Page 80 - Basic Structured Grid Generation
P. 80

Differential geometry of surfaces in E 3  69

                        so also must det(b αβ ) be intrinsic. Moreover, by eqn (3.112) we obtain Gauss’s
                        result
                                                            R 1212
                                                       κ G =     ,                        (3.128)
                                                              a
                        which shows that the Gaussian curvature is also an intrinsic quantity, that is, it depends
                        only on the tensor components a αβ and their derivatives. Such a quantity is sometimes
                        referred to as a bending invariant, as it remains unchanged by any deformation of
                        the surface which involves pure bending without stretching, shrinking, or tearing, thus
                        preserving distances between points, angles between directions at a point, and the
                        coefficients of the first fundamental form and their derivatives.
                          Equation (3.124) may be written in the form

                                     ∂b αβ   µ        µ       ∂b αγ   µ        µ
                                               b
                                         −   αγ βµ −   βγ αµ =     −   b γµ −    b
                                                        b
                                     ∂u γ                     ∂u β    αβ       βγ αµ
                        by subtracting the same term from each side, which gives covariant derivative
                        identities
                                                      b αβ,γ = b αγ,β .                   (3.129)
                          This contains only two non-trivial results:

                                               b 11,2 = b 12,1  and b 21,2 = b 22,1 .     (3.130)
                          These are the Codazzi equations(or the Mainardi-Codazzi equations).

                        Exercise 13. Express the Codazzi equations in the form
                                          ∂L   ∂M       1       2     1       2
                                             −     = L  12  + M(  12  −   ) − N  11
                                                                      11
                                          ∂v   ∂u
                                         ∂M    ∂N       1       2     1       2
                                             −     = L  22  + M(  22  −   ) − N  .        (3.131)
                                                                             12
                                                                      12
                                          ∂v    ∂u
                        Exercise 14. Deduce from eqns (3.131) and (3.52) that when the co-ordinate curves
                        coincide with lines of curvature (so that F = M = 0), the Codazzi equations may be
                        expressed as
                                           ∂L   1 ∂E     L  N     1 ∂E
                                              =          +     =      (κ a + κ b )
                                           ∂v   2 ∂v   E   G      2 ∂v
                                          ∂N    1 ∂G    L  N      1 ∂G
                                              =          +     =      (κ a + κ b ).       (3.132)
                                           ∂u   2 ∂u   E   G      2 ∂u

                          To derive the Gauss equation in the form due to Brioschi, we start with eqns (3.97)
                        and
                                                                                    2
                                                                       2
                                          2
                                                       2
                                    1    ∂ r ∂r ∂r    ∂ r ∂r ∂r       ∂ r   ∂r ∂r
                                2
                        LN − M =            ,  ,         ,   ,    −       ,   ,
                                    a   ∂u 2  ∂u ∂v   ∂v 2  ∂u ∂v     ∂u∂v ∂u ∂v
                                                                                             

                                    1   r uu · r vv  r uu · r u  r uu · r v           r uv · r uv  r uv · r u  r uv · r v     

                                  =       r u · r vv  r u · r u  r u · r v     −     r u · r uv  r u · r u  r u · r v     ,

                                    a                                                        
                                          r v · r vv  r v · r u  r v · r v        r v · r uv  r v · r u  r v · r v
   75   76   77   78   79   80   81   82   83   84   85