ME 2002 (502 20002) Engineering Mathematics (II) Spring, 2024 Topic 1. Orthogonal Functions and Fourier Series Instructor: Kuo-Long Pan Contents 1. Orthogonal Functions 2. Fourier Series 3. Fourier Cosine and Sine Series 4. Complex Fourier Series 5. Sturm-Liouville Problems 6. Bessel and Legendre Series (optional) Reading D. G. Zill, Advanced Engineering Mathematics, 7th Ed., Chap. 12, Sec. 12.1 − 12.6. 1. Orthogonal Functions Definition 12.1.1 Inner Product of Functions The inner product of two functions f1 and f2 on an interval [a, b] is the number b ( f1 , f 2 ) f1 ( x) f 2 ( x)dx a Definition 12.1.2 Orthogonal Functions Two functions f1 and f2 are said to be orthogonal on an interval [a, b] if b ( f1 , f 2 ) f1 ( x) f 2 ( x) dx 0 a Example: 1 Definition 12.1.3 Orthogonal Set A set of real-valued functions {0(x), 1(x), 2(x), …} is said to be orthogonal on an interval [a, b] if b (m , n ) m ( x) n ( x) dx 0, m n a 1 Orthonormal Sets The expression (u, u) = ||u||2 is called the square norm. Thus we can define the square norm of a function as b n ( x) n2 dx and n ( x) 2 a b a n2 ( x)dx If {n(x)} is an orthogonal set on [a, b] with the property that ||n(x)|| = 1 for all n, then {n(x)} is called an orthonormal set on [a, b]. Examples: 2, 3 Vector Analogy Recalling from the vectors in 3-space that u c1v1 c2 v 2 c3 v 3 we have u (u, v1 ) (u, v 2 ) (u, v 3 ) v v2 v3 2 1 2 || v1 || || v 2 || || v 3 ||2 3 (u, v n ) v 2 n n 1 || v n || Thus we can make an analogy between vectors and functions. Orthogonal Series Expansion Suppose {n(x)} is an orthogonal set on [a, b]. If f(x) is defined on [a, b], we first write as f ( x) c00 ( x) c11 ( x) cnn ( x) ? Then b a b b a a b f ( x)m ( x) dx c0 0 ( x)m ( x) dx c1 1 ( x)m ( x) dx c0 (0 , m ) c1 (1 , m ) cn n ( x)m ( x) dx a cn (n , m ) Since {n(x)} is an orthogonal set on [a, b], each term on the right-hand side is 0 except m = n. In this case we have b a b f ( x)n ( x)dx cn (n , n ) cn n 2 ( x)dx cn a b a f ( x)n ( x)dx b a n 2 , n 0, 1, 2,... ( x)dx 2 In other words, f ( x) cnn ( x), cn b a n 0 f ( x)n ( x) dx || n ( x) ||2 Definition 12.1.4 Orthogonal Set/Weight Function A set of real-valued functions {0(x), 1(x), 2(x), …} is said to be orthogonal with respect to a weight function w(x) on [a, b], if b a w( x)m ( x)n ( x)dx 0, m n Under the condition of the above definition, we have cn b f ( x) w( x)n ( x) dx a || n ( x) ||2 b , || n ( x) ||2 w( x)n2 ( x) dx a Complete Sets An orthogonal set is complete if the only continuous function orthogonal to each member of the set is the zero function. 2. Trigonometric Series We can show that the set x 2 x 3 x , cos , cos 1, cos , p p p x 2 x 3 x , sin , sin , sin , p p p is orthogonal on [−p, p]. Thus a function f defined on [−p, p] can be written as f ( x) n x n x a0 an cos bn sin 2 n 1 p p Now we calculate the coefficients. p p Since cos( 𝑛𝜋𝑥 𝑝 f ( x)dx ) and sin( p p n x n x a0 p dx an cos dx bn sin dx p p 2 p n 1 p p 𝑛𝜋𝑥 𝑝 )are orthogonal to 1 on this interval, then (3) becomes 3 p p p a p a f ( x)dx 0 dx 0 x pa0 2 p 2 p Thus we have 1 p f ( x)dx p p a0 In addition, p p m x f ( x)cos dx p p p a0 p m x m x n x m x n x cos dx a cos cos dx b cos n n p p p p p sin p dx 2 p p n 1 By orthogonality we have p m x p cos p dx 0, m 0, And m x n x cos sin dx 0 p p p p 0, m n m x n x cos cos dx