Chapter 8
The wave equation
8.1 Introduction
The principal aim of this chapter is to understand and develop the mathematics required to un-
derstand waves. Intuitively waves can be thought of as oscillations together with some transport
of energy.
We have different types of waves:
1. mechanical waves: sound waves, water waves, elastic waves (these propagate through a
medium);
2. electromagenetic waves: light, radio waves, microwaves, wi-fi, x-rays;
3. gravitational waves.
Electromagnetic and gravitational waves do not require a medium; they travel through the
vacuum.
8.2 Deriving the wave equation
Consider a string, which has mass ρ per unit length, stretched between two points x = 0 and
x = L:
× ×
Now displace it vertically a small amount in the y-direction so its height becomes
y = u(x, t).
```
6 ```
× ` ``` ×
u(x, t)
We will assume that u remains very small relative to L, so that the angle θ between the string
and the horizontal is small.
We will neglect terms of order θ2 . This allows us to approximate all the trigonometric functions,
so we will take
sin θ = tan θ = θ cos θ = 1.
77
CHAPTER 8. THE WAVE EQUATION 78
Let’s look at a small segment of string, from horizontal position x to x + δx. (Note: δx is a
quantity in its own right – this does not mean δ multiplied by x.) This segment has mass ρ δx
and we will apply Newton’s second law to it.
T2
3θ + δθ
6
θ
6 u(x + δx, t)
T1
u(x, t)
? ?
The horizontal component (with no acceleration possible because of the fixed ends) is
T1 cos θ = T2 cos (θ + δθ) T1 = T2
so because the angles are all small, the tension is the same everywhere in the string.
For the vertical component, we do have acceleration:
∂2u
ρ δx = T2 sin (θ + δθ) − T1 sin θ
|{z} ∂t2
|{z}
mass
acceleration
and when we use our approximation for sin and the fact that T1 = T2 = T , say, we get
∂2u
ρ δx = T δθ.
∂t2
(Here, again, δθ does not mean δ times θ but the difference in θ between the two ends of our
string segment.) Now we divide by δx and take the limit δx → 0:
∂2u δθ ∂2u ∂θ
ρ =T ρ =T .
∂t2 δx ∂t2 ∂x
Now all we need is a relation between θ and u. Look again at the diagram, and now focus on
the right-angled triangle whose hypotenuse is our string. Its angle is θ, its base is δx, and its
height is u(x + δx, t) − u(x, t). So we can deduce
u(x + δx, t) − u(x, t)
tan θ = .
δx
Taking the limit δx → 0 and using our small-angle approximation again, this becomes
∂u
θ= .
∂x
Finally we substitute this back into our previous equation to have
∂2u ∂2u
ρ 2
= T 2.
∂t ∂x
We identify T /ρ = c2 to get the standard form of the wave equation:
∂2u 2
2∂ u
= c .
∂t2 ∂x2
CHAPTER 8. THE WAVE EQUATION 79
8.3 Some solutions to the wave equation
8.3.1 Travelling wave
Let’s look at the function
u(x, t) = g(x − ct)
where g is any continuous, twice differentiable function of a single variable. We can use the
chain rule to calculate partial derivatives of u: we introduce y = x − ct so that u(x, t) = g(y) to
make the process clearer
∂u dg ∂y
= = g ′ (x − ct)
∂x dy y=x−ct ∂x
∂u dg ∂y
= = −cg ′ (x − ct)
∂t dy y=x−ct ∂t
∂2u dg ′ ∂y
= = g ′′ (x − ct)
∂x2 dy y=x−ct ∂x
∂2u d(−cg ′ ) ∂y
= = c2 g ′′ (x − ct)
∂t2 dy y=x−ct ∂t
we can see that utt = c2 uxx so this function satisfies the wave equation for any g.
What does this function look like? Suppose for illustration that g(x) looks like this:
Then this is also what u(x, t) looks like at t = 0. At some later time (say t = 1) we have
u(x, t) = g(x − c)
i.e. we take the function value of g from a point to the left of our position x. The shape of the
function remains the same but it is shifted to the right:
This solution is a travelling wave which moves to the right with speed c.
Example 8.1 Exercise: show that the leftwards travelling wave
f (x, t) = g(x + ct)
also satisfies the wave equation.
