THE UNIVERSITY OF MELBOURNE SCHOOL OF MATHEMATICS AND STATISTICS MAST20029 Engineering Mathematics Lecture Notes STUDENT NAME: This compilation has been made in accordance with the provisions of Part VB of the copyright act for the teaching purposes of the University. This booklet is for the use of students of the University of Melbourne enrolled in the subject MAST20029 Engineering Mathematics. © School of Mathematics and Statistics, University of Melbourne, February 2022. These notes have been written by Christine Mangelsdorf and Marcus Brazil at the University of Melbourne. Reproduction of any part of this work other than that authorised by Australian Copyright Law without permission of the copyright owners is unlawful. Edition 17, February 2022. MAST20029 Engineering Maths: Vector Calculus 1.1 VECTOR CALCULUS MAST20029 Engineering Maths: Vector Calculus 1.2 Vector Fields [Kreyszig, p.375-378] Physical examples of vector fields are force fields (eg, due to gravity or electrical charge) and velocity fields (eg, due to air or fluid flow). A vector field is a function Example F : Rn ! Rn n=2 F(x, y) = u(x, y)i + v(x, y)j • 2 independent variables x, y • F is a 2 component vector (u, v) in 2D where each component is a scalar function of x, y. A tank of water is stirred so that the water rotates around a central point. The velocity of the water at any point in the tank is given by the vector field v : R3 ! R3 where v(x, y, z) = yi xj v has no k component because the rotation is planar. Sketching Vector Fields n=3 F(x, y, z) = u(x, y, z)i + v(x, y, z)j + w(x, y, z)k • 3 independent variables x, y, z • F is a 3 component vector (u, v, w) in 3D where each component is a scalar function of x, y, z. The vector field can be sketched by assigning to each point (x, y, z) a vector F represented by an arrow whose tail is at (x, y, z). MAST20029 Engineering Maths: Vector Calculus (1) Using vertical strips: 1.15 MAST20029 Engineering Maths: Vector Calculus (2) Using horizontal strips: 1.16 MAST20029 Engineering Maths: Vector Calculus 1.21 MAST20029 Engineering Maths: Vector Calculus 1.22 Change of Variables in R2 [Kreyszig, p.429 - 432] Suppose we have a two-dimensional change of variable x = x(u, v) y = y(u, v) The Jacobi matrix for this change of variables is 3 2 @x @x 6 @u @v 7 7 6 7 6 4 @y @y 5 @u @v The determinant of the Jacobi matrix is called the Jacobian determinant and is denoted by J(u, v). MAST20029 Engineering Maths: Vector Calculus 1.27 MAST20029 Engineering Maths: Vector Calculus 1.28 MAST20029 Engineering Maths: Vector Calculus 1.35 MAST20029 Engineering Maths: Vector Calculus 1.36 Change of Variables in R3 Suppose we have a three-dimensional change of variable x = x(u, v, w) y = y(u, v, w) z = z(u, v, w) The Jacobi matrix for this change of variables is 2 3 @x @x @x 6 @u @v @w 7 6 7 6 7 6 7 6 @y @y @y 7 6 7 6 @u @v @w 7 6 7 6 7 4 @z @z @z 5 @u @v @w The determinant of the Jacobi matrix is called the Jacobian determinant and is denoted by J(u, v, w). MAST20029 Engineering Maths: Vector Calculus 1.39 Cylindrical coordinates are an extension of polar coordinates to 3 dimensions. MAST20029 Engineering Maths: Vector Calculus The Jacobian determinant is • P 0 is the projection of P (r, ✓, z) onto the xy-plane. J(r, ✓, z) • r and ✓ are measured for P 0 as in polar coordinates. = • z is the z-coordinate of P . @x @r @x @✓ @x @z @y @r @y @✓ @y @z @z @z @z @r @✓ @z cos ✓ r sin ✓ sin ✓ r cos ✓ 0 0 • x2 + y 2 = r 2 . = cos ✓ sin ✓ = r sin ✓ r cos ✓ = r cos2 ✓ + r sin2 ✓ = r The integral becomes ZZZ ZZZ f (x, y, z)dxdydz = V 0 0 1 V⇤ where F (r, ✓, z) = f (r cos ✓, r sin ✓, z). F (r, ✓, z)r drd✓dz 1.40 MAST20029 Engineering Maths: Vector Calculus 1.41 Exercise 12 Find the volume of the solid region V which lies inside the cylinder x2 + y 2 = 1, z 2 R, below the plane z = 4 and above the paraboloid z = 1 x2 y 2 . Solution Using cylindrical coordinates, V is described as MAST20029 Engineering Maths: Vector Calculus 1.42 MAST20029 Engineering Maths: Vector Calculus 1.45 The Jacobian determinant J(r, , ✓) is = = @x @r @x @ @x @✓ @y @r @y @ @y @✓ @z @r @z @ @z @✓ cos ✓ sin sin ✓ sin cos MAST20029 Engineering Maths: Vector Calculus The integral becomes ZZZ ZZZ f (x, y, z)dxdydz = V V F (r, , ✓)r2 sin drd d✓ ⇤ where F (r, , ✓) = f (r cos ✓ sin , r sin ✓ sin , r cos ). r cos ✓ cos r sin ✓ cos r sin r sin ✓ sin r cos ✓ sin 0 = cos r cos ✓ cos r sin ✓ cos r sin ✓ sin r cos ✓ sin + r sin cos ✓ sin sin ✓ sin r sin ✓ sin r cos ✓ sin = cos (r2 cos2 ✓ sin cos + r2 sin2 ✓ sin cos ) + r sin (r cos2 ✓ sin2 = r2 sin cos2 = r2 sin + r sin2 ✓ sin2 ) + r2 sin3 1.46 MAST20029 Engineering Maths: Vector Calculus 1.57 MAST20029 Engineering Maths: Vector Calculus 1.58 MAST20029 Engineering Maths: Vector Calculus Solution 1.63 MAST20029 Engineering Maths: Vector Calculus 1.64 Conservative Fields [Kreyszig, p.421-424] A work integral is path independent if Z Z F · dr = F · dr C1 C2 for any two simple oriented curves C1 , C2 with the same endpoints. Let F(x, y, z) = F1 i + F2 j + F3 k be a vector field such that F1 , F2 and F3 have continuous first-order partial derivatives. R We say that F is a conservative vector field if C F · dr is path independent. This is equivalent to Z F · dr = 0 for any simple closed curve C. 1. C 2. r ⇥ F = 0 3. F = r for some scalar function MAST20029 Engineering Maths: Vector Calculus 1.65 Theorem Let C be a path starting