p p p p, m n p Thus reduces to p p So an n x f ( x)cos dx an p p n x 1 p f ( x)cos dx p p p Likewise m x sin p p dx 0, m 0, p m x n x sin p p cos p dx 0 p And 0, m n m x n x sin sin dx p p p p, m n p we find that bn 1 p n x f ( x)sin dx p p p Definition 12.2.1 Fourier Series The Fourier series of a function f defined on the interval (−p, p) is given by n x n x a0 f ( x) an cos b sin n p 2 n1 p a0 1 p 1 p n 1 p n f ( x ) dx , a f ( x ) cos x dx , b f ( x ) sin x dx n n p p p p p p p p 4 Example: 1 Theorem 12.2.1 Conditions for Convergence Let f and f’ be piecewise continuous on the interval (−p, p); that is, let f and f’ be continuous except at a finite number of points in the interval and have only finite discontinuities at these points. Then the Fourier series of f on the interval converge to f(x) at a point of continuity. At a point of discontinuity, the Fourier series converges to the average f ( x ) f ( x ) 2 where f(x+) and f(x-) denote the limit of f at x from the right and from the left, respectively. Example: 2 Periodic Extension Fig 12.2.2 is the periodic extension of the function f in Example 1. Thus the discontinuity at x = 0, 2, 4, …will converge to f (0 ) f (0) 2 2 and at x = , 3, … will converge to f ( ) f ( ) 0 2 Sequence of Partial Sums Referring to (13), we write the partial sums as (see Fig 12.2.3) S1 S3 4 4 , S2 4 2 cos x sin x, 2 1 cos x sin x sin 2 x 2 12.3 Fourier Cosine and Sine Series Even and Odd Functions even : f(−x) = f(x) odd : f(−x) = −f(x) 5 Theorem 12.3.1 Properties of Even/Odd Functions (a) The product of two even functions is even. (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. (e) The sum (difference) of two odd functions is odd. (f) If f is even then a a a (g) If f is odd then a a f ( x)dx 2 f ( x)dx 0 f ( x)dx 0 Cosine and Sine Series If f is even on (−p, p) then 1 p 2 p f ( x ) dx f ( x)dx p p p 0 1 p n 2 p n an f ( x)cos x dx f ( x)cos xdx p p p p 0 p a0 even bn 1 n f ( x)sin x dx 0 p p p p odd Similarly, if f is odd on (−p, p) then an 0, n 0, 1, 2,... bn 2 p n f ( x)sin xdx p 0 p Definition 12.3.1 Fourier Cosine and Sine Series i) The Fourier series of an even function f on the interval (−p, p) is the cosine series 6 a0 n an cos x 2 n1 p 2 p 2 p n a0 f ( x) dx, an f ( x)cos x dx 0 0 p p p The Fourier series of an odd function f on the interval (−p, p) is the sine series f ( x) ii) f ( x) bn sin i 1 bn n x p 2 p n f ( x)sin x dx 0 p p Examples: 1, 2 Gibbs Phenomenon Fig 12.3.6 shows the partial sums of (7). We can see there are pronounced spikes near the discontinuities. This overshooting of SN does not smooth out but remains fairly constant even when N is large. This is so-called Gibbs phenomenon Half-Range Expansions If a function f is defined only on 0 < x < L, we can make arbitrary definition of the function on −L < x < 0. 