Page 167 - Physical Principles of Sedimentary Basin Analysis
P. 167

6.12 Transient convective heat flow            149

            where the prime denotes differentiation. This expression is rewritten as
                                       U        V     V
                                          =−Pe     +    ,                     (6.174)
                                       U        V     V
            where the left-hand side of the equality is dependent only on ˆ t and the right-hand side is
            dependent only on ˆz. The equality must therefore be equal to a constant, which is written
            −k. The minus sign before k anticipates that it is a positive number. The equation for the
            factor U becomes
                                            U
                                               =−k,                           (6.175)
                                            U
            and the solution for U is (see Exercise 6.5)

                                         U(ˆ t) = U 0 e −k ˆ t  ,             (6.176)
            where U 0 is an arbitrary constant. We cannot allow U(ˆ t) to grow exponentially with time,
            which explains why the constant k has to be a positive number. The equation for V is


                                      V − Pe V + kV = 0,                      (6.177)
            and it is factorized in the following way:

                                      d        d
                                        − r 1     − r 2 V = 0                 (6.178)
                                     dˆz       dˆz
            where r 1,2 are the roots
                                          1
                                                     2
                                    r 1,2 =  Pe ±  Pe − 4k                    (6.179)
                                          2
                                 2
            of the quadratic equation r − Per + k = 0. We see that functions of the form exp(r i ˆz),
            i = 1, 2, are solutions of equation (6.178). The function V can therefore be written as any
            linear combination of these two solutions

                                      V (ˆz) = c 1 e r 1 ˆz  + c 2 e r 2 ˆz .  (6.180)
            The boundary conditions for the transient solution require that V (ˆz=0) = 0 and V (ˆz=1) =
            0, which is the same as c 1 + c 2 = 0 and c 1 exp(−D) + c 2 exp(D) = 0, respectively, where

                        2
            D =   (Pe/2) − k. The boundary conditions therefore require that c 2 =−c 1 and that
            exp(2D) = 1, where the latter equation has the non-trivial solution D = iπn, with i =
            √                                2        2                     2      2
              −1, and where n is any integer. From D =−(πn) it follows that k n = (Pe/2) +(πn)
                        1
            and that r 1,2 = Pe ± iπn. There is not just one k, but one for each integer n. Since D is a
                        2
            complex number it follows from the formula for the exponentiation of a complex number,
            exp(iφ) = cos φ + i sin φ, where φ is real that
                                  V n (ˆz) = c 3 exp(Pe ˆz/2) sin(πnˆz),      (6.181)

            where c 3 is an arbitrary constant and where the subscript n tells us that there is one V for
            each integer n. (Only the complex part i sin φ fulfills the boundary conditions. Only this
   162   163   164   165   166   167   168   169   170   171   172