Page 341 - Computational Fluid Dynamics for Engineers
P. 341

11.4  Artificial  Compressibility  Method:  INS2D                     331



         11.4  Artificial  C o m p r e s s i b i l i t y  M e t h o d :  I N S 2 D

         The  artificial  compressibility  method  INS2D*,  widely  made  popular  by  Roger
         and  Kwak  [8-10],  is  described  in  this  section.  For  simplicity,  the  numerical
         method  is  discussed  with  the  equations  expressed  in  Cartesian  coordinates  for
         steady  flows  rather  than  in  transformed  form  (subsection  2.2.4)  for  steady  and
         unsteady  flows.  In this  case, the  governing  equations  are  given  in Problem  2.17,

                                  £ £
                             +
                                           +
                                                 F F
                            f |( - "» |< - »'=              O            (P2.17.2)
         where
                    \p~              "  0u  '              "  0V  '
               D  =  u    E  =  Ei  =  u 2  +p  F  =  E2  =  uv          (P2.17.2)
                     V  _               uv                 _v 2  +p_

                                   —                                     (P2.17.3)
                                E v           F v
                                                      'xy
                                        u
                                      L  xy
            The  viscous  normal  and  shear  stresses  are  given  by  Eq.  (2.2.£
                         du                 ' du   dv          ~  9v
                &x                  >yx  ~  ^  [  —  +  ~^Z  =  2/i—       (2.2.8)
                                            ^dy   dxj             dy

                                  t
         11.4.1  Discretization  of he  Artificial  Time  Derivatives
         INS2D  uses  an  implicit  scheme  to  solve  Eq.  (P2.17.2)  subject  to  the  boundary
         conditions  to  be  described  in  subsection  11.5.1,  and  the  equation  is written  as

                                       dD_
                                           +  i? =  0                     (11.4.1)
                                        dr
         where
                                                 d_
                              R=^(E-E V)            (F  -  F v)           (11.4.2)
                                                 dy
            Application  of  a  first  order  backward  Euler  formulas  to  Eq.  (11.4.1)  in
         pseudo-time  r  yields
                                   r)n+l  _  j-\n
                                                 -R n+1                    (11.4.3)
                                    r n+l  _  q-n
         where  the  superscript  n  denotes  the  pseudo-time  iteration  count.  Application
         of  Taylor  series  expansion,  and  noting  that  R  is  a  function  of  D,  leads  to
                                                            n
                            Rn+1  =Rn +  f  | ^ V  (D n+1  -  D )          (11.4.4)

           INS2D  and  related  software  can  be  obtained  from  NASA  Ames  by  signing  a  non-
           disclosure  agreement.  For  details,  see  http://people.nasa.goc/~rogers/home.html
   336   337   338   339   340   341   342   343   344   345   346