CE 6102: STRUCTURAL DYNAMICS

Lecture Notes-2: Theory of Vibration-

Free Vibration of SDF Systems (Version 2.1)

Dr. Sukumar Baishya

Professor
Civil Engineering Department
NERIST
July 27, 2019

Contents
1 Preamble 2

2 Formulation of Equation of Motion 2

2.1 d’Alembert’s Principle of Direct Equilibration . . . . . . . . . . . . . . . 2
2.2 Principle of Virtual Work . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.3 Example of formulation of Equation of Motion . . . . . . . . . . . . . . 3
2.3.1 Application of d’Alembert’s Principle . . . . . . . . . . . . . . . 3
2.3.2 Application of Virtual Work Principle . . . . . . . . . . . . . . . 4

3 Types of Vibration 4
3.1 Free Vibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.2 Forced Vibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.2.1 Steady-state and Transient Vibration . . . . . . . . . . . . . . . 5

4 Analysis of Undamped Free Vibration 6

4.1 Physics of Undamped Free Vibration . . . . . . . . . . . . . . . . . . . . 6
4.2 Mathematics of Undamped Free Vibration . . . . . . . . . . . . . . . . . 9
4.3 Vectorial Representation of Harmonic Motion . . . . . . . . . . . . . . . 10
4.3.1 Analysis of Undamped Free Vibration using Vectors . . . . . . . 11

5 Analysis of Damped Free Vibration 11

5.1 Critically Damped System (n = ω) . . . . . . . . . . . . . . . . . . . . . 12
5.1.1 Damping Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
5.2 Over-damped System (n > ω) . . . . . . . . . . . . . . . . . . . . . . . . 13
5.3 Under-damped System (n < ω) . . . . . . . . . . . . . . . . . . . . . . . 13
5.3.1 Damped Frequency and Time Period . . . . . . . . . . . . . . . . 15
5.3.2 Logarithmic Decrement . . . . . . . . . . . . . . . . . . . . . . . 15

6 Problems 16
1 Preamble
In Lecture 1, we have studied the types of dynamic loads, types of dynamic analysis,
Degrees of Freedom and Methods of Structural Idealisation.
In this lecture basic theory of vibration of sdof systems will be studied. We would
first discuss the methods of formulation of equation of motion which would be followed
by classification of vibration and solution of equation of motion. Physical concepts
involved would be highlighted. This will be followed by solution of equation of motion,
for various vibration types, using direct mathematical approach of solution of differential
equations.

2 Formulation of Equation of Motion

In deterministic analysis, the basic objective is to find the deformation-time history
of the structure. In most of the practical cases an approximate analysis, involving
only a limited numbers of dof, gives sufficient accuracy. So, the problem is reduced to
determination of time history of selected displacement components.
In order to determine the time history of the relevant dofs of the idealised structure,
it is important to correctly formulate the equation of motion. The following sub-sections
discuss the methods of formulation of equation of motion.

2.1 d’Alembert’s Principle of Direct Equilibration

Since acceleration is involved in any dynamic system, therefore, its state of motion
involves Newton’s 2nd law. The law states that net force equals the rate of change of
momentum of any mass m. The rate of change of momentum is given by

d2 x
 
d d dx
(mv) = m = m 2 = mẍ
dt dt dt dt

Therefore, the net dynamic force is given by

Fnet = mẍ ⇒ Fnet − mẍ = 0 (1)

Equation (1) indicates that the inertia force is equal and opposite to the net force. In
other words, the inertia force resists the action of the net dynamic force.
The concept that the mass develops an inertia force due to its acceleration which
opposes the acceleration itself is known as d’Alembert’s Principle See sec 3.2 of
lecture 1 for details of the physical concepts involved.

