Вы находитесь на странице: 1из 32

=====j^RUU=======j~íÜÉã~íáÅ~ä=íÉÅÜåáèìÉë=

=
OMMRJOMMS=
=
=
=
=
cçìêáÉê=pÉêáÉë==
=
=
=
aê=dÉçêÖÉ=açÄêÉ=

_~êçå=gçëÉéÜ=cçìêáÉê=
NTSUJNUPM
j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 2.

Introduction to Fourier series

A large number of phenomena and processes have a periodic nature. Consider for example
waves, vibrations, an electrocardiogram, parts of machines, low and high tide, the string of a
violin etc.

Many functions can be written as infinite series involving ascending powers of x. Fourier series
are unusual and particularly useful in as a mathematical technique because they are expansions of
periodic functions f(x) written using just sines and cosines.

Example of a periodic function:

x-p x x+p x+2p

A function f is periodic when f(x+p) = f(x) for all x where p is the period of the function.

When periodic functions are integrated over one period the same result is always obtained
irrespective of where the limits are set:

∫ f (x ) dx ∫ f (x ) dx ∫ f (x ) dx ∫ f (x ) dx
1
+ p
a+ p b+ p 2 p

= = =
a b 1 0
− p
2

Exercise 1. Check that the result above is valid for f(x) = –sin x integrated between 0 and 2π
and again between –π/4 and 7π/4.

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 3.

Even and odd functions

A function is even when: f(–x) = f(x)

Example: cos (–x) = cos (x)


y

-2π 0 2π

A function is odd when: f(–x) = –f(x)

Example: sin (–x) = –sin (x)


y

x
-2π 0 2π

Exercise 2. Show that for any arbitrary f(x)


(i) the function E(x) = f(x) + f(–x) is an even function of x and
(ii) that the function O(x) = f(x) – f(–x) is an odd function of x, where f(x) is an
arbitrary function of x. Hence show that
(iii) the arbitrary function f(x) can be written as a sum of an odd and even function
of x.

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 4.

Integrating even and odd functions:

• Integrating an odd function:

∫ f ( x ) dx = − ∫ f ( x ) dx
b 0

0 −b
Therefore,

∫ f (x ) dx
b
= 0
−b

• Integrating an even function:

∫ f (x ) dx ∫ f ( x ) dx
b 0
=
0 −b
Therefore,

∫ f ( x ) dx = 2 ⋅ ∫ f ( x ) dx
b b

−b 0

• Products of odd and even functions will give either an odd or an even function:

Odd x Odd = Even


Odd x Even = Odd
Even x Odd = Odd
Even x Even = Even

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 5.

Fourier sums

To approximate complicated functions the power series can be used. For periodic functions the
Fourier series are better suited.

The Fourier Sum is a trigonometric polynomial of order n:

1
TPn = a0 + a1 cos x + a2 cos2 x + a3 cos3x ... + an cosnx
2
+ b1 sin x + b2 sin 2 x + b3 sin 3x ... + bn sin nx

The following statement is central to Fourier theory:

“A periodic function f(x), with a period 2π, fulfilling certain conditions,


can be written as the infinite sum of the trigonometric polynomial TPn.”

f ( x ) = a0 + ∑ ( an cos nx + bn sin nx )

1
2 n =1

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 6.

Example: Consider the function A sin(nπx), where A and n are constants. It can be
thought of as the instantaneous amplitude (displacement from horizontal)
at position x along an oscillating string. The maximum amplitude of the
wave is A, and the period will be 2/n, where n is the wavenumber. So, for
example, the function 4 sin(5πx) has amplitude 4 and period 2/5.

In a Fourier series, instead of the power of x increasing by 1 each time, like in the Maclaurin and
Taylor expansions, the wavenumber increases by 1.

If we increase the wavenumber, then the function oscillates more rapidly, so sin(10x) has ten
peaks and troughs between x = 0 and x = 2π whereas sin(x) only has one peak and one trough.

Exercise 3. Plot (to scale) the functions sin(x), sin(2x), sin(5x) and sin(10x)

The terms of a Fourier series represent waves of different periods (or wavelengths), with each
successive term giving a more rapidly oscillating contribution than the previous.

• Large-scale features of any given periodic function - roughly approximated by a sine or


cosine wave, with suitably chosen amplitude.

• Shorter-scale features - more accurately captured by the subsequent terms in the Fourier
series.

