Contents

2 Free Vibration of Single Degree of Freedom Systems 37

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.2 Free Vibration of an Undamped Translational System . . . . . . . . . . . . . . . . . . . 39
2.2.1 Equation of motion using Newton’s second law . . . . . . . . . . . . . . . . . . 39
2.2.2 Equation of motion using other methods . . . . . . . . . . . . . . . . . . . . . . 41
2.2.3 Equation of motion of spring-mass system in vertical position . . . . . . . . . . 42
2.2.4 Equation of motion for an undamped system in the vertical position . . . . . . . 43
2.2.5 System response for an undamped translating system . . . . . . . . . . . . . . . 43
2.3 Free Vibration of an Undamped Torsional System . . . . . . . . . . . . . . . . . . . . . 45
2.3.1 Equation of motion for an undamped torsional system . . . . . . . . . . . . . . 45
2.3.2 System response for an undamped torsional system . . . . . . . . . . . . . . . . 46
2.4 Examples of Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.5 Determining natural frequencies by Rayleigh’s Energy Method . . . . . . . . . . . . . . 51
2.6 Free Vibration with Viscous Damping . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.6.1 Logarithmic decrement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.6.2 Torsional systems with viscous damping. . . . . . . . . . . . . . . . . . . . . . 57
2.7 Stability Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

35
Outline of Chapter 2

Key Concepts:

• Free vibration of undamped translational system

• Newton’s law, conservation of energy method

• Free vibration of undamped torsional system

• Rayleigh’s energy method

• Free vibration with viscous damping

• Logarithmic decrement

• Stability of systems

Textbook Reference:

• Rao, Chapter 2

• Excluding Sections 2.7 to 2.10 (Read only)

Tutorial questions:

The following questions from Rao (Fifth Edition) may be completed as soon as the relevant section has
been covered in the lectures:

1. Review questions 2.1 (up to 20), 2.2, 2.3, 2.4, 2.5 (relevant Q’s)

2. Undamped Translational Systems – Problems 2.1–2.6, 2.8–2.15, 2.19, 2.24, 2.26 – 2.27, 2.31–
2.32, 2.39–2.40, 2.45–2.48, 2.51–2.54

3. Undamped Torsional Systems – Problems 2.66–2.68, 2.70–2.78(a)

4. Energy Method – Problems 2.87, 2.89–2.92, 2.96

5. Systems with viscous damping – Problems 2.98, 2.102–2.105, 2.108–2.110, 2.126, 2.128

6. Stability of systems – Problem 2.156

7. Matlab (for practice) – Problems 2.157–2.159

R EV 5 J ULY 2017

36
Chapter 2

Free Vibration of Single Degree of

Freedom Systems

2.1 Introduction

The study of the vibration of one degree of freedom systems is fundamental to the understanding of
more advanced topics in vibration. A system is said to undergo free vibration when it oscillates only
under the action of an initial disturbance, with no external forces acting thereafter. The oscillations of
a pendulum set into motion, the vertical oscillatory motion felt by a bicyclist after hitting a road bump,
and the motion of a child on a swing after an initial push represent a few examples of free vibration.

Figure 2.1: A simple vibrating system [1].

Figure 2.1(a) shows a spring-mass system that represents the simplest vibratory system. It is a single
degree of freedom system since one coordinate is sufficient to specify the position of the mass at any time.
There is no external force applied to the mass; hence the motion resulting from an initial disturbance will
be a free vibration. Since no dissipative element in the from of a damper is present the system is classed
as an undamped system. It is also known as a conservative system. Note: in real systems there will
always be a small amount of friction and air resistance. In this situation, if damping is important, then
it is included in the form of Viscous or Coulomb Damping, else it is neglected in the interests of rapidly
determining the characteristics of the vibrating system. It may later be included in a more detailed
analysis of the system. Figures 2.1(b) and (c) form part of the analysis of the system and will be covered
in more detail later in the chapter.
Several mechanical and structural systems can be idealized as single degree of freedom systems. In many
practical systems, the mass is distributed, but for a simple analysis it can be approximated by a single
equivalent mass. Similarly, the elasticity of the system, which may be distributed throughout the system,
can also be idealized by a simple equivalent spring. Figure 2.2 shows some more examples of systems
that can be modelled as one degree of freedom systems.

37
 x

Drum
m Water Tank

Trailer Body
x
Cable (k, c)
m

Suspension (k, c)
Support (k, c)

m Elevator Car
x

(a) (b) (c)

Figure 2.2: Some typical systems that may be modelled as simple spring mass systems.

One fundamental approach to derive the differential equations of a vibrating system is to sketch a free
body diagram of the system, and apply Newton’s second law. Alternatively, one can add the inertia effects
to the free body and consider D’Alembert’s principle. Quite often the addition of the inertia effects to the
free body helps to visualize the physical effects an accelerating system is experiencing, particularly for
systems in general plane motion. Energy methods can also be used. These are convenient for use with
systems that consist of several component parts, since the differential equation can be obtained without
having to consider the forces acting between the component parts. One restriction on the use of energy
methods is where the system is non-conservative, i.e. contains damping. Since all real systems contain
damping it implies that the energy method represents an idealisation of the system. Where the damping
is small it is usually possible to neglect it and, as a first approximation, consider an undamped system,
for which the energy method is applicable. Later it will be seen that the loss of energy can be accounted
for in another Energy Method, using Lagrange’s Equation.

Vibration analysis involves the following steps:

+ Determine relevant simplifying assumptions to make the problem tractable (e.g. single DOF sy-
stem, undamped, small angles, etc.).

