Partial Differential Equations

Similar documents
Math Assignment 14

# Points Score Total 100

LECTURE 19: SEPARATION OF VARIABLES, HEAT CONDUCTION IN A ROD

MATH 241 Practice Second Midterm Exam - Fall 2012

Review For the Final: Problem 1 Find the general solutions of the following DEs. a) x 2 y xy y 2 = 0 solution: = 0 : homogeneous equation.

Partial Differential Equations

Sturm-Liouville Theory

Introduction to Sturm-Liouville Theory and the Theory of Generalized Fourier Series

Bessel s Equation. MATH 365 Ordinary Differential Equations. J. Robert Buchanan. Fall Department of Mathematics

Problem set 3: Solutions Math 207B, Winter Suppose that u(x) is a non-zero solution of the eigenvalue problem. (u ) 2 dx, u 2 dx.

Differential Equations

c2 2 x2. (1) t = c2 2 u, (2) 2 = 2 x x 2, (3)

MATH 251 Final Examination August 14, 2015 FORM A. Name: Student Number: Section:

In what follows, we examine the two-dimensional wave equation, since it leads to some interesting and quite visualizable solutions.

Boundary Value Problems in Cylindrical Coordinates

Preparation for the Final

Introduction to Sturm-Liouville Theory

Math 2930 Worksheet Final Exam Review

Problem (p.613) Determine all solutions, if any, to the boundary value problem. y + 9y = 0; 0 < x < π, y(0) = 0, y (π) = 6,

Boundary value problems for partial differential equations

Waves on 2 and 3 dimensional domains

An Introduction to Bessel Functions

Wave Equation With Homogeneous Boundary Conditions

MATH 251 Final Examination August 10, 2011 FORM A. Name: Student Number: Section:

Math 251 December 14, 2005 Answer Key to Final Exam. 1 18pt 2 16pt 3 12pt 4 14pt 5 12pt 6 14pt 7 14pt 8 16pt 9 20pt 10 14pt Total 150pt

Partial Differential Equations Summary

The General Dirichlet Problem on a Rectangle

Boundary-value Problems in Rectangular Coordinates

u tt = a 2 u xx u tt = a 2 (u xx + u yy )

MATH 131P: PRACTICE FINAL SOLUTIONS DECEMBER 12, 2012

Lecture6. Partial Differential Equations

Method of Separation of Variables

Two special equations: Bessel s and Legendre s equations. p Fourier-Bessel and Fourier-Legendre series. p

CHAPTER 10 NOTES DAVID SEAL

MATH 251 Final Examination December 19, 2012 FORM A. Name: Student Number: Section:

Connection to Laplacian in spherical coordinates (Chapter 13)

Solutions to the Sample Problems for the Final Exam UCLA Math 135, Winter 2015, Professor David Levermore

Fourier and Partial Differential Equations

Department of Mathematics

The Wave Equation on a Disk

swapneel/207

1. Solve the boundary-value problems or else show that no solutions exist. y (x) = c 1 e 2x + c 2 e 3x. (3)

MA 441 Advanced Engineering Mathematics I Assignments - Spring 2014

Strauss PDEs 2e: Section Exercise 4 Page 1 of 6

Analysis III Solutions - Serie 12

Mathematics (Course B) Lent Term 2005 Examples Sheet 2

Final Exam May 4, 2016

Math 251 December 14, 2005 Final Exam. 1 18pt 2 16pt 3 12pt 4 14pt 5 12pt 6 14pt 7 14pt 8 16pt 9 20pt 10 14pt Total 150pt

24 Solving planar heat and wave equations in polar coordinates

Spotlight on Laplace s Equation

Introduction and preliminaries

MATH FALL 2014 HOMEWORK 10 SOLUTIONS

Lecture Notes for MAE 3100: Introduction to Applied Mathematics

LEGENDRE POLYNOMIALS AND APPLICATIONS. We construct Legendre polynomials and apply them to solve Dirichlet problems in spherical coordinates.