In this way we can recreate the function as accurately as we choose, by taking more and more
terms, thus including shorter and shorter scales.

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 7.

We saw that Fourier series of a function f(x) looks like this:

f ( x ) = a0 + ∑ ( an cos nx + bn sin nx )

1
2 n =1

The task is then to find all the coefficients a0, a1, b1, and so on, to model the function f(x) as
accurately as possible.

Example: square wave

-2π 0 2π

1 when (2k − 12 )π < x < (2k + 12 )π


f (x) = -1 when (2k + 12 )π < x < (2k + 23 )π k∈Z
0 when x = (k + 12 )π

The longer the Fourier series modelling this function, the more accurate it gets!

Orders 3, 9 and 27 of the trigonometric polynomials are shown below for comparison purposes.
Please accept the coefficients without proof for now – that will come in the next lecture. Also,
notice how the sine terms are conspicuously absent!

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 8.

The third order Fourier sum that approximates the square wave:

TP3 =
4
(cos x − 13 cos 3x )
π
y

-2π 0 2π

The ninth order Fourier sum:

TP9 =
4
(cos x − 13 cos 3x + 15 cos 5x − 17 cos 7 x + 19 cos 9 x )
π

-2π 0 2π

The twenty-seventh order Fourier sum

TP27 =
4
(cos x − 13 cos3x + 15 cos5x − 17 cos7 x + ... − 271 cos27 x )
π

-2π 0 2π

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 9.

Fourier coefficients

To find the coefficients an and bn we calculate the average of f(x) over one period p
= 2π:

∫ ∫ ∫ 2 π −∫π
π π π π
1 1 1 1 1
f ( x ) dx = 2 a 0 dx + a1 cos x dx + a2 cos 2 x dx + ...
2 π −π 2π −π 2π − π

2 π −∫π 2 π −∫π
π π
1 1
+ b1 sin x d x + b2 sin2 x dx + ...

Before we can proceed, we need to evaluate the integrals of cos nx and sin nx with
n = 1,2,3…


1  an 
sin nπx − sin ( −nπx ) = 0
π π

2π  n 
1 an
a cos n x d x = sin n x =
π
n
2π −π −π 2 n

and likewise:

∫ bnsin nx dx = 0 , n = 1, 2,3,...
π
1
2 π −π

Now it is much easier to evaluate the average of f(x):

∫ ∫
π π
1
f ( x ) dx = [2 a 0 x ]π−π
1 1

1
f ( x ) dx = 12 a 0
2π −π 2π 2π −π

Hence,

a0 = ∫ f ( x ) dx
π
1
π −π

Thus 12 a0 is the average of f(x) over its period p = 2π !!!!

To find a1 multiply both sides with cos x and calculate the average over one period
again:

f ( x ) ⋅ cos x = 12 a 0 cos x + a1 cos2 x + a 2 cos 2 x ⋅ cos x + ... + a n cos nx ⋅ cos x


+ b1 sin x ⋅ cos x + b 2 sin 2 x ⋅ cos x + ... + b nsin nx ⋅ cos x

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 10.

And the average over one period of f(x) cos x is:

∫ ∫ a1 ∫ cos2 x dx
π π π
1
f ( x ) cos x dx =
1 1
a cos x d x +
1
2π −π 2π 2π −π
2 0
−π

2π −∫π
π
1
+ a 2 cos 2 x ⋅ cos x dx + ...

b1 ∫ sin x ⋅ cos x dx +
2 π −∫π
π π
1 1
+ b2 sin 2 x ⋅ cos x dx + ...
2π −π

To evaluate the integrals we would like to have some general expressions for the
integrals of

(cos mx · cos nx), (cos mx · sin nx) and (sin mx · sin nx) over a period of 2π:

The average of cos mx · cos nx over a period p = 2π:

 0, m ≠ n

∫−πcos mx ⋅ cos nx dx =  12 , m = n ≠ 0
π
1
2π  1, m = n = 0

The average of cos mx · sin nx over a period p = 2π:

∫ cos mx ⋅ sin nx dx = 0
π
1
2π −π

The average of sin mx · sin nx over a period p = 2π:

 0, m ≠ n


π
sin mx ⋅ sin nx dx =  12 , m = n ≠ 0
1
2π −π  0, m = n = 0

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 11.

