13 Fourier Series (Zill)

.
 FOURIER SERIES 431
27. Consider the set of functions h1, x, x2, x3, Á j defined on the 28. Relate the orthogonal set B9 in Problem 27 with a set of
interval [–1, 1]. Apply the Gram-Schmidt process given in polynomials found in an earlier chapter of this text.
Problem 26 to this set and find ␾0(x), ␾1(x), ␾2(x), and ␾3(x) of
the orthogonal set B9.
11.2 Fourier Series

INTRODUCTION We have just seen that if {␾0(x), ␾1(x), ␾2(x), . . .} is an
orthogonal set on an interval [a, b] and if f is a function defined on the same interval,
then we can formally expand f in an orthogonal series
c0␾0(x) 1 c1␾1(x) 1 c2␾2(x) 1 Á ,
where the coefficients cn are determined by using the inner product concept. The
orthogonal set of trigonometric functions
51, cos ␲p x, cos 2␲p x, cos 3␲p x, . . . , sin ␲p x, sin 2␲p x, sin 3␲p x, . . .6 (1)
will be of particular importance later on in the solution of certain kinds of boundary-

value problems involving linear partial differential equations. The set (1) is orthogo-
nal on the interval [2p, p]. See Problem 12 in Exercises 11.1.
A TRIGONOMETRIC SERIES Suppose that f is a function defined on the interval

(2p, p) and can be expanded in an orthogonal series consisting of the trigonometric
functions in the orthogonal set (1); that is,
o1 2
a0 ` n␲ n␲
f (x) 5 1 an cos x 1 bn sin x . (2)
2 n51 p p
The coefficients a0, a1, a2, . . . , b1, b2, . . . can be determined in exactly the same for-
mal manner as in the general discussion of orthogonal series expansions on page 428.
Before proceeding, note that we have chosen to write the coefficient of 1 in the set (1)
as 12 a0 rather than a0. This is for convenience only; the formula of an will then reduce
to a0 for n 5 0.
Now integrating both sides of (2) from 2p to p gives
# # o1 # # 2
p a0 p ` p n␲ p
n␲
f (x) dx 5 dx 1 an cos x dx 1 bn sin x dx . (3)
2p 2 2p n51 2p p 2p p
Since cos(n␲xyp) and sin(n␲xyp), n $ 1 are orthogonal to 1 on the interval, the right
side of (3) reduces to a single term:
# # u
p p p
a0 a0
f (x) dx 5 dx 5 x 5 pa0.
2p 2 2p 2 2p
Solving for a0 yields
#
1 p
a0 5 f (x) dx. (4)
p 2p
Now we multiply (2) by cos(m␲xyp) and integrate:
# #
p m␲ a0 p m␲
f (x) cos x dx 5 cos x dx
2p p 2 2p p
1
`
o
n51
S#
an
p
2p
cos
m␲
p
x cos
n␲
p
x dx 1 bn #
2p
p
cos
m␲
p
x sin
n␲
p
x dx . D (5)
432 CHAPTER  FOURIER SERIES
By orthogonality we have
# #
p m␲ p m␲ n␲
cos x dx 5 0, m . 0, cos x sin x dx 5 0,
2p p 2p p p
# 0, m Þ n
5
p m␲ n␲
and cos x cos x dx 5
2p p p p, m 5 n.
#
p n␲
Thus (5) reduces to f (x) cos x dx 5 an p,
2p p
#
1 p n␲
and so an 5 f (x) cos x dx. (6)
p 2p p
Finally, if we multiply (2) by sin(m␲xyp), integrate, and make use of the results
# #
p m␲ p m␲ n␲
sin x dx 5 0, m . 0, sin x cos x dx 5 0,
2p p 2p p p
# 0, m Þ n
5
p m␲ n␲
and sin x sin x dx 5
2p p p p, m 5 n,
#
1 p n␲
we find that bn 5 f (x) sin x dx. (7)
p 2p p
The trigonometric series (2) with coefficients a0, an, and bn defined by (4), (6),
and (7), respectively, is said to be the Fourier series of f. Although the French math-
ematical physicist Jean Baptiste Joseph Fourier (1768–1830) did not invent the
series that bears his name, he is at least responsible for sparking the interest of
mathematicians in trigonometric series by his less than rigorous use of them in his
researches on the conduction of heat. The formulas in (4), (6), and (7) that give the
coefficients in a Fourier series are known as the Euler formulas.
DEFINITION 11.2.1 Fourier Series