3 Green s functions in 2 and 3D

1 A complete Fourier series solution

Differential Equations 2 Homework 5 Solutions to the Assigned Exercises

Math 20C Homework 2 Partial Solutions

Solutions to old Exam 3 problems

SAMPLE FINAL EXAM SOLUTIONS

Midterm 2: Sample solutions Math 118A, Fall 2013

Autumn 2015 Practice Final. Time Limit: 1 hour, 50 minutes

MATH-UA 263 Partial Differential Equations Recitation Summary

EXAM 2 MARCH 17, 2004

MA Chapter 10 practice

Partial Differential Equations for Engineering Math 312, Fall 2012

Eigenvalue Problem. 1 The First (Dirichlet) Eigenvalue Problem

Preliminary Exam 2018 Solutions to Morning Exam

MATH 251 Final Examination May 4, 2015 FORM A. Name: Student Number: Section:

Solving the Heat Equation (Sect. 10.5).

Lecture Notes for Ch 10 Fourier Series and Partial Differential Equations

Separation of variables

MATH 251 Final Examination December 16, 2014 FORM A. Name: Student Number: Section:

In this chapter we study elliptical PDEs. That is, PDEs of the form. 2 u = lots,

Separation of variables in two dimensions. Overview of method: Consider linear, homogeneous equation for u(v 1, v 2 )

THE METHOD OF SEPARATION OF VARIABLES

Analysis III (BAUG) Assignment 3 Prof. Dr. Alessandro Sisto Due 13th October 2017

MATH 124B Solution Key HW 03

PDE and Boundary-Value Problems Winter Term 2014/2015

Partial Differential Equations Separation of Variables. 1 Partial Differential Equations and Operators

Math 210, Final Exam, Spring 2012 Problem 1 Solution. (a) Find an equation of the plane passing through the tips of u, v, and w.

FINAL EXAM, MATH 353 SUMMER I 2015

Math 300 Winter 2011 Advanced Boundary Value Problems I. Bessel s Equation and Bessel Functions

McGill University April 20, Advanced Calculus for Engineers

Math 250B Final Exam Review Session Spring 2015 SOLUTIONS

MATH 412 Fourier Series and PDE- Spring 2010 SOLUTIONS to HOMEWORK 5

Series Solutions Near a Regular Singular Point

Math 316/202: Solutions to Assignment 7

Math Homework 2

Lecture 24. Scott Pauls 5/21/07

Separation of Variables. A. Three Famous PDE s

Separation of Variables

THE NATIONAL UNIVERSITY OF IRELAND, CORK COLÁISTE NA hollscoile, CORCAIGH UNIVERSITY COLLEGE, CORK. Summer Examination 2009.

21 Laplace s Equation and Harmonic Functions

G: Uniform Convergence of Fourier Series

A proof for the full Fourier series on [ π, π] is given here.

Chapter 5.8: Bessel s equation

Midterm 1 practice UCLA: Math 32B, Winter 2017

Math 3150 Problems Chapter 3

2.2 Separable Equations

Transcription:

Partial Differential Equations Spring Exam 3 Review Solutions Exercise. We utilize the general solution to the Dirichlet problem in rectangle given in the textbook on page 68. In the notation used there and on page 67 we have a = b = f x = 5x f x = 5 x g y = 5y and g y = 5 y. It follows that the coefficients in the series solution are A n = B n = C n = D n = 5x sin nπx dx = n+ nπ 5 x sin nπx dx = 4 n + n π n 3 π 3 5y sin nπy dy = 4 n + n+ n π n 3 π 3 5 y sin nπy dy = nπ. The complete solution is then given by ux y = A n sin nπx nπ y sinh + nπ x + C n sinh sin nπy + with A n B n C n and D n as above. B n sin nπx D n x sinh nπy sin nπy Exercise. This is simply an exercise in using the chain rule. If z = e x/ then on applying the chain rule we have dy dx = dy dz dz dx = dy dz e x/. Another application of the chain rule with the product rule then yields d y dx = d dy dx dz e x/ = d dy e x/ + dy dx dz dz e x/ = d y dz dz dx e x/ + dy dz e x/ = d y dz e x + dy dz e x/. Since e x/ = z/ this implies that d y dx = z d y 4 dz + z dy 4 dz.

