Fundamentals of Mechanical Vibration - Course Notes 1

Lecture 5
DAMPING

Main Concepts

Following study of this material you should:

 Be able to state several mechanisms of damping and, for viscous damping, its
mathematical model.
 Understand the three types of natural response possible for a viscously damped one-
degree-of-freedom spring-mass-damper system.
 Know that the damped natural frequency decreases by a very small amount compared to
the undamped natural frequency for ‘normal’ engineering quantities of damping.
 Be able to calculate the quantity of damping present from a time-amplitude record of an
under damped vibration.
 Understand that for most engineering situations, the damping ratio is less than about 0.03.

Observation tells us that natural vibrations experience amplitude decay with time so that
eventually the system ceases to move. Clearly the initial energy content of the system has been
lost to forms other than the kinetic and strain (potential) energies that participate in the vibration.
This conversion of energy is termed 'damping'. It means no more than conversion from either
kinetic or strain energy into some other form - usually heat via fluid, surface or solid friction, or
sound.

Our task now is to improve our vibration models by incorporating damping into our
mathematical models in order to better predict the observed decay of natural vibrations. It will
be useful to return to the single degree of freedom system.

A valid question to ask at this stage is "Does damping affect the natural frequencies and mode
shapes that we have spent several lectures developing?". Be assured that our earlier models are
not wasted - as long as the damping is 'light'.

5.1 THE MECHANISMS OF DAMPING

Damping mechanisms can take several forms. For any vibrating system, several of these
mechanisms may be present.

Coulomb Damping

One of the easiest mechanisms to envisage is dry friction damping, also called Coulomb
damping. This is the resistance to motion felt when sliding a book across a desk. To a close

2018 Mechanical Engineering, Curtin University

Fundamentals of Mechanical Vibration - Course Notes 2

approximation, the resistance force is constant with velocity but always acts opposite to the sense
of the velocity.

x

x

Friction force, F
resisting motion

Figure 5.1 Coulomb damping

x
Coulomb damping may be described mathematically as F = − F0 where the ratio simply
x
decides the direction of the force.

Hysteretic Damping

Hysteretic damping is the term that describes the fact that most materials exhibit internal friction
when strained and as such the stress-strain 'curve' on loading is slightly different from that for
unloading. We note that this is true even in the region that we usually identify as being 'truly
elastic'. In actual fact, energy is lost on each stress reversal cycle.

Figure 5.2 depicts a displacement-force curve which could also be understood as a stress-strain
curve. The minor axis of the ellipse is greatly exaggerated. Clearly, the area within the closed
curve represents work done - in this case, energy lost to heat on each vibration cycle.

vibration

Figure 5.2 Hysteretic damping

2018 Mechanical Engineering, Curtin University

Fundamentals of Mechanical Vibration - Course Notes 3

Viscous Damping

Under certain laminar fluid flow conditions the force retarding motion due to the viscous drag is
proportional to the velocity. If the flow is turbulent the velocity squared more closely
approximates the retarding force. The laminar flow conditions can be found in thin fluid layers
such as journal bearings, piston plungers in fluid filled cylinders and flow through orifices.

In general, viscous damping or 'drag', is a very complex phenomenon. However, a simple

representative model serves our purposes very well. Figure 5.3 shows an object experiencing
viscous drag that is proportional to velocity. We shall call this 'viscous damping'. The
mathematical description is:

F = −qx
…(5.1)
T = −qθθ

where q is the constant of proportionality and is called the 'damping coefficient'. Clearly its
dimensions are force per unit velocity, N/(ms-1) for the translational case and Nm/(rad.s-1) for the
rotational case.

Fluid viscous force,

F resisting motion

Viscous fluid

Figure 5.3 Viscous drag

We are going to use viscous damping to investigate its effects in vibration response primarily
because the resulting mathematical models will be linear and therefore reasonably easy to solve
and interpret. It turns out that if the damping forces are small, the viscous damping model
mirrors the actual systems reasonably well. However, if they are not small, it is unlikely that the
linear viscous model will be adequate.

