Differential equations

B. Shibazaki

International Institute of Seismology and Earthquake Engineering,

Building Research Institute

1
 1. Differential equations

1.1 Classification of differential equations

A differential equation is an equation involving functions and their derivatives.

If the functions are real functions of one real variable, the derivatives occurring are
ordinary derivatives, and the equation is called an ordinary differential equation.

If the functions are real functions of more than one real variable, the derivatives
occurring are partial derivatives, and the equation is called a partial differential
equation.

(a) Ordinary differential equations

An ordinary differential equation of order n is a functional equation of the form
F [ x, y ( x), y ′( x ), y ′′( x ), L , y ( n ) ( x )] = 0 (1.1)
which involves the nth-order derivative of the unknown function y = y (x) . The order
of a differential equation is the order of the highest-order derivative appearing in the
equation.
An ordinary differential equation is called linear, if and only if it can be written in
the form

p0 ( x) y ( n ) + p1 ( x) y ( n −1) + L + pn −1 ( x) y′ + pn ( x) y = q( x) (1.2)

where q and the coefficients p0 , p1 ,L, pn are continuous functions of x . All other
differential equations are said to be nonlinear.

Example 1 The ordinary differential equation (1 + x 2 ) y′′ + xy′ + 2 x = 0 is linear, since

it is linear in y , y ′ and y ′′ .

Example 2 The ordinary differential equation y 2 y ′′ − ( y′) 3 = 0 is nonlinear.

2
 (b) Partial differential equations
A partial differential equation of order n is a functional equation of the form
∂φ ∂φ ∂ 2φ
F [ x1 , L , xm ; φ ; ,L, ; , L] = 0 (1.3)
∂x1 ∂xm ∂x1∂x2
which involves at least one nth-order partial derivative of the unknown function
φ = φ ( x1 , L , xm ) of two or more independent variables x1 , L , xm .

Example 3 Many important physical laws are expressed as partial differential

equations. Some of them are special cases of the following type of second-order
differential equations.

∂ 2φ ∂ 2φ ∂ 2φ ∂φ ∂ 2φ
+ + = a φ + b + c
∂x 2 ∂y 2 ∂z 2 ∂t ∂t 2
where φ denotes a function of x , y , z and t .
When a = b = 0 , the equation
∂ 2φ ∂ 2φ ∂ 2φ ∂ 2φ
+ + = c
∂x 2 ∂y 2 ∂z 2 ∂t 2
yields the wave equation that governs wave propagation in various continuous media.

3
1.2 Solutions of differential equations
The solution of a given differential equation implies the determination of the
functions y ( x) that satisfy the differential equation for all values of x in a specified
bounded or unbounded interval. Note that the solutions can be verified by
re-substitution.

(a) General solutions

The general solution of an ordinary differential equation of order n has the form
y = y ( x, c1 , c2 ,L, cn ) (1.4)
where c1 , c2 , L, cn are n essential arbitrary constants.

Example 4 The function y = Ae Bx contains two essential arbitrary constants A and B.

On the other hand, the function y = Ae B + x appears to contain two essential arbitrary
constants, but in fact contains only one, because
Ae B + x = Ae B e x = ce x with c = Ae B

Example 5 The function y = c1 sin x + c2 cos x is the general solution of the

differential equation y ′′ = − y , since it contains two essential arbitrary constants c1
and c2 , and y′′ = −c1 sin x − c2 cos x = − y .

4
(b) Particular solutions
Any solution of a differential equation (1.1) that can be obtained from the general
solution (1.4) by assigning values to the n arbitrary constants c1 , c2 , L, cn is called a
particular solution.

Typical problems involve the determination of a particular solution subject to

n initial conditions or
n boundary conditions
which determine the n arbitrary constants in the general solution.

Problem Solve the differential equation y ′′ = − y , given that y = 0 and y ′ = 1

when x = 0 .

Solution:
The general solution is given by y = c1 sin x + c2 cos x (Example 5).
Substituting x = 0 and y = 0 , we obtain c2 = .