at the point A and finishing at the R point B. If F is conservative and F = r , then C F · dr depends only on the endpoints of C and Z F · dr = (B) (A) C MAST20029 Engineering Maths: Vector Calculus Exercise 21 R Let F(x, y) = xi + yj. Evaluate C F · dr along (a) y = x2 from (0, 0) to (1, 1); (b) line segments joining (0, 0) to (0, 1) to (1, 1); (c) if C is the unit circle. Solution 1.66 MAST20029 Engineering Maths: Vector Calculus 1.67 MAST20029 Engineering Maths: Vector Calculus 1.68 Exercise 22 Consider the vector field F(x, y, z) = (x2 , cos y sin z, sin y cos z). R Evaluate C F · dr where r(t) = (t2 + 1, et , e2t ), 0 t 1 Solution MAST20029 Engineering Maths: Vector Calculus 1.69 MAST20029 Engineering Maths: Vector Calculus 1.70 Surface Integrals [Kreyszig, p.448-450] Let S be a smooth surface defined by z = f (x, y). We define the surface integral of g over S to be ZZ S g(x, y, z)dS = ZZ R q g(x, y, f (x, y)) fx2 + fy2 + 1dydx where R is the region formed by projecting S onto the xyplane. By projecting vertically, we turn the surface integral of g over S into a double integral over the region R in the xy-plane. Note In this subject we only consider surface integrals for surfaces that can be written in the form z = f (x, y). MAST20029 Engineering Maths: Vector Calculus 1.75 MAST20029 Engineering Maths: Vector Calculus 1.76 Flux Across a Surface [Kreyszig, p.443-447] Let S be a smooth oriented surface defined by z = f (x, y). Assume the orientation of S is such that the unit normal vector n̂ on S is upward. Let F(x, y, z) = F1 i + F2 j + F3 k be a vector field. We define the flux integral of F over S to be ZZ S F · n̂ dS = ZZ F1 fx F2 fy + F3 dydx R where R is the region formed by projecting S onto the xyplane. Note In this subject we only consider flux integrals for surfaces that can be written in the form z = f (x, y). MAST20029 Engineering Maths: Vector Calculus 1.83 MAST20029 Engineering Maths: Vector Calculus 1.84 MAST20029 Engineering Maths: Vector Calculus 1.87 MAST20029 Engineering Maths: Vector Calculus Exercise 27 Let F(x, y, z) = z 2 k. Let S be the unit sphere x2 + y 2 + z 2 = 1 Assume S is oriented using the outward unit normal. Use Gauss’ theorem to find the flux of F across S. Solution 1.88 MAST20029 Engineering Maths: Vector Calculus 1.89 MAST20029 Engineering Maths: Vector Calculus 1.90 Stokes’ Theorem [Kreyszig, p.463-468] Let S be an oriented smooth open surface bounded by the curve C. Let C be oriented by the right-hand rule, i.e. anticlockwise if the unit normal n̂ on S is upwards, and clockwise if n̂ is downwards. Let F(x, y, z) = F1 i + F2 j + F3 k be a vector field such that F1 , F2 and F3 have continuous first-order partial derivatives on S. Then Z C F · dr = ZZ S (r ⇥ F) · n̂ dS Note If F is the velocity field of a fluid, Z C F · dr is the circulation of F. It measures the extent to which the corresponding fluid motion is a rotation around C. MAST20029 Engineering Maths: Systems of ODEs 2.1 SYSTEMS OF ORDINARY DIFFERENTIAL EQUATIONS MAST20029 Engineering Maths: Systems of ODEs Solutions of homogeneous constant coefficient systems To solve the system, try [Kreyszig, Ch 4, p124-141] x = we Coupled first order ODEs arise in models of physical systems such as chemical reactions that involve tanks of fluids with di↵erent concentrations and electric circuits with more than one loop. = dx2 dt = a11 x1 (t) + a12 x2 (t) a21 x1 (t) + a22 x2 (t) This can be written in matrix form as ẋ = Ax or " dx1 dt dx2 dt # = " a11 a21 where A is a constant matrix. a12 a22 #" x1 x2 # t where w is a constant vector. Then ẋ = we t Substituting into ẋ = Ax gives A homogeneous constant coefficient system of two first order ODEs has the form dx1 dt 2.2 we Since e t 6= 0 then t = Awe t Aw = w Hence is an eigenvalue of the matrix A and w is the corresponding eigenvector. If A has 2 linearly independent eigenvectors {w1 , w2 } with corresponding eigenvalues { 1 , 2 }, then the general solution to the system is x(t) = ↵1 w1 e 1t where ↵1 and ↵2 are constants. + ↵ 2 w2 e 2t MAST20029 Engineering Maths: Systems of ODEs Exercise 1 Solve the system dx dt dy dt = x + 3y = 2x + 2y subject to the initial conditions x(0) = 0, Solution y(0) = 1 2.3 MAST20029 Engineering Maths: Systems of ODEs 2.4 MAST20029 Engineering Maths: Systems of ODEs 2.5 MAST20029 Engineering Maths: Systems of ODEs 2.6 Real, repeated eigenvalues Suppose A has one repeated eigenvalue (algebraic multiplicity m = 2) and only one linearly independent eigenvector w (geometric multiplicity g = 1). One solution is: x = we t A second linearly independent solution is x = (wt + u)e t Substituting into ẋ = Ax gives: ( tw + u + w)e Since e t t = A(wt + u)e 6= 0 tw + u + w = tAw + Au Since Aw = w, u satisfies (A Au = I)u = u+w w The general solution of the system is x = ↵1 we t + ↵2 (wt + u)e where ↵1 and ↵2 are constants. t t MAST20029 Engineering Maths: Systems of ODEs Exercise 2 Solve the system " Solution dx dt dy dt # = " 3 1 1 1 #" x y # 2.7 MAST20029 Engineering Maths: Systems of ODEs 2.8 MAST20029 Engineering Maths: Systems of ODEs 2.9 MAST20029 Engineering Maths: Systems of ODEs Complex eigenvalues Exercise 3 If A has two complex conjugate eigenvectors {w1 , w2 } with corresponding complex conjugate eigenvalues { 1 , 2 }, then the complex general solution to the system is Solve the system " # x(t) = C1 w1 e 1t + C 2 w2 e 2t Solution where C1 and C2 