CHAPTER 8. THE WAVE EQUATION 80
8.3.2 Simple harmonic wave
A simple harmonic wave has the form
A sin [k(x − ct) − δ].
At a fixed moment, t∗ say, we have
u = A sin [kx − (kct∗ + δ)]
which looks like a sine wave sin kx with a phase shift that depends on time.
At a fixed position in space, x∗ say, we have
u = A sin [−kct + kx∗ − δ] = A sin [kct − (kx∗ − δ + π)]
which looks like a sine wave with frequency kc.
The constants have standard names and physical meanings: A is the amplitude, c is the
wavespeed, k is the wavenumber, δ is the phase, and ω = kc is the (angular) frequency.
8.3.3 Standing wave
The wave equation is linear. If u1 and u2 satisfy the wave equation, then u1 + u2 does too.
Let’s look at a specific case of two matching sinusoidal waves travelling in opposite directions:
u(x, t) = A sin [k(x − ct)] + A sin [k(x + ct)]
We can use the formulae for the sin of the sum of two arguments to write
u(x, t) = A sin kx cos kct − A sin kct cos kx + A sin kx cos kct + A sin kct cos kx
u(x, t) = 2A sin kx cos kct
This is the product of a function of x and a function of t. So every particle oscillates at the
same time with the same frequency ω = kc, with an amplitude that depends on its position x.
Solutions like this, which can be written as
u(x, t) = X(x)T (t)
(and the time-function T (t) has to be oscillatory) are called standing waves.
8.4 Solving the wave equation
8.4.1 Boundary and initial conditions
In order to solve the wave equation, we need boundary and initial conditions.
In the case of the string, it is fixed at x = 0, L. The boundary conditions are therefore
u(0, t) = u(L, t) = 0,
CHAPTER 8. THE WAVE EQUATION 81
for all t.
The initial conditions specify the form of the string and its velocity at t = 0. We can write them
as
u(x, 0) = f (x)
and
∂u
(x, 0) = g(x).
∂t
8.4.2 Separation of variables
There is a five-step method to solve our wave equation.
1. Use separation of variables to find a pair of ordinary differential equations (ODEs) related
by an unknown parameter
2. Solve the x-equation and apply the boundary conditions to determine the possible values
of the parameter
3. For each parameter value, solve the t-equation and therefore find a specific solution to the
wave equation, which satisfies your boundary conditions
4. Create a semi-general solution to the wave equation (given the boundary conditions) from
a linear combination of these ODE solutions
5. Apply the initial conditions to determine the coefficients in your linear combination
STEP 1 Separation of variables or product method.
Let us look for a solution of the form
u(x, t) = X(x)T (t).
Then
∂2u
= X(x)T ′′ (t)
∂t2
and
∂2u
= X ′′ (x)T (t).
∂x2
Substituting into the wave equation, we get
X(x)T ′′ (t) = c2 X ′′ (x)T (t)
and so
T ′′ X ′′
= .
c2 T X
The left hand side is a function of t only, whilst the right hand side is a function of x only.
Therefore they must both be constant, so
T ′′ X ′′
2
= = k.
c T X
The wave equation therefore reduces to two second order linear ODEs:
X ′′ − kX = 0, T ′′ − c2 kT = 0.
CHAPTER 8. THE WAVE EQUATION 82
STEP 2 Boundary conditions and the x-equation.
The boundary conditions were given by u(0, t) ≡ X(0)T (t) = 0 and u(L, t) ≡ X(L)T (t) = 0
for all t. Therefore X(0) = X(L) = 0. The solutions to X ′′ − kX = 0 can only satisfy these
boundary conditions if k is negative, say k = −p2 . Then
X ′′ + p2 X = 0,
so X = A cos(px) + B sin(px), for some constants A and B. The boundary conditions then give
A = 0 and sin (pL) = 0, so pL = nπ, for some integer n. Therefore any constant multiple of
nπx
Xn (x) = sin ,
L
is a solution for n = 1, 2, 3, .... These spatial functions are called modes.
STEP 3 The t-equation.
Our second equation is then
T ′′ + c2 (nπ/L)2 T = 0.
This has solution
Tn (t) = Cn cos (λn t) + Dn sin (λn t),
where λn = nπc/L, for n = 1, 2, 3, ....