In y ′ = c1 cos x − c2 sin x , we set x = 0 and y ′ = 1 , and obtain c1 = .

Hence the required solution is given by y = .

5
 2. First-order differential equations

A first-order differential equation expressible in the form

dy
= f ( x, y ) (2.1)
dx

has a solution y = y (x) through every point within a neighborhood in which f ( x, y )

is continuous.

2.1 Special types of first-order differential equations

The following special types of first-order differential equations are relatively easy to
solve.
(a) The variables are separable: dy / dx = φ ( x) /ψ ( y )
The solution of this type of differential equation can be obtained from

∫ψ ( y)dy =∫ φ ( x)dx + c (2.2)

Problem Solve y′ = ay ( a = constant).

Solution:

Problem Solve y′ + 5 xy = 0 .

Solution:

1
∫ y dy = −∫ 5xdx + c

6
 (b) Linear differential equations.
First-order linear differential equations can be written in the form

y′( x) + p( x) y ( x) = q( x) (2.3)

When q( x) = 0 , the equation is said to be homogeneous. The solution of the

homogeneous equation is given by

y = c exp[− ∫ p( x)dx] (2.4)

Proof: From dy / dx = − p( x) y ,
1
∫ y d y = −∫ p( x)dx + c
ln y = − ∫ p( x)dx + c
y = exp[− ∫ p ( x)dx + c]

Replacing e c with a new arbitrary constant c, we obtain

y = c exp[− ∫ p( x)dx]

7
When q( x) ≠ 0 , the equation is said to be nonhomogeneous.
In order to find the solution to this nonhomogeneous equation, we consider c in Eq.
(2.4) as a function of x . Then,

dy dc
= exp[− ∫ p ( x)dx] − p( x) y (2.5)
dx dx

Substituting of Esq. (2.4) and (2.5) into Eq. (2.3) yields

dc
= q( x) exp[ ∫ p ( x)dx] (2.6)
dx

Solving this differential equation, we get

c( x) = ∫ q( x) exp[ ∫ p ( x)dx]dx + c1 (2.7)

Therefore, from Esq. (2.4) and (2.7), we obtain the general solution of the original
nonhomogeneous equation as

y = exp[− ∫ p( x)dx]{∫ q( x) exp[ ∫ p( x)dx]dx + c} (2.8)

This method is called the variation of parameters.

8
 dy
Problem Solve + y=x.
dx

Solution: The general solution of the associated homogeneous equation, namely

dy / dx + y = 0 , is

y = ce − x

Considering c as a function of x , we substitute the above solution into the original

nonhomogeneous differential equation. Consequently,

dc
= xe x
dx

Solving this differential equation, we obtain

c( x) = ∫ xe x dx = xe x − ∫ e x dx + c1 = xe x − e x + c1

Thus, the general solution of the original differential equation is

y = x − 1 + ce − x

The general solution of a nonhomogeneous linear differential equation can be

expressed as the sum of any particular solution and the general solution of the
associated homogeneous equation.

9
 3 Second-order differential equations

3.1 Linear differential equations

Second-order linear differential equations can be written in the form

y′′( x) + p( x) y′( x) + q( x) y ( x) = r ( x) (3.1)

By using the linear differential operator

d2 d
L≡ 2
+ p ( x) + q ( x ) (3.2)
dx dx

we can express Eq. (3.2) as

Ly ( x) = r ( x) (3.3)

When r ( x) = 0 , the equation is said to be homogeneous.

When r ( x) ≠ 0 , the equation is said to be nonhomogeneous.

Ly = 0 is called the homogeneous equation associated with the original

nonhomogeneous equation Ly = r .

10
Theorem 1 linearity principle

If y1 and y 2 are any two solutions of Ly = 0 , then y3 = c1 y1 + c2 y2 , where c1

and c2 are arbitrary constants, is also a solution of Ly = 0 .

