1

1
2
3
4
5
Linear vibrations 6
7
8
9
DOI: 10.1533/9781782421665.19 10
1
Abstract: Once the fundamental concepts of dynamics are 2
defined, it is important to describe mechanical vibrations. 3
Mechanical vibrations cover a broad field of knowledge;
4
thus, in this chapter we start defining the basic aspects of
5
linear vibrations. The concepts are derived from the
6
equation of motion, which is formulated from the energy
methods introduced in the previous chapter. The equation
7
of motion is a second order differential equation, and the 8
linear term comes from the assumption that the coefficients 9
of this equation are constant. The equation of motion 20
depends on the generalized coordinates, and for each 1
generalized coordinate we set a degree of freedom. The 2
basic concepts are introduced with the analysis of a single 3
degree of freedom system. For this system, three cases are 4
presented: free undamped vibration, forced vibration and 5
damped vibration. From the forced vibration concept, a 6
definition of transmissibility is introduced and finally a 7
system with multiple degrees of freedom is presented. The
8
example included in this chapter is a gear box with four
9
degrees of freedom; this example will be used thoughout
30
the book.
1
Key words: free vibration, forced vibration, damping,
2
transmissibility, multiple degrees of freedom. 3
34R

19

© Elsevier Limited, 2014

Parameter Identification and Monitoring of Mechanical Systems

1.1 Introduction
The dynamics of any machinery is perceived as noise or
vibration, any machine vibrates and if we understand its
vibrations, we can know its conditions. In general, engineering
analysis is done assuming the behavior of a system is linear.
Although it is not true, this approximation is good enough to
solve any operating problem or to design a new machine. In
this sense, the following section will describe the theory and
applications of linear vibrations.
The concepts presented next are derived from previous
definitions, and they will be centered around the general
engineering concepts of the dynamic of elastic systems. It is
assumed that a machine can be represented as a set of masses
connected through elastic elements (springs) and dissipative
elements (dampers). Since we are assuming that the behavior
of such systems is linear, the movement of the masses will be
represented as oscillations.
The development of vibrarion sensors allows the
measurement and collection of vibration data – the most
common sensors are piezoelectic accelerometers, capacitive
accelerometers, velocity sensors and displacement sensors.
Additionally, the development of signal analysis methods
simplifies the diagnosis of machinery. The construction of
frequency spectrums based on the Fast Fourier Transform
(FFT) has become a well known procedure in industrial
applications. Nowadays there are several companies that
offer vibration analysis systems that can be implemented for
the diagnosis of machinery. That is the reason for defining
the math basis of vibration analysis.

20
 Linear vibrations

1.2 Single degree of freedom

A simplified representation of a machine is a mass supported
by a spring and a damper. This representation includes the
main elements and helps understanding the dynamic behavior
of complex systems. Figure 1.1 shows the representation of a
mass m moving in the vertical direction, the movement is
restricted by a spring k and the movement is damped by a
viscous damper with a constant c.
The equation of motion is derived using previous concepts.
In this case, the generalized coordintates will be the
displacement x, and the potential and kinetic energies are:

[1.1]

[1.2]

Figure 1.1 Simplified model of a machine

21
Parameter Identification and Monitoring of Mechanical Systems

The damper dissipates energy, therefore the Lagrange

equation of the system will be:

[1.3]

where

L=T−V [1.4]

and

Q = f(t) − cẋ [1.5]

The equation of motion is found substituting equations 1.4

and 1.5 into 1.3:

mẍ + cẋ+ kx = f(t) [1.6]

There are two solutions, if f(t) = 0 the system will freely

vibrate and it is known as ‘free vibration’. In this case the
solution of the differential equation is

[1.7]

where
[1.8]

[1.9]

and x0 and ẋ0 are the initial conditions.

If we define as the damping coefficient

[1.10]

then the most simple motion occurs when c = 0 in which case

22
 Linear vibrations

p=0

and

and the mass moves according to a simple harmonic motion

[1.11]

Figure 1.2 represents the simple harmonic motion of a system

when ẋ0 = 0. This type of motion is known as ‘Undamped
Free Vibration’.
If c ≠ 0 then the mass displacement damps until it rests.
The form of this type of motion is represented in Figure 1.3.
The function that envelopes the sinusoidal wave is known
as the logarithm decrement,and it is possible to measure the
damping coefficient of a system.
The amplitude of two consecutive peeks x1 and x2 are
separated by a period T. Then, from equation 1.7

[1.12]

[1.13]

Harmonic motion of a single degree system and

Figure 1.2
phase angle

23
Parameter Identification and Monitoring of Mechanical Systems

Thus,

[1.14]

where

[1.15]

and the damping is

[1.16]

Or in a parametric form:

[1.17]

In general, f(t) is assumed to be sinusoidal. In this case, the

solution of the differential equation (Eq. 1.6) will be

mẍ + cẋ + kx = f0cos(Ωt) [1.18]

The solution of equation 1.18 is

[1.19]

Or

[1.20]

24
 Linear vibrations

where

[1.21]

From equation 1.20 the transfer function of the dysnamic

system is definded as

[1.22]

Figure 1.3 shows the amplitude of the transfer function, and

Figure 1.4 describes the variations of the phase angle.

Figure 1.3 Transfer function: forced vibration response

25
Parameter Identification and Monitoring of Mechanical Systems

Phase angle between the input force and the

Figure 1.4
mass displacement: forced vibration

The critical speed is the natural frequency of the system.