2.2 Principle of Virtual Work

Use of d’Alembert’s principle is advantageous for small and simple systems. But when
the system is reasonably complex, involving numbers of interconnected mass points,
direct equilibration becomes difficult. In such cases, Principle of virtual work is advan-
tageous.
The principle states that if a system is in equilibrium under the action of a number
of forces, then, if the system is given a virtual displacement (which is compatible with
system constraints) the total work done by these forces will be zero. The process consists
of

2
 F(t)
(a)
δx

(b)
F(t) Fs
FI Fd

Figure 1: SDOF System (a) Spring-Mass-Dashpot System (b) Free body diagram

• identify the forces acting on the system including d’Alembert’s principle,

• introduce virtual displacement corresponding to each dof of the system,
• find algebraic sum of the works done by each force over the virtual displacement,
• equate the sum to zero.
Virtual works, being scalar quantities, can be added algebraically. The forces, on
the other hand, are vector quantities and can only be added vectorially.

2.3 Example of formulation of Equation of Motion

2.3.1 Application of d’Alembert’s Principle
Consider an sdf system consisting of a spring-mass-dashpot assembly as shown in Figure
1.
The forces acting on the block are i) external dynamic load F (t), ii) restoring force
Fs , and, iii) damping force Fd . The net force on the block is given by

F (t) − Fs − Fd

. Now as per d’Alembert’s Principle, the inertia force Fi would be equal and opposite
to this net force. Therefore,
F (t) − Fs − Fd = Fi (2)
where
Fi = mẍ (3)

3
The restoring force or spring force is given by

Fs = kx (4)

where, k is spring stiffness. The damping force is a function of velocity of the mass. For
viscous damping 1 this is taken as directly proportional to velocity, i.e.,

Fd α ẋ ⇒ Fd = cẋ (5)

where, c is constant of proportionality known as damping coefficient. Thus, the equilib-

rium equation can be written as

F (t) − Fs − Fd = Fi ⇒ F (t) − kx − cẋ = mẍ ⇒ mẍ + cẋ + kx = F (t) (6)

Equation (6) is the equation of motion of an sdof system subjected to dynamic load
F (t).

2.3.2 Application of Virtual Work Principle

Let us impart a virtual displacement δx to the mass of Figure 1 in the direction of
external dynamic load. Since the system is in equilibrium, the total work done by the
forces on the virtual displacement must be zero. The virtual work done by the forces
acting on the system are
• by dynamic load = F (t)δx
• by restoring spring force = −Fs δx
• by damping force = −Fd δx
• by inertia force = −Fi δx
The -ve sign in the above indicates that virtual work is opposite to the direction of
virtual displacement. Now, adding the total works done one gets

− (Fs + Fd + Fi ) δx + F (t)δx = 0 ⇒ Fi + Fs + Fd = F (t).

Using the values of spring, damping and inertia forces from equations (3), (4) and (5)
the above equation leads to mẍ + cẋ + kx = F (t), which is nothing but the equation
(6).
The use of virtual work method is trivial in the above example, which is included
only for illustration purpose. It is apparent that the d’Alembert’s method of direct
equilibration is more advantageous in such simple systems. However, in complicated
systems with large dofs, virtual work method is a better choice.

3 Types of Vibration
Depending on the presence or absence of external dynamic load during oscillation, vi-
brations may be classified as

• Free vibration, and

1 There are other types of damping as well. One such damping type, sparingly used in modelling, is

Coulombic Damping

4
 VIBRATIONS

Free [F(t)=0] Forced [F(t)~=0 ]

Damped (c~=0) Undamped(c=0) Damped Undamped

Figure 2: Classification of Vibration

• Forced vibration

If the damping in a dynamic system is very small and can be neglected, the damping
coefficient c is equated to zero. Such vibration is called undamped vibration. Accord-
ingly, both free and forced vibration may be either damped or undamped.

3.1 Free Vibration

When vibration is initiated by an external force that is immediately withdrawn, vibra-
tion continues in absence of any external force or excitation. In such condition, the
vibration continues by virtue of natural, system dependent forces. This is called Free
vibration. The equation of motion of sdf system undergoing free vibration may be ob-
tained by replacing the right hand side of equation (6) by zero. Thus, equation of free
vibration is given by
mẍ + cẋ + kx = 0 (7)
If the vibration is undamped then its equation of motion will modify as

mẍ + kx = 0 (8)

3.2 Forced Vibration

When the external dynamic excitation continues to act on the system during vibration,
it is called Forced Vibration. Like free vibration, forced vibration may also be either
damped or undamped. The corresponding equations of motion are equation (6) and

mẍ + kx = F (t) (9)