+ Derive the differential equation(s) of motion of the system mass(es) in terms of the displacement
variable that will be used to describe the response of the system.

+ Solve the differential equation(s) of the system to determine the system response (usually in the
form of an equation describing the displacement of the system mass(es), and usually including the
transient and steady state response).

+ Analyse and interpret the response to determine natural frequencies, damping, stability, maximum
displacements, velocities and accelerations etc.

+ Review the assumptions in the light of the system response to determine their validity and the
accuracy of the solution. Revise, if necessary, and re-solve to get a more accurate representation
of the system response (e.g. increase the DOF).

38
It is critical that the differential equations of motion are correctly derived. In the following section the
methods for deriving the differential equations which describe the motion of vibrating systems will be
developed.

2.2 Free Vibration of an Undamped Translational System

2.2.1 Equation of motion using Newton’s second law

Newton’s second law of motion can be used to derive the equation of motion. The following steps outline
a systematic procedure for use in applying Newton’s second law to derive the equation of motion:

1. Select a suitable co-ordinate to describe the displacement of the mass or centroid of a rigid body
in the system. Use a rectilinear co-ordinate to describe the rectilinear motion of a point mass or
the centroid of a rigid body, or an angular co-ordinate to describe the angular motion of the rigid
body.

2. Determine the static equilibrium configuration of the system and measure the displacement of
the mass from its static equilibrium position. Recall that any forces present in springs at the
static equilibrium position balance each other out and do not appear in the resulting differential
equations.

3. Draw a free body diagram of the mass or rigid body when a positive displacement, velocity and
acceleration are given to it from the static equilibrium position. Indicate all the active and reactive
forces acting on the mass or rigid body. In addition to the note in point 2 above about static spring
forces also note whether there are any elastic forces which balance the weight of any masses in
the system at the static equilibrium position (see Section 2.2.3).

4. If necessary, draw the kinetic diagram as well, identifying the resultant translational and rotational
accelerations of the body, mẍ, mÿ, I θ̈ , as appropriate.

5. Apply Newton’s second law.

For a rectilinear system:
+ +
−→ − →
ΣF = mẍ (2.1)
(where force ΣF is assumed to be positive in the positive direction of the initial acceleration),
for the case where the displacement variable is chosen as x.

For a rigid body undergoing rotation:

+ + +
x x x
ΣM = IG θ̈ or JG θ̈ (2.2)

where moment ΣM is assumed to be positive in the positive direction of angular displacement

(shown anti-clockwise in this instance), velocity and acceleration, and the subscript G indicates
the centroid. IG or JG represent the mass moment of inertia about the centroid.
For multiple degree of freedom systems the analysis above is done for each of the system masses.
In Figure 2.3 free body and kinetic diagrams have been drawn for a body experiencing a number
of external forces and moments. The equations given above apply for the case where the sum of
moments relates to the centroid.

39
 F2
M2

G
G

P W P

F3
M1
F1
Free body diagram Kinetic diagram

Figure 2.3: Free body and kinetic diagrams of a rigid body.

F2
M2

G
G
d
P W P

F3
M1
F1
Free body diagram Kinetic diagram

Figure 2.4: Free body and kinetic diagrams of a rigid body with regard to remote point P.

In the case where the sum of moments are taken with regard to another point P (see Figure 2.4)
then the equations must be modified to include the inertial moments as follows:
+ + +
x x x
ΣMP = IG θ̈ + mẍd (2.3)

Alternately, if the acceleration vector ẍ has components, then the moment of these components
about point P can be taken into account using the following alternate form of the equation:

+ + + +
x x x x
ΣM = Σ(Mk )P = IG θ̈ + Σmad (2.4)
where multiple inertial moment terms are represented in the Σmad term.
It is critical that the sign convention is adhered to in the analysis, hence it is always written into the
terms of the equations in Newton’s Law. The sign convention depends on the particular problem
and follows the direction of positive displacement variables chosen in the particular problem.

40
The procedure is now applied to the undamped single degree of freedom system in Figure 2.1. Here the
mass is supported on frictionless rollers and can translate in the horizontal direction. If the mass is distur-
bed such that, at an instant later, it has a positive displacement +x to the right (from the static equilibrium
position), a reaction force kx is developed in the spring, acting to the left [recall that springs always op-
pose applied displacements, regardless of any static spring forces which may exist in the springs at static
equilibrium]. The free body diagram is shown in (c) and the application of Newton’s law yields:
+ +
−→ −
→
ΣF = mẍ
−kx = mẍ
mẍ + kx = 0

2.2.2 Equation of motion using other methods

Equations of motion can be derived using other methods as well, notably D’Alembert’s Principle, Prin-
ciple of Virtual Work and Energy Methods. The reader is referred to the reference for details on the first
two of these methods. Various analytical approaches use the energy method for conservative systems,
and may also be extended to non-conservative systems using Lagrange’s Equation. The simplest Energy
method is discussed here briefly, while Lagrange’s Equation will be dealt with in the chapter on two
degree of freedom systems.

