Page 78 - 6637
P. 78

Solved Problems


                   Now we prove our assertion. First note that if we substitute f = v∇u in the
               divergence theorem,
                                                Z                Z
                                                    ∇ · f dx =        f · n ds,
                                                  R                ∂R
               we obtain the identity,

                                          Z                             Z
                                                                              ∂u
                                             (v∆u + ∇v∇u) dx =              v     ds.                    (10.3)
                                           R                             ∂R   ∂n

               Let u be a solution of Laplace’s equation subject to the Robin boundary condition
               with our restrictions on a. We take v = u in Equation 10.3.

                                       Z                 Z     ∂u           Z
                                                 2
                                                                                   2
                                           (∇u) dx =         u     ds = −       au ds
                                         R                 C   ∂n             C
               Since the first integral is non-negative and the last is non-positive, the integrals
               vanish. This implies that ∇u = 0. u is a constant. In order to satisfy the boundary

               condition where a is non-zero, u must be zero. Thus the unique solution in this
               scenario is u = 0.


               Exercise 204. Solve Laplace’s equation on the surface of a semi-infinite cylinder
               of unit radius, 0 < θ < 2π, z > 0, where the solution, u(θ, z) is prescribed at
               z = 0: u(θ, 0) = f(θ).
               Solution: The mathematical statement of the problem is


                                     ∆u ≡ u + u       zz  = 0,   0 < θ < 2π,       z > 0,
                                               θθ
                                                       u(θ, 0) = f(θ).

               We have the implicit boundary conditions,


                                        u(0, z) = u(2π, z),      u (0, z) = u (0, z)
                                                                                θ
                                                                   θ
               and the boundedness condition,


                                                    u(θ, +∞) bounded.

               We expand the solution in a Fourier series. (This ensures that the boundary condi-

               tions at θ = 0, 2π are satisfied.)

                                                               ∞
                                                             X             inθ
                                                 u(θ, z) =         u (z) e
                                                                     n
                                                            n=−∞

               We substitute the series into the partial differential equation to obtain ordinary
               differential equations for the u .
                                                   n
                                                      2
                                                                   00
                                                  −n u (z) + u (z) = 0
                                                         n
                                                                   n
                                                              74
   73   74   75   76   77   78   79   80   81   82   83