are complex constants. The real-valued solutions to the system may be obtained from ⇥ x(t) = ↵1 Re w1 e 1t ⇤ ⇥ + ↵2 Im w1 e where ↵1 and ↵2 are real constants. dx dt dy dt 1t ⇤ = " 0 2 1 2 #" x y # 2.10 MAST20029 Engineering Maths: Systems of ODEs 2.11 MAST20029 Engineering Maths: Systems of ODEs 2.12 Phase portraits of linear systems [Kreyszig, Ch 4, p 141-151] Consider the linear system " # " dx a11 dt = dy a21 dt a12 a22 #" x y # • The equations define a vector field or a phase plane. At any point x, we have a velocity vector ẋ = Ax. • Each solution x(t) = (x(t), y(t)) defines a curve in the xy-plane called an orbit. The functions x(t) and y(t) are the parametric equations of the curve. • At any point on a solution curve, the velocity vector ẋ is tangential to the curve. • Arrows on the solution curve indicate the direction of t increasing. • The phase portrait is determined by the eigenvalues 1 , 2 and eigenvectors w1 , w2 of the matrix A. MAST20029 Engineering Maths: Systems of ODEs 2.13 Critical points MAST20029 Engineering Maths: Systems of ODEs Case 1: • Any point (x0 , y0 ) where dx =0 dt 1 < 2 2.14 <0 Exercise 4 and dy =0 dt Classify the critical point at the origin and sketch the phase portrait for the system: dx dt dy dt is called a critical point. • The origin (0, 0) is a critical point of every two dimensional homogeneous linear system. • P is an asymptotically stable critical point, if all of the orbits approach P . • P is a stable critical point, if all of the orbits stay near P but do not approach P . • P is an unstable critical point if some or all of the orbits move away from P . = x = 2y Solution The general solution is " # " # x 0 = ↵1 e y 1 2t + ↵2 " 1 0 # e t MAST20029 Engineering Maths: Systems of ODEs Case 2: 1 <0< 2.17 2 Exercise 5 Classify the critical point at the origin and sketch the phase portrait for the system: dx dt dy dt = 2x = x 4y 3y Solution The general solution is " # " # x 1 = ↵1 e y 1 2t + ↵2 " 4 1 # et MAST20029 Engineering Maths: Systems of ODEs 2.18 MAST20029 Engineering Maths: Systems of ODEs 2.21 MAST20029 Engineering Maths: Systems of ODEs Case 4: 1 = 2 2.22 (m = 2, g = 1) Exercise 7 Classify the critical point at the origin and sketch the phase portrait for the system: dx dt dy dt = 4x = x y 2y Solution The general solution is " x y # = ↵1 " 1 1 # e 3t + ↵2 " t t 1 # e 3t MAST20029 Engineering Maths: Systems of ODEs Case 6: 2.27 =± i Exercise 9 Classify the critical point at the origin and sketch the phase portrait for the system: dx dt dy dt = y = 4x cos(2t) 2 sin(2t) # Solution The general solution is " # " x = ↵1 y + ↵2 " sin(2t) 2 cos(2t) # MAST20029 Engineering Maths: Systems of ODEs 2.28 MAST20029 Engineering Maths: Systems of ODEs 2.31 Autonomous nonlinear systems MAST20029 Engineering Maths: Systems of ODEs where [Kreyszig, Ch 4, p 152-156] a= @f (x0 , y0 ), @x b= @f (x0 , y0 ) @y c= @g (x0 , y0 ), @x d= @g (x0 , y0 ) @y Consider the nonlinear system where RHS has no t dependence dx dt dy dt = f (x, y) = g(x, y) In matrix form, the linearised system becomes where f (x, y) and g(x, y) have continuous first order partial derivatives and are not simply linear combinations of x and y. Let (x0 , y0 ) be a critical point of the non linear system, and set X=x x0 , Y =y y0 The linearisation of the system about the critical point (x0 , y0 ) is given by the linear system dX dt dY dt = aX + bY = cX + dY Ẋ = AX where 2 @f 6 @x A = 4 @g @x 3 @f @y 7 @g 5 @y is the Jacobi matrix of f and g, evaluated at (x0 , y0 ). 2.32 MAST20029 Engineering Maths: Systems of ODEs 2.33 Exercise 10 Consider the following nonlinear system dx dt dy dt =1 = x 2xy y = f (x, y) = g(x, y) Find all critical points of this system, and linearise the system about each of its critical points. Solution MAST20029 Engineering Maths: Systems of ODEs 2.34 MAST20029 Engineering Maths: Systems of ODEs 2.35 MAST20029 Engineering Maths: Systems of ODEs 2.36 Phase portraits of nonlinear systems Linearisation Theorem Let (x0 , y0 ) be an isolated critical point of a two dimensional autonomous nonlinear system. Let Ẋ = AX be the linearisation of the system about (x0 , y0 ). If the eigenvalues of A are such that 1. 1 6= 2, and 2. neither eigenvalue has Re( ) = 0, then the orbits of the nonlinear system close to (x0 , y0 ) can be approximated by the orbits of the linearised system close to the origin. MAST20029 Engineering Maths: Laplace Transforms 3.1 LAPLACE TRANSFORMS MAST20029 Engineering Maths: Laplace Transforms Laplace transforms f (t) [Kreyszig, Chapter 6, pp 203-253] Let f be a function of a real variable t 0. The Laplace transform of f , denoted by L[f ], is the function F (s) L[f (t)] = F (s) = Z 1 f (t)e st dt F (s) = Z 1 f (t)e st dt 0 1 s 1. 1 2. tn 3. eat 4. sin(at) 5. cos(at) 6. sinh(at) 7. cosh(at) 8. f 0 (t) sL[f ] 9. f 00 (t) s2 L[f ] n! sn+1 0 where s is a complex variable. A Laplace transform is well defined if 1 s a s 2 + a2 1. f is piecewise continuous, 2. In the limit as t ! 