The Fourier series of a function f defined on the interval (2p, p) is given by
f (x) 5
a0
2
1
`
n51
o
an cos
n␲x
p S
1 bn sin
n␲x
p D (8)
#
1 p
where a0 5 f (x) dx (9)
p 2p
#
1 p n␲x
an 5 f (x) cos dx (10)
p 2p p
#
1 p n␲x
bn 5 f (x) sin dx. (11)
p 2p p
CONVERGENCE OF A FOURIER SERIES In the absence of any stated conditions that

guarantee the validity of the steps leading to the coefficients a0, an, and bn, the equality
sign in (8) should not be taken in a strict or literal sense. Some texts use the symbol , to
signify that (8) is simply the corresponding trigonometric series with coefficients
generated using f in formulas (9)–(11). In view of the fact that most functions in
applications are of the type that guarantee convergence of the series, we shall use the
equality symbol. Is it possible for a series (8) to converge at a number x in the interval
(2p, p), and yet not be equal to f (x)? The answer is an emphatic Yes.
Copyright 2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-203
. FOURIER SERIES 433
PIECEWISECONTINUOUS FUNCTIONS Before stating conditions under which

a Fourier series converges, we need to pause briefly to review two topics from the
first semester of calculus. We shall use the symbols f(x1) and f(x2) to denote the
one-sided limits
f (x1) 5 lim f (x 1 h), f (x2) 5 lim f (x 2 h),

hS0 hS0
h.0 h.0
called, respectively, the right- and left-hand limits of f at x. A function f is said to be

In Section 7.1 we deﬁned piecewise piecewise continuous on a closed interval [a, b] if there are
continuity on an unbounded
interval [0, `). See Figure 7.1.1 on ●a finite number of points x1 , x2 , . . . , xn in [a, b] at which f has a finite
page 282. (or jump) discontinuity, and
● f is continuous on each open interval (xk, x k11).
As a consequence of this definition, the one-sided limits f(x1) and f(x2) must exist
at every x satisfying a , x , b. The limits f(a1) and f(b2) must also exist but it is
not required that f be continuous or even defined at either a or b.
Our first theorem gives sufficient conditions for convergence of a Fourier series
at a point x.
THEOREM 11.2.1 Conditions for Convergence

Let f and f 9 be piecewise continuous on the interval [2p, p]. Then for all x in
the interval (2p, p), the Fourier series of f converges to f (x) at a point continu-
ity. At a point of discontinuity the Fourier series converges to the average
f (x1) 1 f (x2)
,
2
where f (x1) and f (x2) are the right- and left-hand limits of f at x, respectively.
EXAMPLE 1 Expansion in a Fourier Series
Expand f (x) 5 50,␲ 2 x, 2␲ , x , 0

0#x,␲
(12)
in a Fourier series.
y SOLUTION The graph of f is given in Figure 11.2.1. With p 5 ␲ we have from

(9) and (10) that
␲
3# # (␲ 2 x) dx4 5 ␲1 3␲x 2 x2 4
2 ␲
#
2␲ ␲ x 1 ␲ 1 0 ␲ ␲
a0 5 f (x) dx 5 0 dx 1 5
␲ 2␲ ␲ 2␲ 0 0 2
FIGURE 11.2.1 Piecewise-continuous
function f in Example 1
# 3# # (␲ 2 x) cos nx dx4
1 ␲ 1 0 ␲
an 5 f (x) cos nx dx 5 0 dx 1
␲ 2␲ ␲ 2␲ 0
3 # sin nx dx4
␲ ␲
u
1 sin nx 1 ; integration
5 (␲ 2 x) 1
␲ n 0 n 0
by parts
u
␲ 1 2 (21)n
1 cos nx
52 5 ,
n␲ n 0 n2␲
where we have used cos n␲ 5 (21) . In like manner we find from (11) that
n
# (␲ 2 x) sin nx dx 5 n1.
1 ␲
bn 5
␲ 0
o5 6
␲ ` 1 2 (21)n 1
Therefore f (x) 5 1 cos nx 1 sin nx . (13)
4 n51 n2␲ n
Note that an defined by (10) reduces to a0 given by (9) when we set n 5 0. But as
Example 1 shows, this might not be the case after the integral for an is evaluated.
EXAMPLE 2 Example 1 Revisited

y' The equality in (13) of Example 1 is justified because both f and f 9are piecewise
continuous on the interval [2␲, ␲]. See Figures 11.2.1 and 11.2.2. Because f is con-
2␲ ␲ x tinuous for every x in the interval (2␲, ␲), except at x 5 0, the series (13) will con-
21 verge to the value f(x). At x 5 0 the function f is discontinuous, so the series (13) will
converge to
FIGURE 11.2.2 Piecewise-continuous f (01) 1 f (02) ␲
derivative f 9 in Example 2 5 .
2 2
PERIODIC EXTENSION Observe that each of the functions in the basic set (1) has
a different fundamental period*—namely, 2pyn, n $ 1—but since a positive integer
multiple of a period is also a period, we see that all the functions have in common the
period 2p. (Verify.) Hence the right-hand side of (2) is 2p-periodic; indeed, 2p is the
fundamental period of the sum. We conclude that a Fourier series not only repre-
sents the function on the interval (2p, p) but also gives the periodic extension of f
outside this interval. We can now apply Theorem 11.2.1 to the periodic extension of f,
or we may assume from the outset that the given function is periodic with period 2p;
that is, f (x 1 2p) 5 f (x). When f is piecewise continuous and the right- and left-hand
derivatives exist at x 5 2p and x 5 p, respectively, then the series (8) converges to
the average
f (p2) 1 f (2p1)
2
at these endpoints and to this value extended periodically to 63p, 65p, 67p, and
so on.
EXAMPLE 3 Example 1 Revisited

