Page 114 - Computational Fluid Dynamics for Engineers
P. 114

100               4.  Numerical  Methods  for  Model  Parabolic  and  Elliptic  Equations


         4.4  Finite-Difference   Methods    for  Parabolic   Equations

         The  model  equation  used  here  to  illustrate  numerical  methods  for  solving
         parabolic  partial-differential  equations  is  the  one-dimensional  unsteady  heat
         conduction  equation  given  by Eq.  (4.2.4), which  also serves  as  a model  equation
         for  the  boundary-layer  equations,  to  be  discussed  in  detail  in  Chapter  7.
            Equation  (4.2.4)  requires  boundary  and  initial  conditions.  For  simplicity,  as
         shown  in  Fig. 4.3, assume that  the  boundary  conditions  at  x  — 0 and  x  — L  are
         given  by
                          x  =  0,  T  =  Ti(t);  x  =  L,  T  =  T 2(t)    (4.4.1)
         and  the  initial  conditions  by

                                    t  =  0,  T  =  T 0(x)                 (4.4.2)

            The  solution  of  Eq.  (4.2.4)  may  be  obtained  by  using  either  an  explicit  or
         an  implicit  method.  In  an  explicit  method,  the  value  of  T  at  the  next  time
                                                                       n
         step,  £ n + 1 ,  is  expressed  in  terms  of  T  at  the  previous  time  step,  t ,  and  the
         corresponding  equation  is  solved  explicitly  at  each  grid  point.  In  an  implicit
        method,  T  at  the  next  time  step  is expressed  in  terms  of  its  neighboring  points
            n + 1                           n
         at  £  and  the  known  quantities  at  t ',  and  its  solution  for  all  grid  points  on
                       n + 1
        the  time  step,  £  ,  is obtained  simultaneously.
        4.4.1  Explicit  Methods

        In  the  solution  of  Eq.  (4.2.4)  by  an  explicit  method,  dT/dt  may  be  represented
                                                          2
        by  the  forward  difference  formula,  Eq.  (4.3.8),  and  d T/dx 2  by  Eq.  (4.3.10),
        centering  at  the  net  point  (t n,Xi)  (see  Fig.  4.4), that  is,









                                           T=T 2 (t)
         T=T 1(t) (




                                                     Fig.  4.3. Initial and boundary  con-
                                                    ditions  of  Eq.  (4.2.4)  in  the  (x,t)
                                                    plane.  Symbols  x  denote  values
                                                    known  from  initial  conditions  and
         t=0  1                                     symbols  •  denote  values  known
            x=0         T=T 0(x)        x=L
                                                    from  boundary  conditions.
   109   110   111   112   113   114   115   116   117   118   119