Chapter 5

Applications of second-order
differential equations; behavior of
dynamical systems
In the last chapter we have given a substantial effort to study linear
second order differential equations. The main reason for doing this
is that these differential equations form a basis for modeling a large
class of physical problems. If we look back at the behavior of the
solutions of the examples we have presented in the last chapter, we
will observe that there were problems whose solutions were
exponentially growing; problems whose solutions were
exponentially decaying; and there were problems whose solutions
were sinusoidally decaying. A large category of problems in
engineering behaves or designed to behave so that its response is
exponentially decaying or sinusoidally decaying. These include
very simple examples like vibration of blades of a fan, a vehicle in
motion, electrical circuits. We will begin with the study the motion
of a mass on a spring in detail because understanding the behavior
of this simple system is the first step in the investigation of more
complex dynamic systems. Further, the principles involved are
common to many problems.

The model
We start with a helical spring of length l (Figure 5.1). The spring is
hanging freely. We hang a mass m on the spring. As a result, the
 l l+
k Cd
Unstretched
spring
m
System at
rest F(t) y

System in
motion
Figure 5.1 Spring-mass-damper
system
spring elongates by an amount . At this point we apply a time-
dependent force F(t) on the spring. As the mass moves in response
to the force, it also experiences a frictional force. We want to
analyze the displacement of the spring, y(t). We will assume the
displacement is positive downwards, as the acceleration due to
gravity, g acts downwards.

The fundamental tools in our hands to analyze the motion are

Newton’s laws of motion. When the mass is just hanging with the
stretched spring, it is at rest. Therefore, we use the first law of
motion.
(5.1)
where F are the forces acting on the mass. At this state, the forces
acting on the mass are the downward (positive) weight, w, and the
upwards force of spring Fs. Therefore
w – Fs = 0 (5.2)
A simple model for the spring force for small displacements is to
use the Hooke’s law. We will write Fs = k, where k is the spring
constant, and  is the elongation of the spring. Therefore,
mg – k = 0 (5.3)

When the body is in motion, we will apply the second law of

motion,
(5.4)
where a is the acceleration of the mass. The forces acting on the
mass are the weight, w, the spring force, Fs, the damping force
(frictional force), Ff and the external force F(t). Therefore, from
(5.4)
ma = w + Fs + Ff + F(t) (5.5)

Since the position of the mass at any time is y(t). Therefore, the
stretch of the spring will be  + y(t). Therefore the spring force is
Fs = −k[ + y(t)].

The frictional force depends on many factors – the medium of

motion, the shape of the mass, the velocity of the mass. A simple
model is to approximate the friction force as proportional to the
velocity. Therefore, we will propose Ff = Cdv, where Cd is referred
to as the damping coefficient and v is the velocity. Substituting all
these in equation (5.5), we obtain
ma = mg − k[ + y(t)] − Cdv + F(t) (5.6)
The sign of the frictional force has been taken negative, as it
always opposes the motion. From equation (5.3), we have
mg – k = 0. Therefore, we have
 ma = − ky(t) − Cdv + F(t)
or ma + Cdv + ky(t) = F(t) (5.7)
Writing the acceleration and the velocity as derivatives of the
displacement, from (5.7) we will obtain

(5.8)

Equation (5.8) is the model equation for a spring-mass-damper

system. Since the mass m, the damping coefficient Cd and the
spring constant k are all constants, equation (5.8) is a
nonhomogeneous linear second-order differential equation with
constant coefficient. It is important to understand that
equation (5.8) is at best an approximate model. This is because our
model for spring force and friction force should be viewed as
approximations.

To complete the model, we specify the initial conditions

y(0) = y0 y′(0) = y1 (5.9)
Now that we have a physical problem, we can have a better
understanding of the initial conditions. The condition y(0) specifies
the initial position, and y′(0) specifies the initial velocity. It would
be important to investigate how changes in these conditions
changes the dynamic behavior of the system.