The schematic diagram icon used to denote a viscous damping element is a cylinder-piston
arrangement as shown in Figure 5.4. It is often called a 'dashpot'. It represents a fluid-filled
cylinder with a loosely fitted piston. As the piston moves axially along the cylinder, fluid is
forced to flow through the small annular gap between the piston and the cylinder.

 
Figure 5.4 Schematic diagram of a viscous dashpot

2018 Mechanical Engineering, Curtin University

Fundamentals of Mechanical Vibration - Course Notes 4

Note that if the dashpot is 'floating', which means it is mounted between two objects which
themselves can move, the force provided is q(x1 − x 2 ) .

5.2 DAMPED SINGLE DEGREE OF FREEDOM SYSTEM

Let's now modify the undamped system studied x
in Lecture 1 by inserting a viscous damper q
between the mass and the frame. It has a
damping coefficient of q N/(ms-1). Following
the steps presented in section 1.7, ( x, x , x, F m
positive to (say) the right), the free body
diagram is as shown in Figure 5.6. Notice that
the balanced vertical weight and reaction forces k
have been left off as they are perpendicular to
the motion of interest and hence play no part.
Figure 5.5 Damped 1dof system

Applying Newton II: ∑F x = mx

− kx − qx = mx
m
q k
Rearranging: x + x + x = 0 …(5.2)
m m

Figure 5.6 Free body diagram

Equation (5.2) is a linear, second order, homogeneous, constant coefficient, ordinary differential
equation. If you need to review the solution method, refer to mathematics texts 1. Such
equations can be tackled by assuming a solution in the form of a complex exponential. Hence,
let's assume that:

x(t ) = Xe st …(5.3)

where both X and s may be complex. Substituting (5.3) into (5.2) and simplifying yields the
characteristic equation 2.

q k
s2 + s+ =0 …(5.4)
m m

Solutions of (5.4) are those values of s that make (5.3) a solution to (5.2). Solving for s gives:

1
For example: P.V.O'Neil Advanced Engineering Mathematics pub. Wadsworth 1983, sections 2.1-2.5.
2
Later we shall prefer to write this equation as s 2 + 2ζω N s + ω N2 = 0

2018 Mechanical Engineering, Curtin University

Fundamentals of Mechanical Vibration - Course Notes 5

s1  1  − q q
2
k
=  ±   −4 
s2  2  m m m
  …(5.5)
−q  4mk 
= 1 ± 1− 2 
2m  q 

It is always useful to group parameters such as m, k, and q into dimensionless groups. Define:

q
ζ= …(5.6)
2 mk

ζ is a very important dimensionless group called the 'damping ratio'. The reason for this
terminology will become apparent shortly 3.

Furthermore, from (5.6) q = 2ζ mk . Dividing by 2m gives an alternative expression for the

term preceding the square brackets in (5.5).

q 2ζ mk k
= =ζ = ζω N …(5.7)
2m 2m m

where ω N is the natural frequency of the system with all damping removed (q = 0). Substituting
(5.6) and (5.7) into (5.5) gives:

s1   1 
 = −ζω N 1 ± 1 − 2  …(5.8)
s2   ζ 

We see that three cases present themselves, depending on the magnitude of the damping ratio, ζ.
Note that it is preferable to discuss the problem in terms of ζ rather than the damping coefficient
q which is not meaningful by itself. q, k, and m must be considered concurrently. None the less
note that ζ ∝ q .

Case I - Heavy Damping - ζ > 1

When ζ > 1, q > 2 mk . In this case the characteristic equation (5.5) has two, distinct, negative
roots given by (5.8). The general solution to the differential equation (5.2) is:

x(t ) = X 1e s1t + X 2 e s2t …(5.9)

The motion described by (5.9) is not oscillatory. Since both s1 and s2 are negative, the motion is
damped out to zero amplitude - without any overshoot of the equilibrium position.

3
The Greek character here is ζ (zeta). Later we shall make use of ξ (xi). Don't confuse them!

2018 Mechanical Engineering, Curtin University

Fundamentals of Mechanical Vibration - Course Notes 6

5
Initial displacement = 5

Initial velocity = +10, 0, -10 (top to bottom)

4
amplitude

0
0 1 2 3 4 5 6
time

Figure 5.7 Overdamped motion with various initial conditions.