Conservation of Energy Method

Equations of motion of a conservative system can also be obtained by using the principle of conservation
of energy.
In a conservative system there is no energy dissipation1 ; the total mechanical energy is constant and its
time derivative must be zero. This can be expressed as

T +U = constant

or
d
(T +U) = 0 (2.5)
dt
where T is the total system kinetic energy and U is the total change in potential energy from the static
equilibrium position. Considering the spring-mass system shown in Figure 2.1, a displacement x of the
mass results in deformation of the spring and hence stored potential energy of:
1
U = kx2
2
The velocity of the mass is the time derivative of the displacement, ẋ, and hence the kinetic energy is
1
T = mẋ2
2
Thus,
1 1
T + U = mẋ2 + kx2
2 2
Recall from mathematics that, for a function A = A(ẋ, x), where ẋ and x are functions of t that

dA ∂ A d ẋ ∂ A dx
= +
dt ∂ ẋ dt ∂ x dt
1 There is no real system that is non-conservative, but very often systems have low damping and it is useful to consider
them to be undamped in order to get an understanding of the upper limit of the system response.

41
Therefore,
d
(T +U) = mẋẍ + kxẋ = 0
dt
Now ẋ may be eliminated as it cannot be zero (else the solution is trivial), and a positive or negative
displacement or velocity does not affect the solution in this case. The governing differential equation is
thus:
mẍ + kx = 0
Note that the Energy Method did not require a free body diagram to be drawn.

2.2.3 Equation of motion of spring-mass system in vertical position

The issue of gravitational potential energy has not been yet dealt with, and its effect is now reviewed2 .
Consider the spring-mass system shown in Figure 2.5(a). A body of mass m is about to be suspended
from a spring of stiffness k, which which has a particular free spring length and zero residual compressive
force. (Note the small gap between the spring and the mass in (a)). When the body is attached to the
spring and slowly lowered to the static equilibrium position, it results in a static spring deflection δs .

Spring free length

k
kδ s
m Static deflection δs k (δ s + x )
Static equilibrium

&x&
W

W
(a) (b) (c)

Figure 2.5: Vertical spring-mass system

The free body diagram in part (b) of the figure shows that, at the static equilibrium position,

W = mg = kδs (2.6)

Part (c) of the figure shows the free-body diagram of the mass when it has been disturbed and has a
positive displacement x from the static equilibrium position. Applying Newton’s second law, noting the
positive direction downwards, and measuring the displacement of the mass from the point at which the
mass was attached to the spring, gives

W − k(x + δs ) = mẍ
2 The analysis for a suspended mass at static equilibrium is similar to that conducted in Section 1.5.2, for static spring
deflections, and it would be worthwhile reviewing that section in conjunction with this one.

42
But, since from Eq. 2.6, W − kδs = 0, the above equation simplifies to
mẍ + kx = 0 (2.7)
This is the identical equation to Eq. 2.5. It is therefore apparent, just as in Section 1.5.3 that the forces
acting on the system mass at the static equilibrium position (including weight and the elastic spring
force(s) that balance it) will always cancel each other out in the analysis and may be neglected, provided
the displacement is measured from the static equilibrium position.

Important notes:

+ The weight of a mass (a force) may be balanced by one or more spring forces arising from static
spring deflections. These will always balance each other when the displacement of the mass is
measured from the static equilibrium position, and will always cancel each other out in the equation
of motion.

+ Provided the displacement of the mass is made from the static equilibrium position the weight and
the static spring force(s) balancing it may be neglected in the analysis.

+ In the Energy Method the increase in gravitational potential energy of the mass (as it moves up),
as measured from the static equilibrium position, is exactly balanced by the decrease in spring
potential energy (and vice versa), so that the total system energy remains the same. Hence the
affect of both may be neglected.

+ Neglecting the weight does not imply that the mass may be neglected! The mass results in inertia
in the system and is still accounted for in Newton’s Law or the Energy Method to derive the
differential equation of motion.

+ In complex systems consisting of more than one mass, or pendulums hanging vertically, it is
possible that one or more weight terms are not balanced by forces arising from static spring
deflections, and these must be taken into account in the analysis.

+ For the case where the weight of a mass is supported by elastic spring forces at the static equili-
brium position, then the governing differential equation is independent of gravitational accelera-
tion.

2.2.4 Equation of motion for an undamped system in the vertical position

The general equation of motion of a single degree of freedom system is

mẍ + cẋ + kx = 0 (2.8)
which, in the absence of damping (an idealisation), reduces to
mẍ + kx = 0 (2.9)

2.2.5 System response for an undamped translating system

Utilising the Laplace transform approach3 , and applying the general initial conditions x(t = 0) = x0 and
ẋ(t = 0) = ẋ0 , the solution of Eq. 2.9 yields the system response:
ẋ0
x(t) = x0 cos ωnt + sin ωnt (2.10)
ωn
3 See handout “Review of the Solution of Linear Ordinary Differential Equations Found in the Analysis of Mechanical
Vibrations”

43
where the circular natural frequency (denoted “natural frequency” hereafter) is given by
r
k
ωn = (2.11)
m

Eq. 2.10 can also be written in the form

x(t) = A cos(ωnt − φ ) (2.12)

r  ẋ 2  ẋ 
0 0
where A = x02 + and φ = tan−1 represent the amplitude and phase angle of the re-
ωn x0 ωn
sponse, respectively.
Eqs. 2.10 and 2.12 indicate that once the system is set into motion it will vibrate with simple harmonic
motion with a frequency ωn rad/s.
Figure 2.6 shows the graphical representation of the time response of Eqs. 2.10 and 2.12 .

Figure 2.6: Graphical representation of harmonic motion of undamped free vibrating system [1].

Notes:

+ The natural frequency is a property of the system. It is a function of m and k and is independent of
the amplitude of oscillation, or the manner by which the system is set into motion.

+ Unless stability issues arise, and in single DOF systems, the coefficients of the displacement,
velocity and acceleration terms in the differential equation must be positive, and serves as a check
that the derivation has been carried out correctly.