3.2.1 Steady-state and Transient Vibration

Forced vibrations can be further classified depending on the kind of exciting force. If the
exciting force is a periodic force, having constant frequency and amplitude 2 , then the
vibration is called Steady state vibration. If the exciting force changes its frequency and
2 Details of these terms would be discussed in 4.1 and 4.2

5
 x
k
(a) m Fi Fs

KE
Ry B(0,1)

C(−1,0) A(1,0)

PE
Rx

D(0,−1)
(b)

Figure 3: Undamped Free Vibration a) Spring-mass system and Free body diagram b)
Plot of KE versus PE of the system

amplitude then vibration is called Transient. Examples of steady state forced vibrations
are those due to machinery. Vibration due blast, earthquake or winds, etc. are examples
of transient vibrations.

4 Analysis of Undamped Free Vibration

The solution of equation of motion in undamped free vibration (equation (8) is dis-
cussed. At first the physical significance is discussed in section 4.1, followed by direct
mathematical treatment in section 4.2

4.1 Physics of Undamped Free Vibration

The dynamic system and the free body diagram of the mass is shown in figure 3. To
have the physical feel of the problem, the following discussion is presented.
In absence of any energy absorbing mechanism, the total energy of the system is
constant. Therefore, KE + P E = Constant. Now

KE = 1/2mẋ2

and
P E = 1/2Fs x = 1/2kx2 .
Therefore,
1 1
mẋ2 + kx2 = Constant (10)
2 2

6
This conservation of energy is satisfied at any point of time, including the instant of
starting of vibration (t = 0). Assuming initial displacement at t = 0 as x = xo and
initial velocity as ẋ = 0, the total energy at t = 0 is given by
1 2
(T E)t=0 = kx
2 o
Therefore,
1 1 1
mẋ2 + kx2 = kx2o
2 2 2
or,
m 2
ẋ + x2 = x2o
k
or,
 2  2
m ẋ x
+ =1
k xo xo
or,
!2  2
ẋ x
p + =1 (11)
k/m xo xo
Equation (11) is an equation of circle having unit radius and centered at origin. The
first term of this equation is proportional to KE and the second term is proportional
to PE of the system. Figure 3 b) shows this circle where the first term is plotted as Ry
and second term as Rx . The following points may be observed from this figure:
• At point A(1,0), the PE is maximum and KE is minimum (zero)
• The mass starts moving towards mean position of rest (corresponding to point B
in fig 3). PE reduces with corresponding increase in KE.
• At B (0,1), all the PE converts to KE; PE = 0; KE is maximum. The spring is
undeformed, but mass has its highest velocity.
• Beyond point B, spring is compressed and KE is converted into PE. This continues
up to point C(-1,0), where PE is again maximum but the velocity is in opposite
direction.
• From C to D(0,-1), the mass moves in +ve direction and the spring starts extending
again.
• From D to A, the KE again converts to PE and mass moves to extend the spring
to the maximum limit at A. Thus the cycle starts again.
The representative unit of time for describing the behaviour of the system is the
time required to cover the loop ABCDA. This is called the Time Period of the system
and is denoted by T .
Thus, the motion is repeated after time T which has a period of 2π. Therefore, the
displacement x of the system can be expressed by either sin or cosine functions, or by
their sum. Thus, solution of equation (8) can be written as

x(t) = A sin ωt + B cos ωt

where,
2π
ω= (12)
T

7
The constants A and B are to determined from the known initial conditions. The initial
conditions are
(ẋ)t=0 = 0
and
(x)t=0 = xo .
Therefore, A = 0 and B = xo . Using the values of A and B, the solution of equation of
motion can be written as

x = xo cos ωt ⇒ ẋ = −xo ω sin ωt ⇒ ẍ = −xo ω 2 cos ωt (13)