Proof: If y1 and y 2 are solutions of Ly = 0 , then, by definition,

Ly1 = 0 and Ly 2 = 0 .

Since L is a linear operator,

Ly3 = L(c1 y1 + c2 y2 ) = c1 Ly1 + c2 Ly2 = 0 .

Theorem 2 General solution of homogeneous equations

If y1 and y 2 are any two linearly independent solutions of Ly = 0 , then every
solution of Ly = 0 can be expressed in the form
y3 = c1 y1 + c2 y2 (3.4)
where c1 and c2 are arbitrary constants.

This theorem states that a linear combination of any two linearly independent solutions
of Ly = 0 yields the general solution of Ly = 0 .

11
For testing whether two solutions y1 and y2 of Ly = 0 are linearly independent or
not, we can use the function W of x defined by

W ( y1 , y2 ) ≡ y1 y2′ − y2 y1′ (3.5)

The quantity W ( y1 , y2 ) is called the Wronskian of the functions y1 and y 2 .

If W ( y1 , y2 ) is not identically zero in an interval (a, b) , where y1 and y 2 are

differentiable functions defined in (a, b) , then y1 and y2 are linearly independent in
( a, b) .

Example 1 Find the general solution of d 2 y / dx 2 + y = 0 .

Solution: The functions y1 = sin x and y2 = cos x are solutions of the given
differential equations, since

d 2 y1 d 2 (sin x)
= = − sin x = − y1
dx 2 dx 2
d 2 y2 d 2 (cos x)
= = − cos x = − y2
dx 2 dx 2

These two functions are linearly independent on any interval, since

W ( y1 , y2 ) = sin x(− sin x) − cos x cos x = −(sin 2 x + cos 2 x) = −1 ≠ 0

Hence the general solution is given by a linear combination of y1 and y 2 ;

y = c1 sin x + c2 cos x

12
Theorem 3 General solution of nonhomogeneous equations

If y3 = c1 y1 ( x) + c2 y2 ( x) is the general solution of Ly = 0 , and y = y (x) is any

particular solution of Ly = r , then the general solution of Ly = r is given by

y = y3 + y = c1 y1 ( x) + c2 y2 ( x) + y ( x) (3.6)

Proof: Since Ly3 = 0 and Ly = r ,

L( y3 + y ) = Ly3 + Ly = 0 + r = r

To obtain the general solution of a nonhomogeneous linear differential equation

Ly = r ,

(1) we must find two linearly independent solutions y1 and y2 of the associated
homogeneous equation Ly = 0 .

(2) a linear combination of these two solutions, y3 = c1 y1 + c2 y2 , gives the general

solution of Ly = 0 .

(3) then, adding any particular solution y = y (x) of Ly = r to the general solution of
Ly = 0 , we obtain the general solution of Ly = r .

13
3.2 Linear differential equations with constant coefficients
Homogeneous linear differential equations with constant coefficients can be written
in the form

y′′ + ay′ + by = 0 (3.7)

where a and b are constants. There is a method for completely solving this type of
differential equation.
The substitution of a trial solution of the form y = e tx , where t is a constant to be
determined, into Eq. (3.7) yields the quadratic equation

t 2 + at + b = 0 (3.8)

which is called the characteristic equation of the differential equation in Eq. (3.7).

The roots t1 and t2 of Eq. (3.8) are given by

− a ± a 2 − 4b
t1 , t 2 = (3.9)
2

It is necessary to consider three separate cases:

(a) a 2 − 4b > 0 ; t1 and t2 are real and unequal,

(b) a 2 − 4b = 0 ; t1 and t2 are real and equal,
(c) a 2 − 4b < 0 ; t1 and t2 are complex numbers.

14
Case (a) a 2 − 4b > 0

When t1 and t2 are distinct real roots of Eq. (3.8), the solutions y1 = e t1 x and
y 2 = e t 2 x are linearly independent, since

W ( y1 , y2 ) = (t 2 − t1 )e ( t1 +t2 ) x ≠ 0
Hence,
y = c1et1x + c2 et2 x (3.10)
defines the general solution of the differential equation in Eq. (3.7).