The Fourier series (13) in Example 1 converges to the periodic extension of the func-
tion (12) on the entire x-axis. See Figure 11.2.3. At 0, 62␲, 64␲, . . . and at
6␲, 63␲, 65␲, . . . the series converges to the values
f (01) 1 f (02) ␲ f (␲2) 1 f (2␲1)

5 and 5 0,
2 2 2
respectively. The solid black dots in Figure 11.2.3 represent the value ␲y2.
24␲ 23␲ 22␲ 2␲

x
␲ 2␲ 3␲ 4␲
FIGURE 11.2.3 Periodic extension of function f shown in Figure 11.2.1
*
See Problems 19–24 in Exercises 11.1.
. FOURIER SERIES 435
SEQUENCE OF PARTIAL SUMS It is interesting to see how the sequence of par-

tial sums {SN(x)} of a Fourier series approximates a function. For example, the first
three partial sums of (13) in Example 1 are
␲ ␲ 2 ␲ 2 1
S1 (x) 5 , S 2 (x) 5 1 cos x 1 sin x, and S 3 (x) 5 1 cos x 1 sin x 1 sin 2x.
4 4 ␲ 4 ␲ 2
In Figure 11.2.4 we have used a CAS to graph the partial sums S3(x), S8(x), and S15(x)
of (13) on the interval (2␲, ␲). Figure 11.2.4(d) shows the periodic extension using
S15(x) on (24␲, 4␲).
y y
3 3
2 2
1 1
x x
23 22 21 1 2 3 23 22 21 1 2 3
(a) S3(x) (b) S8(x)
y y
3 3
2 2
1 1
x x
23 22 21 1 2 3 210 25 5 10
(c) S15(x) (d) S15(x)
FIGURE 11.2.4 Partial sums of Fourier series (13) in Example 1
EXERCISES 11.2 Answers to selected odd-numbered problems begin on page ANS-18.
5␲␲ ,2 x ,
In Problems 1–16 find the Fourier series of f on the given interval. 2
2␲ , x , 0
Give the number to which the Fourier series converges at a point of 6. f (x) 5 2 2
0#x,␲
discontinuity of f.
7. f (x) 5 x 1 ␲, 2␲ , x , ␲
1. f (x) 5 5
0, 2␲ , x , 0
1, 0#x,␲
8. f (x) 5 3 2 2x, 2␲ , x , ␲
2. f (x) 5 5
21, 2␲ , x , 0
2, 0#x,␲
9. f (x) 5 50,sin x, 2␲ , x , 0
0#x,␲
3. f (x) 5 5
50,cos x,
1, 21 , x , 0 2␲y2 , x , 0
10. f (x) 5
x, 0#x,1 0 # x , ␲y2
4. f (x) 5 5
0, 21 , x , 0
5
0, 22 , x , 21
x, 0#x,1 22, 21 # x , 0
11. f (x) 5
5. f (x) 5 5
0, 2␲ , x , 0 1, 0#x,1
x , 0#x,␲
2 0, 1#x,2
5
0, 22 , x , 0 21. Use the result of Problem 7 to show that
12. f (x) 5 x, 0#x,1 ␲ 1 1 1
5 1 2 1 2 1 Á.
1, 1#x,2 4 3 5 7
22. Use the result of Problem 9 to show that
13. f (x) 5 5 1,
1 1 x,
25 , x , 0
0#x,5 ␲ 1
5 1
1
2
1
1
1
2
1
1 Á.
4 2 1?3 3?5 5?7 7?9
14. f (x) 5 5
2 1 x, 22 , x , 0
23. (a) Use the complex exponential form of the cosine and sine,
2, 0#x,2
n␲ ein␲x/p 1 e2in␲x/p
15. f (x) 5 e x, 2␲ , x , ␲ cos x5
p 2
16. f (x) 5 50,e 2 1,
x
2␲ , x , 0
0#x,␲ sin
n␲
p
x5
ein␲x/p 2 e2in␲x/p
2i
,
to show that (8) can be written in the complex form

In Problems 17 and 18 sketch the periodic extension of the indicated
`
function.
f (x) 5
n52`
o cn ein␲x/p,
17. The function f in Problem 9
where
18. The function f in Problem 14
a0 an 2 ibn an 1 ibn
c0 5 , cn 5 , and c2n 5 ,
19. Use the result of Problem 5 to show that 2 2 2
where n 5 1, 2, 3, . . . .
␲2 1 1 1
511 21 21 21 Á
6 2 3 4 (b) Show that c0, cn, and c2n of part (a) can be written as one
integral
␲2 1 1 1
5 1 2 2 1 2 2 2 1 Á.
#
1 p
and cn 5 f (x) e2in␲x/p dx, n 5 0, 61, 62, . . . .
12 2 3 4 2p 2p
20. Use Problem 19 to find a series that gives the numerical value 24. Use the results of Problem 23 to find the complex form of the
of ␲2y8. Fourier series of f (x) 5 e2x on the interval [2␲, ␲].
11.3 Fourier Cosine and Sine Series

