Page 94 - 6099
P. 94

The only way a function of x can be identically equal to a function of t is if both of the
                   functions are equal to the same constant; therefore similarly, as for the wave equation, we
                   can write
                                                     0
                                                               00
                                                   T (t)     X (x)
                                                          =         = −λ.                            (9.4)
                                                    2
                                                   c T(t)    X(x)
                       The reason for using a negative constant is that it will give us a Sturm-Liouville
                   equation to solve for X(x) that looks like one we have already seen. The reason for calling
                   the constant λ is that λ turns out to be an eigenvalue and eigenvalues are conventionally
                   called λ.
                       Equation (9.4) can be written in the form of two ordinary differential equations:
                                                      0       2
                                                     T (t) + λc T(t) = 0
                                                       00
                                                     X (x) + λX(x) = 0

                   To obtain boundary conditions on X(x), we use the given boundary conditions (9.2) on
                   u(x, t), that is, since u(x, t) = X(x)T(t)

                                                   0 = u(0, t) ≡ X(0)T(t)
                                                   0 = u(l, t) ≡ X(l)T(t)

                   has to be true for all t > 0. The only way this can be true, without making T(t) identically
                   zero, is to require
                                                     X(0) = X(L) = 0.

                   This leads to a Sturm-Liouville problem for X of the following form:
                                               00
                                             X + λX = 0,      X(0) = X(l) = 0.

                   This is one of the problems we solved in the previous class, and it was shown there that
                   there exists an infinite set of solutions of the form
                                                                   πnx
                                                  X n (x) = A n sin                                  (9.5)
                                                                      l
                   corresponding to the eigenvalues

                                                        2 2
                                                       π n
                                                 λ n =      ,  n = 1, 2, . . . .
                                                        l 2
                   To find the function T n (t) corresponding to X n (x), it is necessary to solve the first-order
                   ODE
                                                                      2 2 2
                                                                    π n c
                                             0           2
                                           T (t) = −λ n c T n (t) = −      T n (t).
                                             n
                                                                       l 2
                   This is a separable first-order equation with general solution
                                                                  2 2 2
                                                                  π n c
                                                    T n (t) = a n e −  l 2  t .                      (9.6)
                   We now have an infinite family of solutions of our heat equation; namely

                                                                                 2 2 2
                                                                      πnx     π n c
                                      u n (x, t) = X n (x)T n (t) = b n sin  e −  l 2  t ,
                                                                         l

                                                             87
   89   90   91   92   93   94   95   96   97   98   99