7 If y = f(x) is defined on 0 < x < L, (i) reflect the graph about the y-axis onto −L < x < 0; the function is now even (Fig 12.3.7). (ii) reflect the graph through the origin onto −L < x < 0; the function is now odd (Fig 12.3.8). (iii) define f on −L < x < 0 by f(x) = f(x + L). See Fig 12.3.9. Example: 3 Periodic Driving Force Consider the following physical system d 2x n m 2 k x f (t ), x p (t ) Bn sin t dt p n 1 is a half-range sine expansion. Example: 4 12.4 Complex Fourier Series Euler’s formula eix cos( x) i sin( x) e ix cos( x) i sin( x) Complex Fourier Series From Euler’s formula, we have cos x Using this relation to replacecos( 𝑛𝜋𝑥 𝑝 eix e ix eix e ix , sin x 2 2i ) and sin( 𝑛𝜋𝑥 𝑝 8 ), then a0 ein x /p e in x /p ein x /p e in x /p an bn 2 n1 2 2i a0 1 1 (an ibn )ein x /p (an ibn )e in x /p 2 n1 2 2 n 1 n 1 c0 cnein x /p c ne in x /p When the function f is real, cn and c-n are complex conjugates. We have 1 1 p c0 f ( x) dx 2 p p 1 (an ibn ) 2 n x n x 1 1 p 1 p f ( x) cos dx i f ( x) sin dx 2 p p p p p p cn n x n x 1 p f ( x ) cos i sin dx 2 p p p p 1 p f ( x)e in x / p dx p 2p 1 (an ibn ) 2 n x n x 1 1 p 1 p f ( x) cos dx i f ( x) sin dx 2 p p p p p p c n n x n x 1 p f ( x) cos i sin dx 2 p p p p 1 p f ( x)ein x / p dx p 2p Definition 12.4.1 Complex Fourier Series The Complex Fourier Series of function f defined on an interval (p, p) is given by 1 p f ( x) cnein x / p , cn f ( x)e in x / p dx, n 0, 1, 2, p 2p n If f satisfies the hypotheses of Theorem 12.2.1, a complex Fourier series converges to f(x) at a point of continuity and to the average 9 at a point of discontinuity. f ( x ) f ( x ) 2 Example: 1 Fundamental Frequency The fundamental period is T = 2p and then p = T/2. The Fourier series becomes a0 (an cos n x bn sin n x) and cnein x 2 n1 n where = 2/T is called the fundamental angular frequency. Frequency Spectrum If f is periodic and has fundamental period T, the plot of the points (n, |cn|) is called the frequency spectrum of f. Examples: 2, 3 (correction!) 12.5 Sturm-Liouville Problem Eigenvalue and Eigenfunctions Recall the ODE with boundary condition y y 0, y (0) 0, y ( L) 0 This equation possesses nontrivial solutions only when takes on the values n 2 2 n 2 , n 1, 2,3... called eigenvalues. The corresponding nontrivial solutions L n x n x y c2 sin are called the eigenfunctions. or simply y sin L L Example: 1 Regular Sturm-Liouville Problem Let p, q, r and r be real-valued functions continuous on [a, b], and let r(x) > 0 and p(x) > 0 for every x in the interval. Then solve 10 d [r ( x) y] (q ( x) p ( x)) y 0 dx (3) Subject to A1 y (a) B1 y(a) 0 A2 y (b) B2 y(b) 0 is said to be a regular Sturm-Liouville problem. The coefficients are assumed to be real and independent of . Theorem 12.5.1 Properties of the Regular Sturm–Liouville Problem (a) There exist an infinite number of real eigenvalues that can be arranged in increasing order 1 < 2 < 3 < … < n < … such that n → as n → . (b) For each eigenvalue there is only one eigenfunction (except for nonzero constant multiples). (c) Eigenfunctions corresponding to different eigenvalues are linearly independent. (d) The set of eigenfunctions corresponding to the set of eigenvalues is orthogonal with respect to the weight function p(x) on the interval [a, b]. Proof of (d): Let ym and yn be eigenfunctions corresponding to eigenvalues m and n. Then d [r ( x) ym ] ( q( x) m p( x)) ym 0 dx d [r ( x) yn ] (q ( x) n p ( x)) yn 0 dx we have (m n ) p ( x) ym yn ym d d r ( x) yn yn r ( x) ym dx dx Integrating the above equation from a to b, then b (m n ) p ( x) ym yn dx a r (b)[ ym (b) yn (b) yn (b) ym (b)] r (a )[ ym (a ) yn (a ) yn (a ) ym (a )] Since all solutions must satisfy the boundary condition, we have 11 (8) A1 ym (a ) B1 ym (a ) 0 A1 yn (a ) B1 yn (a ) 0 For this system to be satisfied by