Fourier Series:) ) Sin Cos ( (
Fourier Series:) ) Sin Cos ( (
Fourier Series:) ) Sin Cos ( (
doc
Fourier Series
2.1 Introduction
f(x) = a
n 0
n xn
where ao, a1, a2, … are constants, called the coefficients of the series. A Taylor series does not
include terms with negative powers. A Fourier series is an infinite series expansion in terms of
trigonometric functions
f(x) = ao +
n 1
( a n cos(nx ) bn sin( nx ))
(2.1-1)
Any piecewise smooth function defined on a finite interval has a Fourier series expansion.
Periodic Functions
A function satisfying the identity f(x) = f(x + T) for all x, where T > 0, is called periodic or T-
periodic as shown in Figure 2.1-1.
y
T
If T is a period then nT is also a period for any integer n > 0. T is called a fundamental period.
The definite integral of a T-periodic function is the same over any interval of length T. Example
2.1-1 will use this property to integrate a 2-periodic function shown in Figure 2.1-2.
19
491422464.doc
______________________________________________________________________________
Example 2.1-1. Let f be the 2-periodic function and N is a positive integer. Compute
N
N
f 2 ( x )dx if f(x) = x + 1 on the interval 0 x 2
y
1
x
-3 -2 -1 1 2 3
-1
Figure 2.1-1. A 2-periodic function.
Solution
N N 2 N 4 N
N
f 2 ( x )dx = N
f 2 ( x )dx +
N 2
f 2 ( x ) dx + + N 2
f 2 ( x ) dx
2
N N 2 1
N
f 2 ( x )dx = N
N 2
f 2 ( x )dx = N ( x 1) 2 dx = N
0 3
( x 1) 3
0
N N 2
N
f 2 ( x )dx =
3
[ 1 1] =
3
N
______________________________________________________________________________
The most important periodic functions are those in the (2-period) trigonometric system
Orthogonal functions
b
If
a
f ( x ) g ( x )dx = 0 then f and g are orthogonal over the interval [a, b].
20
491422464.doc
Fourier series are special expansions of functions of the form
f(x) = ao +
n 1
( a n cos(nx ) bn sin( nx )) (2.1-1)
where the coefficients a0, a1, a2, , b1, b2, must be evaluated.
The coefficient a0 is determined by integrating both sides of Eq. (2.1-1) over the interval [, ].
f ( x )dx =
a 0 dx + ( a n
n 1
cos( nx ) bn sin( nx )) dx
Since cos nxdx = sin nxdx = 0 for n = 1, 2,
1
f ( x )dx = a
0 dx = 2 a0 a0 =
2
f ( x ) dx
The coefficient an is determined by multiplying both sides of Eq. (2.1-1) with cos mx and
integrating the resulting equation over the interval [, ].
f ( x ) cos(mx )dx =
a 0 cos( mx ) dx + a
n 1
n cos( nx ) cos( mx ) dx
+ b
n 1
n sin( nx ) cos( mx )dx
Since cos mxdx = 0, cos mx sin nxdx = 0 for all m and cos mx cos nxdx = 0 for m
n
(cos nx ) dx = an
2
f ( x ) cos(mx )dx = an
1 1 2
an =
f ( x ) cos( mx )dx =
0
f ( x ) cos(mx )dx
Similarly the coefficient bn is determined by multiplying both sides of Eq. (2.1-1) with sin mx
and integrating the resulting equation over the interval [, ].
1 1 2
bn =
f ( x ) sin( mx )dx =
0
f ( x ) sin(mx )dx
21
491422464.doc
_____________________________________________________________________________
Example 2.1-2. Solve the one dimensional heat equation with no heat generation, zero boundary
conditions (0oC), and constant initial temperature of 100oC.
Solution
T 2T
Cp =k + Q(x, t)
t x 2
nx n 2
T(x, t) = b n sin
L
exp t
n 1 L
nx
At t = 0, T(x, 0) = 100 = b
n 1
n sin
L
This is the Fourier sine series (with a0, a1, a2, = 0) where the coefficients b1, b2, can be
determined by by multiplying both sides of the above equation with sin mx and integrating the
resulting equation over the interval [0, L].
mx
nx mx
b
L L
100 0 sin dx = n sin sin dx
L n 1
0 L L
L nx m x
Since
0
sin
L
sin
L
dx = 0 for m n
n x
2
n x
dx = bn sin
L L
100 0 sin dx
L 0
L
1 cos 2 x
Using the identity sin2x = , the above equation becomes
2
L
L n x L 1 2 n x
100 cos = bn 0 1 cos dx
n L 0 2 L
22
491422464.doc
L
L b L 2nx bn
100 [ cos n + 1] = n x sin = L
n 2 2 n L 0 2
200
bn = [1 cos n ]
n
400
For n = odd, cos(n) = 1 bn =
n
( 2n 1)x 51
The plot of f(x) =
n 1 L
b sin
for 51 terms is a good approximation of 100 away
2 n 1
from the end points as shown in Figure 2.1-1. There is a 18 % overshoot called Gibbs
phenomenon near the end points. Gibbs phenomenon occurs only when a finite series of
eigenfunctions approximates a discontinuous function.
120
100
Fourier expansion of 100
80
60
40
20
0
0 0.2 0.4 0.6 0.8 1
x/L
23
491422464.doc
( 2n 1)x
Figure 2.1-2. 51 points approximation of f(x) = 100 = b
n 1
2 n 1 sin
L
51
400 ( 2n 1)x
_________Table 2.1-1 Matlab program to plot f(x) = n
n 1
sin
L
___________
24