Using (13) in (11) one gets

 2
 2
 −xo ωqsin ωt  + xo cos ωt
=1
xo m k xo

or,
ω2
sin2 ωt + cos2 ωt = 1
k/m
or,
ω2
k
=1
m
or, r
k
ω= (14)
m
The quantity ω is called Natural Frequency or Circular Frequency of the sdf system; its
unit is radian/sec (written as rad/s). Using the value of ω from equation (12) in (14),
r
2π k
=
T m

or,
2π
T = q (15)
k
m

Equations (14) and (15) give the values of natural frequency and Natural Time Period
of the sdf system. The inverse of time period is called Frequency of vibration of the
system and is denoted by f . Thus,
r
1 1 k
f= = (16)
T 2π m
Unit of f is cycles/sec or Hertz (also written as c/s or cps and Hz, respectively).

In an sdf system, calculation of Natural frequency and Time period is the essential
part of analysis.

8
4.2 Mathematics of Undamped Free Vibration
In section 4.1, the solution of equation (8) is obtained emphasizing the physics of un-
damped free vibration. The approach is lengthy however. In practice, for solving
equation (8) a more rigorous but direct approach is used as discussed below. Equation
(8) is repeated here for convenience.

mẍ + kx = 0

or,
k
ẍ + x=0
m
Using equation (14) in above,
ẍ + ω 2 x = 0 (17)
The characteristic equation of (17) is given by

λ2 + ω 2 = 0 (18)

Solving for λ, we get,

λ = ±iω
Therefore, the displacement of the system is given by

x = A1 exp (iωt) + A2 exp (−iωt) (19)

Using Euler’s equation, exp (±iωt) = cos ωt ± i sin ωt, in equation (19), x can be ex-
pressed as
x = C1 cos ωt + C2 sin ωt (20)
where, C1 = (A1 + A2 ) and C2 = i (A1 − A2 ). The values of C1 and C2 can be obtained
from initial conditions of the system.

Evaluation of Integration constants

Let the initial conditions at time t = 0 be (x)t=0 = xo . and (ẋ)t=0 = ẋo . Using these in
equation (20) one gets C1 = xo and ẋo = C2 ω ⇒ C2 = ẋωo . Thus, solution of equation
of motion is obtained as
ẋo
x = xo cos ωt + sin ωt (21)
ω
Equation (21) shows that the displacement of the system is expressed as sum of sine
and cosine functions with natural frequency as one of the arguments of the functions.
Equation (21) can also be expressed as a sine function as follows. Let xo = R sin φ
and ẋωo = R cos φ. Therefore,
s  2
ẋo
R = x2o + (22)
ω
and  
xo ω
φ = tan−1 (23)
ẋo
Using the above equations x can be expressed as

x = R sin (ωt + φ) (24)

9
 C B

B
A

|OA|= A
A C |OB|= ωA
φ
ωt 2
|OC|=ω A
O
(a) (b)

Figure 4: Vector Representation of Simple Harmonic Motion (SHM) a) Resultant of

two SHM b) Magnitude and Phase relations among Displacement (OA), Velocity (OB)
and Acceleration (OC) Vectors in SHM

The quantity R, given by Eq (22) is known as the Amplitude of the motion. Amplitude
of any dynamic quantity may be defined as the peak value it attains during a cycle of
motion. 3 The angle φ is known as Phase Angle or Phase Difference of the displacement
x. Since this is +ve, it is also termed as Phase Lead. Similarly a -ve phase angle is
called Phase Lag.2

4.3 Vectorial Representation of Harmonic Motion

Let us consider the displacement given by equation (21). This may be written as
ẋo
x = xo sin (ωt + π/2) + sin ωt (25)
ω
The 2nd term of this equation can be represented by a vector of length ẋωo , rotating
CCW with a circular velocity ω, making an angle ωt with the reference axis at any
instant t. The projection of this vector on y axis gives the 2nd term of eqn (21), as
shown in Fig 4.
Similarly, the first term of Eqn. (21) can be represented by another vector having
modulus xo , rotating with the same circular frequency, but with a phase difference of
π/2 with the previous vector. Their vector sum R gives the resultant motion which
has a magnitude as given by equation (22) and phase difference φ as given by eqn.
(23). So, the resultant displacement is given by the sine component of R, which is
already indicated by equation (24). Thus, any simple harmonic motion (SHM) cn be
represented by a vector rotating CCW (or CW, depending on sign convention) with a
constant circular velocity ω.
As an example, let us consider an SHM given by x = A sin ωt. The velocity and
acceleration are given by
 π
ẋ = ωA cos ωt = ωA sin ωt + (26)
2

ẍ = −ω 2 A sin ωt = ω 2 A sin (ωt + π) (27)

3 Strictly speaking, Amplitude is defined only for a harmonic quantity. For transient case, instead of

amplitude, the peak of the dynamic quantity during the complete transient motion might be of interest.