Substituting this into the differential equation y + e x y = and again using the fact that e x/ = z/ we get z d y 4 dz + z dy 4 dz + z 4 y =. Multiplying through by 4 finally gives us which is Bessel s equation of order. z d y dz + z dy dz + z y = Exercise 3. The central idea in both parts of this exercise is to use equation 6 of section 4.8 namely J p+ = p x J px J p x to express the given Bessel function in terms of those of lower order. a. Equation with p = 3/ tells us that J 5/ x = 3 x J 3/x J / x. Our goal is to get formulas for J / x and J 3/ x and plug them in to. According to Example in section 4.7 J / x = sin x. 3 πx We now need a similar formula for J 3/ x. For this we turn to identity 4 of section 4.8. If we divide by x and set p = / we find that J 3/ x = J /x + x J /x = πx cos x + x πx sin x + x πx sin x = πx cos x + sin x x πx where we have used the expression 3 for J / x. If we plug this and 3 into we obtain J 5/ x = 3 x πx cos x + 3 x πx sin x πx sin x 3 = πx x sin x 3x cos x which is what we were asked to show. b. This is simply a repeated use of. Starting with p = 4 we get J 5 x = 8 x J 4x J 3 x. 4

Taking p = 3 gives J 4 x = 6 x J 3x J x and plugging this into 4 and simplifying we find that 48 J 5 x = x J 3 x 8 x J x. If we repeat this process using p = and then p = in we eventually find that J 5 x = 7 x + 384 J x 4 x + x 9 J x 3 x. [Remark: There is actually a nice way to make the procedure just described much more efficient i.e. amenable to hand computation. The idea is to write the relationship in matrix form as If we let then this becomes It then follows that That is Jp+ x J p x A p = = p/x p/x Jp x J p x Jp x v p = J p x v p+ = A p v p.. v 5 = A 4 v 4 = A 4 A 3 v 3 = A 4 A 3 A v = A 4 A 3 A A v. J5 x J 4 x = A 4 A 3 A A J x J x Computing the matrix product A 4 A 3 A A a relatively straightforward procedure this immediately expresses J 5 x and J 4 x too in terms of J x and J x. This same idea can be used to quickly express any given Bessel function of the first kind in terms of Bessel functions of lower order.]. Exercise 4. As with Exercise we simply refer to the general solution to the vibrating membrane problem as given on pages and 3 in our textbook. The statement of the problem implies that the initial shape of the membrane is fr θ = i.e. the membrane is initially flat which according to the integral formulae on page immediately tells us that a mn = b mn = for all m and n. Moreover using the given radius and initial velocity the integral factor appearing in the formula for a n is π 5 r r 3 cos 3θJ λ n rr dθ dr = π dr cos 3θ dθ = }{{}

we have simply written the r-integrand as since it is clearly immaterial to this computation. This implies that a n = for all n. Likewise the integral factor appearing in b mn separates as π dr cos 3θ sin mθ dθ = }{{} which implies that b mn = for all m and n the fact that the θ integral vanishes can be observed in one of several ways: by direct antidifferentiation; by observing that the integrand is an odd π-periodic function and we are integrating over a complete period; or by utilizing previously established orthogonality results. We now turn to the computation of the a mn coefficients. integral factor we find that it separates to become 5 r r 4 J m λ mn r dr cos 3θ cos mθ dθ. }{{} } {{ } A π B Again focusing only on the The integral B is equal to unless m = 3 as above there are a few ways to verify this the simplest probably being orthogonality in which case it is equal to π. This means a mn = unless m = 3. Setting m = 3 we can quickly evaluate the integral A by taking advantage of the useful identity on page of the textbook. It gives 5 r r 4 J 3 λ 3n r dr = 5 r r 4 J 3 λ 3n r dr = 57 J α3n 5 α 3n. Putting these computations together we find that taking a quick glance at equation 8 on page 3 a 3n = 5πcα 3n J 4 α 3n AB = 4 56 J 5 α 3n cα 3 3nJ 4 α 3n. This can be further simplified although this isn t strictly necessary using identity with p = 4 which tells us that J 5 α 3n = 8 α 3n J 4 α 3n J 3 α 3n = 8J 4α 3n α 3n since J 3 α 3n =. Using this in our previous expression and simplifying slightly we end up with a 3n = 5 cα 4 3nJ 4 α 3n.