We can interpret equation (5.8) with a broader perspective. We

have a spring-mass-damper system. We give an input F(t) to the
system, initiate the system with some initial conditions, and ask for
the response y(t) from the system. So, in broad terms,
equation (5.8) can be seen as an input-output system.
Equation (5.8) is referred to as the equation of forced vibration, as
there is an external force F(t) acting on the system. Eventually we
will investigate the equation in its full form, but we will get there
gradually. In the first step, we will assume that no external forces
are acting on the system, F(t) = 0. In other word we will assume
the homogeneous form of equation (5.8).

(5.9)

The perspective of equation (5.9) is something like this: there was

a system that was responding to the external force F(t). At some
point, the external force is removed. We want to investigate the
motion of the system after that. Equation (5.9) is referred to as the
equation of damped free vibration. We will investigate one more
simpler system where we will assume that there is no damping in
the system, Cd = 0.

(5.10)

This system will be referred to as undamped free vibration.

Therefore, our step-by-step investigation of equation (5.8) will
proceed as follows: first we will investigate undamped free
vibration, then damped free vibration, finally forced vibration.

Undamped free vibration

The equation of motion is (5.10).

(5.10)

The characteristics equation is

mr2 + k = 0 (5.11)
Since m and are always positive, the roots of this equation are
complex.

(5.12)

The general solution is

(5.13)
In discussing the solution (5.13) it is often convenient to write it as
(5.14)
or (5.15)
Comparing (5.14) and (5.15) we can write
and
Therefore
and
For a typical set of boundary conditions, the solution of (5.14) is
shown in Figure 5.2. The graph is a displaced cosine wave that
describes a periodic, or simple harmonic motion of the mass. The
period of the motion is

R cos δ δ δ+ δ+2

0 t

−R

Figure 5.2 Undamped free vibration

 (5.16)

The circular frequency 0 measured in radians per unit time, is

called the natural frequency of the vibration. It is a permeant
property of the system and is an extremely important entity during
design of an engineering system. The maximum displacement R of
the mass from equilibrium is the amplitude of the motion. The
dimensionless parameter  is called the phase, or phase angle, and
measures the displacement of the wave from its normal position
corresponding to  = 0.

Note that the motion described by equation (5.14) has a constant

amplitude that does not diminish with time. This reflects the fact
that, in the absence of damping, there is no way for the system to
dissipate the energy imparted to it by the initial displacement and
velocity. Therefore, the motion is perpetual. Further, for a given
mass m and spring constant k, the system always vibrates at the
same frequency , regardless of the initial conditions. However,
the initial conditions do help to determine the amplitude of the
motion. Finally, observe from equation (5.16) that the period T
increases as the mass m increases, so larger masses vibrate more
slowly. On the other hand, the period T decreases as the spring
constant k increases, which means that stiffer springs cause the
system to vibrate more rapidly.

We have not discussed the effect of the initial conditions on the

solution. We are doing this through an example.

Example 5.1
A body of weight w = 9.81 N is hanged on a spring. As it is
hanged, it is stretched by 5 cm. The body is stretched further by
5 cm. The initial velocity of the body will have three different
cases. Assuming that there is no friction, discuss the behavior of
the motion of the body under the three different cases.

The motion of the body will follow equation (5.10) whose solution
is (5.13) or (5.14). With the given information, the mass of the
body, m = w/g. Assuming the acceleration due to gravity
g = 9.81 m/sec2, the mass m = 9.81/9.81 = 1 kg.

The spring stretches by 5 cm (0.05 m) when the weight is hanged

on the spring. Therefore, the spring constant is
k = 9.81/0.05 ≈ 196 N/m. Therefore, the natural frequency of
vibration is

Following (5.13), the general solution is

(5.17)
To find the values of A and B we need to assign boundary
conditions. We have been told that the initial displacement is
0.05 m. Therefore, y(0) = 0.05. In this problem, we intend to
investigate the effect of the initial velocity on the final solution. We
will solve the problem with three initial velocities, (i) y′(0) = −0.6,
(ii) y′(0) = 0.0, and (iii) y′(0) = 0.6.