A couple of things to note: 1

0.5

(i) the initial amplitude is X 1 + X 2 . 0

(ii) if the initial amplitude is small but the initial velocity -0.5

is large and negative, the motion will cross the axis, -1

as shown, and then decay back to the equilibrium

-1.5

position.
-2
0 1 2 3 4 5 6

Case II - Critical Damping - ζ = 1

When ζ = 1, q = 2 mk . The value of the damping coefficient that corresponds to this condition
is known as the 'critical damping coefficient'. From (5.6) we get:

qcritical = 2 mk …(5.10)

The roots of the characteristic equation are then two equal negative roots which, from (5.8), are
s1 = s 2 = −ω N . The general solution of (5.2) is:

x(t ) = ( X 1 + X 2 t )e −ω N t …(5.11)

The motion is still decaying and not oscillating - but only just. Any slight reduction in the
damping coefficient will result in the mass just overshooting the equilibrium position.

2018 Mechanical Engineering, Curtin University

Fundamentals of Mechanical Vibration - Course Notes 7

It is now possible to give another interpretation to the meaning of the damping ratio, ζ. Let q
and qcritical be the damping coefficient and the critical damping coefficient respectively, for the
same system. Then we have:

q
=ζ …(5.12)
qcritical

This is why ζ is called the damping ratio.

A critically damped system will have the smallest damping required for non-oscillatory motion.
Hence, the mass returns to the position of rest in the shortest possible time without overshoot.

Some examples where this motion is desirable are analogue meter movements and the recoil
mechanism on artillery guns.

Case III - Light Damping - ζ < 1

When ζ < 1, q < 2 mk . In this case the characteristic equation (5.5) has two, distinct, complex
roots given by (5.8). From (5.8):

s1 
 = −ω N ζ ± ζ − 1
s2 
2
[ ]

Now since ζ < 1,

s1 
[
 = −ω N ζ ± j 1 − ζ
s2 
2
] …(5.13)

For ease of manipulation, put a = −ω N ζ and b = ω N 1 − ζ 2 . Substituting into our assumed

solution (5.3) gives:

x(t ) = X 1e (a + jb )t + X 2 e (a − jb )t …(5.14)

where X1 and X2 may be complex (which means they contain the phase angle information
X 1 = X 1 e jφ ). Now we'll manipulate (5.14) into a form that is more easily interpreted.

[
x(t ) = e at X 1e jbt + X 2 e − jbt ]
= e at [X 1 cos(bt ) + X 1 j sin (bt ) + X 2 cos(bt ) − X 2 j sin (bt )]
= e at [( X 1 + X 2 )cos(bt ) + j ( X 1 − X 2 )sin (bt )]

Since x(t ) is real, the RHS must also be real, implying that ( X 1 + X 2 ) and j ( X 1 − X 2 ) must be
real. This can only be possible if X1 and X2 are complex conjugates. So, putting
( X 1 + X 2 ) = B1 and j ( X 1 − X 2 ) = B2 ,

2018 Mechanical Engineering, Curtin University

Fundamentals of Mechanical Vibration - Course Notes 8

x(t ) = e at [B1 cos(bt ) + B2 sin (bt )]

…(5.15)
= Ce at cos(bt + φ )

Note that C is real and the phase angle is now contained in φ.

(
x(t ) = Ce −ω Nζ t cos ω N 1 − ζ 2 t + φ ) …(5.16)

(
This result describes an oscillatory motion, cos ω N 1 − ζ 2 t + φ , which has its amplitude )
modulated in time by the decaying exponential multiplier, e −ω Nζ t . Figure 5.8 provides a typical
plot of a lightly damped (also termed 'under damped') response.

This motion is oscillatory but is not periodic in the ordinary sense of f (t ) = f (t + nT ) since the
amplitudes on successive pulses are different. However, the cosine term in (5.16) passes through
a complete cycle and becomes zero at intervals of

2π
T= …(5.17)
ω N 1−ζ 2

The frequency may then be interpreted as the number of these periodic zero crossings occurring
in unit time. Hence we say that the frequency is:

ω N 1−ζ 2
f = …(5.18)
2π

0.8
←y = exp(-0.05t)*cos(0.5t+5.2)
0.6

0.4

0.2
Amplitude

↓y = exp(-0.05t)
0

↑y = -exp(-0.05t)
-0.2

-0.4

-0.6

-0.8

-1
0 10 20 30 40 50 60 70 80 90 100
Time

Figure 5.8 Natural motion of an underdamped single-degree-of-freedom

spring-mass-damper system

2018 Mechanical Engineering, Curtin University

Fundamentals of Mechanical Vibration - Course Notes 9

The effect of viscous damping is therefore seen to reduce the natural frequency, compared to the
undamped natural frequency, by a factor of 1 − ζ 2 .