INTRODUCTION The effort that is expended in evaluation of the definite integrals
that define the coefficients the a0, an, and bn in the expansion of a function f in a
Fourier series is reduced significantly when f is either an even or an odd function.
Recall that a function f is said to be
even if f (2x) 5 f (x) and odd if f (2x) 5 2f (x).
On a symmetric interval such as (2p, p) the graph of an even function possesses
symmetry with respect to the y-axis, whereas the graph of an odd function possesses
symmetry with respect to the origin.
EVEN AND ODD FUNCTIONS It is likely that the origin of the terms even and odd
derives from the fact that the graphs of polynomial functions that consist of all even
powers of x are symmetric with respect to the y-axis, whereas graphs of polynomials
that consist of all odd powers of x are symmetric with respect to origin. For example,
even integer
f(x) 5 x2 is even since f(2x) 5 (2x)2 5 x2 5 f (x)
odd integer
f(x) 5 x3 is odd since f(2x) 5 (2x)3 5 2x3 5 2f(x).
. FOURIER COSINE AND SINE SERIES 437
y See Figures 11.3.1 and 11.3.2. The trigonometric cosine and sine functions are even
y 5 x2
and odd functions, respectively, since cos(2x) 5 cos x and sin(2x) 5 2sin x. The
exponential functions f (x) 5 e x and f (x) 5 e2x are neither odd nor even.
f (2x) f (x)
2x x x PROPERTIES The following theorem lists some properties of even and odd
functions.
FIGURE 11.3.1 Even function; graph

symmetric with respect to y-axis THEOREM 11.3.1 Properties of Even/Odd Functions
(a)(a) The product
The product of even
of two two even functions
functions is even.
is even.
y y5 x3 (b) The product of two odd functions is even.
(c) The product of an even function and an odd function is odd.
(d) The sum (difference) of two even functions is even.
2x f (x)
(e) The sum (difference) of two odd functions is odd.
f(2x) x x (f ) If f is even, then e2a
a
f (x) dx 5 2e0a f (x) dx.
(g) If f is odd, then e2a f (x) dx 5 0.
a
FIGURE 11.3.2 Odd function; graph

symmetric with respect to origin
PROOF OF b Let us suppose that f and g are odd functions. Then we
have f (2x) 5 2f (x) and g(2x) 5 2g(x). If we define the product of f and g as
F(x) 5 f (x)g(x), then
F(2x) 5 f (2x) g(2x) 5 (2f (x))(2g(x)) 5 f (x) g(x) 5 F(x).
This shows that the product F of two odd functions is an even function. The proofs of
the remaining properties are left as exercises. See Problem 52 in Exercises 11.3.
COSINE AND SINE SERIES If f is an even function on (2p, p), then in view of the
foregoing properties the coefficients (9), (10), and (11) of Section 11.2 become
1
a0 5 –
p E p
2p
2
f(x) dx 5 –
p E p
0
f(x) dx
1
an 5 –
p E p
2p
n␲ 2
f(x) cos ––– x dx 5 –
p p E p
0
n␲
f(x) cos –––
p x dx
even
1
bn 5 –
p E p
2p
n␲
f(x) sin ––– x dx 5 0.
p
odd
Similarly, when f is odd on the interval (2p, p),
# f (x) sin n␲p x dx.

2 p
an 5 0, n 5 0, 1, 2, . . . , bn 5
p 0
We summarize the results in the following definition.
DEFINITION 11.3.1 Fourier Cosine and Sine Series

(i) The Fourier series of an even function f defined on the interval (2p, p) is
the cosine series
a0 ` n␲
f (x) 5
2
1
n51
o
an cos
p
x, (1)
(continued)
# f (x) dx
2 p
where a0 5 (2)
p 0
# f (x) cos n␲p x dx.

2 p
an 5 (3)
p 0
(ii) The Fourier series of an odd function f defined on the interval (2p, p) is
the sine series
` n␲
f (x) 5 o bn sin
n51 p
x, (4)
# f (x) sin n␲p x dx.

2 p
where bn 5 (5)
p 0
Because the term sin(n␲xyp) is 0 at x 5 2p, x 5 0, and x 5 p, a sine series (4)

converges to 0 at those points regardless of whether f is defined at these points.
EXAMPLE 1 Expansion in a Sine Series

Expand f (x) 5 x, 22 , x , 2 in a Fourier series.
y SOLUTION Inspection of Figure 11.3.3 shows that the given function is odd on the
interval (22, 2), and so we expand f in a sine series. With the identification 2p 5 4
we have p 5 2. Thus (5), after integration by parts, is
x
# x sin n␲2 x dx 5 4(21)

2 n11
bn 5 .
y 5 x, 22 , x , 2 0 n␲
FIGURE 11.3.3 Odd function in Example 1 4 `(21)n11 n␲

