Page 211 - Introduction to Computational Fluid Dynamics
P. 211

P1: IWV
                           CB908/Date
                                        0 521 85326 5
            0521853265c06
                     190
                                                                2D CONVECTION – COMPLEX DOMAINS
                            6.3.11 Pressure-Correction Equation                    May 25, 2005  11:10
                            In Chapter 5, the total pressure-correction equation in Cartesian coordinates was
                            derived to read as
                                                                   
   l


                                                  ∂     p   ∂p    ∂ ρ u i   ∂ρ
                                                       
       =         +    ,               (6.128)
                                                        i
                                                 ∂x i     ∂x i      ∂x i    ∂t
                            where
                                                               ρα  V
                                                           p
                                                         
 =          .                       (6.129)
                                                           i
                                                                 AP  u i
                                               p

                            In this definition of 
 , α and AP are, respectively, the underrelaxation factor and
                                                         u i
                            the AP coefficient used in the momentum equations. Invoking the Gauss theorem
                            again, the discretised version of Equation 6.128 will read as
                                             NK            NK                       NK
                                                                               V
                                                                    
      o             p

                                    AP p =      AE k p Ek  −  C ck − ρ P − ρ P    +    D ,    (6.130)
                                         P
                                                                                         k
                                             k=1           k=1                 t    k=1
                                           NK
                            where AP =        AE k and
                                           k=1
                                                                     p
                                                              p     (
 A) ck
                                                     AE k = d   =         .                   (6.131)
                                                             ck
                                                                    l P 2 E 2
                               Two comments are now important:
                                     p
                            1. The D term in Equation 6.130 will contain Cartesian gradients of p . However,

                                     k
                               duringiterativecalculation,sincethepressure-correctionequationistreatedonly
                                                    p
                               as an estimator of p , D is set to zero.

                                                    k
                                             p
                            2. Evaluation of 
 ck  in Equation 6.131 will require evaluation of  V and AP  u i
                               at the cell face (see Equation 6.129). The evaluation of cell-face volume can
                               be accomplished via a fresh construction at the cell face as shown in Fig-
                               ure 6.13. The construction involves drawing lines parallel to ab passing through
                               P 2 and E 2 . Then, two lines parallel to normal  n (and, hence, parallel to line
                               P 2 E 2 ) are drawn through a and b. The resulting rectangle c 1 –c 2 –c 3 –c 4 will have
                               volume
                                                                             .             (6.132)
                                                V ck = l ab × l P 2 E 2  × 1 = A ck l P 2 E 2
                               Using this equation therefore gives
                                                                   2
                                                             α (ρ A ) ck
                                                      AE k =       u  ,                    (6.133)
                                                                AP ck
                                        u
                               where AP = AP    u 1  = AP u 2  can be evaluated from formula (6.66). 6
                                        ck      ck     ck
                            6  Alternatively, one may evaluate AP exactly by carrying out a structured-grid-like discretisation
                                                        u
                                                        ck
                              over the control volume c 1 –c 2 –c 3 –c 4 . This is left as an exercise.
   206   207   208   209   210   211   212   213   214   215   216