Question 1 [12 marks in total]

(a) [2 marks] Suppose that a function f (t ) is a periodic function with a period of 2. State the
Fourier expansion theorem for this function.

The Fourier expansion theorem states that f (t ) can be expanded as:

f (t ) a0 a1 cos(t ) b1 sin(t ) a2 cos(2t ) b2 sin(2t ) ...

or

f (t ) a0 a
n 1
n cos(nt ) b sin(nt )
n 1
n or as

f (t ) a0 {a
n 1
n cos(nt ) bn sin(nt )}

This series converges to the value of f (t ) at all points where f (t ) is continuous. At points where
f (t ) has a jump discontinuity the Fourier series converges to the average of f (t ) at this point.

State the formulas used to compute the a and b coefficients in the expansion based on an
integral over the interval 0 t 2 .

The formulas for the a's and bs are

2
1
a0
2
0
f (t )dt

2
1
an

0
f (t )cos(nt )dt for n 1, 2, 3...

2
1
bn

0
f (t )sin(nt )dt for n 1, 2, 3...

(b) One period of a function g (t ) is defined by

0, 0 t
g (t )
2, t 2
Draw a graph of this function between 2 and +4.

-2 - 2 3 4

1 2203BPS
(c) Calculate the Fourier coefficient a0 for g (t ) .

2 2
1 1 1
a0
2 0 g (t )dt
2 0 g (t )dt
2
g (t )dt [

Put in the formulas for g(t) on each interval:

2
1 2 2 2
a0
2 2dt
2
[t ]

1

Another approach is to use the fact that a0 = the average of function g(t) over one period
Because the function is very simple it is obvious that a0 = 1.

(d) Find a formula for the Fourier coefficients an of g (t ) for n > 0.

Use this to find the values of a1 , a2 and a3 .

2 2
1 1 1
an
0
g (t )cos(nt )dt
g (t )cos(nt )dt
0 g (t )cos(nt )dt

2
1 2 sin(nt ) 2 2
an
2 cos(nt )dt

[
n
]
n
{sin(n 2 ) sin(n )}

an 0
therefore a1 0, a2 0 and a3 0

(e) Find a formula for the Fourier coefficient bn of g (t ) for n > 0.

Use this to find the values of b1 , b2 and b3 . [3 marks]

2 2
1 1 1
bn
g (t )sin(nt )dt
0

g (t )sin(nt )dt
0

g (t )sin(nt )dt

2
1 2 cos(nt ) 2 2
bn
2 sin(nt )dt



[
n
]
n
{cos(n ) cos(n 2 )}

2 2
bn {cos(n ) 1} or bn { 1 1}
n

n n
4 4
therefore b1 , b2 0 and b3
3
[ b1 1.273, b2 0 and b3 0.424 ]

(f) Write down the terms of the Fourier expansion of g (t ) up to a third harmonic. [2 marks]

4 4
g (t ) 1 sin(t ) sin(3t ) ...
3

Question 2 [12 marks in total]

(a) [2 marks] State the definition of the Fourier transform f (k ) of a function f (t ) .

The Fourier transform f (k ) of a function f (t ) is defined by:

2 2203BPS


f (t ) e
ikt
f (k ) dt

What condition on f (t ) is needed for f (k ) to be finite ?

The condition on f (t ) is that the integral | f (t ) | dt

is finite.

(b) Write a formula for f (t ) in terms of f (k ) .[1 mark]

1
f (t ) e
ikt
f (k ) dk
2

(c) A function g (t ) is defined by

1, 1 t 0

g (t ) 1, 0 t 1
0,
otherwise

Draw the graph of g(t). [1 mark]

-1 1

(d) Calculate the Fourier transform of g. [3 marks]

0 1
g (k ) g (t ) e
ikt
dt 1 e
ikt
dt 1 e i k t dt
1 0

0 1
eikt eikt 1 1
g (k ) [e0 eik ] [e ik e0 ]
ik 1 ik 0 ik ik
1
g (k ) [eik e ik 2]
ik
Use eik cos(k ) i sin(k ) and e ik cos(k ) i sin(k )
So eik e ik 2cos(k )
1 2i 1 cos(k )
Then g (k ) {2cos(k ) 2}
ik k
3 2203BPS
(e) What is the value of g (k ) when k = 0? Explain your reasoning carefully. [3 marks]

2i 1 cos(k )
If we put k = 0 in the formula g (k ) get 0/0 which is not defined.
k
Therefore need to work out the limit as k 0 using lHopitals rule
d
2i 1 cos(k ) 2i 1 cos(k ) 2i sin(k )
g (0) = lim lim dk lim
k 0 k k 0 d k 0 1
k
dk
2i sin(0)
So g (0) = 0
1

g (t ) e
ikt
Another way is to put k = 0 into g (k ) dt this gives:

0 1
g (0) g (t ) dt 1dt 1dt 1 1 0
1 0

1 k
(f) If c>0 is a constant and h(t ) f (ct ) prove that h(k ) f( ). [2 marks]
c c

Write down definition and put in the formula for h(t):

h(t )e dt
ikt
h( k ) f (ct )e ikt dt

Then make change of variable u = c t t = u/c and dt = du/c. Put these into the integral:

1
h(k ) f (u )eik u / c du
c
k
1 i u
h( k ) f (u)e c du
c
1 k
Therefore: h(k ) f ( )
c c

Question 3 [12 marks in total]

2u 1 2u
(a)[6 marks] If the one dimensional wave equation represents the vibration of a
x 2 c 2 t 2
guitar string that is tied down at both ends, what does the function u( x, t ) represent physically?