Therefore f (x) 5
␲ o
n51 n
sin
2
x. (6)
The function in Example 1 satisfies the conditions of Theorem 11.2.1. Hence

the series (6) converges to the function on (22, 2) and the periodic extension
(of period 4) given in Figure 11.3.4.
210 28 26 24 22 2 4 6 8 10 x
FIGURE 11.3.4 Periodic extension of function shown in Figure 11.3.3
EXAMPLE 2 Expansion in a Sine Series
1
The function f (x) 5 521,1, 2␲ , x , 0
0 # x , ␲,
shown in Figure 11.3.5 is odd on the
interval (2␲, ␲). With p 5 ␲ we have, from (5),
# (1) sin nx dx 5 ␲2 1 2 (21)

2 ␲ n
2␲ ␲ x
bn 5 ,
␲ 0 n
21
2 `1 2 (21)n
FIGURE 11.3.5 Odd function in Example 2
and so f (x) 5
␲ o
n51 n
sin nx. (7)
GIBBS PHENOMENON With the aid of a CAS we have plotted the graphs S1(x),
S2(x), S3(x), and S15(x) of the partial sums of nonzero terms of (7) in Figure 11.3.6.
As seen in Figure 11.3.6(d), the graph of S15(x) has pronounced spikes near the dis-
continuities at x 5 0, x 5 ␲, x 5 2␲, and so on. This “overshooting” by the partial
sums SN from the functional values near a point of discontinuity does not smooth out
but remains fairly constant, even when the value N is taken to be large. This behav-
ior of a Fourier series near a point at which f is discontinuous is known as the Gibbs
phenomenon.
The periodic extension of f in Example 2 onto the entire x-axis is a meander
function (see page 316).
y y
1 1
0.5 0.5
x x
20.5 20.5
21
23 22 21 1 2 3 23 22 21 1 2 3
(a) S1(x) (b) S2(x)
y y
1 1
0.5 0.5
x x
y 20.5 20.5
21 21
23 22 21 1 2 3 23 22 21 1 2 3
x (c) S3(x) (d) S15(x)
2L L
FIGURE 11.3.6 Partial sums of sine series (7)
FIGURE 11.3.7 Even reflection
y
HALFRANGE EXPANSIONS Throughout the preceding discussion it was under-
stood that a function f was defined on an interval with the origin as its midpoint—that
2L
is, (2p, p). However, in many instances we are interested in representing a function
L
x that is defined only for 0 , x , L by a trigonometric series. This can be done in
many different ways by supplying an arbitrary definition of f (x) for 2L , x , 0.
For brevity we consider the three most important cases. If y 5 f (x) is defined on the
interval (0, L), then
FIGURE 11.3.8 Odd reflection
(i) reflect the graph of f about the y-axis onto (2L, 0); the function is now
y even on (2L, L) (see Figure 11.3.7); or
(ii) reflect the graph of f through the origin onto (2L, 0); the function is now
odd on (2L, L) (see Figure 11.3.8); or
(iii) define f on (2L, 0) by y 5 f (x 1 L) (see Figure 11.3.9).
x
2L L
Note that the coefficients of the series (1) and (4) utilize only the definition of
the function on (0, p) (that is, half of the interval (2p, p)). Hence in practice there
f(x) 5 f(x 1 L) is no actual need to make the reflections described in (i) and (ii). If f is defined for
0 , x , L, we simply identify the half-period as the length of the interval p 5 L. The
FIGURE 11.3.9 Identity reflection coefficient formulas (2), (3), and (5) and the corresponding series yield either an
even or an odd periodic extension of period 2L of the original function. The cosine
and sine series that are obtained in this manner are known as half-range expan-
sions. Finally, in case (iii) we are defining the function values on the interval (2L, 0)
to be same as the values on (0, L). As in the previous two cases there is no real need
to do this. It can be shown that the set of functions in (1) of Section 11.2 is orthogo-
nal on the interval [a, a 1 2p] for any real number a. Choosing a 5 2p, we obtain
the limits of integration in (9), (10), and (11) of that section. But for a 5 0 the limits
of integration are from x 5 0 to x 5 2p. Thus if f is defined on the interval (0, L), we
identify 2p 5 L or p 5 Ly2. The resulting Fourier series will give the periodic exten-
sion of f with period L. In this manner the values to which the series converges will
be the same on (2L, 0) as on (0, L).
EXAMPLE 3 Expansion in Three Series

Expand f (x) 5 x2, 0 , x , L,
(a) in a cosine series (b) in a sine series (c) in a Fourier series.
y SOLUTION The graph of the function is given in Figure 11.3.10.

y 5x , 0 , x , L (a) We have
# x dx 5 32 L , # x cos n␲L x dx 5 4Ln(21)

