Lecture 7

Weston Barger

July 13, 2016

1 Superposition of standing waves

In last lecture, we examined standing wave solutions to the problem

 u(0, t ) = 0
utt = c 2 uxx , . (1.1)
 u(L, t ) = 0

We found that all standing wave solutions to (1.1) had the form Therefore, our solutions
are
cnπt cnπt nπx
      
un (x , t ) = A cos + B sin sin , n = 1, 2, 3, · · · .
L L L
We would like to show that we can superimpose solutions of (1.1). Suppose that u1
and u2 are solutions of (1.1). We would like to show that v = α1 u1 + α2 u2 , where α1 and
α2 are constants, is also a solution of (1.1). Note that

v (0, t ) = α1 · 0 + α2 · 0 = 0
v (L, t ) = α1 · 0 + α2 · 0 = 0.

So, v satisfies the boundary condition. Furthermore,

vtt = α1 (u1 )tt + α2 (u2 )tt = α1 c 2 (u1 )xx + α2 c 2 (u2 )xx = c 2 vxx .

Thus, v is a solution to (1.1).

1
Note: We were able to superimpose solutions to (1.1) because it is a homogeneous equa-
tion. Suppose that we were interested in solving

 u(0, t ) = 0
utt = c 2 uxx , . (1.2)
 u(L, t ) = 1

Problem (1.2) is not a homogeneous problem, because u = 0 does not satisfy

the boundary condition at x = L. Suppose that u1 and u2 solve (1.2) and let
v = α1 u1 + α2 u2 . Then

v (L, t ) = α1 · 1 + α2 · 1 = α1 + α2 .

Hence, solutions to (1.2) cannot be superimposed to create additional solutions.

Okay, so we assume that by superposition we can write solutions to (1.1) as

∞ 
X cnπt
 
cnπt
  
nπx

u(x , t ) = An cos + Bn sin sin .
n=1 L L L

For a given f and g, how do we compute An and Bn ?

2 Fourier Series
Suppose that f (x ) is a periodic function defined on [–L, L]. We wish to represent f (x ) as
∞
X nπx
 
nπ
 
f (x ) = a0 + an cos + bn sin x
n=1 L L

The above series is known as a Fourier series.

Note: Many phenomenon arising in engineering and the physical sciences are periodic.
Representing these waves as a decomposition of their constituent components can
help illuminate the underlying phenomenon. The Fourier series is named for Jean
Baptiste Joseph Fourier, for his work on solving the heat equation on a plane with
arbitrary heat source.

The Fourier series is analogous in some ways to the Taylor series. Instead of representing
a function as a sum of monomials, we use sine and cosine waves with varying frequencies.
We have a formula for the coefficients of the Taylor series, and we would like one for the
coefficients a0 , an and bn . First, we need a definition:

2
Definition 2.1. Let f (x ) and g(x ) be two functions defined on [–L, L]. The functions f
and g are said to be orthogonal if
Z L
f (x )g(x ) dx = 0. (2.1)
–L

We can think of the functions f and g as like vectors in a vector space. Recall that
for x , y ∈ Rn , x and y are orthogonal if

x · y = x1 y1 + x2 y2 + x3 y3 + · · · + xn yn = 0. (2.2)

The definition in (2.1) is like the continuous version (2.2). Here, f (x ) and g(x ) are multi-
plied and summed for each index x ∈ [–L, L].

Proposition 2.2. For positive integers n and m, we have


Z L
nπx
  
mπ
  L n =m
sin sin x dx = = Lδn,m , (2.3)
–L L L  0 n 6= m

Z L
nπx
  
mπ

L n =m

cos cos x dx =  = Lδn,m , (2.4)
–L L L 0 n=6 m
Z L
nπx mπ
   
cos sin x dx = 0, (2.5)
–L L L
where δn,m is known as the Kronecker delta function and is defined as

 1 n =m
δn,m := .
 0 n 6= m

Proof. Let us show (2.3). Recall the identities

sin α sin β – cos α cos β = – cos (α + β)

sin α sin β + cos α cos β = cos (α – β) .

Adding these equations together yield

cos(α – β) – cos(α + β)
sin α sin β = .
2
Let us assume that n 6= m.
Z L
nπx mπ
   
sin sin x dx
–L L L

3
 " ! !#
1Z L (n – m)πx (n + m)πx
= cos – cos dx
2 –L L L
(n – m)πx L (n + m)πx L
! !
1 1
= sin – sin
2(n – m)π/L L 
–L 2(n + m)π/L L 
–L
=0.

Now, assume that n = m. In this case,

Z L Z L
nπx mπ nπx
     
sin sin x dx = sin2 dx
–L L L –L L
Z L
1 2nπx
 
= 1 – cos dx
2 –L L
=L.

The proofs for (2.4) and (2.5) are analogous.

Exercise 2.3. Prove (2.4) and (2.5).

Note: The above proposition gives orthogonality relationships.