1, |f | Kect , for some finite constants K, c. For many functions, the integral only converges for a limited range of s. Applications of Laplace transforms include problems involving heat conduction, fluid flow, chemical reactions and electrical circuits. a s2 s + a2 a s2 a2 s s2 a2 f (0) s f (0) f 0 (0) 3.2 MAST20029 Engineering Maths: Laplace Transforms 3.3 MAST20029 Engineering Maths: Laplace Transforms Exercise 1 Exercise 2 Using the definition of the Laplace transform, prove that Using the definition of the Laplace transform, prove that L[1] = Solution 1 s L[t] = Solution 1 s2 3.4 MAST20029 Engineering Maths: Laplace Transforms 3.5 MAST20029 Engineering Maths: Laplace Transforms Linearity Exercise 4 The Laplace transform is linear, so that Using the definition of the Laplace transform, prove that L[eat ] = f (t) = c1 f1 (t) + c2 f2 (t) then Solution L[f ] = c1 L[f1 ] + c2 L[f2 ] where c1 , c2 2 C. Exercise 3 Compute the Laplace transform of f (t) = t2 Solution 2t + 1 1 s a , a2C 3.6 MAST20029 Engineering Maths: Laplace Transforms 3.7 MAST20029 Engineering Maths: Laplace Transforms Exercise 5 Exercise 6 Compute the Laplace transform of Compute the Laplace transform of f (t) = cosh(at) for a 2 C, using the exponential definition of cosh(at). Solution f (t) = sin(at) for a 2 C, using the exponential definition of sin(at). Solution 3.8 MAST20029 Engineering Maths: Laplace Transforms 3.9 Laplace transforms of derivatives L[f ] = Z 1 0 = lim b!1 0 f (t) e Z b st f 0 (t) e h i L f (n) = sn L[f ] ⇥ = lim f (b)e sb ⇤ st f (0) + s lim dt b!1 Z t=0 b f (t)e 0 Similarly = = s L[f 0 ] s2 L[f ] L[f t=b f (0) + sL[f ] L [f 00 ] f (0) sn 2 0 f (0) ··· 0 b!1 = 1 f (n If L[f ] = F (s) dt Integration by parts Z st = lim f (t)e + sf (t)e b!1 sn dt st 3.10 The Laplace transform of the n-th derivative of f The Laplace transform of f 0 (t) 0 MAST20029 Engineering Maths: Laplace Transforms f 0 (0) s f (0) f 0 (0) st dt (n) n ] = s F (s) n X1 k=0 sn 1 k (k) f (0) 1) (0) MAST20029 Engineering Maths: Laplace Transforms 3.11 Inverse Laplace transform If L[f (t)] = F (s) is the Laplace transform of f , then f is the inverse Laplace transform of F and is denoted by L 1 [F ]. It is defined by the complex contour integral f (t) = L 1 1 [F ] = 2⇡i Z MAST20029 Engineering Maths: Laplace Transforms 3.12 Partial Fractions Revision Let f (x) and g(x) be polynomials, then f (x) ! degree n g(x) ! degree d can be written as the sum of partial fractions if n < d. c+i1 F (s)e st ds c i1 Denominator Factor Note • This integral is beyond the scope of this subject. We will use tables to invert Laplace transforms. (x a) (x a)r • Linearity also holds for inverse transforms. Hence if F (s) = c1 F1 (s) + c2 F2 (s) Partial Fraction Expansion A x a A1 A2 Ar + + ··· + x a (x a)2 (x a)r then L 1 [F ] = c1 L where c1 , c2 2 C. 1 [F1 ] + c2 L 1 [F2 ] Ax + B + bx + c (x2 + bx + c) (x2 + bx + c)r x2 A1 x+B1 x2 +bx+c + A2 x+B2 (x2 +bx+c)2 + ··· + Ar x+Br (x2 +bx+c)r MAST20029 Engineering Maths: Laplace Transforms Exercise 7 MAST20029 Engineering Maths: Laplace Transforms Using Laplace transforms to solve linear ODEs Compute the inverse Laplace transform of F (s) = Solution 3.13 (s 1 2)(s Exercise 8 Solve the second order initial value problem 1) x00 for x(t). Solution x = 1, x(0) = 1, x0 (0) = 2 3.14 MAST20029 Engineering Maths: Laplace Transforms 3.15 MAST20029 Engineering Maths: Laplace Transforms 3.16 Using Laplace transforms to solve linear systems of 1st order ODEs Exercise 9 Solve the system of inhomogeneous di↵erential equations x0 + y 0 + x + y y0 x+y = = 1 (1) t for x(t) and y(t), given that x(0) = 0 and y(0) = 0. Solution (2) MAST20029 Engineering Maths: Laplace Transforms 3.17 MAST20029 Engineering Maths: Laplace Transforms 3.18 MAST20029 Engineering Maths: Laplace Transforms 3.19 MAST20029 Engineering Maths: Laplace Transforms The s-shifting theorem Exercise 10 If L[f (t)] = F (s) then Find the Laplace transform of L[e at g(t) = t2 e f (t)] = F (s + a) Solution or L 1 [F (s + a)] = e at f (t) for any real number a. If s is a real number, the exponential shifts the variable s to the left by a. 4t 3.20 MAST20029 Engineering Maths: Laplace Transforms 3.21 MAST20029 Engineering Maths: Laplace Transforms Exercise 11 Exercise 12 Find the inverse Laplace transform of Find the inverse Laplace transform of G(s) = Solution 4s s2 12 6s + 18 H(s) = Solution 4s + 12 (s 3)2 + 9 3.22 MAST20029 Engineering Maths: Laplace Transforms 3.25 MAST20029 Engineering Maths: Laplace Transforms The t-shifting theorem in terms of step functions Exercise 13 Using step functions, we can write the t-shifting theorem as Find the inverse Laplace transform of If L[f (t)] = F (s), then L[f (t or L 1 ⇥ e a)u(t as G(s) = a)] = e ⇤ F (s) = f (t for any real number a > 0. as F (s) a)u(t a) Solution e 2s s5 3.26 MAST20029 Engineering Maths: Laplace Transforms 3.27 Exercise 14 Find the Laplace transform of the function 8 > if 0 t < ⇡ < 2 g(t) = 0 if ⇡ t < 2⇡ > : sin t if t 2⇡ g 2 O t ⇡ 2⇡ 3⇡ 4⇡ Solution Write g as a linear combination of regular functions multiplied by unit step functions. MAST20029 Engineering Maths: Laplace Transforms 3.28 MAST20029 Engineering Maths: Laplace Transforms 3.31 MAST20029 Engineering Maths: Laplace Transforms 3.32 MAST20029 Engineering Maths: Laplace Transforms 3.33 MAST20029 Engineering Maths: Laplace Transforms Impulse The Dirac delta function Some physical phenomena involve the action of large forces over short intervals of time, such as a tennis ball hit by a racquet or a voltage applied to an electric circuit. Define the Dirac -function for a For any a 0 and ✏ > 0, let ( f✏ (t a) = lim f✏ (t ✏!0 if a t a + ✏ otherwise 0 Z t (⌧ a) d⌧ = u(t 0 1/✏ Z 0 a a+✏ t The impulse of a force acting over a short time interval a t a + ✏ is the integral of f✏ from a to a + ✏. I✏ = Z 