Page 88 - Essentials of applied mathematics for scientists and engineers
P. 88

book   Mobk070    March 22, 2007  11:7








                     78  ESSENTIALS OF APPLIED MATHEMATICS FOR SCIENTISTS AND ENGINEERS

                            The length and time scales are clearly the same as the problem in Chapter 2. Hence, τ =
                            2
                       tα/L and ξ = x/L. If we choose U = (u − u 0 )/(u 1 − u 0 ) we make the boundary condition
                       at x = 0 homogeneous but the condition at x = L is not. We have the same situation that we
                       had in Section 2.3 of Chapter 2. The differential equation, one boundary condition, and the
                       initial condition are homogeneous. Proceeding, we find


                                                  U τ = U ξξ
                                                  U(τ, 0) = U(0,ξ) = 0                              (5.20)

                                                  U ξ (τ, 1) + B[U(τ, 1) − 1] = 0

                       where B = hL/k. It is useful to relocate the nonhomogeneous condition as the initial condition.
                       As in the previous problem we assume U(τ, ξ) = V (τ, ξ) + W(ξ).

                                                    V τ = V ξξ + W ξξ
                                                    W(0) = 0
                                                    W ξ (1) + B[W(1) − 1] = 0
                                                                                                    (5.21)
                                                    V (τ, 0) = 0
                                                    V ξ (τ, 1) + BV (τ, 1) = 0

                                                    V (0,ξ) =−W(ξ)

                       Set W ξξ = 0. Integrating twice and using the two boundary conditions on W,

                                                                   Bξ
                                                         W(ξ) =                                     (5.22)
                                                                 B + 1
                            The initial condition on V becomes

                                                     V (0,ξ) =−Bξ/(B + 1) .                         (5.23)

                       Assume V (τ, ξ) = P(τ)Q(ξ), substitute into the partial differential equation for V , and divide
                       by PQ as usual.

                                                         P     Q       2
                                                            =     =±λ                               (5.24)
                                                         P     Q
                       We must choose the minus sign for the solution to be bounded. Hence,

                                                            2
                                                   P = Ae −λ τ
                                                                                                    (5.25)
                                                   Q = C 1 sin(λξ) + C 2 cos(λξ)
   83   84   85   86   87   88   89   90   91   92   93