10
 Acceleration
2
ωA

Restoring force
kA
O 01
0
1 O
Displacement
A

Motion Vectors
Inertia Force
2
mω A

Force Vectors

Figure 5: Vectorial Analysis of Undamped Free Vibration

Above equations indicate that velocity vector leads displacement vector by an angle π/2
while acceleration vector leads displacement vector by angle π. Also the amplitudes are
increased by the factors ω and ω 2 respectively. Thus, all these SH quantities can be
represented by vectors as shown in Fig 4 (b).

4.3.1 Analysis of Undamped Free Vibration using Vectors

Using vectorial representation of harmonic dynamic quantities, the analysis of vibration
can be achieved more easily. Consider the sdf system shown in 3(a). Using equations
(26) and (27) the motion vectors and the force vectors of the system can be drawn as
shown in Fig 5. For equilibrium, the force vector diagram gives
r
2 2 k
mω A = kA ⇒ ω = k/m ⇒ ω =
m
Thus, the expression of natural freqency is obtained more easily.

5 Analysis of Damped Free Vibration

We know that the equation of motion of damped free vibration is given by eqn. (6),
with its rhs replaced by zero, i.e.,

mẍ + cẋ + kx = 0. (28)

The auxiliary equation is given by

mλ2 + cλ + k = 0

Its solution can be obtained as

s
−c c2 4mk
λ= ± −
2m (2m)2 (2m)2

11
or, r
−c c 2 k
λ= ± −
2m 2m m
Let us define
c/2m = n (29)
Using eqn (29) and (14), the solutions become
p
λ1,2 = −n ± n2 − ω 2 (30)
Therefore, general solution of (28) may be obtained as
x = A1 exp (λ1 t) + A2 exp (λ2 t) (31)
where λ1,2 are given by eqn (30)
Now, from eqn (30), three cases may arise:
• n = ω; the quantity under radical is zero. The system is called Critically-
damped
• n > ω; the quantity under radical is +ve. The system is called Over-damped
• n < ω; the quantity under radical is -ve. The system is called Under-damped
They are analysed in the following sections.

5.1 Critically Damped System (n = ω)

In this case the system is said to be subjected to Critical Damping and the coefficient
of damping is denoted by cc . For critically damped system, we have
n2 = ω 2
. Using eqn (14) and (29) in the above,
 c 2
c k √
= ⇒ c2c = 4km ⇒ cc = 2 km (32)
2m m
Also  
cc k 2k
= ω ⇒ cc = 2mω = 2 ω= (33)
2m ω2 ω
Combining the last two equations, the critical damping coefficient is obtained as
√ k
cc = 2 k m = 2mω = 2 (34)
ω

5.1.1 Damping Ratio

It is advantageous to express other values of damping coefficient as a fraction of critical
damping coefficient. This ratio is termed as Damping ratio, denoted either by ζ or ξ
and is usually expressed as a percentage. Thus,
c
ζ=ξ= × 100% (35)
cc
When a system vibrates in critical damping, its motion is non-oscillatory. It crosses
the mean-position-of-rest once and then comes back slowly, without any further oscil-
lation. This is shown in Figure 6.

12
 x(t) x(t)

t t

(a) (b)

Figure 6: Deformation Pattern in Damped Free Vibration (a) Critically damped system
(b) Over-damped system

5.2 Over-damped System (n > ω)

In over-damped system, the damping coefficient is larger that critical damping coefficient
i.e., n > ω and c > cc . The roots of equation (30) are real and -ve. The displacement
of the system is such that (See fig 6) when it is disturbed from position of equilibrium,
it creeps back to its original position slowly.