Case (b) a 2 − 4b = 0
When the roots t1 and t2 of the characteristic equation are real and equal, y1 = e t1 x
gives a solution for Eq. (3.7), but a second solution y 2 for which y1 and y2 are
linearly independent is absent.

The second solution y 2 may be obtained by using the method of reduction of order.
We obtain

y 2 = xet1 x (3.11)

Hence, the general solution of Eq. (3.7) is given by

y = (c1 + c2 x)e t1 x (3.12)

15
Case (c) a 2 − 4b < 0
When the roots of the characteristic equation are not real, then they must be complex
conjugates, since a and b are real numbers.
Let these complex roots be denoted by
t1 = α + iβ
and
t 2 = α − iβ ,

where α and β are real, and i = − 1 is the complex imaginary unit.

Then, the two linearly independent solutions of Eq. (3.7) can be written as

y1 = e (α +iβ ) x = eαx eiβx = eαx (cos βx + i sin βx) (3.13)

y2 = e (α −iβ ) x = eαx e −iβx = eαx (cos βx − i sin βx) (3.14)

iθ
by applying Euler’s identity e = cos θ + i sin θ .

The linear combinations

y1 − y2 y1 + y2
y3 = and y 4 = (3.15)
2i 2
yields

y3 = eαx sin βx and y4 = eαx cos βx (3.16)

Hence, the general solution of Eq. (3.7) is given by

y = eαx (c1 sin βx + c2 cos βx) (3.17)

16
 4. Vibrations
4.1 Undamped mechanical system
We consider a vertically suspended spring with a mass m attached to its lower end.
According to Hooke’s law, the force F exerted by the spring is proportional to the
extension x :

F = − kx

The force is acting along the upward direction,

and x increases as the spring is stretched
along the downward direction.

m x

d 2x
By using Newton's second law m × acceleration = m = −kx , we get
dt 2
d 2x
m + kx = 0
dt 2

17
Problem Obtain a general solution by assuming x = e pt .

The general solution is

x = c1 cos ω0t + c2 sin ω0t

where ω0 = k / m . ω0 is called the angular frequency

Problem
(a) Obtain the periodic time T required to complete one oscillation.
(b)Obtain the frequency f of an oscillation: the number of complete cycles per second,
namely, 1 / T .

18
We consider a mass suspended from a spring with the mass subject to a force that
changes with time.

The resistance force F is

d2 f
F = −V − kx
dt 2

We consider the simple harmonic form of f (t ) = cos ωt .

F = Vω 2 cos ωt − kx
Hence, applying Newton’s second law, we get
d 2x
m = Vω 2 cos ωt − kx
dt 2

d 2x k Vω 2
+ x = cos ωt
dt 2 m m

d 2x k
The general solution of the homogeneous equation + x = 0 is
dt 2 m

x = c1 cos ω0t + c2 sin ω0t

d 2x k Vω 2
For the particular solution of + x = cos ωt , we can try a solution of the
dt 2 m m
form:
x = b1 cos ωt + b2 sin ωt

Problem Determine b1 and b2 by substituting a trial solution into the differential

d 2x k Vω 2
equation + x = cos ωt .
dt 2 m m

19
 Vω 2
We obtain b1 = .
m(ω 02 − ω 2 )

Therefore, the particular solution is

Vω 2
xp = cos ωt
m(ω02 − ω 2 )
The general solution is
Vω 2
x = c1 cos ω0t + c2 sin ω0t + cos ωt
m(ω02 − ω 2 )

20
4.2 Damped vibrations
Homogeneous linear differential equations with constant coefficients can be written in
the form

&x& + 2γω0 x& + ω0 2 x = 0 (4.1)

where γ and ω0 denote the damping factor and natural frequency respectively.

Problem Find the general solution of Eq. (4.1).