So, we’d like to find the a0 , an and bn such that

∞
X 
nπx
 
nπ

f (x ) = a0 + an cos + bn sin x . (2.6)
n=1 L L

Now, let us integrating both sides gives

 
1 ZL 1 Z L X∞ nπx
 
nπ 
 
f (x ) dx = a0 + an cos + bn sin x dx
2L –L 2L –L n=1 L L
∞
X1 ZL nπx
   
nπ

= a0 + an cos + bn sin x dx
n=1 2L –L L L
= a0 ,

where the last equality was achieved by recalling that the average value of sine and cosine
over their periods is zero. We now have an expression for a0 . Now, let choose an arbitrary
positive integer m, multiply both sides of (2.6) by sin(nπx /L) and integrate to get
Z L
mπx
 
sin f (x ) dx
–L L

4
  
Z L
mπx
  ∞
nπx
X nπ 
   
= sin a0 + an cos + bn sin x dx
–L L n=1 L L
Z L
mπx
 
= a0 sin dx
–L L
X∞ Z L 
mπx
 
nπx
 
mπx
 
nπx

+ an sin cos + bn sin sin dx
n=1 –L L L L L
∞
X
= bn Lδm,n
n=1
=Lbm

We have an expression now for bm . Repeating this process with cos(mπx /L) yields
Z L
mπx
 
cos f (x ) dx
–L L  
Z L 
mπx 
 X∞ 
nπx
 
nπ 

= cos a0 + an cos + bn sin x dx
–L L n=1 L L
Z L
mπx
 
= a0 cos dx
–L L
∞ Z L
X 
mπx
 
nπx
 
mπx
 
nπx

+ an cos cos + bn cos sin dx
n=1 –L L L L L
∞
X
= an Lδm,n
n=1
=Lam .

Now, we have an expression for am .

Theorem 2.4. The Fourier series for a periodic function f (x ) defined on [–L, L] is
∞
X nπx
 
nπ
 
f (x ) = a0 + an cos + bn sin x ,
n=1 L L
where
1 ZL
a0 = f (x ) dx
2L –L
1ZL nπx
 
an = f (x ) cos dx
L –L L
1ZL nπx
 
bn = f (x ) sin dx
L –L L

5
Theorem 2.5. Let f (x ) be a twice differentiable function for which f (x ), f 0 (x ) and
f 00 (x ) are pointwise continuous on [–L, L]. Suppose that a0 , an and bn are the Fourier
coefficients for f (x ). Then, for any x ∈ [–L, L],
 
N
X 
nπx
 
nπx 

f (x– ) + f (x+ )
lim a0 + an cos + bn sin = .
N →∞ n=1 L L 2

Example 2.6. Find the Fourier series of f (x ) = x 2 on [–1, 1]. For this example L = 1.
So,
1Z 1 2 1
a0 = x dx = ,
2 –1 3
and
Z 1
4(–1)n
an = x 2 cos (nπx ) dx = ,
–1 π2n 2

where the integral was performed by parts. Now, note that x 2 sin(nπx ) is an odd function,
therefore,
Z L
bn = x 2 sin (nπx ) dx = 0.
–L

Thus,

1 ∞ 4(–1)n
x2 =
X
+ cos (nπx ) .
3 n=1 π 2 n 2

This example leads to a few remarks:

 When we constructed the Fourier series for x 2 , we constructed only over [–1, 1] i.e.
we computed the series representation of the periodic extension of x 2 on [–1, 1].
This extension is pictured in Figure 1

6
 Figure 1: The periodic extension of x 2 on [–1, 1].

 Suppose that f (x ) is an even function. Then

1 ZL 1ZL
a0 = f (x ) dx = f (x ) dx
2L –L L 0

1ZL nπx 2ZL nπx

   
an = f (x ) cos dx = f (x ) cos dx
L –L L L –L L
and
1ZL nπx
 
bn = f (x ) sin dx = 0.
L –L L
This leads to the following definition:

Definition 2.7. If f (x ) is an even function, then the Fourier series on [–L, L] collapses
and we have the Fourier cosine series
∞
X nπx
 
f (x ) = a0 + an cos ,
n=1 L

with
1ZL nπx 2ZL nπx
   
a0 = f (x ) cos dx , an = f (x ) cos dx .
L 0 L L –L L

 Suppose now that f (x ) is an odd function. Then,

1 ZL 1ZL nπx
 
a0 = f (x ) dx = 0, an = f (x ) cos dx = 0,
2L –L L –L L

7
 and
1ZL nπx 2ZL nπx
   
bn = f (x ) sin dx = f (x ) sin dx .
L –L L L 0 L
This motivates the definition

Definition 2.8. If f (x ) is an odd function, then the Fourier series on [–L, L] collapses
and we have the Fourier sine series
∞
X nπx
 
f (x ) = bn sin ,
n=1 L

with
2ZL nπx
 
bn = f (x ) sin dx .
L 0 L

3 Solution to the wave equation on finite string

Let us return to the problem



 x ∈ (0, L), t ∈ [0, ∞)





 u(x , 0) = f (x )

utt = c 2u xx , ut (x , 0) = g(x ) . (3.1)


u(0, t ) = 0








u(L, t ) = 0

We know solutions have the form

∞ 
X 
cnπt
 
cnπt
 
nπx

u(x , t ) = An cos + Bn sin sin .
n=1 L L L

We have already applied the boundary conditions to derive the above equation. Applying
the initial conditions give
∞
X 
nπx

u(x , 0) = f (x ) = An sin ,
n=1 L
X∞ cnπ nπx
 
ut (x , 0) = g(x ) = Bn sin (3.2)
n=1 L L

8
Let us consider the odd extension of f (x ), fˆ(x ). Then on [–L, L],
∞ nπx
 
fˆ(x ) =
X
An sin
n=1 L

where
2ZL nπx
 
An = f (x ) sin dx .
L 0 L
Now, to compute Bn . We see that (3.2) is the sine series of the odd extension of g(x ),
ĝ(x ). Thus,

cnπ 2ZL nπx

 
Bn = g(x ) sin dx
L L 0 L
2 ZL nπx
 
⇒Bn = g(x ) sin ddx .
cnπ 0 L
Proposition 3.1. The full solution to (3.1) is
∞ 
X 
cnπt
 
cnπt
 
nπx

u(x , t ) = An cos + Bn sin sin ,
n=1 L L L

where
2ZL nπx 2 ZL nπx
   
An = f (x ) sin dx , Bn = g(x ) sin ddx .
L 0 L cnπ 0 L