Case (i), y(0) = 0.05, y′(0) = −0.6

We differentiate (5.17) to obtain
 (5.18)
Substituting y(0) = 0.05 in (5.17), we obtain A = 0.05. Substituting
y′(0) = −0.6 in (5.18), we obtain B = −0.057. Substituting these
values in (5.17), we obtain the final solution.
(5.19)
For better understanding of the solution, we want to write the
solution in the form (5.14). For this

Therefore,
(5.20)
We can do the algebra similarly and find the solutions for other
cases. For case (ii)
(5.21)
For case (iii)
(5.22)

These three solutions are shown in Figure 5.3 on the same graph
for comparison. We observe that in all three cases the body
executes a simple harmonic motion. We should also notice that
when we impart an initial velocity to the body – in any direction,
the amplitude of vibration increases; and the increased amplitude
persists. The reason for this is that the introduced velocity can be
interpreted as ‘momentum per unit mass’. As a result, the
amplitude rises, and because there is no damping in the system, the
increased amplitude persists.
 Case (iii), y′(0) = 0.6
Case (ii), y′(0) = 0.0
y(0) =
0.05

Case (i), y′(0) = −0.6,

Figure 5.3 Effect of initial conditions.

Damped free vibration

The equation of motion is (5.9)

(5.9)

The characteristics equation is

(5.23)
The roots of this equation are

(5.24)

Depending upon the sign of , the roots are of three

cases.
Case I. . Real roots. The solution is of type.
Case II. . Equal roots. The solution is of type.

Case III. . Complex conjugate roots. The solution

is of type.

/( )
,

We will discuss each of these cases separately.

Case I.
This is a case of overdamping. Both roots r1 and r2 are real. Since
; m, Cd and k are all positive, .
Therefore, both the roots r1 and r2 would be negative. Therefore,
both exponentials in the solution would be decaying exponentials.
This means that as t → ∞, the solution would approach zero.
Intuitively, this makes sense, because with high damping, any
momentum that is imparted to the system with initial conditions
will dissipate quickly. We see further properties of this case
through an example.

Example 5.2
A spring-mass-damper system has the governing equation
4y + 8y′ + 3y = 0 (5.25)
We will investigate this system for different initial conditions.

The general solution of (5.25) is

(5.26)
To find the constants c1 and c2, we will need to differentiate (5.26).
(5.27)
We set initial displacement y(0) = 1. We are looking into four
initial velocities. Intuitively what the situation means is that we
will extend the mass on the spring by one unit, then give a sudden
pull (positive velocity), or a sudden push (negative velocity). The
results are as follows.
y′(0) = 1 (5.28)

y′(0) = 0 (5.29)

y′(0) = −1 (5.30)

y′(0) = −2 (5.31)
For comparison of their dynamic behavior, all the solutions are
shown in Figure 5.4.

Case II
This is a case of critical damping. This is the transition between
overdamping and underdamping and is more of theoretical interest
rather than any practical.

Case III
This is the most interesting case. This is a case of underdamping.
The roots are complex conjugates, and the solution is
/( )
, (5.32)
 y′(0) = 1

y(0) = 1 y′(0) = 0

y′(0) = −1
y′(0) = −2

Figure 5.4. Underdamped free vibration. Effect

of initial conditions.
where

(5.33)

For the sake of discussion, we write equation (5.32) as

/( ) (5.34)
As stated earlier
and
Equation (5.34) indicates that the solution is a cosine curve bound
/( )
between the amplitudes

Although the motion is not periodic, the parameter  determines

the frequency with which the mass oscillates back and forth;
 /( )
𝑅𝑒

R cos 
   



/( )
−𝑅𝑒

Figure 5.5 General nature of overdamped

vibration.
consequently,  is called the quasi-frequency. Let us recall, the
natural frequency of the system is 0. By comparing  with the
frequency 0 of undamped motion, we find that

(5.35)

By analogy Td = 2/ is called the quasi-period of the motion.