L L 2 n
2 2
a0 5 2 2
an 5 2
,
L
x L 0 L 0 ␲ 2 2
FIGURE 11.3.10 Function f in Example 3

where integration by parts was used twice in the evaluation of an. Thus
is neither odd nor even.
L2 4L2 ` (21)n n␲
f (x) 5
3
1 2
␲ o
n51 n
2
cos
L
x. (8)
(b) In this case we must again integrate by parts twice:
# 2L2(21)n11 4L2
2 L n␲
bn 5 x2 sin x dx 5 1 3 3 [(21)n 2 1].
L 0 L n␲ n␲
o5 6
2L2 ` (21)n11 2 n␲
Hence f (x) 5 1 3 2 [(21)n 2 1] sin x. (9)
␲ n51 n n␲ L
(c) With p 5 Ly2, 1yp 5 2yL, and n␲yp 5 2n␲yL we have
# x dx 5 32 L , # x cos 2n␲
L L 2
2 2 L
a0 5 2 2
an 5 2
x dx 5 ,
L 0 L 0 L n␲ 2 2
# L2
2 L 2n␲
and bn 5 x2 sin x dx 5 2 .
L 0 L n␲
o 5n2␲ cos 6
L2 L 2 ` 1 2n␲ 1 2n␲
Therefore f (x) 5 1 x 2 sin x . (10)
3 ␲ n51 L n L
The series (8), (9), and (10) converge to the 2L-periodic even extension of f, the
2L-periodic odd extension of f, and the L-periodic extension of f, respectively. The
graphs of these periodic extensions are shown in Figure 11.3.11.
24L 23L 22L 2L L 2L 3L 4L x
(a) cosine series
24L 23L 22L 2L L 2L 3L 4L x
(b) sine series
24L 23L 22L 2L L 2L 3L 4L x
(c) Fourier series
FIGURE 11.3.11 Same function on (0, L) but different periodic extensions
PERIODIC DRIVING FORCE Fourier series are sometimes useful in determining a

particular solution of a differential equation describing a physical system in which
the input or driving force f (t) is periodic. In the next example we find a periodic par-
ticular solution of the nonhomogeneous linear differential equation
d 2x
m 1 kx 5 f (t) (11)
dt2
by first representing f by a half-range sine expansion and then assuming a partic-

ular solution of the form
` n␲
xp (t) 5 o Bn sin
n51 p
t. (12)
EXAMPLE 4 Particular Solution of a DE

1
An undamped spring/mass system, in which the mass m 5 16 slug and the
spring constant k 5 4 lb/ft, is driven by the 2-periodic external force f (t) shown in
f (t) Figure 11.3.12. Although the force f (t) acts on the system for t . 0, note that if we
␲ extend the graph of the function in a 2-periodic manner to the negative t-axis, we
obtain an odd function. In practical terms this means that we need only find the half-
range sine expansion of f (t) 5 ␲t, 0 , t , 1. With p 5 1 it follows from (5) and
integration by parts that
1 2 3 4 5 t
# 2(21)n11
1
bn 5 2 ␲t sin n␲t dt 5 .
0 n
2␲
From (11) the differential equation of motion is seen to be
FIGURE 11.3.12 Periodic forcing `
1 d 2x 2(21) n11
function for spring/mass system in
Example 4 16 dt 2
1 4x 5 o
n51 n
sin n␲t. (13)
To find a particular solution xp(t) of (13), we substitute (12) into the differential equa-
tion and equate coefficients of sin n␲t. This yields
S 2
1 2 2
16
n ␲ 1 4 Bn 5D2(21)n11
n
or Bn 5
32(21)n11
n(64 2 n2␲2)
.
` 32(21) n11
Thus xp (t) 5 o n(64 2 n2␲2) sin n␲t.
n51
(14)
Observe in the solution (14) that there is no integer n $ 1 for which the
denominator 64 2 n2␲2 of Bn is zero. In general, if there is a value of n, say N, for
which N␲yp 5 ␻, where ␻ 5 Ïkym, then the system described by (11) is in a state
of pure resonance. In other words, we have pure resonance if the Fourier series
expansion of the driving force f (t) contains a term sin(N␲yL)t (or cos(N␲yL)t) that
has the same frequency as the free vibrations.
Of course, if the 2p-periodic extension of the driving force f onto the negative
t-axis yields an even function, then we expand f in a cosine series.
EXERCISES 11.3 Answers to selected odd-numbered problems begin on page ANS-18.
5xx 12 1,1,
In Problems 1–10 determine whether the function is even, odd, or 21 , x , 0
neither. 20. f (x) 5
0#x,1
1. f (x) 5 sin 3x 2. f (x) 5 x cos x
5
1, 22 , x , 21
3. f (x) 5 x2 1 x 4. f (x) 5 x3 2 4x 2x, 21 # x , 0
21. f (x) 5
5. f (x) 5 euxu 6. f (x) 5 e x 2 e2x x, 0#x,1
1, 1#x,2
52xx ,, 21 , x , 0
2
7. f (x) 5
5
2
0#x,1 2␲, 22␲ , x , 2␲
22. f (x) 5 x, 2␲ # x , ␲
8. f (x) 5 52xx 11 5,5, 22 , x , 0
0#x,2
␲, ␲ # x , 2␲
9. f (x) 5 x3, 0#x#2 23. f (x) 5 usin x u, 2␲ , x , ␲