0 1 f✏ (t a) dt = 0 1 a) = ( 0 1 and f✏ O a) = ( Z a+✏ a if t 6= a if t = a So that 1 ✏ a) = (t 0 as 1 1 dt = [(a + ✏ ✏ ✏ a) = 1 1 (⌧ a) d⌧ = 1. if t < a if t a 3.34 MAST20029 Engineering Maths: Laplace Transforms 3.35 Laplace transform of Dirac delta function MAST20029 Engineering Maths: Laplace Transforms Exercise 16 Solve L[ (t a)] y 00 + 4y 0 + 4y = (t = L[lim f✏ (t = lim L[f✏ (t a)] ✏!0 Z 1 lim f✏ (t a)e ✏!0 0 Z a+✏ 1 st lim e dt ✏!0 a ✏ t=a+✏ e st lim ✏!0 s✏ t=a = = = = = = ✏!0 lim e as e s✏ ✏!0 lim se ✏!0 as as s a)] e as e 1) with y(0) = 0, st dt s✏ s✏ by L’Hôpital’s rule e Note Putting a = 0 gives L[ (t)] = 1 and L 1 [1] = (t) Solution y 0 (0) = 0 3.36 MAST20029 Engineering Maths: Laplace Transforms 3.37 MAST20029 Engineering Maths: Laplace Transforms 3.38 Laplace transforms of integrals If f has a Laplace transform L[f ] = F (s), then L = = = = = = Z Z t f (⌧ ) d⌧ 0 ✓Z 1 0 lim b!1 lim b!1 lim b!1 Z f (⌧ ) d⌧ 0 b 0 " ✓Z ◆ e t f (⌧ ) d⌧ 0 st ◆ e dt st dt Integration by parts Z Z 1 1 st t e f (t)e f (⌧ ) d⌧ + s s 0 1 lim s b!1 F (s) s t 1 e s Z sb Z b 0 1 f (⌧ ) d⌧ + s b f (t)e 0 st dt Z t=b st dt t=0 b f (t)e 0 st dt # MAST20029 Engineering Maths: Laplace Transforms 3.39 MAST20029 Engineering Maths: Laplace Transforms 3.40 Exercise 17 Convolution Solve the integral equation We define the (Laplace) convolution of two functions f and g as follows y(t) Z t y(⌧ ) d⌧ = 3 0 (f ⇤ g)(t) = Solution Z t f (⌧ )g(t ⌧ ) d⌧ 0 The convolution of two functions is symmetric, i.e. f ⇤g =g⇤f or Z t f (⌧ )g(t ⌧ ) d⌧ = 0 Z t f (t ⌧ )g(⌧ ) d⌧ 0 The convolution theorem If L[f (t)] = F (s) and L[g(t)] = G(s), then L[(f ⇤ g)(t)] = F (s)G(s) or (f ⇤ g)(t) = L 1 [F (s)G(s)] MAST20029 Engineering Maths: Laplace Transforms 3.41 Exercise 18 Use the convolution theorem to find the inverse Laplace transform of 1 H(s) = 2 2 s (s + 1) Solution MAST20029 Engineering Maths: Laplace Transforms 3.42 MAST20029 Engineering Maths: Laplace Transforms Exercise 19 Solve the integral equation y(t) = Solution 1 2 t 2 Z t y(⌧ )(t 0 ⌧ ) d⌧ 3.43 MAST20029 Engineering Maths: Laplace Transforms 3.44 MAST20029 Engineering Maths: Sequences and Series 4.1 SEQUENCES AND SERIES Sequences MAST20029 Engineering Maths: Sequences and Series Standard Limits 1. lim 1 =0 n!1 np (p > 0) 2. lim rn = 0 (|r| < 1) A sequence is an ordered list of numbers a1 , a2 , a3 , a4 . . . a n . . . n!1 denoted by {an } where an is the nth term (n 1). A sequence {an } has the limit L (L is finite) if an approaches L as n approaches infinity. We write: lim an = L, Theorem Let f be a real function and {an } be a sequence of real numbers such that an = f (n). If x!1 lim an = L n!1 Note lim an = L n!1 6) lim f (x) = L x!1 (a > 0) 4. lim n1/n = 1 n!1 5. lim If the limit exists the sequence converges. Otherwise, the sequence diverges. then n!1 an =0 n!1 n! n!1 lim f (x) = L 3. lim a1/n = 1 (a 2 R) loge n =0 (p > 0) np ⇣ a ⌘n 7. lim 1 + = ea (a 2 R) n!1 n 6. lim n!1 np =0 n!1 an 8. lim (p 2 R; a > 1) 4.2 MAST20029 Engineering Maths: Sequences and Series 4.3 Series S 1 = a1 A series is denoted by S 2 = a1 + a 2 S 3 = a1 + a 2 + a 3 a n = a1 + a 2 + a 3 + a 4 . . . .. . n=1 S n = a1 + a 2 + a 3 + . . . + a n Example The sequence {n} = 1, 2, 3, 4, . . .. 1 X The series n = 1 + 2 + 3 + 4... The sequence {Sn } is called the sequence of partial sums. n=1 If {Sn } converges, that is lim Sn = L, where L is a finite n!1 The sequence and series diverge to infinity. value, then we say that the series 1 X ai converges. i=1 Example The sequence The series 4.4 Given a particular sequence {an }, we can formulate a sequence of sums, A series is the sum of all terms in the sequence {an }. 1 X MAST20029 Engineering Maths: Sequences and Series ⇢ 1 10n If a series does not converge, then it diverges. = 0.1, 0.01, 0.001, . . .. 1 X 1 = 0.1 + 0.01 + 0.001 + . . . = 0.11111111 n 10 n=1 The sequence converges to 0 while the series converges to 1 . 9 MAST20029 Engineering Maths: Sequences and Series Exercise 1 4.5 MAST20029 Engineering Maths: Sequences and Series Exercise 2 th n Find the k Solution Solution n=1 ⇢ 1 1 1 1, , , , . . . . 2 4 8 1 1 1 Evaluate the sum 1 + + + + . . .. 2 4 8 partial sum of the sequence {( 1) }. 1 X Determine if ( 1)n converges or diverges. Find the k th partial sum of 4.6 MAST20029 Engineering Maths: Sequences and Series Geometric Series MAST20029 Engineering Maths: Sequences and Series The Divergence Test 1 X an diverges. If lim an 6= 0 then n!1 The sum of a geometric series 1 X 4.7 arn = a + ar + ar2 + . . . Note n=0 is convergent if |r| < 1 and diverges if |r| If |r| < 1, we have 1 X n=0 arn = 1. a 1 r Exercise 3 1 1+ + 4 Solution n!1 ✓ ◆2 ✓ ◆ 3 1 1 + + ... 4 4 1 X an may converge or diverge. n=1 Exercise 4 1 X n+1 converge or diverge? Does n n=1 Solution What does the series converge to? If lim an = 0 then n=1 4.8 MAST20029 Engineering Maths: Sequences and Series 4.9 MAST20029 Engineering Maths: Sequences and Series 4.10 Integral Test Exercise 5 If f is a continuous, positive, decreasing function on [1, 1) 1 X and an = f (n). Then the series an For what values of p > 0 does n=1 1. converges if 2. diverges if R1 1 R1 1 f (x) dx converges. f (x) dx diverges. Solution Z 1 1 1 dx converge? xp MAST20029 Engineering Maths: Sequences and Series 4.11 MAST20029 Engineering Maths: Sequences and Series 4.12 Harmonic p Series The harmonic p series 1 X 1 p n n=1 converges if p > 1 and diverges if p 1. Proof p>0 • f (x) = • Z 1 1 1 is continuous, positive and decreasing on [1, 1). xp 1 dx is convergent if p > 1 and divergent if p 1. xp 1 X 1 • Hence, converges if p > 1 and diverges if p 1, by np n=1 the integral test. p0 1 X 1 1 = 6 0, diverges by the divergence test. p n!1 np n n=1 • Since lim Combining cases: 1 X 1 converges if p > 1 and diverges if p 1. np n=1 MAST20029 Engineering Maths: Sequences and Series 4.13 MAST20029 Engineering Maths: Sequences and Series 4.14 Exercise 6 1 X 1 Does converge or diverge? (n 3)2 n=4 Comparison Test 1 1 X X an and bn be positive term series. Let n=1 n=1 Solution 1. If an bn for all n and converges. 