We can now find the average of f(x) · cos x over one period p = 2π:

∫ a1 ∫ cos2 x dx
π π
f ( x ) cos x dx =
1 1
2π −π 2π −π

2 π −∫π π −∫π
π π
1
f ( x ) cos x d x = 1
a
2 1 → a 1 =
1
f ( x ) cos x dx

Now we are looking for a general formula for an:

2 π −∫π ∫ 2π −∫π
π π π
1
f ( x ) cos nx d x =
1 1
a cos nx dx +
1
a1 cos x ⋅ cos nx dx

2 0
−π

a 2 ∫ cos 2 x ⋅ cos nx dx + ... + a n ∫ cos2 nx dx


π π
1 1
+
2π −π 2π −π

b1 ∫ sin x ⋅ cos nx dx +
2π −∫π
π π
1 1
+ b2 sin 2 x ⋅ cos nx dx + ...
2π −π

All terms on the right hand side are zero except:


π
1 1
an cos2 nx dx = an
2π − π 2

Solving for an we have:

an = ∫ f ( x ) ⋅cos nx dx
π
1
π −π

The same procedure can be followed to find the coefficients bn:

 Multiply both sides by sin nx


 Calculate the average of f(x) · sin nx over one period p = 2π:

Which results in:

π −∫π
π
1
bn = f ( x ) ⋅sin nx dx

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 12.

Hence, the complete Fourier Series for a function with period p = 2π is:

f ( x ) = 12 a0 + ∑ ( ancos nx + bnsin nx )

n =1

π −∫π
π
1
a0 = f ( x ) dx

an = ∫ f ( x ) ⋅cos nx dx bn = ∫ f ( x ) ⋅sin nx dx
π π
1 1
π −π π −π

Exercise 4: For the square wave defined below, calculate the Fourier coefficients

y
1

x
-2π -π 0 π 2π

0 when (2k − 1)π < x < (2k )π


f (x) = 1 when 2kπ < x < (2k + 1)π k ∈ Z (integer)
1
2 when x = kπ

Solution: Calculate a0, by calculating the average of f (x)

a 0 = ∫ f ( x ) dx = ∫ 0 dx + ∫ 1 dx = 1
π 0 π
1 1 1
π −π π −π π0

The cosine coefficients an:

a n = ∫ f ( x ) ⋅ cos nx dx
π
1
π −π

a n = ∫ 0 ⋅ cos nx dx + ∫ 1 ⋅ cos nx dx = ∫ cos nx dx


0 π π
1 1 1
π −π π0 π0

⋅ [sin nx ] 0π = 0
1 1
an =
π n

The sine coefficients bn:


Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk
j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 13.

b n = ∫ f ( x ) ⋅ sin nx dx
π
1
π −π

b n = ∫ 0 ⋅ sin nx dx + ∫ 1 ⋅ sin nx dx = ∫ sin nx dx


0 π π
1 1 1
π −π π0 π0

bn = ⋅ [- cos nx ] 0π = ⋅ [- cos nπ + cos 0]=


1 1
π n
1 1
π n
−1

[
(- 1)n − 1 ]

The formula for the sine coefficients becomes:

 0 for even n

bn =  2
 nπ for odd n

And the Fourier series that describes the square wave:

1 2  sin x sin 3x sin 5 x sin 7 x 


f( x ) = +  + + + + ...
2 π 1 3 5 7 

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 14.

 x 2 [− π ,π]
Given the function f( x ) = 
 f (x + 2π )
Exercise 5.

Draw a sketch of f(x), is f(x) even or odd?

Show that the Fourier series of f(x) is given by


 cos x cos 2x cos 3x cos 4x cos 5x cos 6x 
− 4 2 − + ...
π2
f ( x) =
 1 
2
+ 2
− 2
+ 2
− 2
3 2 3 4 5 6
Calculate:
+ ... and ∑ 2

1 1 1 1 1 1
f ( x) = 1 − + − + −
4 9 16 25 36 n =1 n

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 15.

Dirichlet Conditions

What are the conditions a function needs to fulfil in order to be able to write it as a Fourier
Series?

If f (x) is:

 periodic of period 2π

 single valued between –π, π

 has a finite number of minima and maxima between –π, π

 has a finite number of discontinuities between –π, π (such functions are said to be
“piecewise continuous”)

 has a finite number of jumps between –π, π (jumps = points where the lateral limits f(x+)
and f(x-) are well defined and finite).