The substitution of a trial solution of the form x = e pt , where p is a constant to be
determined, into Eq. (4.1) yields the quadratic equation

p 2 + 2γω0 p + ω0 = 0
2
(*)

The roots p1 and p 2 of Eq. (*) are given by

p1 , p2 = (−γ ± r 2 − 1)ω0

Case (a); γ 2 − 1 > 0 (overdamping)

x = c1e ( −γ + γ 2 −1 )ω0 t
+ c2 e ( −γ − γ 2 −1 )ω0t

Case (b); γ 2 − 1 = 0 (critical damping)

x = e −ω0t (c1 + c2t )

Case (c); γ 2 − 1 < 0 (underdamping)

p1 , p2 = (−γ ± i 1 − γ 2 )ω0

x = e −γω0t {c1 cos( 1 − γ 2 ω0t ) + c2 sin( 1 − γ 2 ω0t )}

21
Problem Find the particular solution of Eq. (4.1) that satisfies the initial conditions
x = 1 and x& = 0 when t = 0 .

22
Fig. 4. Changes in displacement with time for three cases: under-damping, critical
damping, and over-damping, when the initial conditions are x(0) = 1 and x& (0) = 0 .

23
4.3 Forced vibrations

In this section, we investigate the equation of seismometers

&x& + 2γω0 x& + ω0 2 x = −V&f&(t ) (4.2)

where γ and ω0 are constants. We consider the simple harmonic form of

f (t ) = cos ωt . In this case (4.2) becomes

&x& + 2γω0 x& + ω0 2 x = Vω 2 cos ωt (4.3)

For solving this equation, we introduce y that satisfies the conjugate equation

&y& + 2γω0 y& + ω0 2 y = Vω 2 sin ωt (4.4)

so that, if z = x + iy , we have

&z& + 2γω0 z& + ω0 2 z = Vω 2 eiωt (4.5)

Putting z = Aeiωt into Eq. (4.5) as a trial solution yields

[−ω 2 + 2iγω0ω + ω02 ] A = Vω 2 (4.6)

Therefore,
Vω 2 ω02 − ω 2 − 2iγω0ω
A= = Vω 2

ω02 − ω 2 + 2iγω0ω (ω02 − ω 2 ) 2 + 4γ 2ω02ω 2

Vω 2
= e iφ (4.7)
[(ω 0 − ω ) + 4γ ω 0 ω ]
2 2 2 2 2 2 1/ 2

where
− 2γω0ω
tan φ = (4.8)
ω02 − ω 2

24
Problem

⎛ a b ⎞
a + ib = a 2 + b 2 ⎜⎜ +i ⎟ = a 2 + b 2 e iφ tan φ = b / a
⎟
⎝ a +b a2 + b2 ⎠
2 2

Using the above formula,

ω02 − ω 2 − 2iγω0ω =

25
Thus, a particular solution of Eq. (4.2) is

Vω 2
z= e i (ωt +φ ) (4.9)
[(ω0 − ω ) + 4γ ω0 ω ]
2 2 2 2 2 2 1/ 2

We then deduce that a particular solution of Eq. (4.9) is

Vω 2
x = Re{z} = cos(ωt + φ ) (4.10)
[(ω02 − ω 2 ) 2 + 4γ 2ω02ω 2 ]1/ 2

The general solution of Eq. (4.2) is given by the sum of a general solution of Eq. (4.1)
and a particular solution Eq. (4.10).

Amplitude spectrum
Vω 2
A(ω ) =
[(ω02 − ω 2 ) 2 + 4γ 2ω02ω 2 ]1/ 2
V
=
⎛ω ⎞ ⎛ω2 ⎞
2

[(⎜ 0 ⎟ − 1) 2 + 4γ 2 ⎜⎜ 02 ⎟⎟]1/ 2
⎝ω ⎠ ⎝ω ⎠

Phase spectrum
ω0
− 2γ
tan φ = ω
⎛ ω0 ⎞
2

⎜ ⎟ −1
⎝ω ⎠

26
Amplitude spectrum

Phase spectrum

27