2. If an diverges. 1 X bn converges, then n=1 bn for all n and 1 X n=1 1 X an 1 X an n=1 bn diverges, then n=1 To apply the comparison test we compare a given series to a harmonic p series or a geometric series. MAST20029 Engineering Maths: Sequences and Series 4.15 MAST20029 Engineering Maths: Sequences and Series 4.16 Exercise 7 Exercise 8 Determine whether converges or diverges. Solution 1 X n2 n3 + 1 n=1 Determine whether 1 X n=2 converges or diverges. Solution 5 n3 1 MAST20029 Engineering Maths: Sequences and Series 4.17 MAST20029 Engineering Maths: Sequences and Series 4.18 Ratio Test 1 X Let an be a positive term series and Exercise 9 1 X nn Does converge or diverge? n! n=1 n=1 L = lim n!1 1. If L < 1, 1 X an+1 . an an converges. n=1 2. If L > 1, 1 X an diverges. n=1 3. If L = 1, the ratio test is inconclusive. The ratio test is particularly useful if an contains an exponential or factorial function of n. Solution MAST20029 Engineering Maths: Sequences and Series 4.19 MAST20029 Engineering Maths: Sequences and Series 4.20 Exercise 10 Alternating Series Does converge or diverge? Solution 1 X n!n! (2n)! n=1 An alternating series has terms which have alternating signs (+, , +, . . .) or ( , +, , + . . .). Example 1 X ( 1)n n n=1 1 =1 1 1 + 2 3 1 1 + ... 4 5 Alternating Series Test (Leibniz Test) If the series 1 X ( 1)n an satisfies n=1 1. an > 0 for all n 1 2. lim an = 0 n!1 3. an+1 an for all n N , for some integer N then the series converges. If condition 2 is violated, the series diverges. MAST20029 Engineering Maths: Sequences and Series 4.21 MAST20029 Engineering Maths: Sequences and Series 4.22 Exercise 11 1 X ( 1)n Does converge or diverge? n n=1 Exercise 12 1 X 3n( 1)n Does 4n 1 n=1 Solution Solution 1 converge or diverge? MAST20029 Engineering Maths: Sequences and Series 4.23 MAST20029 Engineering Maths: Sequences and Series 4.24 Power Series Theorem 1 X cn (x a)n are of three types: A general power series about x = a has the form n=0 1 X cn (x a)n = c0 + c1 (x a) + c2 (x a)2 + c3 (x a)3 . . . 1. convergent only when x = a n=0 where x 2 R is variable and cn 2 R are constants for n 0. 2. convergent for all x 3. convergent for all |x |x a| > R a| < R and divergent for all R is called the radius of convergence. • In case (1), R = 0. • In case (2), R = 1. The interval of convergence is the interval of x for which the series converges. MAST20029 Engineering Maths: Sequences and Series 4.25 MAST20029 Engineering Maths: Sequences and Series 4.26 Generalised Ratio Test 1 X Let an and Exercise 13 n=1 L = lim n!1 1. If L < 1, 1 X an+1 . an Solution an converges. n=1 2. If L > 1, 1 X Find the radius of convergence and interval of convergence of 1 X ( 3)n xn . the series n2 + 1 n=1 an diverges. n=1 3. If L = 1, the generalised ratio test is inconclusive. MAST20029 Engineering Maths: Sequences and Series 4.27 MAST20029 Engineering Maths: Sequences and Series 4.28 MAST20029 Engineering Maths: Sequences and Series 4.29 Exercise 14 Using the generalised ratio test, find the radius of convergence 1 X (x 1)n . and interval of convergence of the series 2n n=1 Solution MAST20029 Engineering Maths: Sequences and Series 4.30 MAST20029 Engineering Maths: Sequences and Series 4.33 MAST20029 Engineering Maths: Sequences and Series 4.34 The approximation improves as we take higher order terms. In general Exercise 15 f (x) =f (a) + f 0 (a)(x f 00 (a)(x 2! a)2 a)+ + ⇡ Find the 4th order Taylor polynomial about x = for 2 f (x) = cos x Solution f 000 (a)(x 3! a)3 + ... This is called a Taylor series. If a = 0, then the Taylor series is called a Maclaurin series. We can approximate a di↵erentiable function f by a truncated Taylor series. We call the polynomial approximation a Taylor polynomial of degree n. f (x) ⇡ Pn (x) = n X f (k) (a)(x k=0 k! a)k ⇣⇡⌘ ⇣⇡ ⌘ ⇣ ⇡ ⌘ f 00 ⇡2 x +f x + P4 (x) =f 2 2 2 2! 4 ⇡ 3 000 ⇡ iv ⇡ f x 2 x ⇡2 f 2 2 + + 3! 4! 0 ⇡ 2 2 MAST20029 Engineering Maths: Sequences and Series 4.35 MAST20029 Engineering Maths: Sequences and Series 4.36 Exercise 16 Error in Taylor Polynomials Find the Maclaurin series for f (x) = loge (1 + x) Taylor’s Theorem Solution The nth order Taylor Polynomial for f around x = a is Pn (x) = f (a) + f 0 (a)(x a) + . . . + f n (a)(x n! a)n The truncation error or remainder for the nth order Taylor Polynomial is So, P (x) = f (0) + f 0 (0)x + Rn (x) = f (x) 00 2 000 3 iv Pn (x) 4 f (0)x f (0)x f (0)x + + +. . . 2! 3! 4! Rn (x) = f n+1 (c)(x a)n+1 (n + 1)! where c lies between a and x. If Rn ! 0 as n ! 