next up previous
Next: About this document

to 0.9inRead:Sections 11.8, 12.8, 12.9 to 0.9 inProblems:page 707 #1,3,7; page 711 #1, 5, 7, 8; to 0.9inpage 717 #1, 2, 14, 16; This page #1, 2; to 0.9inpage 768 #1; page 778 # 1, 13, 14, 15, 16, 17


Fall, 1996Chemical Engineering 535, Assignment 7page Problems: This page

1.
Show that the Green's function, tex2html_wrap_inline103 , satisfies the end conditions and

displaymath81

where tex2html_wrap_inline105 is the Dirac delta function, (See page 218) and that, therefore, tex2html_wrap_inline103 , as a function of x for fixed tex2html_wrap_inline111 , can be found by solving the problem

displaymath82

with the end conditions satisfied. Use this to find the Green's function for

displaymath83

2.
Consider the problem

displaymath84

where L is a Sturm-Liouville operator, L(u)=(ru')'+pu. Find f(x) and a Sturm-Liouville operator tex2html_wrap_inline119 so that the problem becomes equivalent to finding a solution v(x) to

displaymath85

(i.e. Convert a problem with non-homogeneous boundary conditions to one in standard form, i.e. with homogeneous boundary conditions.) Apply this method to find a solution to

displaymath86


Correction: In Assignment 6 (on page 2) the definition of the inner product for its use in the Statements of the theorems should read

displaymath87

i.e. r(x) should be replaced by q(x). Please make this correction.

Let the Sturm-Liouville operator, L(u) be defined by

displaymath88

with end conditions given by

eqnarray35

Then consider the problem, L(u)=f(x) with u satisfying the end conditions. For given q(x), one can find the eigenvalues, tex2html_wrap_inline135 and the eigenfunctions, tex2html_wrap_inline137 of the problem tex2html_wrap_inline139 , with u satisfying the end conditions. Then, if none of the eigenvalues is zero (Theorem 1 tex2html_wrap_inline143 , page 715 addresses the case where an eigenvalue is zero) the solution is given in equations (7) and (8), page 713, which we paraphrase as

displaymath89

It appears, then, that the solution can be given in terms of a Green's function: Let tex2html_wrap_inline145 so that tex2html_wrap_inline147 is a normalized eigenfunction: tex2html_wrap_inline149 . Then

displaymath90

and the solution to L(u)=f(x), with the end conditions satisfied, is

displaymath91

Laplace's equation in more than one dimension:

The Laplacian operator, tex2html_wrap_inline153 , is defined by

displaymath155

In cylindrical coordinates,

displaymath157

For spherical coordinates, see problem 8, page 778. Laplace's equation, namely tex2html_wrap_inline159 , governs steady state distribution of temperature, T.

In the plane, we may use separation of variables to solve Laplace's equation in a rectangle where the unknown, T, is given on one of the edges of the rectangle and is zero on the others. Adding the solutions of four of such problems will solve the problem where T is given on all four edges.

In polar coordiates, separation of variables will lead to a solution inside a disk, where T is assigned on the edge of the disk. This will give rise to Fourier series.

In Example 1, page 773, there is solved the problem of LaPlace's equation in a cylinder, tex2html_wrap_inline169 , tex2html_wrap_inline171 . Boundary conditions consist of setting T=0 at the ends, z=0 and tex2html_wrap_inline177 , and assigning the value of T on the lateral face: tex2html_wrap_inline181 . The separation of variables method then leads to a solution of the form

displaymath183

(See equation (13), page 777.) See the rest of page 777 for the evaluation of the coefficients tex2html_wrap_inline185 and tex2html_wrap_inline187 .




next up previous
Next: About this document

Barry Bernstein
Wed Oct 30 14:29:41 CST 1996