+ The amplitude A and the phase angle φ are dependent of the initial conditions.
r
2 mg g
+ From Eqs. 2.6 and 2.11, mωn = k = . Therefore ωn = , which means that the natural
δs δs
frequency of a system may be determined directly by measuring the static spring deflection when
the mass is attached, without necessarily knowing the stiffness k or the mass m. Note that this is
only valid if there is no residual compressive force in the spring.

44
2.3 Free Vibration of an Undamped Torsional System

If a rigid body oscillates about a specific reference axis, the resulting motion is called torsional vibration.
In this case the displacement is measured in terms of an angular coordinate. In a torsional system the
restoring moment may be due to torsion of an elastic member, or the unbalanced moment of a force
or couple. An identical procedure to that of translational systems, based on Newton’s Law, is used for
analyzing torsional systems. The only differences are that the co-ordinate system is angular, say θ , and
Newton’s Law is written as ΣM = J0 θ̈ (where ΣM is the summation of the moments about the rotational
axis, and J0 is the mass moment of inertia of the system about the rotational axis). The moment M is
assumed to be positive in the positive direction of the angular acceleration.

Figure 2.7: Torsional vibration of a disc [1].

Figure 2.7 shows a disc, which has a mass moment of inertia J0 , mounted onto the end of a solid circular
shaft, the other end of which is fixed (assume that the mass moment of inertia of the shaft is small).
Let the angular rotation of the disc about the axis of the shaft be θ , which also represents the shaft’s
angle of twist. From solid mechanics theory the angle of twist is

Ml
θ= (2.13)
GJ
where M is the torque or moment applied to the end of the shaft, l is the length of the shaft, G is the shear
modulus, and J is the polar area moment of inertia (πd 4 /32 for a circular shaft of diameter d). If the
disc in Figure 2.7 is displaced by θ from its equilibrium position, the shaft provides a restoring torque of
magnitude Mt . Thus the shaft acts as a torsional spring with a spring constant

Mt GJ πGd 4
kt = = = (2.14)
θ l 32l

2.3.1 Equation of motion for an undamped torsional system

The equation of motion of the disc can be derived using Newton’s second law by considering the free-
body diagram of the disc in Figure 2.7. As before a datum is chosen and a positive angular displacement

45
direction is chosen. The disc is given an initial positive displacement and the rod acts as a torsional
spring which always resists the applied displacement.

+ +
y y
ΣM = J0 θ̈
J0 θ̈ + kt θ = 0 (2.15)

2.3.2 System response for an undamped torsional system

The solution of Eq.2.15 has the same form as Eq. 2.10.

θ˙0
θ (t) = θ0 cos ωnt + sin ωnt (2.16)
ωn
where ωn is the natural frequency and is given by Eq. 2.17
s
kt
ωn = (2.17)
J0

and the initial conditions have been applied, as before.

Eq. 2.16 may also be written as
θ (t) = A cos(ωnt − φ ) (2.18)
s
 θ̇ 2  θ̇ 
0 0
where A = θ02 + and φ = tan−1 .
ωn θ0 ωn

2.4 Examples of Problem Formulation

Example 1:
A pendulum consists of a sphere of mass m1 = 3 kg and radius r = 40 mm. It is suspended from a pivot
O by means of a rigid slender rod of mass m2 = 1.2 kg and length l = 1.5 m. It has two springs (each of
stiffness k = 500 N/m) which are attached at the centre of the rod. The pendulum hangs vertically at
static equilibrium. See Figure 2.8.

1. Derive the differential equation of motion of the system and determine the period of oscillation of
the pendulum.

2. If the pendulum is given an initial displacement of θ0 = 0.1 rad, and zero initial velocity, determine
the system response θ (t).

3. Determine the maximum angular velocity of the pendlum.

46
 O Rx O

Ry
mass m2
l/2 l/2

k k
2k(l/2)q

m2g
l/2 l/2
+
q
r static
r
equil.
mass m1
m1g
q, q, q +
(a) System (b) FBD

Figure 2.8: Pendulum example

Procedure:

1. After reading the problem statement carefully, draw the free body diagram indicating:

• the static equilibrium position,

• all forces and moments acting on the system when it is displaced from the static equilibrium
position, noting if there are weight components that are balanced by static spring deflections
(if so, these are not drawn on the FBD and a clear statement is made in the assumptions),
• all reaction forces,
• relevant variables such as stiffnesses, lengths etc.,
• the direction of positive displacement, velocity and acceleration of the system variable which
is to be used to derive the equation of motion.

2. State relevant assumptions.

3. Derive the equation of motion using Newton’s Law or other means. Indicate direction of moment
and positive acceleration in derivation. Check units of expressions and variables! Only substitute
the values of the parameters into equations at the end to get final results.

4. Determine the system response x(t) or θ (t) by applying initial conditions.

5. Answer all the questions that were asked!

6. Check the validity of the answers in the light of simplifying assumptions.

47
Solution:

• Free-body diagram — see Figure 2.8(b)

• Assumptions:
– θ small such that sin θ ≈ θ and cos θ ≈ 1.
– The pendulum hangs vertically at static equilibrium, so the weights of the rod and bob
are not balanced by static spring deflections, therefore the weights must be taken into
account.
• Determine the mass moment of inertia of the sphere and rod about the axis of rotation O.
2 1
Isphere,0 = m1 r2 + m1 (r + l)2 and Irod,0 = m2 l 2
5 3
2 1
∴ I0 = m1 r2 + m1 (r + l)2 + m2 l 2
5 3
• Apply Newton’s Law to derive the differential equation of motion. Do not substitute any values!
+ +
x x
ΣM0 = I0 θ̈ (anti-clockwise positive)
     
l l l
−2k θ. − m2 g. θ − m1 g.(l + r)θ = I0 θ̈
2 2 2
 2   !
l l
∴ I0 θ̈ + 2k + m2 g + m1 g(l + r) θ = 0
2 2

• Check units: all terms have units of Nm X.