1 the Taylor series converges to f . MAST20029 Engineering Maths: Sequences and Series 4.37 Exercise 17 Find an upper bound on the error when loge (1 + x) is approximated by a 3rd order Maclaurin polynomial for |x| 0.1. Solution Since loge (1 + x) ⇡ x and f iv (x) = the error is 6 (1 + x)4 x3 x2 + 2 3 MAST20029 Engineering Maths: Sequences and Series 4.38 MAST20029 Engineering Maths: Sequences and Series 4.39 Exercise 18 Approximate (1.2)7/2 using a 2nd order Maclaurin polynomial for (1 + 2x)7/2 . Give an upper bound on the error. Solution MAST20029 Engineering Maths: Sequences and Series 4.40 MAST20029 Engineering Maths: Sequences and Series 4.41 Exercise 19 (a) Determine the Maclaurin series for ex and find its radius of convergence. 1 X 1 = e. (b) Show that n! n=0 Solution The Maclaurin series is P (x) = f (0) + f 0 (0)x + f 000 (0)x3 f 00 (0)x2 + + ... 2! 3! MAST20029 Engineering Maths: Sequences and Series 4.42 MAST20029 Engineering Maths: Fourier Series 5.1 FOURIER SERIES MAST20029 Engineering Maths: Fourier Series 5.2 Many periodic functions can be expressed as a Fourier series. Assume that a function f has period T = 2L. Let ! = [Kreyszig, p.474-494] A function f is periodic if there is a positive number T such that f (t) = f (t + T ) for all t 2⇡ T . Then f can be represented by a Fourier series of the form: f (t) = a0 + 1 X (an cos(n!t) + bn sin(n!t)) n=1 T is called the period of f . The graph of f is obtained by periodic repetition of its graph in any interval of length T . Example The coefficients a0 , an and bn are known as the Fourier coefficients and may be calculated from Euler’s formulae: a0 = Example an = Periodic functions describe rotating machines, sound waves or seasonal phenomena. f bn = sin t has period 2⇡, 3T 2T T cos(2t) has period ⇡ O T 2T 3T t Z L 1 f (t) dt 2L L Z 1 L f (t) cos(n!t) dt, L L Z 1 L f (t) sin(n!t) dt, L L n = 1, 2, ... n = 1, 2, ... MAST20029 Engineering Maths: Fourier Series 5.5 MAST20029 Engineering Maths: Fourier Series 5.6 MAST20029 Engineering Maths: Fourier Series 5.7 MAST20029 Engineering Maths: Fourier Series 5.8 MAST20029 Engineering Maths: Fourier Series 5.9 MAST20029 Engineering Maths: Fourier Series 5.10 Fourier Series Theorem If a periodic function f with period T = 2L is di↵erentiable at all except finitely many points in the interval [ L, L] then the Fourier series with coefficients given by Euler’s formulae is convergent. If f is continuous at a point t0 2 [ L, L] then the value of the Fourier series is f (t0 ). If f is discontinuous at a point t0 2 [ L, L] then the value of the Fourier series is the average value of the left and right limits of f at t0 . Example MAST20029 Engineering Maths: Fourier Series 5.13 MAST20029 Engineering Maths: Fourier Series 5.14 Fourier cosine series for even functions Fourier sine series for odd functions Suppose f is an even function with period T = 2L. Suppose f is an odd function with period T = 2L. Then f (t) cos(n!t) is an even function, and f (t) sin(n!t) is an odd function. Then f (t) cos(n!t) is an odd function, and f (t) sin(n!t) is an even function. So f may be represented by a Fourier cosine series So f may be represented by a Fourier sine series f (t) where a0 an = = = a0 + 1 L 2 L Z Z 1 X an cos(n!t) n=1 L f (t) = where L f (t) cos(n!t) dt 0 bn sin(n!t) n=1 f (t) dt 0 1 X bn = 2 L Z L f (t) sin(n!t) dt 0 MAST20029 Engineering Maths: Fourier Series 5.17 MAST20029 Engineering Maths: Fourier Series 5.18 MAST20029 Engineering Maths: Fourier Series 5.19 Half Range Expansions Suppose a function f is defined only for the interval 0 t L. MAST20029 Engineering Maths: Fourier Series 5.20 We can extend f to an even periodic function fe with period T = 4. Then fe will have a Fourier cosine series. fe The function may be represented in terms of either a cosine or sine series, by extending its definition to the interval L t L. 3 O Exercise 3 6 4 2 Consider the function 4 2 t 6 1 f (t) = t2 1 Similarly, we can extend f to an odd periodic function fo with period T = 4. Then fo will have a Fourier sine series. for 0 < t < 2. Sketch f . fo Solution 3 f 3 1 6 O 1 2 t 4 2 O 1 3 2 4 6 t MAST20029 Engineering Maths: Fourier Series 5.21 Finding Particular Solutions of Ordinary Di↵erential Equations using Fourier Series MAST20029 Engineering Maths: Fourier Series Solution From Exercise 2, the Fourier series for f is Exercise 4 f (t) = Suppose the motion of a constrained object subject to an external force f is given by the di↵erential equation 0 y + 3y = f where y is the displacement of the body at time t and ( t + ⇡2 , ⇡t<0 f (t) = ⇡ 0t<⇡ t + 2, with f (t) = f (t + 2⇡). 1 X an cos(nt) n=1 where an = ( 0, 4 n2 ⇡ , n even n odd Assume that yp can be represented by a Fourier series with the same period as f , so that yp (t) = ↵0 + 1 X (↵n cos(nt) + n=1 Di↵erentiating with respect to t gives Find a particular solution yp to the di↵erential equation using Fourier series. 