Beyond this value, the phase angle between the excitation
force and the mass displacement is closed to 180°. If the
excitation force comes from the foundation, then the
movement of the mass is the result of the trasmissibility of
the supports, in this case we represent the supports as a linear
spring and a linear damper. The force transmitted by these
elements is

fk = kxb [1.23]

fc = cẋb [1.24]

Then, the equation of motion becomes

mẍ + cẋ + kx = kxb + cẋb [1.25]

26
 Linear vibrations

Assuming that the base has an armonic motion, then

xb = Acos(Ωt) and ẋb = − AΩsin(Ωt). Thus, the equation of
motion is

mẍ + cẋ + kx = A(kcos(Ωt) − cΩsin(Ωt)) [1.26]

or

ẍ + 2ξωnẋ + ωnx = A(ωncos(Ωt) − 2ξωnΩsin(Ωt)) [1.27]

The solution of the equation of motion will be x = Xcos ×

(Ωt − φ).
The relation between the excitation amplitude and the
mass displacement amplitude is the trasmissibility of the
foundation, and it is

[1.28]

This equation is the basis for a foundation design, and the

design criterion will be to have the minimum transmitted
force. Thus, it can be shown that the magnitude of the
transmitted force |T|>1 will occur if . That means
that the excitation frequency must be greater than the natural
frequency. But in many cases this type of design is not
possible and the operating frequencies are a lot lower than
the natural frequencies. Thus, the stiffness and damping
characteristics should be modified by adding a soft material
between the vibration source and the object that will be
isolated. Since there are two design parameters, the problem
is underdetermined; therefore, a range of frequencies and an
amplitude threshold must be defined. Figure 1.5 shows a
design chart for a support. The desired amplitude is the input

27
Parameter Identification and Monitoring of Mechanical Systems

Figure 1.5 Design chart for a soft support

data, and the output is the range of frequencies that the

isolation will absorb.

1.3 Multiple degrees of freedom

Although many engineering problems can be solved with a
single degree of freedom model, complex systems have to be
analyzed as a multibody system. The system is divided into a
set of point inertias (masses or moments of inertia) connected
through springs and dampers. Each inertia represents a
generalized coordinate and the equation of motion is derived
using Lagrange’s equation. The concept is introduced with
an example. Figure 1.6 shows a gear transmission, where the
gear teeth and the rolling bearings are represented by linear
springs. In the following chapters they will be modeled as
nonlinear elements.

28
 Linear vibrations

A pair of gears represented as a system with

Figure 1.6
four degrees of freedom

Representing the gear box as a lumped-mass system, four

generalized coordinates are identified: {x1, x2, θ1, θ2}. The
equation of motion is obtained using Lagrange’s equation
and the dynamic behavior of the pair of gears consists of
four differential equations:

29
Parameter Identification and Monitoring of Mechanical Systems

[1.29]

[1.30]

[1.31]

[1.32]

The system of equations can be represented in a matrix form

as:

[1.33]

or in a simplified form
–¨ –
[I]{X } + [K ]{X } = {p– } [1.34]

The homogenous solution of equation (1.34) is found

assuming that {p– } = 0 and xi = Xi cos(ωit − φi). Thus, the matrix
equation will be:

[−ω2[I] + [K]]{X} = 0 [1.35]

The solution is found solving the eignevalues problem, and

the roots of the polynomial correspond to the natural
frequencies of each generalized coordinates

det|[K] − ω2[I]| = 0 [1.36]

Solving the eigenvalue problem, a fourth order polynomial is

found:

30
 Linear vibrations

[1.37]

where λ = ω2.
The particular solution is found numerically using the
Runge-Kutta method. The external forces are the excitation
forces caused by the rotation of the rolling bearings and the
gear teeth action, they are approximated as:

Pe(t) = (Fp + Fr)cos(Ωt) + (fp + Fϕ)cos(NΩt) [1.38]

where Fr is the gear run out, Fp is the accumulated pitch error,

Fϕ is the profile error, and fp is the pitch error.

Pj(t) = Fircos(Ωir) + Forcos(Ωor) + Fccos(Ωc)

+ Frecos(Ωre) [1.39]

where Pj(t) are the bearing excitation forces ( j = 1,2).

The four frequencies are calculated with the following
equations.
Contact frequency between the roller element and the
internal track Ωir:

[1.40]

Contact frequency between the roller element and the

external track Ωor:

31
Parameter Identification and Monitoring of Mechanical Systems

[1.41]

The casing has a frequency Ωc:

[1.42]

and the roller spin frequency is Ωre:

[1.43]

where db is the roller diameter and Dp is the pitch diameter,

αb is the axial contact angle.
As an example, a gear box is simulated to previous
concepts. In the first part, the natural frequencies are
found using the eigenvalue solution of equation 1.37.
Then, the forced vibration solution of the multiple degree
of freedom problem is solved using the Runge-Kutta
algorithm.
The gear data listed in the table below and with these
values, the natural frequencies are ω1 = 937.0 rad/s, ω2 = 662.5
rad/s, ω3 = 371.9 rad/s.

Gear 1 Gear 2
Mass (m) 4.87 kg 10.57 kg
Moment of inertia (J) 0.0011 kg-m2 0.004652 kg-m2
Base radius (r) 0.031853 m 0.0483285 m
6
Bearing stiffness (K) 3.66 × 10 N-m 4.14106 N-m
Gear mesh stiffness (Ke) 4.37 × 105

32
 Linear vibrations

The solution for the forced vibration part is calculated

assuming that all the excitation forces are acting on the gear
box. In this way, the characteristic spectrum of an industrial
gear box is illustrated, and the particular frequencies are
identified.

33