∫ f (x )
π
 dx is finite
−π

Then

 the Fourier Series converge to f (x)

f ( x ) = T∞ ( x )

 at jumps u the Fourier Series converges to the midpoint of the jump

T∞ ( u ) =
1
( )
lim f ( x ) + lim f ( x )
2 x ↑u x ↓u

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 16.

Exercise 6. Of the following two periodic functions only one violates a Dirichlet condition.
Which Dirichlet condition is violated and show with calculations on both functions which one
can be represented by a Fourier series and which one can’t.

1 / x 0 ≤ x < 1 1/ x 0 ≤ x <1


f( x ) =  g( x ) = 
 f ( x + 1)  g ( x + 1)

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 17.

Fourier Series for a periodic function with p ≠ 2π

Function Period
sin x, cos x, eix 2π

2π 2π i x
sin x, cos x, e p p
p p
π
π π i x
sin x, cos x, e l 2l
l l

If we want to obtain the Fourier Series for a function with an arbitrary period p = 2l we need:

 sines and cosines with the same period 2l

 integration limits from (-l, l) or (0, 2l)

πx 2πx 3πx
f( x ) = 12 a 0 + a1 cos + a 2 cos + a 3 cos + ...
l l l
πx 2πx 3πx
+ b1 sin + b2 sin + b3 sin
l l l
 nπx 
= 12 a 0 + ∑  a n cos 

nπx
+ bn sin
n =1  l l 

The Fourier coefficients can be found by calculating the average value of f (x), f (x) · cos x, and

f (x) · sin x over the period p=2l

l −∫l
l
f ( x ) dx
1
a0 =

a n = ∫ f ( x ) ⋅ cos
nπ x
l
1
dx
l −l l


nπ x
l
f ( x ) ⋅ sin
1
bn = dx
l −l l

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 18.

y
Example: Saw Function:
1

|x| for x ∈[-1, 1]


f (x) = x
f (x+2) = f (x) -2 -1 0 1 2

-1

The period p (= 2l) of this function is 2l = 2

First, the coefficient a0 is calculated:

∫ x d x = 2 ⋅ ∫ x dx = 2 ⋅ [x]
1 1
a0 = 1
2
2 1
0 =1
−1 0

The calculation of the Fourier coefficients an:

∫ x ⋅ cos nπx dx = 2 ⋅ ∫ x ⋅ cos nπx dx


1 1
an =
−1

1 
0

= 2 ⋅  ⋅ x ⋅ sin nπx + ⋅ cos nπx  10


1
 nπ (nπ ) 
2

= 2 2 (cos nπ − cos0) = 2 2 (− 1) − 1

2

2 n
( )

0 for even n
an =
−4
n 2 π2 for odd n

Since f (x) is an even function all the sine terms bn are zero (check this yourself by calculating bn)

1 4  
− 2  cos πx + + ...
cos 3πx cos 5πx cos 7πx
f( x ) = + +
2 π  9 25 49 

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 19.

 πx
sin 2 [0 , 2]
Given the function f( x ) =  0 [2 , 4]

Exercise 7.
 f ( x + 4 )

Sketch the graph of this function for at least 3 periods. Is the function even or odd or
neither? Calculate the first Fourier coefficient a0

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 20.

 4, 0 < x ≤ 1
2 , 1 < x ≤ 3
Given the following periodic function f( x ) = 

Exercise 8.
0, 3 < x ≤ 4
 f ( x + 4 )
a) Sketch a graph of the function. Is it even, odd or neither?

b) Derive the Fourier Series for this function and write down the expansion up to the term
with n=8

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 21.

Exercise 9. Sketch a graph and find the Fourier Series for:

 3 cos 2 x − 1 x ∈ [0 , π / 2]
f( x ) = 
 f( x + π/ 2)

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 22.

Exercise 10. Given the following functions:

 0, −1 < x < 0 − x −1 < x < 0


 2 , x = n n ∈ Ζ  x [− 1, 1] 1 − 1 (− 1)n x = n, n ∈ Ζ
 2
f( x ) =  g( x ) =  , and h( x ) = 
1

  g( x + 2 ) 1 + x 0 < x < 1
,
1, 0 < x < 1
 f ( x + 2 )  h( x + 2 )

and the Fourier series for f (x) and g (x):

1 2  sin πx sin 3πx sin 5πx cos 7πx 


f( x ) = −  + + + + ...
2 π 1 3 5 7 

1 4  cos πx cos 3πx cos 5πx cos 7πx 


g( x ) = − 2 2 + + + + ...
2 π  1 3 2
5 2
7 2

 Sketch a graph of f (x), g (x), and h (x)


 Find the Fourier series for h (x)
 Calculate a0 for h (x) using the formula for a0

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 23.