• Check that coefficients of θ̈ and θ are positive X.
• Determine expression for natural frequency and calculate ωn and τ. Do not simplify ωn unless
asked to do so as it is easier to trace mistakes!
s

2 r
2k 2l + m1 g(l + r) + m2 g 2l


616.6512
ωn = = = 8.770 rad/s
I0 8.0167
2π
τ = = 0.72 s
ωn
• System response:
θ˙0
θ (t) = θ0 cos ωnt + sin ωnt
ωn
where θ0 = 0.1 rad and θ̇0 = 0. Therefore the system response is:

θ (t) = 0.1 cos 8.770t

• Check assumptions: maximum angular displacement is 0.1 rad which is consistent with small
angle approximation; the weight terms at static equilibrium (vertical) were not balanced by static
spring forces so were included in the analysis.
• Maximum angular velocity. To determine maximum angular velocity it would generally be re-
quired to differentiate the response θ (t) twice to get acceleration, then solve for the time when
θ̈ = 0, then substitute that time into the expression for velocity. However, in an undamped case
the velocity response is θ̇ (t) = −ωn θ0 sin ωnt so the maximum velocity is simply the amplitude of
this harmonic response, or θmax = ωn θ0 = 8.770 × 0.1 = 0.877 rad/s.

48
Example 2:
A rigid slender bar of length l = 1.8 m and negligible mass is pivoted at 0 (located 600 mm from one end),
and has masses m1 = 5 kg and m2 = 12 kg attached to its ends (assume point masses for this example),
as indicated in Figure 2.9. A torsional spring of stiffness kt = 250 Nm/rad is attached to the bar at the
pivot point, and a tension spring of stiffness k = 1500 N/m is attached 600 mm from the other end of the
bar. The bar is at static equilibrium when it is horizontal. Derive the differential equation of motion and
determine the natural frequency.

(a) System

kt k
m1 l/3 l/3 l/3 m2
O
�

(b) FBD kt � k(l/3)�

static equil.
Ox � �� �� �
O
Oy

Figure 2.9: Bar example

Solution:

• Free-body diagram — see Figure 2.9(b). (Note: no weight term, see assumptions).

• Assumptions:

– θ small such that sin θ ≈ θ and cos θ ≈ 1.

– The bar is horizontal at the static equilibrium position, therefore the weight of the two masses
is balanced by the spring forces due to their static deflection. Thus the weight terms m1 g and
m2 g and the spring forces that balance these weights at the static equilibrium position may
be ignored.
– The physical geometry of the masses has not been given, therefore assume point masses4 .

• Although this case is reasonably simple, it is necessary when determining the differential equation
of motion to determine the term on the RHS of Eq. 2.2. Either I0 can be determined directly as in
the this simple case, or Eq. 2.4 can be used. It would require knowledge of where the centre of
gravity is located so will not be used in this example.
The mass moments of inertia of the two masses about the axis of rotation 0 can be determined and
substituted in Eq. 2.2 as follows:
 2  2
l 2l
I1,0 = m1 and I2,0 = m2
3 3
 2  2
l 2l
∴ I0 = I1,0 + I2,0 = m1 + m2
3 3
4 Repeat the exercise if the masses have defined geometries such as spheres or blocks with assumed geometries and
see how this affects the solution.

49
• Apply Newton’s Law to derive differential equation of motion.

+ +
y y
ΣM0 = I0 θ̈ (clockwise positive)
   
l l
−k θ. − kt θ = I0 θ̈
3 3
 2 !
l
∴ I0 θ̈ + k + kt θ = 0
3

• Check units: all terms have units of Nm X.

• Check that coefficients of θ̈ and θ are positive X.

• Determine expression for natural frequency and calculate ωn

s
2
k 3l + kt
r
790
ωn = = = 6.4346 rad/s
I0 19.08

Exercise:
A pendulum bob of small size and mass m is attached to a rigid massless link of
length l. The link is attached to a frictionless pivot. A spring of stiffness k is
attached to the link a distance a from the pivot. Determine the differential
equations of motion of the three configurations (note static equilibrium positions).
S.E.
q
m

a k k
k
l m a l
S.E. a
l
m q
q
S.E.
(a) (b) (c)

Answers:
2
.. 2 2
.. 2 2
.. 2
ml q+(mgl+ka )q=0 ml q+ka q=0 ml q+(ka -mgl)q=0
Additional Questions:
If the static equilibrium position is at a slant (i.e. not at vertical or horizontal),
what effect will it have on the equations of motion?
What can be observed from the coefficient of q in part (c)?