Example 5.3
The motion of a certain spring-mass-damper is governed by the
equation
(5.36)

This is equation of a damped motion. For later comparison, first

we find the solution of the undamped motion. The undamped
equation is
 (5.37)
Since we have solved similar problems earlier, we are skipping the
details. The final solution is
(5.38)
or (5.39)
We have also obtained solutions to problems similar to
problem (5.36) earlier. The final solution is

(5.40)

We write this as
(5.41)
The graph of equation (5.41) along with equation (5.39) are shown
in Figure 5.6.

y + 0.125y′ + y = 0
y + y = 0

Figure 5.6 Example 5.3. Overdamped

vibration.
From equation (5.39), the natural frequency of vibration of the
system is 1 rad/sec. With damping, the quasi-frequency has
reduced to 0.988 rad/sec. Intuitively, it is expected that as damping
will dissipate momentum, the mass will vibrate slower.

Forced vibration
This is what we have been preparing ourselves for. This class of
vibration covers a large variety of problems. Just to mention a few:
(i) vibration of a deck of a bridge because of vehicular traffic over
it, (ii) vibration of a river bridge because of wind passing across it
as well as the water flow underneath creating Kármán vortices
(iii) three-directional vibration of a pedestrian overbridge because
of the walking motion that involves movements in all three
directions, (iv) vibration of the blades of a ceiling fan or propeller
of an engine primarily because of the Kármán vortex behind it,
(v) vibration of tall buildings because of wind flowing across it and
many other similar problems can be modelled using spring-mass-
damper model with a forcing function.

In general, we will be seeking the behavior of the

nonhomogeneous equation (5.8).

(5.8)

where F(t) is an external forcing function. As we can envision,

depending on a physical problem, the external forcing function can
be of variety of algebraic forms. We will investigate a wide range
of forcing functions in the next chapter after we study Laplace
Transform. For the time being, we consider a simpler forcing
function of F(t) = F0 cos t. Though seemingly simple, this
 Vertical

Lateral

Longitudinal

Figure 5.7 Vibration of a pedestrian overbridge.

periodic forcing function can still bring out the critical details of
the behavior of a dynamic system. Nevertheless, the dynamic
behavior of complicated systems can also display sufficiently
periodic nature. As an example, Figure 5.7 shows the vibration
nature of a pedestrian overbridge. We can clearly see that the
vertical and lateral components are substantially periodic.

However, even for simplest cases also, the algebra can be quite
overwhelming. Therefore, we will pass over the algebraic details
and present the final results. We will, however, discuss the
behavior of the problem through examples.

Example 5.4
Find and discuss the solution of the initial value problem
 y(0) =0, y′(0) = 3 (5.42)

Recalling our earlier knowledge, the solution of equation (5.42)

will have a homogeneous part and a particular part.
y(t) = yh(t) + yp(t) (5.43)
The homogeneous solution can be found as
yh(t) = c1e−t/2cos t + c2e−t/2sin t (5.44)
Assuming the particular solution to be
yp(t) = A cos t + B sin t
Going through the exercise we have gone through before, we find
the solution
(5.45)
The final solution is

(5.46)

After substituting the initial conditions, the final solution is

(5.47)

The solution is shown graphically in Figure 5.8.

We observe the nature of the solution. The homogeneous portion of

the solution, being enveloped with a negative exponential, decays
 Total solution

Steady state
solution

Transient
solution

Figure 5.8 Transient solution, steady state

solution and total solution of a forced
vibration problem.
as time increases. However, the particular portion of the solution
survives and continues.

Let us formalize the solution. The differential equation is

(5.48)
The solution of (5.48) will be
y(t) = yh(t) + yp(t) (5.49)
The homogeneous solution yh(t) will depend on the roots of the
homogeneous equation
my + Cdy′ + ky = 0 (5.50)
We have discussed the solution of (5.50) in great details. If the
roots of (5.50) are real, the solution would be exponentials; if the
roots are complex conjugates, the solutions are sinusoidal. In both
cases the solutions decay to zero as time increases. Therefore, yh(t)
decays and ‘dies-out’. This is referred to as the transient
response. The particular solution yp(t) survives, and is referred to
as the steady state response. Therefore, we will concentrate on the
nature and behavior of the particular solution. To find the particular
solution yp(t), we assume
yp(t) = A cos t + B sin t (5.51)
We go through the usual exercise of computing and ,
substitute in equation (5.48), collect terms, equate coefficients, the
values of A and B are obtained as follows.

(5.52a)

(5.52b)

where is the natural frequency of vibration.