In many engineering applications, ζ << 1 so that the reduction in frequency of free oscillations is
generally very small. This is the reason that we initially ignored damping when considering
vibrating systems.

Realise though that damping is very important - no matter how small - in controlling resonant
conditions. This will become clear when forced vibrations are studied.

The effect of damping on frequency can be demonstrated with the aid of a dimensionless plot.
Squaring (5.18) and substituting ω N = 2πf N gives:

(2πf N ) (1 − ζ 2 )
2
f 2
=
(2π )2
so that:

2
 f 
  + ζ 2 = 1 …(5.19)
 fN 

which is the equation of a circle having unit radius. Plotting (5.19) gives a quadrant of a circle.

f
It is seen that when ζ is small, the ratio remains
1 fN
substantially equal to one. A typical range for the damping
ratio in an engineering application is 0.002-0.05. Values
higher than about 0.2 need to be deliberately designed.
1
Lastly, we note that the period and hence the implied
frequency of viscously damped free vibrations does not vary
0 ζ
1 with time or as the amplitude decays.

5.3 METHODS OF QUANTIFYING DAMPING

LOGARITHMIC DECREMENT

Several methods exist for describing the amount of damping a system contains. The damping
ratio has been defined above. The Quality Factor is applicable to forced systems and is a
measure of the 'sharpness' of the resonant peak. We shall not consider it in detail here.

The Logarithmic Decrement measures damping by the rate of decay of a freely vibrating system.
Consider a portion of a viscously damped one dof free vibration response, Figure 5.9.

2018 Mechanical Engineering, Curtin University

Fundamentals of Mechanical Vibration - Course Notes 10

Recall that the damped frequency is ω D = ω N 1 − ζ 2 . Using equation (5.16), the ratio of
amplitudes on two successive cycles, one period, T, apart is:

y0 (
Ce −ζω N t cos ω N 1 − ζ 2 t + φ )
=
(
y1 Ce −ζω N (t +T ) cos ω N 1 − ζ 2 (t + T ) + φ ) …(5.20)

= e +ζω N T

since the cosine term cancels as it is itself periodic. It is clear that in general, the ratio of any two
consecutive displacements, one period apart, is

yj
= e +ζω N T = eδ …(5.21)
y j +1

0.8

0.6

0.4
y0
0.2
y1
y2
0

-0.2

-0.4

-0.6

-0.8
T T
-1
0 5 10 15 20 25 30 35 40
Time

Figure 5.9 Logarithmic decrement

where δ = ζω N T …(5.22)

2018 Mechanical Engineering, Curtin University

Fundamentals of Mechanical Vibration - Course Notes 11

2π 2π
Note that T ≠ but T = . The quantity δ is called the logarithmic decrement and
ωN ω N 1−ζ 2
is equal to the dimensionless group δ = ζω N T . It is a measure of the amount of damping
present.

The ratio of displacements, n periods apart, can be found by forming a product of n ratios, each
equal to eδ :

y0 y1 y 2 y y y
× × ×  × n−2 × n−1 = 0 = e nδ
y1 y 2 y3 y n−1 y n yn

Solving for δ gives:

1 y 
δ = ln 0  …(5.23)
n  yn 

We also note that the logarithmic decrement can be expressed in terms of ζ alone by substituting
(5.17) into (5.22):

2π 2π
δ = ζω N ⋅ = …(5.24)
ω N 1−ζ 2 1
−1
ζ2

Note that δ is undefined for ζ ≥ 1 .

