Page 265 - Introduction to Computational Fluid Dynamics
P. 265

P1: KsF/ICD
                                                                                                16:28
                                                                                   May 10, 2005
                                        0 521 85326 5
                           CB908/Date
            0521853265c08
                     244
                                                                       NUMERICAL GRID GENERATION
                               Identical expressions again emerge for P (ξ 1 ,ξ 2max ) and Q (ξ 1 ,ξ 2max ). One can
                            thus prescribe P and Q functions over the whole domain using Equations 8.39 and
                            8.40.
                            8.4.3 Discretisation
                            Equations 8.26 and 8.27 can be written in the following general form:
                                                  2
                                                                    2
                                         ∂       ∂         ∂       ∂            ∂    ∂
                                      P  ∗   − α    2  + Q ∗   − γ   2  =−2β             ,     (8.45)
                                         ∂ξ 1    ∂ξ 1      ∂ξ 2    ∂ξ 2        ∂ξ 1  ∂ξ 2
                                                                         2
                                                        2
                                               ∗
                                                                ∗
                            where   = x 1 , x 2 , P =−PJ , and Q =−QJ . Equation 8.45, being of the
                            conduction–convection type, can be discretised using the UDS to yield
                                        AP   P = AE   E + AW   W + AN   N + AS   S + S,        (8.46)
                            where
                                                          1
                                                                     ∗
                                                               ∗
                                               AE = α P +   (| P |− P ),
                                                                     P
                                                               P
                                                          2
                                                          1
                                                               ∗
                                                                     ∗
                                              AW = α P +    (| P |+ P ),
                                                                     P
                                                               P
                                                          2
                                                          1
                                                               ∗
                                                                     ∗
                                               AN = γ P +  (| Q |− Q ),
                                                                     P
                                                               P
                                                          2
                                                          1
                                                                     ∗
                                                               ∗
                                               AS = γ P +  (| Q |+ Q ),
                                                               P
                                                                     P
                                                          2
                                               AP = AE + AW + AN + AS,
                                                 S =−2β P (  ne −   nw −   se +   sw ).        (8.47)
                            Equation 8.46 can be solved using the ADI method.
                            8.4.4 Solution Procedure
                            Sorenson’s method can be implemented through the following steps.
                            Initialisation
                             1.Choose coordinates x 1 (ξ 1 , 0), x 2 (ξ 1 , 0), x 1 (ξ 1 ,ξ 2max ), and x 2 (ξ 1 ,ξ 2max ) on the
                               south and north boundaries, respectively. Also specify x 1 (0,ξ 2 ) (west) and
                               x 1 (ξ 1max ,ξ 2 ) (east).
                                      1
                             2.Specify  s 0 and  s max and θ 0 and θ max . For orthogonal intersection, θ = π/2.
                             3.Let P (ξ 1 ,ξ 2 ) = Q (ξ 1 ,ξ 2 ) = 0.

                            1  It will be appreciated that this liberty to specify  s 0 and  s max can be very useful when south
                              and north boundaries are walls and the HRE e–  turbulence model is employed. One can therefore
                              place the first node away from the wall in the range 30 < y < 100.
                                                                        +
   260   261   262   263   264   265   266   267   268   269   270