It is worthwhile to write equation (5.51) in the form

yp(t) = R cos(t – ) (5.53)
where, still
and
Combining (5.52a) and (5.52b), we obtain

(5.54a)

(5.54b)
 0

0.25
0.5
Cd = 2 1

/0
Figure 5.9 Amplitude amplification factor
as a function of the frequency ratio.
We are particularly interested to investigate the behavior of the
steady state amplitude R varies with the frequency of the forcing
function . After some algebra, we write (5.54a) as

(5.55)

R is the amplitude of steady state vibration, k is the spring constant,

F0 is amplitude of the external sinusoidal force. Therefore, F0/k is
the initial stretch of the string. The ratio Rk/F0 is referred to as the
amplitude amplification factor. For the purpose of
demonstration, we choose k = 1, and plot the amplitude function as
a function of the frequency ratio /0. Figure 5.9 shows the result.
We notice when the damping factor is not negligible, with increase
in the frequency ratio, the amplitude amplification rises and
decays. This is an experience we may have experienced in cars. As
we increase the speed of the engine from start, the vibration of the
engine increases, and reduces beyond a certain point. However, if
the damping is negligible, the amplitude amplification becomes
unbounded as /→1. This is the phenomenon of resonance.
Resonance is always a design issue in all mechanical and physical
structures.

We further investigate the dynamic behavior of a vibrating system

under different conditions through an example.

Example 5.5
We will take this problem from Example (5.3) and find the solution
for different frequencies of the forcing function.
(5.56)

For this system, the natural frequency of vibration 0 = 1 rad/sec.

From Example (5.3), the homogeneous solution is

(5.51)

We assume the particular solution to be

(5.52)
After some algebra, we find

With this we can construct the solution

(5.53)
 (a)  = 0.3 (b)  = 1.0

(c)  = 2.0 (d)  = 4.0

Figure 5.10 Dynamic behavior of the spring-mass-

damper system for four values of forcing frequencies.
Substituting the initial conditions, with some more algebra, we can
obtain the constants c1 and c2

The dynamic nature of the system for three frequencies,

 = 0.3 rad/sec,  = 1 rad/sec,  = 2 rads/sec and  = 4 rads/sec
are shown respectively in Figure 5.10a, b, c, d. In each graph, the
vibration of the mass is shown in blue and the forcing function is
shown in red.

Figure 5.10a shows the low frequency case, /0 = 0.3. After the
initial transient response is substantially damped out, the remaining
steady-state response is essentially in phase with the excitation,
and the amplitude of the response is somewhat larger than the
static displacement.
The resonant case, /0 = 1.0, is shown in Figure 5.10b. Here, the
amplitude of the steady-state response is eight times the static
displacement, and the figure also shows the predicted phase lag of
/2 relative to the external force.

The case of comparatively high frequency excitation for  = 2.0 is

shown in Figure 5.10c. Observe that the amplitude of the steady
forced response is approximately one-third the static displacement
and that the phase difference between the excitation and the
response is approximately .

The case of even higher frequency for  = 4.0 is shown in

Figure 5.10d. In this case the amplitude of the steady state
response has reduced further to almost one-sixth. It is almost as if
the vibration of the mass is not able to keep up with the frequency
of the forcing function.

Forced vibration with no damping

This will be our last topic before we move away from the topic of
forced vibration. With no damping, Cd = 0. Therefore, the
differential equation is
(5.54)
The natural frequency of vibration is still . From our
previous discussion, we understand that two cases would be of
interest: (i) if the frequency of the forcing function  ≠ 0, and
(ii) if the frequency of the forcing function  = 0.
First, we will look into the case of  ≠ 0. The general solution
would be of the form
y(t) = c1cos 0t + c2sin 0t + A cos t + B sin t (5.55)
Of particular interest would be the case of no initial displacement,
and no initial velocity. Therefore, y(0) = 0 and y′(0) = 0. After
going through all the algebra, The solution would be

(5.56)

This is the sum of two periodic functions of same amplitude, but

different frequencies. Using trigonometric identities of cos(A ± B),
from (5.56), we write

(5.57)
Pause for a while and take some time to ponder over (5.57). If
 ≠ 0, but  ≈ 0, consequently will be a
rapidly fluctuating function compared to .

