Page 97 - Computational Fluid Dynamics for Engineers
P. 97

3.2  Zero-Equation  Models                                             83



         3.2  Zero-Equation    Models

         The  zero-equation,  often  referred  to  as  algebraic  eddy  viscosity  and/or  mixing
         length  models,  are  used  to  model the  Reynolds  shear  stress term  in the  momen-
         tum  equations.  For  a  two-dimensional  incompressible  laminar  and  turbulent
         flow,  the  continuity  and  momentum  equations  (2.4.33)  and  (2.4.34),  respec-
         tively,  are  solved  subject  to  given  initial  conditions  and  boundary  conditions
         with  — u'v'  represented  by  Eq.  (3.1.2).  For  a  limited  range  of  flows,  it  is  possi-
         ble  to  specify  the  turbulent  viscosity  of  Eq.  (3.1.2)  as the  mixing  length  of  Eq.
         (3.1.1)  in  closed  algebraic  expressions.  Examples  include:
                £/6  =  0.09      for  a  plane  jet
                   =  0.075       for  a  round  jet
                   =  0.16        for  a  plane  wake
                   =  0Aly/6      for  the  near  wall  region  of  a  boundary  layer
                   =  0.09        for  the  outer  region  of  a  boundary  layer
        The  restricted  range  of  the  correlations  should  be  noted,  together  with  the
        simplicity  of  the  approach.  No  differential  equations  are  solved  for  turbulence
        quantities,  and  turbulence  transport  is,  consequently,  not  represented.


        3.2.1  Cebeci-Smith   Model

        For  external  flows, a  popular  zero-equation  model  is  the  Cebeci-Smith  (CS)
        model.  This  model  assumes that  a turbulent  boundary  layer  can  be  represented
        in  terms  of  eddy  viscosities  for  inner  and  outer  layers. The  corresponding  func-
        tions  are empirical  and  based  on  limited  ranges  of experimental  data;  the  range
        of data  is,  however,  extensive  so that  the  full  algebraic  formulation  is the  most
        general  available  at  present.
           In the  inner  and  outer  regions  of  a boundary  layer  on  a smooth  surface,  with
        or  without  mass  transfer,  the  eddy  viscosity  s m  is written  as  [2,3]
           Inner  region:  0  <  y  <  y c

                                               du
                                    \£m)i  — '     7tr                      (3.2.1)
                                               dy
        Here  the  mixing  length  /  is  given  by

                                  l  =  K,y\l-  exp - - | ) 1             (3.2.2a)
                                                 (
        where  K — 0.40  and  A  is  a damping-length  constant,  which  may  be  represented
        by
                                              1
               A  =  2 6 ^ " \  JV =  (l-11.8p+) /2  P +  =  ^ ^ r ,      (3.2.2b)
           Outer  region:  y c  <  y  < 6
   92   93   94   95   96   97   98   99   100   101   102