Page 54 - Computational Modeling in Biomedical Engineering and Medical Physics
P. 54

40    Computational Modeling in Biomedical Engineering and Medical Physics


                and if on the surface Σ either its normal component or its tangential component are
                specified
                                         nUFnðÞ 5 F n ; n 3 FnðÞ 5 F t :             ðA1:4Þ

                   The curl and divergence are local physical quantities and the mathematical models,
                to which they lead, may be higher order partial differential equations (PDEs) of differ-
                ent types—Laplace, Helmholtz, and so on.
                   The integral form of the theorem (A.1.3) requires that the circulation, R(r), and
                the flux, Q(r), of the vector field F(r) are specified
                           I         ð                  I        ð
                              FUds 5    ð r 3 FÞUdA 5 R;  FdA 5    ð rUFÞdv 5 Q:     ðA1:5Þ
                             Γ        A Γ                Σ        v Σ

                   The two integral relations, Stokes and Gauss, that relate the circulation and
                flux integrals to surface and volume integrals, respectively, are used as discussed
                earlier.
                   The PDEs are solved numerically by using domain (interior) methods, such as
                finite differences and finite element (Bathe and Wilson, 1976; Peyret and Taylor,
                1983; Morega, 1998). The mathematical models derived using the integral forms
                may be solved numerically using boundary numerical methods (Brebia et al.,
                1984).
                   Remarkably, the integral forms (3) are used to define the divergence and the
                curl, which are invariant with respect to the system of coordinates. The curl is
                given by
                                        H                       H
                                         Γ Fds                   Σ ð n 3 FÞdA
                        r 3 F 5 n lim        ; or r 3 F 5 n lim             ;        ðA:1:6Þ
                                  ΔA Γ -0 ΔA Γ             Δv Σ -0  Δv Σ
                                                                     j
                where n is the unit vector of the direction along the which r 3 Fj is maximum, A Γ
                is the area of an open surface A that is bounded by the closed curve Γ, and v A is the
                volume bounded by A. The divergence is defined by

                                                        H
                                                          F ndA
                                                         Σ
                                           r  F 5 lim                                ðA1:7Þ
                                                  Δv Σ -0  Δv Σ
                   The special forms of the curl and divergence for interfaces between media with
                different material properties are as follows:
                               r S 3 F 5 n 12 3 F 2 2 F 1 Þ; r S   F 5 n 12   F 2 2 F 1 Þ;  ðA1:8Þ
                                                                   ð
                                             ð
                where (   ) 1,2 denote the two different media separated by the interface, and n 12 is
                the normal to the interface, oriented from (   ) 1 to (   ) 2 . If a superficial source
   49   50   51   52   53   54   55   56   57   58   59