Page 90 - Introduction to Computational Fluid Dynamics
P. 90

P2: IWV
            P1: KsF/ICD/GKJ
                                                                                   May 25, 2005
                                        0 521 85326 5
                           CB908/Date
            0521853265c03
                        EXERCISES
                         6. Consider the steady 1D conduction–convection problem discussed in this chap- 10:54  69
                            ter. Assume a nonuniform grid (i.e.,  x e  =  x w ). Hence, derive expressions for
                            AE, AW, and AP using the power-law scheme. If   E = 1 and   W = 0, calcu-
                                         = u  x e /α =−10, −5, −1, 0, 1, 5, and 10 when  x e / x w =
                            late   P for P c e
                                                                                           ).]
                            1.2. [Hint: Start with Equation 3.20 with ψ e = F(P c e  ) and ψ w = F(P c w
                         7. Show that if f in Equation 3.40 equals its associated ξ, Equations 3.38 and 3.39
                            will yield the UDS formula. Hence, derive Equations 3.42 and 3.43 and the
                            expression for the S TVD term in Equation 3.44.
                         8. Use   E = 1 and   W = 0 and determine the variation of   P with P c for the
                            TVD scheme when   EE = 5 and   WW =−0.1. Assume −200 < P c < 200
                            and use the Lin–Lin and HLPA schemes. Assume a uniform grid. Compare
                            your results with those given in Table 3.1 and comment on the result. (Hint:
                            Iterations are required.)

                         9. Show that for a general differencing scheme, the false conductivity is given
                            by k false = ρ C p u  x (ψ − 0.5), where ψ is defined by Equation 3.20. Hence,
                            compare k false for UDS and HDS and comment on the result. Assume a uniform
                            grid.
                        10. Runchal [61] developed a controlled numerical diffusion with internal feed-
                            back (CONDIF) scheme capable of sensing the shape of the local   profile.
                            According to this scheme, AE and AW in Equation 3.14 are given by
                                            1     |P c |− P c                        |P c |+ P c

                             AE = 1 + 1 +                   ,    AW = 1 + (1 + R)              ,
                                            R        4                                  4
                            where

                                                   ∂ /∂ X | e  (  E −   P ) X w
                                              R =           =                 .
                                                  ∂ /∂ X | w   (  P −   W ) X e
                            Further, the values of R are constrained as follows: If R < 1/R max then R =
                            1/R max ;if R > R max then R = R max . Typical values assigned to R max vary
                            between 4 and 10. Assuming a uniform grid, show that

                            (a) If R = 1, the CONDIF scheme is the same as the UDS.
                            (b) CONDIF represents both convection and diffusion terms to second-order
                                accuracy irrespective of the sign and the magnitude of the Peclet number.
                            (c) Taking   W = 0 and   E = 1, compare values of   P for | P c | < 20 with the
                                exact solution given in Table 3.1. Carry out this comparison for R max = 4
                                and 10.

                        11. Derive Equations 3.55, 3.57, 3.63, and 3.64.
                        12. Starting with Equation 3.59, show the correctness of Equations 3.60.
                        13. Verify that T = T 0 exp(−τ)sin(X) is an exact solution to the unsteady heat
                                                               2
                                                         2
                            conduction equation ∂T /∂τ = ∂ T /∂ X .
   85   86   87   88   89   90   91   92   93   94   95