10. f (x) 5 ux u 5
24. f (x) 5 cos x, 2␲y2 , x , ␲y2
In Problems 11–24 expand the given function in an appropriate

In Problems 25–34 find the half-range cosine and sine expansions of
cosine or sine series.
the given function.
11. f (x) 5 521,1, 2␲ , x , 0
0#x,␲ 25. f (x) 5 51, 0 , x , 12
0, 12 # x , 1
5
1, 22 , x , 21
12. f (x) 5 0, 21 , x , 1 26. f (x) 5 50,1, 0 , x , 12
1
2 #x,1
1, 1,x,2
27. f (x) 5 cos x, 0 , x , ␲y2
13. f (x) 5 u x u, 2␲ , x , ␲
28. f (x) 5 sin x, 0,x,␲
14. f (x) 5 x, 2␲ , x , ␲
15. f (x) 5 x2, 21 , x , 1 29. f (x) 5 5x,␲ 2 x, 0 , x , ␲y2
␲y2 # x , ␲
16. f (x) 5 x u x u, 21 , x , 1
17. f (x) 5 ␲2 2 x2, 2␲ , x , ␲
30. f (x) 5 50,x 2 ␲, 0,x,␲
␲ # x , 2␲
18. f (x) 5 x3, 2␲ , x , ␲
31. f (x) 5 51,x, 0,x,1
19. f (x) 5 5
x 2 1, 2␲ , x , 0
x 1 1, 0#x,␲
1#x,2
51,22 x, 0,x,1 48. (a) Solve the differential equation in Problem 45,
32. f (x) 5 1
1#x,2 4 x 0 1 12x 5 f (t), subject to the initial conditions
x(0) 5 1, x9(0) 5 0.
33. f (x) 5 x2 1 x, 0,x,1
(b) Use a CAS to plot the graph of the solution x (t) in part (a).
34. f (x) 5 x(2 2 x), 0,x,2
49. Suppose a uniform beam of length L is simply supported at
x 5 0 and at x 5 L. If the load per unit length is given by
In Problems 35–38 expand the given function in a Fourier series. w(x) 5 w0xyL, 0 , x , L, then the differential equation for the
35. f (x) 5 x2, 0 , x , 2␲ deflection y(x) is
36. f (x) 5 x, 0,x,␲ d 4y w0 x

EI 4
5 ,
37. f (x) 5 x 1 1, 0,x,1 dx L
38. f (x) 5 2 2 x, 0,x,2 where E, I, and w0 are constants. (See (4) in Section 5.2.)
(a) Expand w(x) in a half-range sine series.
In Problems 39–42 suppose the function y 5 f (x), 0 , x , L,
given in the figure is expanded in a cosine series, in a sine series, (b) Use the method of Example 4 to find a particular solution
and in a Fourier series. Sketch the periodic extension to which each yp(x) of the differential equation.
series converges. 50. Proceed as in Problem 49 to find a particular solution yp(x)
39. y 40. y y 5 f(x)
when the load per unit length is as given in Figure 11.3.17.
y 5 f(x)
w (x)
x x w0
L L
FIGURE 11.3.13 Graph for FIGURE 11.3.14 Graph for

Problem 39 Problem 40 L/3 2L/3 L x
41. y 42. FIGURE 11.3.17 Graph for Problem 50

y y 5 f(x)
y 5 f(x) 51. When a uniform beam is supported by an elastic foundation and
subject to a load per unit length w(x), the differential equation
x x for its deflection y(x) is
L L
FIGURE 11.3.15 Graph for FIGURE 11.3.16 Graph for d 4y

Problem 41 Problem 42 EI 1 ky 5 w(x),
dx 4
In Problems 43 and 44 proceed as in Example 4 to find a particular where k is the modulus of the foundation. Suppose that
solution xp(t) of equation (11) when m 5 1, k 5 10, and the driving the beam and elastic foundation are infinite in length
force f (t) is as given. Assume that when f (t) is extended to the nega- (that is, 2` , x , `) and that the load per unit length is
tive t-axis in a periodic manner, the resulting function is odd. the periodic function
5 25,5, 0,t,␲
5
43. f (t) 5 ; f (t 1 2␲) 5 f (t) 0, 2␲ , x , 2␲y2
␲ , t , 2␲
w(x) 5 w0, 2␲y2 # x # ␲y2; w(x 1 2␲) 5 w(x).
44. f (t) 5 1 2 t, 0 , t , 2; f (t 1 2) 5 f (t) 0 ␲y2 , x , ␲
In Problems 45 and 46 proceed as in Example 4 to find a particular Use the method of Example 4 to find a particular solution yp(x)
solution xp(t) of equation (11) when m 5 14, k 5 12, and the driving of the differential equation.
force f (t) is as given. Assume that when f (t) is extended to the nega-
tive t-axis in a periodic manner, the resulting function is even.
Discussion Problems
45. f (t) 5 2␲t 2 t2, 0 , t , 2␲; f (t 1 2␲) 5 f (t) 52. Prove properties (a), (c), (d), (f), and (g) in Theorem 11.3.1.
5t,1 2 t,
1
0,t, 53. There is only one function that is both even and odd. What
46. f (t) 5 ; f (t 1 1) 5 f (t)
2
is it?
2,t,1
1
54. As we know from Chapter 4, the general solution of the

