Applications

Parseval’s Identity

MA 201: Differentiation and Integration of

Fourier Series
Applications of Fourier Series
Lecture - 10

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Fourier Series: Orthogonal Sets

We begin our treatment with some observations: For m, n = 1, 2, 3, . . .

L L
nπx nπx
Z Z
cos dx = 0 & sin dx = 0, (1)
−L L −L L
Z L
mπx nπx
sin cos dx = 0 (2)
−L L L
Z L (
mπx nπx 0, m 6= n
sin sin dx = (3)
−L L L L, m=n
Z L (
mπx nπx 0, m 6= n
cos cos dx = (4)
−L L L L, m=n

Setting x = t − L, observe that

nπx nπx
{1, cos , sin : n = 1, 2 3, . . .},
L L
is orthogonal on [0, 2L].
MA 201 (2016): PDE
 Applications
Parseval’s Identity

Thus, if we have
∞
a0 X n nπx nπx o
f (x) = + an cos + bn sin , x ∈ [0, 2L]
2 n=1
L L

then all coefficients an and bn can be determined as

2L
1 nπx
Z
an = f (x) cos dx, n = 0, 1, 2, . . . and (5)
L 0 L
2L
1 nπx
Z
bn = f (x) sin dx, n = 1, 2, . . . (6)
L 0 L

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Fourier Series in [0, 2L]

• The infinite series
∞
a0 X h nπx nπx i
+ an cos + bn sin , (7)
2 n=1
L L

with
2L
1 nπx
Z
an = f (x) cos dx, n = 0, 1, 2, . . . and (8)
L 0 L
2L
1 nπx
Z
bn = f (x) sin dx, n = 1, 2, . . . (9)
L 0 L

is called Fourier series (FS) of f (x) in [0, 2L].

Remark
The Fourier series of a function is defined whenever the integrals in (8)
and (9) have meaning.
MA 201 (2016): PDE
 Applications
Parseval’s Identity

Fourier Series in [c, c + 2L]

• The infinite series
∞
a0 X h nπx nπx i
+ an cos + bn sin , (10)
2 n=1
L L

with
c+2L
1 nπx
Z
an = f (x) cos dx, n = 0, 1, 2, . . . and (11)
L c L
c+2L
1 nπx
Z
bn = f (x) sin dx, n = 1, 2, . . . (12)
L c L

is called Fourier series (FS) of f (x) in [c, c + 2L].

Remark
The Fourier series of a function is defined whenever the integrals in (11)
and (12) have meaning.
MA 201 (2016): PDE
 Applications
Parseval’s Identity

Gibbs’ Phenomenon
• This is about how the Fourier series of a piecewise continuously
differentiable periodic function behaves at a jump discontinuity. It is
named after the American mathematical physicist J. W Gibbs.
• Consider the following periodic function whose definition in one
period is:

1, 0 < x < 1,
f (x) = (13)
0, 1 < x < 2.
• This function can be represented as
∞
1 2 X sin(2n − 1)πx
f (x) = + , x ∈?. (14)
2 π n=1 2n − 1
• To see how well this infinite series represents the function, let us
truncate the series after N terms.
• Let the sum of these first N terms of the infinite series be denoted
by SN :
N
1 2 X sin(2n − 1)πx
SN = + . (15)
2 π n=1 2n − 1
MA 201 (2016): PDE
 Applications
Parseval’s Identity

Gibbs’ Phenomenon

• Graphs of equation (15) are shown for N = 2, 4, 8, 16 and 32.

• In the figure the overshoot (overshoot is the occurrence of a signal

or function exceeding its target) at x = 1− and the undershoot at
x = 1+ are characteristics of Fourier series at the points of
discontinuity.

• This phenomenon is known as Gibbs’ phenomenon

• This phenomenon persists even though a large number of terms are

considered in the partial sum.
• In the approximation of functions, overshoot/undershoot is one term
describing quality of approximation. Here, the convergence is in
point-wise sense.

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Gibbs’ Phenomenon

Figure : Gibbs Phenomenon with 2 terms

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Gibbs’ Phenomenon

Figure : Gibbs Phenomenon with 4 terms

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Gibbs’ Phenomenon

Figure : Gibbs Phenomenon with 8 terms

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Gibbs’ Phenomenon

Figure : Gibbs Phenomenon with 16 terms

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Gibbs’ Phenomenon

Figure : Gibbs Phenomenon with 32 terms

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Convergence of Fourier series for piecewise continuous

functions

Definition
Let f (t) : [−L, L] → R be a piecewise C 1 -function. Define the adjusted function g (t)
as follows:
 1
 2 [f (t + ) + f (t − )], −L < t < L,
g (t) = (16)
 1 + −
2 [f (−L ) + f (L )], t = ±L.