Finally we conclude that the shape of the membrane at time t is given by ur θ t = = a 3nJ 3 λ 3n r cos 3θ sin cλ 3n t 5 cos 3θ c α3nj 4 4 α 3n J α3n 3 5 r cα3n t sin 5 the elasticity constant c was unspecified so it must appear in our solution as a parameter. Exercise 5. This one s extra credit. No hints! Exercise 6. a. The given differential equation is not in S-L form. However if we multiply it by x / it becomes which is the same as x 3/ y 3x x / y + nn + x / y = x 3/ y + nn + x / y =. This is in S-L form with px = x 3/ qx = and rx = x / with parameter λ = nn +. The given problem is singular because: The boundary conditions are not of the correct form see equation of section 6.. The function px is not positive for x since p = p =. The function rx is not positive for x since r = r =. b. Suppose that y and y are eigenfunctions of the problem in question with distinct eigenvalues. In class we showed that y and y are guaranteed to be orthogonal on the interval [ ] with respect to the weight function rx = x / provided that px y xy x y xy x =. 5 Because p = p = as noted above this is immediate thereby verifying orthogonality. Exercise 7. Separating variables we assume that ux t = XxT t and plug into the given PDE yielding X T xxt =. Dividing by xxt this becomes X xx T T =

or X xx = T T = λ for some constant λ since the first two terms depend on different independent variables. Reorganizing slightly we obtain the two ODEs X + λxx = 6 T + λt =. 7 Because we seek nontrivial solutions the given boundary conditions also tell us that X = X =. The equation for X is in S-L form with px = qx = and rx = x. The conditions X = and X = are regular see equation of section 6. but the fact that r = means that the boundary value problem for X is actually singular. Nonetheless the fact that the boundary conditions are regular implies the eigenfunctions for distinct eigenvalues are orthogonal on [ ] with respect to the weight rx = x. Moreover one can show directly that the eigenvalues form an increasing sequence λ < λ < λ 3 < that tends to infinity that up to a scalar multiple the eigenfunction X n x associated with λ n is unique and that the Fourier convergence theorem holds for these eigenfunctions. Taking this as given equation 7 tells us that T t = T n t = e λnt so that the normal modes of the original boundary value problem are and the general solution is given by The initial condition gives u n x t = X n xe λnt ux t = c n X n xe λnt. fx = ux = c n X n x with according to the Fourier convergence result quoted above c n = fxx n xx dx. Xnxx dx According to my notes I originally intended to use the interval x. This would have kept the problem regular since then p and r would have been positive throughout the interval. This is automatically true of any regular S-L problem and a brief sketch of the derivation of these facts in this particular singular case is given below.

[Remark: Changing variables specifically by first substituting y = x / X followed by setting s = x 3/ /3 one can transform the equation X + λxx = into the parametric Bessel equation of order /3. Imposing the boundary conditions one can then show that nonzero solutions are possible only for λ > and after back substitution one can show that in this case the eigenvalues and eigenfunctions are λ n = 9α /3n 4 X n x = x / J /3 α/3n x 3/ for n = 3.... Aside from the completeness statement i.e. that Fourier convergence holds this proves the claims made above.]