Page 479 - Electromagnetics
P. 479

into (A.61), we find that only the first two terms give non-zero contributions to the
                        integral and


                                           ∂F(r, t)            F(r, t) · dS −  F(r, t) · dS

                                ψ(t) =            · dS + lim  S 2          S 1        .        (A.62)
                                        S(t)  ∂t        t→0              t
                        The second term on the right can be evaluated with the help of Figure A.5. As the surface
                        moves through a displacement v t it sweeps out a volume region  V that is bounded
                        on the back by S 1 , on the front by S 2 , and on the side by a surface S 3 =  S. We can
                        thus compute the two surface integrals in (A.62) as the difference between contributions
                        from the surface enclosing  V and the side surface  S (remembering that the normal
                        to S 1 in (A.62) points into  V ). Thus

                                          ∂F(r, t)           S 1 +S 2 + S  F(r, t) · dS −   S  F(r, t) · dS 3

                                ψ(t) =           · dS + lim
                                             ∂t         t→0                 t
                                       S(t)

                                          ∂F(r, t)           V  ∇· F(r, t) dV 3 −   S  F(r, t) · dS 3

                                    =            · dS + lim
                                       S(t)  ∂t         t→0                t
                        by the divergence theorem. To compute the integrals over  S and  V we note from
                        Figure A.5 that the incremental surface and volume elements are just
                                             dS 3 = dl × (v t),  dV 3 = (v t) · dS.
                        Then, since F · [dl × (v t)] =  t(v × F) · dl, we have

                                   ∂F(r, t)             S(t)                     t     [v × F(r, t)] · dl
                                                      t    [v∇· F(r, t)] · dS
                         ψ(t) =           · dS + lim                      − lim                    .
                                S(t)  ∂t         t→0           t             t→0         t
                        Taking the limit and using Stokes’s theorem on the last integral we have finally
                                       d                  ∂F
                                             F · dS =        + v∇· F −∇ × (v × F) · dS,        (A.63)
                                       dt  S(t)       S(t)  ∂t
                        which is the Helmholtz transport theorem [190, 43].
                          In case the surface corresponds to a moving physical material, we may wish to write
                        the Helmholtz transport theorem in terms of the material derivative. We can set v = u
                        and use

                                      ∇× (u × F) = u(∇· F) − F(∇· u) + (F ·∇)u − (u ·∇)F
                        and (A.60) to obtain
                                        d                   DF
                                              F · dS =        + F(∇· u) − (F ·∇)u · dS.
                                       dt  S(t)       S(t)  Dt
                        If S(t) in (A.63) is closed, enclosing a volume region V (t), then

                                          [∇× (v × F)] · dS =   ∇· [∇× (v × F)] dV = 0
                                        S(t)                 V (t)
                        by the divergence theorem and (B.49). In this case the Helmholtz transport theorem
                        becomes

                                             d                   ∂F
                                                   F · dS =        + v∇· F · dS.               (A.64)
                                             dt  S(t)       S(t)  ∂t



                        © 2001 by CRC Press LLC
   474   475   476   477   478   479   480   481   482   483   484