This means that for n = 1, 2, . . . and any coefficients Cn and Dn ,
nπx
nπct
nπct
un (x, t) = Xn (x)Tn (t) = sin Cn cos + Dn sin
L L L
is a solution of the wave equation.
STEP 4 Linear combination.
We can build a much more general solution to the wave equation (in fact it is the most general
solution satisfying the boundary conditions) by simply adding up these functions:
∞
X nπx
nπct
nπct
u(x, t) = sin Cn cos + Dn sin .
L L L
n=1
STEP 5 Can this function satisfy the initial conditions?
We are looking to satisfy
u(0, x) = f (x)
and
∂u
(0, x) = g(x)
∂t
Substituting in the formula for u(x, t) with t = 0 gives
∞
X nπx
Cn sin = f (x)
L
n=0
and
∞
X nπc nπx
Dn sin = g(x).
L L
n=0
CHAPTER 8. THE WAVE EQUATION 83
We can solve these for the coefficients Cn and Dn using Fourier series.
Quick review of Fourier sine series results
If we have a Fourier sine series valid over 0 ≤ x ≤ L
∞
X nπx
h(x) = bn sin
L
n=1
then we can calculate the coefficients bn using
Z L nπx
2
bn = h(x) sin dx.
L 0 L
We use our Fourier sine series formula to tell us
Z nπx
2 L
Cn = f (x) sin dx.
L 0 L
Z nπx Z L nπx
nπc 2 L 2
Dn = g(x) sin dx Dn = g(x) sin dx.
L L 0 L nπc 0 L
8.5 Example string calculation with Fourier series
For a string fixed at x = 0 and x = L, we have found the general solution
∞
X nπx
nπct
nπct
u(x, t) = sin Cn cos + Dn sin
L L L
n=1
with the formulae Z
2 L nπx
Cn = f (x) sin dx
L 0 L
Z L nπx
2
Dn = g(x) sin dx.
ncπ 0 L
Suppose we start with the string flat (f (x) = 0) but given an instantaneous velocity in a small
segment in the middle:
0 0 ≤ x < (1 − δ)L/2
g(x) = U (1 − δ)L/2 ≤ x ≤ (1 + δ)L/2
0 (1 + δ)L/2 < x ≤ L
This is what happens when a hammer hits a piano string. Then since f (x) = 0 we have Cn = 0,
and our calculation for Dn becomes
Z L nπx Z (1+δ)L/2 nπx
2 2U
Dn = g(x) sin dx = sin dx
ncπ 0 L ncπ (1−δ)L/2 L
nπx (1+δ)L/2
2U L 2U L (1 − δ)nπ (1 + δ)nπ
= − cos = 2 2 cos − cos
ncπ nπ L (1−δ)L/2 cn π 2 2
CHAPTER 8. THE WAVE EQUATION 84
We can simplify:
nπ nπ
(1 − δ)nπ δnπ δnπ
cos = cos cos + sin sin
2 2 2 2 2
so nπ
(1 − δ)nπ (1 + δ)nπ δnπ
cos − cos = 2 sin sin
2 2 2 2
which is zero if n is even.
The final solution is
X nπ nπx
4U L δnπ nπct
u(x, t) = sin sin sin sin .
cn2 π 2 2 2 L L
n odd
8.6 Special case: string released from rest
Recall: our general solution for a string fixed at x = 0 and x = L is
∞
X nπx
nπct
nπct
u(x, t) = sin Cn cos + Dn sin
L L L
n=1
and if we want it to satisfy the intial conditions
∂u
u(x, 0) = f (x) and (x, 0) = g(x)
∂t
then we need Z
2 L nπx
Cn = f (x) sin dx
L 0 L
Z L nπx
2
Dn = g(x) sin dx.
ncπ 0 L
Let us consider the particular case where the string is released from rest, so that g(x) ≡ 0. Then
Dn = 0 and without calculating the coefficients we can see that
∞
X nπx
nπct
u(x, t) = Cn cos sin
L L
n=1
1
∞
X nπ 1X ∞ nπ
= Cn sin (x − ct) + Cn sin (x + ct)
2 L 2 L
n=1 n=1
1
= (f (x − ct) + f (x + ct)).
2
Therefore the solution to the wave equation for a string released from rest with form f (x) is the
mean of a travelling wave with form f travelling to the right with speed c and a travelling wave
with form f travelling to the left with speed c.