5.3 Under-damped System (n < ω)

Most of the real structures vibrates as an underdamped system. In this case, the roots
of equation (30) are complex. They can be expressed as
p
λ1,2 = −n ± i ω 2 − n2 = −n ± iωD (36)
√
where, ωD = ω 2 − n2 is called circular Frequency of Damped Free Vibration.
Using the value of ωD in equation (31), displacement is given by
x = A1 e(−n+iωD )t + A2 e(−n−iωD )t = e−nt A1 eiωD t + A2 e−iωD t
 
(37)
Using Euler’s equation, the above equation can be simplified to
x = e−nt [A cos ωD t + B sin ωD t] (38)
The integration constants A and B can be obtained from initial conditions.
Let the initial conditions be (x)t=0 = xo and (ẋ)t=0 = ẋ0 . Using these in equation
(38), the constants are obtained as
ẋo + nxo
A = xo and B = (39)
ωD
Using eqn (39) in eqn (38)
 
ẋo + nxo
x=e −nt
xo cos ωD t + sin ωD t (40)
ωD
This is the solution of equation of motion for underdamped free vibration of an sdf
system. The plot of this equation is as shown in fig 7. It is seen that the amplitude of
displacement decays exponentially.

13
 6.0
5.0
4.0
3.0
2.0
1.0
0.0
x(t)

−1.0
−2.0
−3.0
−4.0
−5.0
−6.0
−7.0

0 5 10 15 20 25 30
t

Figure 7: Displacement in Damped Free Oscillation (ω = 0.7071 rad/s, ωD = 0.7062

rad/s, ξ = 0.05, xo = 1.5 m, ẋo = −5 m/s)

14
5.3.1 Damped Frequency and Time Period
From eqn (36) we have,
r  n 2
p
ωD = ω 2 − n2 = ω 1−
ω
Now, by definition of n,
 2
 n 2  c 2 m c2 c 2
= × = = = (ξ)
ω 2m k 4km cc
Therefore, p
ωD = ω 1 − ξ2 (41)
Similarly, damped time period is given by
2π 2π
TD = = p (42)
ωD ω 1 − ξ2

Practical Significance of Damped Natural Frequency

For all practical purposes, the damping ratio lies between 2 - 20%. Using ξ = 20%
damped natural frequency = ωD = 0.98ω ≈ ω
Thus, for all practical purposes, the damped natural frequency can be approximated
to natural frequency; so also in case of damped time period. However, it is clear from fig
7 that the same is not at all true in case of amplitude of vibration.

5.3.2 Logarithmic Decrement

We have seen that amplitude of vibration decreases between consecutive cycles due to
damping. This decrease can be used as a measure of damping present in the system.

Logarithmic decrement is defined as the natural logarithm of the ratio of

amplitudes in two consecutive cycles, in damped free vibration. Let the amplitudes
of vibration in consecutive cycles be x1 , x2 , x3 , .... The logarithmic decrement δ is defined
as
x1 x2
δ = ln = ln ... (43)
x2 x3
As discussed in 4.2, eqn (40) can be expressed as

x(t) = Re−nt cos (ωD t − φ) (44)

where, s  2
ẋo + nxo
R= x2o + , (45)
ωD
and
(ẋo + nxo )
φ = tan−1 (46)
xo ωD
Now, let x1 and x2 be two consecutive amplitudes at times t1 and (t1 + TD ) respec-
tively. Using eqn (44)
x1 = R e−nt1 and

15
 x2 = R e−(nt1 +TD )
Therefore, logarithmic decrement is obtained as

x1 e−nt1
δ = ln = ln −n(t +T ) = ln e−n(t1 −t1 −TD ) ⇒ δ = ln enTD = nTD (47)
x2 e 1 D

Using n = ξω, the above eqn yields

δ = ξωTD

Using value of TD from eqn (42),

2πξ
δ= p (48)
1 − ξ2

For small damping ratio, this can be approximated as

δ ≈ 2πξ (49)

6 Problems
Refer to relevant problems solved and/discussed in Tutorial Classes.

16