5.4 EXPERIMENTAL DETERMINATION OF THE

DAMPING COEFFICIENT

When the decay trace of y against time is obtained from experiment, the damping coefficient q
can be deduced from it by measuring values of displacements y0 and yn. Equating (5.23) and
(5.24), substituting (5.6) and solving for q gives:

4mk
q= 2
…(5.25)
 2πn 
  +1
 ln ( y0 y n )

2018 Mechanical Engineering, Curtin University

Fundamentals of Mechanical Vibration - Course Notes 12

5.5 TYPICAL VALUES OF DAMPING

In mechanical engineering systems, damping is normally quite a small fraction of critical
damping. As stated earlier, damping ratios above about 0.2 will be by design. The table below 4
shows some typical values.

Component or Material ζ
Car shock absorbers 0.1 to 0.5
Rubber 0.04
Riveted steel structures 0.03
Concrete 0.02
Cast Iron 0.002 to 0.01
Wood 0.003
Cold rolled steel 0.0006
Cold rolled aluminium 0.0002
Phosphor bronze 0.00007

Note the high damping of cast iron relative to steel. This is one reason why cast iron is used for
machine tool structures.

5.6 ENERGY DISSIPATED IN VISCOUS DAMPING

Let's consider the energy that is lost per cycle in viscously damped 'harmonic motion'. To do this
we are going to assume that the damping is sufficiently 'light' that over one cycle the motion is
very nearly simple harmonic. The instantaneous rate of energy dissipation with time (i.e. power)
is:

dU
Power loss = = Fq y = qy 2
dt

For simple harmonic motion, y = Y sin (ωt ) , and hence the energy dissipated over a complete
cycle will be:
2π
ω

∆U = ∫ qy 2 dt where y = −ωY cos ωt

0
2π
ω

= qω Y ∫ cos ωtdt
2 2 2

= πqωY 2 …(5.26)

Note that this is a function of both frequency and the square of the amplitude.

Specific Damping

4
R.F.Steidel An Introduction to Mechanical Vibrations, Wiley, 1979.

2018 Mechanical Engineering, Curtin University

Fundamentals of Mechanical Vibration - Course Notes 13

Specific damping capacity is the fractional part of total energy of the vibrating system which is
dissipated during each complete cycle, ∆U U . For a simple system with one degree of
freedom, the specific damping is directly related to the logarithmic decrement and the damping
ratio:

∆U πqωY 2 2πqω
= 1 2 = …(5.27)
U 2 kY k
2 q k q
= × 2π = 4π = 4πζ
2 k m 2 mk

2π
Recalling that δ = , which for small ζ is approximately δ = 2πζ , we see that :
1
−1
ζ2

∆U
= 4πζ ≅ 2δ …(5.28)
U

2018 Mechanical Engineering, Curtin University

Fundamentals of Mechanical Vibration - Course Notes 14

Review Questions

1. Explain the term ‘damping’ as it relates to mechanical vibration.

2. Is Coulomb damping represented by a constant magnitude force

opposing the motion?

3. What are the units of hysteretic damping?

4. The primary agent that dissipates energy in viscous damping is the

viscosity of the sheared fluid. True/False.

5. What are the units of the torsional damping coefficient, qθ ?

6. A linear, second order, homogeneous, constant coefficient, ordinary

differential equation always has a solution of the form ….

7. What is the difference between heavy, critical and light damping?

8. What magnitude of the damping coefficient leads to critical system

damping?

9. Does the presence of a small amount of damping alter the oscillation

frequency by a significant quantity compared to the undamped
frequency? Define ‘small amount’.

10. Explain in words the meaning of the ‘logarithmic decrement’.

11. Give one reason why machine tool structures are often made from cast
iron.

12. Our model for viscously damping free vibration of a single degree of
freedom system yields equation (5.16). Does this model indicate that
the motion eventually comes to a halt? If not, what comment can you
make concerning the model?

13. When modelling natural motion up to this point we have ignored

damping. Was this valid and if so why? State any constraints that
apply to the natural frequencies obtained by ignoring damping.

14. Are the ‘shock absorbers’ on motor vehicles vibration dampers? Are
they viscous dampers with the characteristic of equation (5.1)?

15. If a system has ζ = 0.02 , approximately how much of the remaining

system energy is dissipated during each cycle?

2018 Mechanical Engineering, Curtin University