The function u( x, t ) is the displacement of the guitar string from equilibrium at position x and
time t.

2u 2u
what is the physical interpretations of and ?
x 2 t 2
2u
gives the curvature of the string.
x 2
4 2203BPS
 1 2u
is related to the acceleration of string at position x, time t.
c 2 t 2

Use your above explanations and an appropriate physics law to give an explanation of why the
wave equation makes sense.

The more the curvature of the string the greater the vertical force on the section of the string.
The wave equation is based on Newtons law: force = mass acceleration.

(b) [6 marks] The separated solutions of the 1D wave equation are solutions of the form
X ( x)T (t ) . The possible separated solutions are listed below. Suppose that a 1D wave equation is

solved on 0 x and that the boundary conditions are u ( x, t ) 0, at x 0 and
x

u ( x , t ) 0 at x .
x

(i) Write the boundary conditions for X ( x) .

The boundary conditions for X ( x) are:X = 0 when x = 0 and X = 0 when x = . [

(ii) [5 marks] Find the separated solutions that can match these boundary conditions. Explain
carefully why you reject solutions and how you include the boundary conditions.

Look at the first solution which is that X(x) = cos(kx), sin(kx) or a mixture.
At x = 0, only cos(kx) has zero derivative when x = 0
X(x) = sin(kx) and at x = , we have X()=sin(k) = 0
This means that k = , 2, 3 .
So k = 1,2,3

Look at the 2nd solution for X(x) this is X ( x) e k x , e k x or a mixture of these terms
The only combination derivative of which can be zero when x = 0 is proportional to
X ( x) e k x e k x , as X '( x) ke k x ke k x
But X '( x) ke k x ke k x is not zero when x =
So the 2nd option does not satisfy BC.
Look at the 3d solution for X(x):
The only solution derivative of which is zero when x = 0 is X(x)= 1
Derivative of X(x)= 1 is zero at x = as well,
so the X(x) solution from the third list is X(x)= 1.

Conclusion the separated solution that matches the boundary conditions are:
X(x)T(t) = cos(kx)cos(ckt), X(x)T(t) = cos(kx)sin(ckt) for k = 1,2,3
and X(x)T(t) = 1 and X(x)T(t) = t

Or could give a list of the possible solutions:

1, t , cos( x)cos(ct ), cos( x)sin(ct );cos(2 x)sin(2ct ), cos(2x)cos(2ct );cos(3x)sin(3ct ), cos(3x)cos(3ct )

5 2203BPS
 Alternativly:
Students can use X(x) = Acos(kx) + Bsin(kx), X ( x) Ae k x B e k x , X ( x) A Bx and then
find possible A and B from BC. If all solutions are found correctly this results in full marks.

Question 4 [12 marks in total]

u 2u
(a) [2 marks] The one dimensional heat equation is D 2.
t x
The shape of the temperature distribution at time t = 0 is shown below.

u(x,0)

x=0 x=1 x=2

Describe what happens to u 1, t just after t = 0. Here x = 1 is in the middle of the graph.
Explain your answer carefully.

2u
At x = 1 the curve is concave up so that 0
x 2
u
Then from the heat equation we get 0
t
This means that u(1,t) increases as t increases.

(b) [6 marks] The separated solutions of the 1D heat equation are solutions of the form X ( x)T (t ) .
The possible separated solutions are listed below. Suppose that u is a solution of the 1D heat
equation for 0x2 and that the boundary conditions are u( x, t ) 0 at x 0 and
u( x, t ) 0 at x 2 . Find the separated solutions that can match these boundary conditions.
Explain why you reject solutions and how you use the boundary conditions.

The boundary conditions for X ( x) are:X = 0 when x = 0 and X = 0 when x = 2.

Look at the first solution which is that X(x) = cos(kx), sin(kx) or a mixture.
The only solution that is zero when x = 0 is X(x) = sin(kx)
Use second BC: X(2) = sin(k2) = 0
This means that 2k , 2 , 3 , ... so k = , 1, 3/2, 2 etc

Look at the second solution which is X(x) = exp(kx), exp(kx) or a mixture.

The only solution that is zero when x = 0 is proportional to
X(x) = exp(kx)exp(kx)
But X(x) = exp(kx)exp(kx) is not zero at x = 2 ,
therefore second solution is not possible.

Alternativly:
From the 2nd list there are no options because it is not physically possible for the solution to
increase exponentially with time.

From the 3rd list the only option for X(x) that is zero when x = 0 is X(x) = x
6 2203BPS
 This solution cannot match the second BC that X 0 at x 2

The solutions that match the two boundary conditions are:

1 3
sin( x)e D t / 4 , sin( x)e D t , sin( x)e 9 D t / 4 , ...
2 2

(c) [4 marks]Give an intuitive interpretation of the Dirac delta function (t c) and

calculate the following integrals:

An intuitive interpretation of the Dirac delta function (t c) is a sharp impulse (or infinite
spike) at time t = c.
It is zero away from t = c and its total integral is 1.

t { (t 1) (t 1) (t 7)}dt
2
(i)
0
5 5 5

t (t 1) dt t (t 1) dt t (t 7) dt
2 2 2

0 0 0

1 0 0 1
2

[give full 1 mark if correct answer but no expansion]

e (t ) dt e0 1
i k t
(ii)

e (t a) dt ei k a
i k t
(iii)


(iv)

f (t ) (t a) dt f (a)

7 2203BPS