5.22 n sin(nt)) MAST20029 Engineering Maths: Fourier Series 5.23 MAST20029 Engineering Maths: Fourier Series 5.24 MAST20029 Engineering Maths: Fourier Series 5.27 MAST20029 Engineering Maths: Fourier Series 5.28 MAST20029 Engineering Maths: Fourier Series 5.33 MAST20029 Engineering Maths: Fourier Series 5.34 Finding Particular Solutions of Ordinary Di↵erential Equations using Fourier Integrals Exercise 7 Find a particular solution to the di↵erential equation y 00 + y 0 + 2y = f where f (t) = ( for |t| 1 for |t| > 1 1 0 Solution From Exercise 5, the Fourier cosine integral for f is: 2 f (t) = ⇡ Z 1 0 sin(!) cos(!t) d! ! Assume that yp can be represented as a Fourier integral: yp (t) = Z 1 0 ↵(!) cos(!t) + (!) sin(!t) d! MAST20029 Engineering Maths: Fourier Series 5.35 MAST20029 Engineering Maths: Fourier Series 5.36 MAST20029 Engineering Maths: Fourier Series 5.37 MAST20029 Engineering Maths: Second order PDEs 6.1 SECOND ORDER PDES [Kreyszig, p.540-552, 558-567] A partial di↵erential equation (PDE) is an equation involving one or more partial derivatives of an unknown function . The dependent variable depends on two or more independent variables which are often time t and one or several space variables x, y and z, e.g. (x, t), (x, y, z). Order The order of a PDE is the order of the highest derivative in the equation. Linearity A PDE is linear if and its partial derivatives appear only linearly (to the first power). The independent variables can appear in any way. Homogeneity A PDE is homogeneous if each of its terms contains either or one of its partial derivatives. Otherwise we call it inhomogeneous. MAST20029 Engineering Maths: Second order PDEs 6.6 MAST20029 Engineering Maths: Second order PDEs Exercise 2 Solution Solve Laplace’s equation Assume the solution has the form @2 @2 + =0 @x2 @y 2 (6) for {(x, y) : 0 < x < 1, 0 < y < 1} subject to the boundary conditions (0, y) = 0 (1, y) = 0 (x, 0) = 0 (x, 1) = sin(⇡x) + 2 sin(3⇡x) (x, y) = X(x)Y (y) Di↵erentiating twice gives @2 @x2 = X (x)Y (y) @2 @y 2 = X(x)Y (y) 00 00 Substitute into (6) y 00 1 = sin(⇡x) + 2 sin(3⇡x) 00 X (x)Y (y) + X(x)Y (y) = 0 00 X (x) = X(x) =0 =0 00 Y (y) Y (y) This occurs only if both sides are equal to a constant: O =0 1 x 00 X (x) = X(x) 00 Y (y) = Y (y) 6.7 MAST20029 Engineering Maths: Second order PDEs 6.8 Case 1: This gives two second order ODEs 00 X (x) X(x) = 0 00 Y (y) + Y (y) = 0 The boundary conditions can be written as (0, y) = X(0)Y (y) = 0 so X(0) = 0 (1, y) = X(1)Y (y) = 0 so X(1) = 0 (x, 0) = X(x)Y (0) = 0 Y (0) = 0 so (x, 1) = sin(⇡x) + 2 sin(3⇡x) Solve the ODEs for the three possible cases of the separation constant : 1. >0 2. =0 3. <0 MAST20029 Engineering Maths: Second order PDEs >0 6.9 MAST20029 Engineering Maths: Second order PDEs Case 2: =0 6.10 MAST20029 Engineering Maths: Second order PDEs Case 3: <0 6.11 MAST20029 Engineering Maths: Second order PDEs 6.12 MAST20029 Engineering Maths: Second order PDEs 6.13 MAST20029 Engineering Maths: Second order PDEs 6.14 MAST20029 Engineering Maths: Second order PDEs 6.15 Exercise 3 Solve the wave equation @2 @x2 @2 =0 @t2 (7) for {(x, t) : 0 < x < L, t > 0}, subject to the boundary and initial conditions @ (0, t) @x @ (L, t) @x @ (x, 0) @t (x, 0) = 0 = 0 = 0 = 1 + cos ✓ 2⇡x L t @ =0 @x O @ =0 @x @ L =0 @t 0 1 2⇡x C A = 1 + cos B@ L x ◆ MAST20029 Engineering Maths: Second order PDEs Solution 6.16 MAST20029 Engineering Maths: Second order PDEs 6.17 This gives two second order ODEs Assume the solution has the form 00 X (x) 00 T (t) (x, t) = X(x)T (t) X(x) = 0 T (t) = 0 Di↵erentiating gives The boundary and initial conditions can be written as @2 @x2 = X (x)T (t) @2 @t2 = X(x)T (t) 00 @ (0, t) = X 0 (0)T (t) = 0 so @x @ (L, t) = X 0 (L)T (t) = 0 so X 0 (L) @x @ (x, 0) = X(x)T 0 (0) = 0 so T 0 (0) @t ◆ ✓ 2⇡x (x, 0) = 1 + cos L 00 Substitute into (7) 00 X (x)T (t) = 0 = 0 = 0 00 X(x)T (t) = 0 00 00 X (x) T (t) = X(x) T (t) This occurs only if both sides are equal to a constant: 00 X 0 (0) 00 X (x) T (t) = = X(x) T (t) Solve the ODEs for the three possible cases of the separation constant : 1. > 0 - this gives only trivial solutions 2. =0 3. <0 MAST20029 Engineering Maths: Second order PDEs Case 2: =0 6.18 MAST20029 Engineering Maths: Second order PDEs Case 3: <0 6.19 MAST20029 Engineering Maths: Second order PDEs 6.20 MAST20029 Engineering Maths: Second order PDEs 6.21 MAST20029 Engineering Maths: Second order PDEs Solution 6.24 MAST20029 Engineering Maths: Second order PDEs 6.25 This gives two ODEs (one first order and one second order) Assume the solution has the form 00 X (x) 0 T (t) (x, t) = X(x)T (t) X(x) = 0 T (t) = 0 Di↵erentiating gives The boundary and initial conditions can be written as @2 @x2 = X (x)T (t) @ @t = X(x)T (t) 00 0 (0, t) = X(0)T (t) = 0 so X(0) = 0 (L, t) = X(L)T (t) = 0 so X(L) = 0 (x, 0) = f (x) Substitute into (8) Solve the ODEs for the three possible cases of the separation constant : 00 0 X (x)T (t) = X(x)T (t) 00 0 T (t) X (x) = X(x) T (t) This occurs only if both sides are equal to a constant: 00 0 X (x) T (t) = = X(x) T (t) 1. > 0 - this gives only trivial solutions 2. = 0 - this gives only trivial solutions 3. <0 MAST20029 Engineering Maths: Second order PDEs Case 3: <0 6.26 MAST20029 Engineering Maths: Second order PDEs 6.27 MAST20029 Engineering Maths: Second order PDEs 6.28 Exercise 5 MAST20029 Engineering Maths: Second order PDEs We now need to compare (x, 0) = f (x) with equation (10). To do this, represent f (x) as a Fourier sine series. Solve the heat equation @ @2 = 2 @x @t (9) We obtain the odd periodic extension (period T = 2L) for {(x, t) : 0 < x < L, t > 0}, subject to the boundary and initial conditions (0, t) = 0 (L, t) = 0 (x, 0) = (x, 0) = f (x) = ( for 0 x L/2 for L/2 < x L x L x Solution bn sin n=1 bn = = Exercise 4 gives the general solution accounting for the homogeneous boundary conditions. (x, t) = 1 X ⇣ n⇡x ⌘ L where the Fourier coefficients are f (x) = 1 X n=1 6.29 En exp ✓ n2 ⇡ 2 t L2 ◆ sin ⇣ n⇡x ⌘ L 2 L 8 > < > : Z L f (x) sin 0 0 4L n2 ⇡ 2 4L n2 ⇡ 2 ⇣ n⇡x ⌘ L dx for even n for n = 1, 5, 9, ... for n = 3, 7, 11, ... (11)