Page 93 - 6099
P. 93

across that end. We will assume the simplest condition: that the temperature at each
                                             0
                   end of the rod is held at 0 for all t > 0.
                       2. Initial Conditions.
                   The initial temperature, that is u(x, 0), must be specified as a function f(x) on the
                   interval 0 ≤ x ≤ l, where l is the length of the rod. The function f(x) needs to be at
                   least piecewise continuous on 0 ≤ x ≤ l.



                   9.1      Homogeneous Dirichlet conditions


                   The problem we are going to solve can be summarized as follows. Find a function u(x, t)
                   such that:
                                                               2
                                                         u t = c u xx                                (9.1)
                                                u(0, t) = u(l, t) = 0,  t > 0                        (9.2)

                                                 u(x, 0) = f(x),   x ∈ [0; l]                        (9.3)
                       The method of solution we will use is called the Method of Separation of Variables.
                   It is first assumed that there exist solutions of the form

                                                    u(x, t) = X(x)T(t),

                   that is, solutions which are products of a function of x and a function of t. There will
                   turn out to be an infinity of such product solutions

                                                   u n (x, t) = X n (x)T n (t),

                   each of which satisfies the two boundary conditions. Using the fact that the pde is linear,
                   any linear combination of these,
                                                     X
                                                         a n X n (x)T n (t)
                                                       n
                   will also be a solution. The constants a n will then be chosen to make the infinite sum
                   satisfy the initial condition.
                       If u(x, t) = X(x)T(t), then it is easy to find its partial derivatives. We need u t and
                   u xx to substitute into the PDE:

                                                                       00
                                                        0
                                            u t = X(x)T (t),   u xx = X (x)T(t).
                   Substituting these derivatives into the PDE (9.1),
                                                        0
                                                                  00
                                                               2
                                                 X(x)T (t) = c X (x)T(t),
                   must be true for 00 ≤ x ≤ l and t > 0. The next step is to get everything involving x on
                   one side and everything involving t on the other. This can be done by dividing both sides
                                       2
                   of the equation by c X(x)T(t).
                                                                                  00
                                                         00
                                              0
                                                                        0
                                                      2
                                       X(x)T (t)     c X (x)T(t)       T (t)    X (x)
                                                   =               ⇒         =
                                                                       2
                                       2
                                                       2
                                      c X(x)T(t)      c X(x)T(t)      c T(t)     X(x)
                                                             86
   88   89   90   91   92   93   94   95   96   97   98