47. (a) Solve the differential equation in Problem 43,
differential equation in Problem 51 is y 5 yc 1 yp. Discuss
x0 1 10x 5 f (t), subject to the initial conditions
why we can argue on physical grounds that the solution of
x(0) 5 0, x9(0) 5 0.
Problem 51 is simply yp. [Hint: Consider y 5 yc 1 yp as
(b) Use a CAS to plot the graph of the solution x (t) in part (a). x S 6`.]

13 Fourier Series (Zill)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

13 Fourier Series (Zill)

Uploaded by

Copyright:

Available Formats

.

 FOURIER SERIES 431

11.2 Fourier Series

will be of particular importance later on in the solution of certain kinds of boundary-

A TRIGONOMETRIC SERIES Suppose that f is a function defined on the interval

Solving for a0 yields

Now we multiply (2) by cos(m␲xyp) and integrate:

DEFINITION 11.2.1 Fourier Series

CONVERGENCE OF A FOURIER SERIES In the absence of any stated conditions that

PIECEWISECONTINUOUS FUNCTIONS Before stating conditions under which

f (x1) 5 lim f (x 1 h), f (x2) 5 lim f (x 2 h),

called, respectively, the right- and left-hand limits of f at x. A function f is said to be

THEOREM 11.2.1 Conditions for Convergence

EXAMPLE 1 Expansion in a Fourier Series

Expand f (x) 5 50,␲ 2 x, 2␲ , x , 0

y SOLUTION The graph of f is given in Figure 11.2.1. With p 5 ␲ we have from

EXAMPLE 2 Example 1 Revisited

EXAMPLE 3 Example 1 Revisited

f (01) 1 f (02) ␲ f (␲2) 1 f (2␲1)

24␲ 23␲ 22␲ 2␲

FIGURE 11.2.3 Periodic extension of function f shown in Figure 11.2.1

SEQUENCE OF PARTIAL SUMS It is interesting to see how the sequence of par-

FIGURE 11.2.4 Partial sums of Fourier series (13) in Example 1

EXERCISES 11.2 Answers to selected odd-numbered problems begin on page ANS-18.

to show that (8) can be written in the complex form

11.3 Fourier Cosine and Sine Series

FIGURE 11.3.1 Even function; graph

FIGURE 11.3.2 Odd function; graph

F(2x) 5 f (2x) g(2x) 5 (2f (x))(2g(x)) 5 f (x) g(x) 5 F(x).

Similarly, when f is odd on the interval (2p, p),

# f (x) sin n␲p x dx.

We summarize the results in the following definition.

DEFINITION 11.3.1 Fourier Cosine and Sine Series

# f (x) cos n␲p x dx.

# f (x) sin n␲p x dx.

Because the term sin(n␲xyp) is 0 at x 5 2p, x 5 0, and x 5 p, a sine series (4)

EXAMPLE 1 Expansion in a Sine Series

# x sin n␲2 x dx 5 4(21)

FIGURE 11.3.3 Odd function in Example 1 4 `(21)n11 n␲

The function in Example 1 satisfies the conditions of Theorem 11.2.1. Hence

FIGURE 11.3.4 Periodic extension of function shown in Figure 11.3.3

EXAMPLE 2 Expansion in a Sine Series

interval (2␲, ␲). With p 5 ␲ we have, from (5),

# (1) sin nx dx 5 ␲2 1 2 (21)

FIGURE 11.3.6 Partial sums of sine series (7)

FIGURE 11.3.7 Even reflection

EXAMPLE 3 Expansion in Three Series

y SOLUTION The graph of the function is given in Figure 11.3.10.

# x dx 5 32 L , # x cos n␲L x dx 5 4Ln(21)

FIGURE 11.3.10 Function f in Example 3

(b) In this case we must again integrate by parts twice:

(c) With p 5 Ly2, 1yp 5 2yL, and n␲yp 5 2n␲yL we have

24L 23L 22L 2L L 2L 3L 4L x

(a) cosine series

24L 23L 22L 2L L 2L 3L 4L x

(b) sine series

24L 23L 22L 2L L 2L 3L 4L x

(c) Fourier series

FIGURE 11.3.11 Same function on (0, L) but different periodic extensions

PERIODIC DRIVING FORCE Fourier series are sometimes useful in determining a

by first representing f by a half-range sine expansion and then assuming a partic-

EXAMPLE 4 Particular Solution of a DE

EXERCISES 11.3 Answers to selected odd-numbered problems begin on page ANS-18.

9. f (x) 5 x3, 0#x#2 23. f (x) 5 usin x u, 2␲ , x , ␲