A1 and B1 not both zero, the determinant must be zero ym (a ) yn (a ) yn (a ) ym (a ) 0 ym (b) yn (b) yn (b) ym (b) 0 Similarly, we have Hence we have the orthogonality relation b a p( x) ym ( x) yn ( x) dx 0, m n (9) Example: 2 Singular Sturm-Liouville problem There are several important conditions for the boundary condition of Sturm-Liouville problem i) r(a) = 0 and a boundary condition of the type is specified at x = b. (Singular BVP) ii) r(b) = 0 and a boundary condition of the type is specified at x = a. (Singular BVP) iii) r(a) = r(b) = 0 and no boundary condition is specified at either x = a or at x = b; iv) r(a) = r(b) and boundary conditions y(a) = y(b), y′(a) = y′(b). (Periodic BVP) By assuming the solutions of (3) are bounded on [a, b], from (8) we have If r(a) = 0, then the orthogonality relation (9) holds with no boundary condition at x = a; If r(b) = 0, then the orthogonality relation (9) holds with no boundary condition at x = b; If r(a) = r(b) = 0, then the orthogonality relation (9) holds with no boundary conditions specified at either x = a or x = b; If r(a) = r(b), then the orthogonality relation (9) holds with periodic boundary conditions y(a) = y(b), y′(a) = y′(b). Self-Adjoint Form r ( x) y r ( x) y (q( x) p( x)) y 0 Thus we can write the Legendre’s differential equation (1 x 2 ) y " 2 xy ' n(n 1) y 0 as d [(1 x 2 ) y] n(n 1) y 0 dx 12 Here we find that the coefficient of y is the derivative of the coefficient of y. In addition, if the coefficients are continuous and a(x) 0 for all x in some interval, then any second-order differential equation a( x) y b( x) y (c( x) d ( x)) y 0 can be recast into the so-called self-adjoint form (3). To understand the above fact, we start from a1(x)y + a0(x)y = 0 Let P = a0(x)/a1(x), then y + Py = 0, y + Py = 0, μ′ = Pμ, = e P(x)dx d(y)/dx = 0. Now, let Y = y, and the integrating factor be e [b(x)/a(x)] dx, then e e Y b ( x )/ a ( x ) dx ( b / a ) dx y b ( x )/ a ( x ) dx b( x) b ( x )/ a ( x ) dx d e Y ... e Y ... a( x) dx b( x) ( b / a ) dx d ( x) ( b / a ) dx c( x) ( b / a ) dx e y e e y 0 a( x) a( x) a( x) d ( b / a ) dx c( x) (b / a ) dx d ( x) (b / a ) dx e y e e y 0 a ( x) dx a ( x) r ( x) e ( b / a ) dx , q( x) c( x) (b / a ) dx d ( x) (b / a ) dx e , p( x) e a( x) a( x) Examples: 3 (correction!), 4 12.6 Bessel and Legendre Series (for your own reading and optional exploration) Fourier-Bessel Series We have shown that {Jn(ix)}, i = 1, 2, 3, … is orthogonal w.s.t. p(x) = x on [0, b] when the i are defined by A2 J n ( b) B2 J n ( b) 0 This orthogonal series expansion or generalized Fourier series of a function f defined on (0, b) in b terms of this orthogonal set is xJ n ( i x) f ( x) dx f ( x) ci J n ( i x), ci 0 2 J n ( i x) i 1 The square norm of the function Jn(ix) is defined by b J n ( i x) xJ n2 ( i x)dx 2 0 f(x) is called a Fourier-Bessel series. 