Complex form of the Fourier Series

sin nx =
2i
e −e [
1 inx −inx
]

cos nx =
2
[
1 inx
e + e −inx ]
By substituting the complex exponentials for the sines and cosines one arrives at the following
formula:

f ( x ) = c0 + c1eix + c-1e −ix + c2e2ix + c-2e −2ix + ...

where:

c0 = 12 a 0

cn = 12 (a n − i bn )

c-n = 1
2 (a n + i bn )

Calculating c0:

⋅ ∫ f ( x ) dx = ∫ 2π −∫π
π π π
f ( x ) ⋅ 1 dx = f ( x ) ⋅ e −i⋅0⋅x dx
1 1 1 1
c0 =
2 π −π 2π −π

2π −∫π
π
f ( x ) ⋅ e −inx dx
1
c0 = n=0

Calculating cn:

1 1 
c n =  ∫ f ( x ) ⋅ cos nx dx − ∫ f ( x ) ⋅ sin nx dx 
π π
i
2  π −π π −π 

1  
 ∫ f ( x ) ⋅ (cos nx − i ⋅ sin nx ) dx 
π
=
2π  −π 


π
f ( x ) ⋅ e −inx dx
1
= n = 1, 2 , 3,...
2π −π

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 24.

Calculating c-n:


π
f ( x ) ⋅ einx dx
1
c-n = n = 1, 2 , 3,...
2π −π

Which is equivalent to:


π
f ( x ) ⋅ e −inx dx
1
cn = n = −1, − 2 , − 3,...
2π −π

Fourier Series in complex notation:

∑c
n = +∞
f(x)= n einx
n = −∞


π
f ( x ) ⋅ e −inx dx
1
cn =
2π −π

Example: Slightly different square wave:


y

1 when (2k − 12 )π < x < (2k + 12 )π


-1 when (2k + 12 )π < x < (2k + 23 )π
x
f (x) = k∈Ζ
-2π 0 2π
0 when x = (k + 12 )π

We can write this function as a Fourier Series:

∑c
n = +∞
f(x)= n einx with cn:
n = −∞

∫ ∫ 2 π − π∫/ 2
π π/2 3 / 2π
f ( x ) ⋅ e −inx dx =
1 1 1
cn = 1 ⋅ e −inx dx + − 1 ⋅ e −inx dx
2π −π 2π −π / 2

=
− 1 −inx
2πin
e [ ] π/2
−π / 2 +
2πin
e [ ]
− 1 −inx 3 / 2π
π/2 =
1
2πin
[
e −in3/2 π − e −inπ / 2 − e −inπ / 2 + einπ / 2 ]

=
1
2 π in
[
2einπ / 2 − 2e −inπ / 2 =
2

]
sin n
π
2

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 25.

0 for even n

cn =
2
(n + 1) / 2 odd, c0 = 0

−2
(n + 1) / 2 even

2  (eix + e − ix ) (ei 3 x + e − i 3 x ) (ei 5 x + e − i 5 x ) (ei 7 x + e − i 7 x ) 


f( x ) =  − + − + ...
π 1 3 5 7 

Rewriting as sines and cosines:

4  cos x cos 3x cos 5 x cos 7 x 


f( x ) =  − + − + ...
π 1 3 5 7 

This formula we already know from pages 7-8.

Question:

1 1 1 1 1 1
How much is 1 − + − + − + − ... ?
3 5 7 9 11 13

Answer:

f (0) = 1

4  cos 0 cos 0 cos 0 cos 0 cos 0 


=  − + − + − ...
π 1 3 5 7 9 
Hence:

1 1 1 1 1 1 π
1 − + − + − + − ... =
3 5 7 9 11 13 4

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 26.

Complex Notation of Fourier Series p = 2l

Find the complex form of a periodic function with period 2l:

 integration limits from –1 to 1

 exponential function that is periodic with period 2l: einπx / l

Which results in:

∑c
n = +∞
f(x)= n einπx / l
n = −∞

c n = ∫ f ( x ) ⋅ e −inπx / l dx
l
1
2l − l

 x (2 π − x ) [0 , 2 π]
Given f( x ) = 
 f ( x + 2π )
Exercise 11.

a) obtain the complex Fourier Series for f (x) by using the complex notation

b) rewrite the complex form as the regular sine and cosine series

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 27.

 e 3 x x ∈ [0 , ln 2]
Sketch a graph and find the Fourier series for f( x ) = 
 f ( x + ln 2 )
Exercise 12.