50
2.5 Determining natural frequencies by Rayleigh’s Energy Method

The natural frequency of a conservative system can be deduced by Rayleigh’s energy method without
obtaining the differential equations. For a conservative system,

T +U = const

where T is the kinetic energy of the system, and U the change in potential energy of the system from its
potential energy in the static equilibrium position. This equation can also be represented as follows:

T1 +U1 = T2 +U2 (2.19)

where the subscripts 1 and 2 denote two different instants of time. It is convenient to use the points at
which there is maximum potential energy (and hence zero kinetic energy), and maximum kinetic energy
(and hence zero potential energy). Thus for a system undergoing harmonic motion

Umax = Tmax (2.20)

This is known as Rayleigh’s energy method and can be used to determine the natural frequency of the
system. For example, consider again the pendulum system shown mass-spring system where the mass of
the spring is not negligible, shown in Fig. 2.8.
The maximum kinetic energy is
1
Tmax = IO θ̇ 2
2
and the maximum potential energy (assuming small angles, and noting again that the weight, or in this
case gravitational potential energy term, must be taken into account) is

1  l 2 l
Umax = (2k) θ + m1 g(l + r)(1 − cos θ ) + m2 g (1 − cos θ )
2 2 2
The response is known to be harmonic and may be written as

θ (t) = Θ cos ωnt

from which
θ̇ = −ωn Θ sin ωnt
It is evident that the displacement magnitude is Θ and the angular velocity magnitude is ωn Θ. Writing
these into the equations for kinetic and potential energies (so as to yield magnitudes of the energies),
θ2
utilising the approximation cos θ = 1 − , and solving for ωn :
2
Tmax = Umax
1 1  l 2  Θ2   l  Θ2 
IO (ωn Θ)2 = (2k) Θ + m1 g(l + r) + m2 g
2 2
v 2 2 2 2
u 2  
u 2k l + m1 g(l + r) + m2 g l
t 4 2
ωn = , as before
IO

Note that in some instances only the natural frequency is of interest, so this method may be used without
deriving the differential equation of motion (provided it is a conservative system). If, however, the system
response was required, then it would be necessary to derive the differential equation.

51
2.6 Free Vibration with Viscous Damping

As stated in Chapter 1 the viscous damping force is proportional to velocity ẋ and acts in the opposite
direction of the velocity. A single degree of freedom system with a viscous damper is shown in Figure
2.10(a) with its free body diagram shown in part (b).

Figure 2.10: Single degree of freedom system with viscous damper [1].

Applying Newton’s Law the equation of motion is

mẍ + cẋ + kx = 0 (2.21)

in which m, c and k are all positive constants. Following the procedure in the handout “Review of the
Solution of Linear Ordinary Differential Equations Found in the Analysis of Mechanical Vibrations”, the
characteristic equation can be written as
ms2 + cs + k = 0 (2.22)
This is also known as the auxiliary equation of the system. The roots of the characteristic equation are
√
−c ± c2 − 4km
s1,2 = (2.23)
2m
c c p
Alternatively5 , writing ζ = = √ and ωd = ωn 1 − ζ 2 , the roots may be written as:
2mωn 2 km
p
or s1,2 = −ζ ωn ± ωn ζ 2 − 1 (2.24)
5 Eq.2.23 may be rewritten as follows:
r
c c2 k
s1,2 = − ± −
2m 4m2 m
r  r
c c 2 2 k
∴ s1,2 = − ± − ωn since ωn =
2m 2m m
s
 2
c c
∴ s1,2 = − ± ωn −1
2m 2mωn

The terms inside the square-root are dimensionless, so we may define a new parameter
c c
ζ= = √
2mωn 2 km
ζ is called the damping factor or damping ratio. Since m, c and k are positive, ζ is a positive, dimensionless number.

52
and the general solution to Eq. 2.21 is given by

x(t) = a1 es1t + a2 es2t (2.25)

The nature of the roots s1 and s2 , and hence the behaviour of the solution, Eq. 2.25, depends upon the
magnitude of damping ratio ζ , as shown in the handout “Review of the Solution of Linear Ordinary Dif-
ferential Equations Found in the Analysis of Mechanical Vibrations”. The various cases are summarised
here.

The system may be classified as:

+ Undamped, ζ = 0, yielding imaginary roots, with resulting motion purely oscillatory with
constant amplitude. See Sections 2.2 and 2.3.

+ Underdamped, ζ < 1, yielding complex roots, with resulting motion oscillatory but decaying in
amplitude.

+ Critically damped, ζ = 1, yielding repeated roots, with resulting motion non-oscillatory.

+ Overdamped, ζ > 1, yielding real roots, with resulting motion non-oscillatory.

• ζ < 1. Underdamped System

The roots described by Eq. 2.24 are complex conjugates and the response is given by6 :
h  ẋ + ζ ω x  i
0 n 0
x(t) = e−ζ ωnt x0 cos ωd t + sin ωd t (2.26)
ωd

The system described in Eq. 2.26 is a damped harmonic motion of circular frequency ωd , but be-
cause of the factor e−ζ ωnt , the amplitude decreases exponentially with time, as shown in Fig. 2.11.
p
The quantity ωd = ωn 1 − ζ 2 is called the frequency of damped vibration or the damped natu-
ral frequency. It can be seen that ωd is always less than the undamped natural frequency ωn . Also,
from the equation for period of vibration τ = 2π/ω, it is clear that τd > τn .

• ζ = 1. Critically Damped System

In this case, the two roots s1 and s2 of Eq. 2.24 are equal to −ωn . Following the process in the
handout, the general solution of Eq. 2.21 is given by

x(t) = x0 + (ẋ0 + ωn x0 )t e−ωnt

 
(2.27)

It can be seen that the motion represented by Eq. 2.27 is aperiodic (i.e., non-oscillatory). Since
e−ωnt → 0 and te−ωnt → 0 as t → ∞, the motion will eventually diminish to zero as shown in
Fig. 2.11
The property of critical damping is used in many practical applications. For example, the recoil
mechanism of large guns are usually critically damped so that they return to their original position
after firing in the minimum time without oscillation. If the damping provided was more than the
critical value, some delay would be caused before the next firing.
√
The critical damping coefficient cc is given by cc = 2 km. Hence the damping factor ζ can be
defined as ζ = c/cc .