Note:
The above definition tells that g (t) coincides with f (t) at all points in (−L, L), where
f (t) is continuous, but g (t) is the average of the left-hand and right-hand limits of
f (t) at points of discontinuity in (−L, L).

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Convergence of Fourier series for piecewise continuous

functions

Theorem
Let f : [−L, L] → R be a piecewise C 1 function and let g (t) be the
adjusted function as defined in (16). Then Fourier series of f (t) = g (t),
for all t ∈ [−L, L].

Note:
• The existence of Fourier series depends on the evaluation of Fourier
coefficients.
• On the other hand, convergence of the Fourier series is done via
adjusted function g .
• Thus, a highly oscillating function may be decomposed as sum of a
(possibly infinite) set of simple oscillating trigonometric functions.
• A discontinuous function is approximated by a continuous function.

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Applications of Fourier Series

Square wave-high frequencies
One application of Fourier series, the analysis of a square wave in terms of its Fourier
components, occurs in electronic circuits designed to handle sharply rising pulses.

Physically, square wave contains many high-frequency components.

Figure : Square wave

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Suppose that our modified square wave is defined by


0, −π < x < 0,
f (x) = (17)
h, 0 < x < π.

We can easily calculate the Fourier coefficients to be

A0 = h, An = 0, n = 1, 2, 3, . . . , (18)
 2h
Bn = nπ , n odd, (19)
0, n even.

So, the resulting series is

 
h 2h sin x sin 3x sin 5x
g (x) = + + + + ··· , (20)
2 π 1 3 5
g is the adjusted function.

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Figure : Square wave

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Parseval’s Identity for Fourier Series

• There is a very useful identity obtained from Fourier series which has found
applications in electrical engineering.
• But before knowing Parseval’s identity, we need the following theorem which is
more general:
Theorem RL
If f (t) and g (t) are continuous in (−L, L), and provided −L |f (t)|2 dt < ∞ and
RL 2
−L |g (t)| dt < ∞, and if An , Bn are Fourier coefficients of f (t) and Cn , Dn are
Fourier coefficients of g (t), then
L ∞
L
Z X
f (t)g (t)dt = A0 C0 + L (An Cn + Bn Dn ).
−L 2 n=1

• Proof: We can express f (t) and g (t) in terms of Fourier series as

∞
A0 X h nπt nπt i
f (t) = + An cos + Bn sin ,
2 n=1
L L
∞
C0 X h nπt nπt i
g (t) = + Cn cos + Dn sin .
2 n=1
L L

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Parseval’s Identity for Fourier Series

• Taking product of f (t) with g (t) we obtain

∞ h
A0 X nπt nπt i
f (t)g (t) = g (t) + An g (t) cos + Bn g (t) sin .
2 n=1
L L

• Integrating this series from −L to L gives

∞
" #
L L L L
A0 nπt nπt
Z Z X Z Z
f (t)g (t)dt = g (t)dt+ An g (t) cos dt + Bn g (t) sin dt
−L 2 −L n=1 −L L −L L

• Putting back the values of the Fourier coefficients Cn and Dn :

L ∞
1 A0 C0 X
Z
f (t)g (t)dt = + [An Cn + Bn Dn ].
L −L 2 n=1

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Parseval’s Identity for Fourier Series

• Theorem
(Parseval’s Identity) If f (t) is continuous in the range (−L, L) and is square
RL
integrable (i.e. −L |f (t)|2 dt < ∞) and has Fourier coefficients An and Bn , then

L ∞
1 A20 X 2
Z
[f (t)]2 dt = + [An + Bn2 ].
L −L 2 n=1

• This result can obtained easily from the previous theorem by taking g (t) = f (t).
• The left-hand side represents the mean square value of f (t).
• It can, therefore, be thought of in terms of energy if f (t) represents a signal.
• What Parseval’s theorem states therefore is that the energy of a signal expressed
as a waveform is proportional to the sum of the squares of its Fourier coefficients.

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Parseval’s Identity for Fourier Series

• Parseval’s identity can be used to determine the power delivered by an electric
current, I (t), flowing under a voltage, E (t), through a resistor of resistance R:

P = EI = RI 2 .

• In most applications I (t) is a periodic function.