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 28.

 ex
 −x
− ln 2 ≤ x ≤ 0
Given f( x ) = e 0 ≤ x ≤ ln 2
 f ( x + 2 ln 2 )
Exercise 13.

a) Sketch the periodic function for at least 3 periods, is it even or odd or neither?

b) Obtain the complex form of the Fourier series for this function and write down all terms
for –3 ≤ n ≤ 3

c) Rewrite these terms into the sine and cosine representation of the Fourier Series and
comment.

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 29.

Parseval’s Theorem:

∫ ∑
π
1 ∞ 2
1
2π −π
( f ( x ))2
d x = ( 1
2 a 0 )2
+
2 n =1
(a n + b 2n )

Derive Parseval’s equation by squaring f (x) and calculating the average (f (x))2:

f ( x ) = 12 a 0 + ∑ (a n cos nx + b n sin nx )

n =1

 
=  12 a 0 + ∑ (a n cos nx + b n sin nx )
∞ 2

( f (x ))
 
2

n =1


1 1
∫ ∫  (a n cos nx + bn sin nx ) dx
π π ∞ 2
1
( ( )) = +
2 π − π 
2
f x dx a
2π − π
2 0
n =1

Evaluating all the cross terms

 
Average  = ∫ Term dx 
π


1
Term 2π
−π

( 12 a 0 )2 ( 12 a 0 )2
(a n cos nx )2 1
2 a 2n

(bn sin nx )2 1
2 b 2n
1
2 a 0a n cos nx 0
1
2 a 0 b n sin nx 0

a n a m cos nx cos mx m≠n 0


b n b m sin nx sin mx m≠n 0
a n b m cos nx sin mx 0

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 30.

Example:

1
π π
− <x<


2 2
The square wave f ( x ) = − 1
π 3π
<x<
 2 2
 f ( x + 2π )


can be written as the following Fourier series:

4  cos x cos 3x cos 5 x cos 7 x 


f (x ) =  − + − + ...
π 1 3 5 7 

Evaluating the integral on the left hand side of Parseval:

∫ ∫ ∫ (− 1)
+π 3π
π
1 22 2
1
( f ( x ))2
d x = 1 dx +
1 2
dx = 1
2π −π 2π − π 2 2π π
2

The same integral we now know can be expressed as:

( 12 a 0 )2 + 1 ∑ (a 2n + b2n ) = 1 ⋅ 162 1 + 


+ 2 + 2 + 2 + 2 + ...

1 1 1 1 1
2 n =1 2 π  2
3 5 7 9 11 

after substituting for an and bn.

Hence:

  π
1 + 2 + 2 + 2 + 2 + 2 + ... =
2
1 1 1 1 1
 3 5 7 9 11  8

 sin x [− π / 2 , π / 2]
Given the function f( x ) = 
 f( x + π)
Exercise 14.

a) Draw a sketch of f (x), is f (x) even or odd?

b) Show that the Fourier Series of f (x) is given by:

2 4  cos 2 x cos 4 x cos 6 x cos 8 x 


f (x ) = −  + + + + ...
π π  1⋅ 3 3⋅ 5 5⋅ 7 7⋅9 

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 31.

c) By selecting an appropriate value of x calculate:

1 1 1 1 1
− 2 + 2 − 2 + 2 − ...
2 − 1 4 − 1 6 − 1 8 − 1 10 − 1
2

d) Using Parseval’s theorem calculate:

 1 
∑  2 
2

n = even  (n − 1) 

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk


j^RUU=j~íÜÉã~íáÅ~ä=qÉÅÜåáèìÉëW===clrofbo=pbofbp 32.

Exercise 15. Using Parseval’s Theorem, previous questions and some common sense (c),
calculate:

∑n

1
a) 4
n =1

1 1 1 1 1 1
b) 4
+ 4 + 4 + 4 + 4 + 4 + ...
1 3 5 7 9 11

1 1 1 1 1 1
c) 4
+ 4 + 4 + 4 + 4 + 4 + ...
2 4 6 8 10 12

Any questions? Ask at http://webct.kent.ac.uk or email George at G.Dobre@kent.ac.uk

Вам также может понравиться