6 See handout “Review of the Solution of Linear Ordinary Differential Equations Found in the Analysis of Mechanical
Vibrations”.

53
 • ζ > 1. Overdamped System
In this case, the two roots s1 and s2 from Eq. 2.24 are real, distinct and negative. Following the
process in the handout, the general solution of Eq. 2.21 is given by
h  ẋ + ζ ω x  i
0
p n 0 sinh(ωn ζ 2 − 1t)
p p
x(t) = e−ζ ωnt x0 cosh(ωn ζ 2 − 1t) + (2.28)
ωn ζ 2 − 1

As with critically damped systems the motion is also aperiodic, and diminishes exponentially with
time, but takes longer for the motion to reach zero displacement than for the critically damped
case. Fig. 2.11 shows the typical aperiodic response.

Summary:
For the three damped vibration cases considered above, the behaviour of the solution depends on the
value of ζ . Vibratory (oscillatory) motion is possible only in the undamped and underdamped cases.
In all three damped cases, the displacement of the mass will eventually diminish until the mass returns
to the static equilibrium position, but the critically damped system takes the minimum time possible to
return to its equilibrium position.
Note that ζ = 0 is a special case of the underdamped system. When ζ = 0, the system has no damping,
and the amplitude will not diminish with time. The undamped case and the three damped cases are
represented in Figure 2.11. In each case the system parameters and initial conditions were identical (m =
1 kg, and k = 100 N/m, x0 = 0.1 m, ẋ0 = 2 m/s), but the damping ratio ζ was varied (zero for undamped
case, 0.2 for underdamped case, 1 for critically damped case, and 2 for overdamped case).

0.25

Undamped
0.2

0.15

x0 Critically
0.1 Underdamped
−1
Damped
tan ẋ0
Displacement (m)

0.05 Overdamped

−0.05

−0.1
2π
ωd
−0.15

−0.2
2π
ωn
−0.25
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
Time (s)

Figure 2.11: Comparison of motions with different types of damping (Adapted from [1]).

54
The following points relate to Figure 2.11:

1. The initial displacement is the same, and the initial slope (or velocity) is the same for each case
(since x0 and ẋ0 remained the same for all the cases).

2. As the damping ratio increases the maximum displacement reached by the mass in the four cases
decreases since some energy is being dissipated while the displacement of the mass is still increa-
sing after the initial disturbance. The higher the damping ratio, the more energy is dissipated and
the lower the maximum displacement.

3. The period of vibration for the undamped case is τn = 0.6283 s, while the period for the underdam-
ped case is τd =0.6413 s, which is greater, as expected.

4. The critically damped case and the overdamped case are both aperiodic, but the critically damped
case reaches steady state the quickest. While these do not adhere to the definition of “vibration” or
“oscillation” they are special cases where the mass has been disturbed and displaced away from the
static equilibrium position, and return under the action of restoring forces generated in the system
components.

In mechanical engineering systems inherent damping is usually small. Where damping is specifically
designed (such as a car suspension) the damping ratio is deliberately chosen. The equivalent viscous
damping ratio of typical engineering materials or structures7 is usually fairly small. The table below
gives some indication of values that could be encountered in various materials, with a comparison to
vehicle shock absorber damping ranges. Note that while shock absorbers are viscous dampers the other
forms of damping in the table are not and should be seen as estimates of equivalent viscous damping
only.

Table 2.1: Typical damping values

Application or material ζ
Vehicle shock absorbers 0.1 – 0.5
Rubber 0.04
Riveted steel structures 0.03
Concrete 0.02
Wood 0.003
Cold rolled steel 0.0006
Cold rolled aluminium 0.0002
Phosphur bronze 0.00007

2.6.1 Logarithmic decrement.

The mass and stiffness of a one degree of freedom system can be determined from static tests; however,
damping requires a dynamic test to measure. The following analysis relates to an underdamped system
(which has oscillatory motion), and enables the determination of the damping ratio from a record of the
displacement-time response.
With reference to Figure 2.12, let t1 and t2 denote the times corresponding to two consecutive displace-
ment amplitudes, measured one complete cycle apart, that is t2 = t1 + τd where τd = 2π/ωd .
7 Most materials actually have hysteretic damping but when the damping is small it is not unreasonable to use an
equivalent viscous damping ratio.

55
 x0
x1
x2
x3

Displacement x(t) t1 t2 t3 t

Figure 2.12: Response of an underdamped system (adapted from [2]).

Eq. 2.26 may also be written as x(t) = Ae−ζ ωnt cos(ωd t − φ ) from which the ratio of successive amplitu-
des may be determined from:

x1 Ae−ζ ωnt1 cos(ωd t1 − φ )

= −ζ ω (t +τ )
x2 Ae n 1 d cos(ω t + 2π − φ )
d 1

Mathematically, the cos terms cancel as they represent the same value, one cycle apart. The equation
above therefore simplifies to
x1
= eζ ωn τd (2.29)
x2
which defines the ratio of successive displacements and depends on the nature of the damping and the
natural frequency.
The logarithmic decrement δ is defined8 as the natural logarithm of the ratio of any two successive
amplitudes:
x1 2πζ
δ = ln = ζ ωn τd = p (2.30)
x2 1−ζ2

The above equation can also be solved for ζ to yield

δ
ζ=p (2.31)
(2π)2 + δ 2

The logarithmic decrement is dimensionless and is actually another form of the dimensionless damping
ratio ζ . Once δ is known, ζ can be calculated by solving Eq. 2.31. Thus the damping can be determined
experimentally by measuring any two consecutive displacement amplitudes x1 and x2 . By taking the
natural logarithm of the ratio x1 /x2 , δ may be obtained. By using Eq. 2.31, the damping ratio ζ may be
determined. In fact, the damping ratio can also be found by measuring two displacements separated by
any number of complete cycles.

