Page 224 - Computational Fluid Dynamics for Engineers
P. 224

7.2  Standard,  Inverse  and  Interaction  Problems                   213












         C P



                                             Fig.  7.1. Pressure  distribution  around  a  cir-
                                             cular  cylinder  in  the  subcritical  and  super-
               0   60  120 180 240 300 360
                                             critical  range  of  Reynolds  numbers  qoo =
                                  8°         iA?v£.


         approximation  to the  actual  flow.  In the  case  of streamlined  bodies, where  sepa-
         ration  takes  place  close to  the  rear  stagnation  point,  the  solution  of the  inviscid
         flow  equations  can  serve  as  a  good  approximation  to  real  flow  as  discussed  in
         detail  in  [4]. For  completeness,  it  is  briefly  discussed  below.
            The  boundary-layer  equations  are  not  singular  at  separation  when  the  ex-
         ternal  velocity  or  pressure  is  computed  as  part  of  the  solution.  Catherall  and
         Mangier  [5] were the  first  to  show that  in two-dimensional  steady  laminar  flows,
         the  modification  of  the  external  velocity  distribution  near  the  region  of  flow
         separation  leads  to  solutions  free  of  numerical  difficulties.  Prescribing  the  dis-
         placement  thickness as a boundary  condition  at the boundary-layer  edge, that  is,

                                   y  =  (5,  8*(x)  — given               (7.2.3)

         in  addition  to  those  given  by  Eq.  (2.7.3)  with  no  mass  transfer,  they  were  able
        to  integrate  the  boundary-layer  equations  through  the  separation  location  and
         into  a  region  of  reverse  flow  without  any  evidence  of  singularity  at  the  sep-
        aration  point.  This  procedure  for  solving  the  boundary-layer  equations  for  a
        prescribed  displacement  thickness  distribution,  with  the  external  velocity  or
        pressure  computed  as part  of the  solution  is known  as the  inverse  problem.  This
        observation  of  Catherall  and  Mangier  has  led  to  other  studies  by  various  inves-
        tigators  of inverse  solutions  of the  boundary-layer  equations;  these  are  obtained
         by  prescribing  distributions  of  displacement  thickness  or  wall  shear.  Further-
         more,  it  has  been  demonstrated  in  [6] that  for  flows  with  separation  bubbles,
        these  solutions  are  in  good  agreement  with  the  solutions  of  the  Navier-Stokes
        equations.
            A  problem  associated  with  the  use  of  these  inverse  techniques  for  external
         flows  is  the  lack  of  a  priori  knowledge  of  the  required  displacement  thickness
        or  wall  shear.  The  appropriate  value  must  be  obtained  as  part  of  the  overall
        problem  from  interaction  between  the  boundary  layer  and  the  inviscid  flow.  In
   219   220   221   222   223   224   225   226   227   228   229