13 Differential Recurrence Relations d n d n [ x J n ( x)] x n J n1 ( x), [ x J n ( x)] x n J n1 ( x) dx dx Square Norm b J n ( i x) xJ n2 ( i x)dx is dependent on i i 2 2 0 If y = Jn(x) we have 2 n2 d xy x y 0 dx x After we multiply by 2xy’, then d d xy2 2 x 2 n 2 [ y ]2 0 dx dx Integrating by parts on [0, b], then 2 2 xy 2 dx xy ( 2 x 2 n 2 ) y 2 b 0 2 b 0 Since y = Jn(x), the lower limit is 0 for n > 0, because Jn(0) = 0. For n = 0, at x = 0. Thus b 2 2 xJ n2 ( x) dx 2b 2 [ J n ( b)]2 ( 2b 2 n 2 )[ J n b)]2 0 where y ' J n ( x). Now we consider three cases for the condition A2 J n ( b) B2 J n ( b) 0 Case I: If we choose A2 = 1 and B2 = 0, then J n ( b) 0 There are an infinite number of positive roots xi = ib of (8) (see Fig 5.3.1), which defines i = xi/b. The eigenvalues are positive and then i = i2 = xi2/b2. No new eigenvalues result from the negative roots of (8) since Jn(−x) = (−1)nJn(x). The number 0 is not an eigenvalue for any n since Jn(0) = 0, n= 1, 2, 3, … and J0(0) = 1. b2 2 || J n i ( x) || J n1 ( ib) 2 2 Case II: If we choose A2 = h 0 and B2 = b, then hJ n ( b) bJ n ( b) 0 14 There are an infinite number of positive roots xi = ib for n = 1, 2, 3, …. As before i = i2 = xi2/b2. = 0 is not an eigenvalue for n = 1, 2, 3, …. || J n ( i x) ||2 i2b 2 n 2 h 2 2 J n ( ib) 2 i2 Case III: If h = 0 and n = 0 in hJ n ( b) bJ n ( b) 0 i are defined from the roots of J 0 ( b) 0 Though this is a special case, it is the only situation for which = 0 is an eigenvalue. For n = 0, the result implies J0(b) = 0 is equivalent to J1(b) = 0. Since x1 = ib = 0 is a root of the last equation and because J0(0) = 1 is nontrivial, we conclude that from 1 = 12 = x12/b2 that 1 = 0 is an eigenvalue. But we cannot use i2b 2 n 2 h 2 2 || J n ( i x) || J n ( ib) 2 i2 2 when 1 = 0, h = 0, n = 0, and n = 0. However from b J n ( i x) xJ n2 ( i x)dx 2 0 we have b ||1||2 x dx 0 b2 2 For i > 0 we can use || J n ( i x) ||2 with h = 0 and n = 0: i2b 2 n 2 h 2 2 J n ( ib) 2 i2 b2 2 || J 0 ( i x) || J 0 ( ib) 2 2 Definition 12.6.1 Fourier–Bessel Series The Fourier-Bessel series of a function f defined on the interval (0, b) is given by i) f ( x) ci J n ( i x), ci i 1 b 2 xJ n ( i x) f ( x) dx b 2 J n21 ( ib) 0 where the i are defined by Jn(b) = 0. ii) f ( x) ci J n ( i x), ci i 1 b 2 i2 xJ n ( i x) f ( x) dx 2 2 2 2 2 ( i b n h ) J n ( ib) 0 15 where the i are defined by hJn(b) + bJ’n(b) = 0. iii) b 2 b 2 x f ( x ) dx, ci 2 2 xJ 0 ( i x ) f ( x ) dx 2 0 b b J 0 ( ib) 0 i 2 where the i are defined by J′0(b) = 0. f ( x) c1 ci J 0 ( i x), c1 Theorem 12.6.1 Conditions for Convergence If f and f ′ are piecewise continuous in the open interval (0, b), then a Fourier-Bessel expansion of f converges to f(x) at any point where f is continuous and to the average f ( x ) f ( x ) 2 at a point where f is discontinuous. Examples: 1, 2 Definition 12.6.2 Fourier–Legendre Series The Fourier-Legendre series of a function f defined on the interval (–1, 1) is given by f ( x) cn Pn ( x), cn n 0 2n 1 1 f ( x) Pn ( x) dx 2 1 Theorem 12.6.2 Conditions for Convergence If f and f ′ are piecewise continuous in the open interval (–1, 1), then a Fourier-Legendre series converges to f(x) at a point of continuity ant to the average f ( x ) f ( x ) 2 at a point of discontinuity. Example: 3 Alternative Form of Series If we let x = cos , the x = 1 implies = 0, x = −1 implies = . Since dx = −sin d, FourierLegendre series become, respectively, 16 F ( ) cn Pn (cos ), cn n 0 2n 1 F ( ) Pn (cos )sin d 2 0 where f(cos ) has been replaced by F(). Exercises ․ ․ ․ ․ ․ ․ 12.1: 16, 19, 22, 25, 27, 28, 30 12.2: 21, 22, 24, 25 12.3: 37, 39, 40, 42, 43, 46, 47, 49, 50 (try 53, 54, 55) 12.4: for self-practice 12.5: 3, 4, 9, 10, 11 12.6: (optional) for your own study as much as you wish. 17