8 Note that in vibration analysis there are at least three uses of the Greek letter δ : static spring deflection, logarithmic
decrement and impulse.

56
Let x1 denote the amplitude at time t1 , and let x p+1 denote the amplitude after p complete cycles corre-
sponding to time t p+1 = t1 + pτd where p is an integer representing the number of complete cycles.
x1 x1 x2 xp
= ···
x p+1 x2 x3 x p+1
= eζ ωn τd eζ ωn τd · · · eζ ωn τd
= e p(ζ ωn τd ) (2.32)

Taking natural logarithm yields  x 

1
ln = p(ζ ωn τd ) = pδ (2.33)
x p+1
Thus
1  x1 
δ= ln (2.34)
p x p+1

which can be substituted into Eq. 2.31 to obtain ζ . Note that p is the number of complete cycles after
the first measurement of displacement.
The damping
p in many real systems is relatively small, with ζ often less than 0.2. In such instances the
quantity 1 − ζ 2 ' 1, so that Eq. 2.30 is written as

δ
δ ' 2πζ or ζ= (2.35)
2π

Figure 2.13: Variation of δ with ζ (adapted from [1])

Figure 2.13 shows the variation of δ with ζ as given by Eqs. 2.30 and 2.35. It can be noticed that for
values up to ζ ≈ 0.3, the two curves are very similar. In field calculations, Eq. 2.35 is often used to
calculate ζ .

2.6.2 Torsional systems with viscous damping.

The methods presented in previous sections for translational vibrations with viscous damping can be
extended directly to viscously damped torsional vibrations. Consider a single degree of freedom torsional

57
 Figure 2.14: Torsional system with damping[1].

system with a viscous damper as shown in Figure 2.14. For the torsional vibration of the shaft, the
viscous damping torque is T = −ct θ̇ where ct is the torsional viscous damping constant and θ̇ is the
angular velocity of the shaft. The equation of motion is

J0 θ̈ = −kt θ − ct θ̇

or
J0 θ̈ + ct θ̇ + kt θ = 0 (2.36)
where J0 is the mass moment of inertia of the disc and kt is the spring constant for angular motion.
The solution to the equation of motion is determined in the same way as before where, for an underdam-
ped case
h  θ̇ + ζ ω θ  i
0 n 0
θ (t) = e−ζ ωnt θ0 cos ωd t + sin ωd t (2.37)
ωd
p
ωd = ωn 1 − ζ 2 (2.38)
s
kt
ωn = (2.39)
J0
ct ct ct
ζ = = = √ (2.40)
ctc 2J0 ωn 2 kt J0

2.7 Stability Conditions

Mechanical systems do exist where lateral vibration occurs about a fixed rotational axis, where a moment
due to the gravitational force about the fixed axis exists. In cases where springs apply restoring forces to
components undergoing this form of oscillation a condition may exist where the moment of the gravita-
tional force about the rotational axis is not able to be resisted by the spring forces acting about the same
axis. In such a case the system is said to become unstable. Consider Figure 2.15 which shows a uniform
slender rod pivoted at one end and connected symmetrically by two springs. Assume that the mass of
the rod is m, and that the rod is in the vertical position at static equilibrium9 . When the rod is displaced
9 If the rod is in the vertical position at static equilibrium the weight vector does not exert a moment about the pivot
and therefore does not develop reactive forces in the springs to balance it. Hence the weight must be taken into
account.

58
by an angle θ the spring force in each spring is kl sin θ (one in tension and the other in compression,
but their reactive forces in the same direction). The weight W = mg acts vertically downward through
the centre of gravity G and results in a moment about the rotational axis O of (mgl sin θ )/2. The mass
moment of inertia Io = ml 2 /3.

Figure 2.15: Stability of rigid rod [1].

The equation of motion of the rod rotating about point O is

1 2 mgl
ml θ̈ + (2kl sin θ )l cos θ − sin θ = 0 (2.41)
3 2
For small oscillations sin θ ≈ θ and cos θ ≈ 1, this reduces to
1 2  mgl 
ml θ̈ + 2kl 2 − θ =0 (2.42)
3 2
Thus v
u 2kl 2 − mgl
u r
12kl 2 − 3mgl
ωn = t 1 2 2 = (2.43)
3 ml
2ml 2

The sign of the numerator in the square-root indicates the type of roots that the characteristic equation
has and whether the system is stable. If 12kl 2 − 3mgl > 0 the oscillations are stable and θ = A1 cos ωnt +
A2 sin ωnt.
If the term is zero it indicates a static position where the weight term is exactly balanced by the spring
restoring forces and the rod is not moving. If 12kl 2 − 3mgl < 0 the system is unstable and θ increases
with time without oscillation (i.e. the rod moves to one side and continues to move until it has fallen flat
on the surface).
The condition for instability is thus k ≤ mg/4l. The variables in this stability criterion are all system
parameters, and even with a small disturbance from the vertical the system will become unstable.
As a general rule, for inverted pendulums where the weight term does not tend to return the pendulum
to the static equilibrium position, the stiffness coefficient in the differential equation should always be
checked in the light of instability.

59