conditions are

p(t )

mu(t ) ku(t )

0
u (0)

and

0 t

t1<<T

t1

(2.41)

t t1
u (0) 0

t1
(2.42)

Fig.2.14

Integrating Eq.(2.41) with respect to time and incorporating the initial conditions, we have
t1

mu (t1 )
Hence u (t1 )

u (t t1 )

1
m

u (t1 )

t1
0

[ p(t ) ku(t )]dt

p (t )dt and u (t1 )

t1
0

p(t )dt

( t1

0)

0 when t1 0. The impulse response is

t1

p(t )dt

sin (t t1 ) u (t1 ) cos (t t1 )

sin (t t1 ) (t

t1 ) (2.43)

It can be proved that the impulse response in damped system is

t1

u (t t1 )
If

t1
0

p (t )dt

p(t )dt

( t t1 )

sin

(t t1 )

(t

t1 )

(2.44)

1, Eqs.(2.43) and (2.44) are called unit impulse response functions which are shown

respectively for undamped and damped systems as

h(t t1 )

h(t t1 )

1
sin (t t1 ) (t t1 )
m
1
e (t t1 ) sin D (t t1 )
m D

(2.45)

(t

t1 )

(2.46)

10

2.4 Response to Arbitrary Loading*

General arbitrary loading, Duhamel integral (Time-domain)
For an arbitrary loading as shown in Fig.2.15. it can be considered as a combination of numerous
(infinite in fact) successive unit impulses (loading). At time t= , the loading acting during the interval
of time d represents a very short-duration impulse p( )d on the system, which leads a displacement
response increment effected from that time t- 0 .
Hence the response increment due to p( )d can be
p(t)
calculated based on either Eq.(2.43) or Eq.(2.45)
for undamped system

p( )d
sin (t
m

du (t )

) (t

p( )

) (2.47)

t
d

For linearly elastic system, the total response can

be obtained by summing all these differential
response increment developed during the load
history, that is a Duhamel integral

u (t )

1
m

t
0

p( ) sin (t

)d

(t

du(t)
t-

0) (2.48)

t-

It is seen that Eq.(2.48) suggests the loading was

initiated at t=0 and the system was at rest at that
time. If the initial conditions are not zeros, the
additional free vibration response must be added to
this solution.

Response increment

Fig.2.15

u (t )

u (0) cos t

u (0)

1
m

sin t

p( ) sin (t

)d

(t

0)

(2.49)

Should the nonzero initial conditions be produced by loading p(t) for t<0, the response can also be
found through Eq.(2.48) by changing the lower limit of the integral (from zero to minus infinity).
Similarly, the under critically damped response can be derived

u (t )

t
0

(t

p ( )e

sin

(t

)dt

(2.50)

Arbitrary periodic loading, Fourier series expansion

Any periodic loading can be express as a series of harmonic loading via Fourier expansion

p(t )

a0

a n cos

n 1

in which

a0

1
T

T
0

p (t )dt

(2.51)

2
and T is the period of the loading p(t).
T
2 T
2
an
p (t ) cos n t dt and bn
T 0
T

Undamped steady-state response (superposition)

u (t )

bn sin
n 1

1
a0
k

1
n 1

2
n

(an cos

T
0

p (t ) sin

t dt (2.52)

/ )

t bn sin

t)

(2.53)

Damped steady-state response (superposition)

u(t )

1
a0
k

a n (1
n 1

2
n

) 2

b cos

n n

(1

2 2
n

bn (1
(2

2
n
2

) 2

a n sin

t)

(2.54)

11

References
[1] R. W. Clough and J Penzien. Dynamics of Structures. McGraw-Hill Inc., 2nd edition, 1993,
Singapore.
[2] M. Paz. Structural Dynamics--theory and computation. Van Nostrand Reinhold Company Inc., 2nd
edition, 1985, New York.
[3] R. R. Craig. Structural Dynamics--an introduction to computer methods. John Wiley & Sons, 1981,
Canada.
[4] G. Vertes. Structural Dynamics. Developments in civil engineering; V.11. Elsevier, 1985,
Amsterdam.
[5] Wu Z. J., Han F. and Wu H. J. Elasticity, University Press of Beijing Institute of Technology, 2010,
ISBN 978-7-5640-3267-8. Beijing.

12