Example 5.6
Find the solution of the initial value problem
y(0) = 0, y′(0) = 0 (5.57)
The solution is
(5.58)
Figure 5.11 shows the wave form. This type of motion, possessing
a periodic variation of amplitude, exhibits what is called a beat.
 y(t) = 2.778 sin (0.1t)
y(t) = 2.778 sin (0.1t) sin(0.8t)

Figure 5.11 A beat.

The last sine function is what the musicians listen to when they
tune their instruments.

We return to equation (5.54) and consider the case of resonance

where  = 0, that is the natural frequency of vibration, and the
frequency of the forcing function are same. From our discussion in
chapter 4, this is a case of repeated roots. Let us return to problem
(5.54), the solution in this case would be

(5.59)

Example 5.7
y(0) = 0, y′(0) = 0 (5.60)
The general solution would be
(5.61)
 1
𝑡 sin 𝑡
4
1
𝑡
4

Figure 5.12 Wave form of resonance.

To satisfy the initial conditions, c1 = 0 and c2 = 0. Thus, the
solution to the problem is
(5.62)
Figure 5.12 shows the wave form. This is the waveform of
resonance. The amplitude of vibration becomes unbounded. We
observed this behavior in Figure 5.9 for Cd = 0, where the
amplitude amplification factor rose unbounded for /0 =1.

Even though this dynamic behavior has been obtained for the case
when damping coefficient is zero, this behavior can be seen for
low damping also. We have seen a similar behavior in
Figure 5.10(b) where the damping was low, and /0 = 1. In
mechanical systems this dynamic behavior can be disastrous
sometimes.

Modelling Electrical Circuits

 Resistance, R Having discussed the dynamic
behavior of mechanical systems
in considerable detail, we will
now briefly look into electrical
E(t) = E0 sin t Inductor, L systems, where we will see
remarkable similarities between
Capacitor, C these two systems.

The elements of an electrical

Figure 5.13 A series RLC circuit. circuit we are considering are a
resistance (R), a capacitor (C)
and an inductor (L). The circuit can be constructed by placing them
in several combinations – series and parallel – each serving a
different specific design purpose. As an example, we will discuss
the dynamic behavior of a series RLC circuit as shown in
Figure 5.13. The current I, measured in amperes (A), is a function
of time t. The resistance R in ohms (), the capacitance C in
farads (F), and the inductance L in henrys are all positive and are
assumed to be known constants. The impressed voltage E(t) in
volts (V) is a given function of time. Another physical quantity that
enters the discussion is the total charge Q in coulombs (C) on the
capacitor at time t. The relationship between the charge Q and
current I is

(5.63)

The flow of current in the circuit is governed by Kirchhoff’s

second law, the Kirchhoff Voltage Law (KVL), which states that in
a closed circuit the impressed voltage is equal to the sum of the
voltage drop in rest of the circuit. According to the elementary
laws of electricity, we know that
 The voltage drop across the resistance is

The voltage drop across a capacitor is

The voltage drop across an inductor is

Therefore, by Kirchhoff’s law

(5.64)

Substituting I from equation (5.63), we obtain the differential

equation

(5.65)

for the charge Q. The initial conditions are

Q(t0) = Q0, Q′(t0) = I0 (5.66)
Alternatively, we can obtain a differential equation for the current I
by differentiating equation (5.64) with respect to t and substituting
dQ/dt from equation (5.63)

(5.67)

With the initial condition

I(t0) = I0, I′(t0) = I′0 (5.68)
The set of equations (5.65) along with initial conditions (5.66) or
equation (5.67) along with initial conditions (5.68) are the
differential equation for an RLC circuit in series. We can see that
these are second-order differential equations just like the governing
equations for a spring-mass-damper system. Therefore, everything
we have learnt about undamped system, natural frequency, quasi
natural frequency, resonance, transient response, steady-state
response are all carried over here.

An RLC circuits are conceptually very important in electrical

system design and analysis. These elements, placed in different
configurations form the backbone for many entities such as
frequency filters, amplifiers and oscillators.