•
L
1
Z
Average Power = Pav = RI 2 (t) dt
2L −L
Z L
R
= I 2 (t) dt
2L −L
" ∞
#
A20 1X 2 2
= R + (A + Bn ) ,
4 2 n=1 n

• Here we have made use of the Fourier expansion of I (t):

∞
A0 X h nπt nπt i
I (t) = + An cos + Bn sin .
2 n=1
L L

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Parseval’s Identity for Fourier Series

• Mean square of the current is:

L ∞
1 A20 1X 2
Z
Iav = I 2 (t)dt = + (A + Bn2 ).
2L −L 4 2 n=1 n

• Root mean square of the current is:

v
u 2 ∞
uA 1X 2
Irms =t 0 + (A + Bn2 ).
4 2 n=1 n

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Application of Parseval’s Identity

∞
π2 X (−1)n
• Example: Given the Fourier series t 2 = +4 cos nt, −π < t < π,
3 n=1
n2
∞
X 1 π4
deduce that 4
= .
n=1
n 90
2π 2 4(−1)n
• Here A0 = , An = , Bn = 0.
3 n2
• Parseval’s identity is

L ∞
1 A20 X 2
Z
[f (t)]2 dt = + [An + Bn2 ].
L −L 2 n=1

so that here L = π.
• We get
∞
1 π
2π 4 X 16
Z
t 4 dx = +
π −π 9 n=1
n4
giving us
∞
2π 4 2π 4 X 16
= +
5 9 n=1
n4
• Hence the result follows.
MA 201 (2016): PDE
 Applications
Parseval’s Identity

Finite Vibrating String Problem

• The IBVP under consideration consists of the following:

• The governing equation:

utt = c 2 uxx , (x, t) ∈ (0, L) × (0, ∞). (21)

The boundary conditions for all t > 0:

u(0, t) = 0, u(L, t) = 0. (22)

The initial conditions for 0 ≤ x ≤ L are

u(x, 0) = φ(x), ut (x, 0) = ψ(x). (23)

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Bernoulli’s Solution
• In [0, π], Bernoulli gave the solution of (21) as a series of the form
u = b1 sin x cos ct + b2 sin 2x cos 2ct + . . . . (24)
• When t = 0, we should have
φ(x) = b1 sin x + b2 sin 2x + . . . . (25)
This is possible and can be treated as the Fourier sine series of φ(x),
and which converges to φ(x).
Few Facts:
As we are expecting u(x, t) as a solution of (21), term-wise
differentiation of the series given by (24) should exist. In other sense, the
limit function u of the infinite series (24) should be differentiable.
Term-wise differentiation of an infinite series is not always possible.
In fact, infinite sum of continuous functions may have discontinuous
limit.
1 1 1 + x2
1+ 2
+ 2 2
+ ... = .
1+x (1 + x ) x2
MA 201 (2016): PDE
 Applications
Parseval’s Identity

Differentiation and integration of Fourier series

The term-by-term differentiation of a Fourier series is not always permissible.
Example
Recall that Fourier series for f (x) = x, −π < x < π is
∞
X sin nx
(−1)n+1 ,
n=1
n

which converges to f (x) for all x ∈ (−π, π), that is

∞
X sin nx
x= (−1)n+1 .
n=1
n

Term-by-term differentiation leads to

∞
X
1= (−1)n+1 cos nx, x ∈ (−π, π).
n=1

which fails at x = 0. In fact, the RHS series diverges for all x(?)

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Differentiation of Fourier series

Theorem (Differentiation of Fourier series)

Let f (x) : R → R be continuous and f (x + 2L) = f (x). Let f 0 (x) and
f 00 (x) be piecewise continuous on [−L, L]. Then, The Fourier series of
f 0 (x) can be obtained from the Fourier series for f (x) by termwise
differentiation. In particular, if
∞
A0 X n nπx nπx o
f (x) = + An cos + Bn sin ,
2 n=1
L L

then ∞
X nπ n nπx nπx o
f 0 (x) = −An sin + Bn cos .
n=1
L L L

MA 201 (2016): PDE

 Applications
Parseval’s Identity

Integration of Fourier series

Termwise integration of a Fourier series is permissible under
much weaker conditions.

Theorem (Integration of Fourier series)

Let f (x) : [−L, L] → R be piecewise continuous function with Fourier
series
∞
A0 X n nπx nπx o
f (x) = + An cos + Bn sin .
2 n=1
L L

Then, for any x ∈ [−L, L], we have

x x ∞ x
A0 nπτ nπτ o
Z Z X Z n
f (τ )dτ = dτ + An cos + Bn sin dτ.
−L −L 2 n=1 −L L L

MA 201 (2016): PDE