Page 86 - Computational Fluid Dynamics for Engineers
P. 86

2.7  Boundary  Conditions                                              71



         stream  extending  to y =  — oo  (Fig.  2.5b),  in which  case  the  viscous  region is
         called  a  "free  shear  layer."  The  variation  of the  static  pressure p with  x  within
         the  shear  layer  depends  on the  shape  of the  solid  body,  and,  additionally  on  the
         displacement  effect  of the  shear  layer,  which  may  either  be  small  enough to be
         neglected  or  can  be  accounted  for,  for  example,  by the  interactive  procedures  to
         be  described  in Chapter  7.  In  this  case  there  are  three  boundary  conditions  for
         the  velocity  field  that  must  be  specified,  two at the  wall,  y =  0,  and  the  other
         at  the  boundary-layer  edge,  y = 6.  The  conditions at the  wall  are  usually  the
         specification  of normal  (v) and  tangential  (u)  components  of  velocity,  and  that
         at  the  edge  specifies  u as the  external  velocity  u e(x).  They  can  be  summarized
         as
                                y = 0,  u = 0,  v = v w(x)                (2.7.3a)

                                    2/-><5,  u = u e{x)                   (2.7.3b)
         where  6 is sufficiently  large  so that  du/dy  at the  boundary-layer  edge is small,
                      - 4
         say  around  10 .  The  transpiration  velocity,  v w(x),  may  be either  suction or
         injection.  On  a nonporous  surface  it is equal to zero.
            For  boundary-layer  flows,  the energy  equation,  being  second  order, re-
         quires  two  boundary  conditions,  one at the  wall, y = 0,  and  one at the  thermal
         boundary-layer  edge,  ^ , which  may  be smaller  or  larger than  the  hydrodynamic
         boundary-layer  thickness,  6.  For  a Prandtl  number,  Pr,  less than  unity,  6t >  <5,
         and  conversely  for Pr greater  than  unity,  6t < 6.  The  boundary  conditions at
         the  wall  may  correspond  to either  specified  wall  temperature,  T w(x),  or wall
         heat  flux,  q w(x),  and  with  the  requirement  at the  boundary-layer  edge,  they
         can  be  summarized as


                     = 0,  T = T w(x),  or   (?p\   =-q w(x)/k             (2.7.4a)
                   y
                                                                          (2.7.4b)
                                     y = S u  T = T e
         where  T e is constant  for an incompressible  flow  but is a  function  of x  for  a
         compressible  flow  with  pressure  gradient.
            In  free  shear  layers  (Figs.  2.5b and  c), the external  velocity  must  be  specified
         on both  edges. The  difficulties  associated  with  the v boundary  condition  in  free
         shear  layers  are  less  obvious. If the  flow is  symmetrical  (Fig.  2.5a),  no  problem
         arises:  the  initial  symmetrical  velocity  profile  is specified  and  v is required to
         be  zero  on the  centerline  and u = u e at one  edge.  If the  flow  is not  symmetrical
         (Fig.  2.5b),  a boundary  condition  for v  cannot  be found  from  consideration  of
         the  shear  layer  and  the  boundary-layer  equations  alone.  In the  real  flow the
         behavior  of v outside  the  shear  layer  depends  on  the  ^-component  equation  of
         motion  and the continuity  equation,  applied throughout  the  flow and  not  merely
         in the  shear  layer.  Of  course, a similar  problem  occurs  in  determining  u e  either
         in  the  boundary  layer  or wake  of Fig.  2.5a;  the  latter  problem  is both  more
   81   82   83   84   85   86   87   88   89   90   91