Page 262 - Fundamentals of The Finite Element Method for Heat and Fluid Flow
P. 262

CONVECTION IN POROUS MEDIA
                        254
                        8.4 Non-isothermal Flows
                        Several examples of porous medium flow problems are non-isothermal in nature. The main
                        focus in this case will be to demonstrate non-isothermal flow through a porous medium. As
                        mentioned previously, an energy equation needs to be solved, in addition to the momentum
                        and pressure equations if the flow is non-isothermal. For steady state problems, if no
                        coupling exists between the momentum and energy equation, the temperature field can be
                        established after calculation of the velocity fields. The temporal discretization of the energy
                        equation can be written in a similar form to the momentum equation and is given as
                                                                           2    n+θ 2
                                         n+1
                                                n
                                        T    − T         ∂T  n+θ 1   k  ∗  ∂ T


                                      σ           =− u i         +                          (8.66)
                                            t            ∂x i      ReP r  ∂x 2
                                                                            i
                        where θ 1 and θ 2 have the same meaning as previously discussed in Section 8.3. The variable
                        involved in this case is temperature and can be spatially approximated as
                                                        T = [N]{T}                          (8.67)
                        The weak form of the energy equation can be written (assuming θ 1 and θ 2 are both equal
                        to zero) as
                                                        ∂T         k  t     ∂[N] ∂T
                                                             n      ∗           T     n
                                                     T
                                 T
                             σ[N]  T d
 =− t      [N] u i    d
 −                      d
 + b.t.
                           
                    
       ∂x i       ReP r  
  ∂x i ∂x i
                                                                                            (8.68)
                        where
                                                      T = T  n+1  − T  n                    (8.69)
                        The substitution of Equation 8.67 into Equation 8.68 yields the final global matrix form of
                        the energy equation, that is,
                                                                                 n
                                        σ[M p ]{ T}=− t [C p ]{T}+ [K T ]{T}− {f 4 }        (8.70)
                        where the elemental matrices are
                                                 b   b i b j b i b k        c  c i c j c i c k
                                                 2                       2          
                                         k  ∗     i    2            k  ∗    i    2
                               [K Te ] =        b j b i  b j  b j b k   +   c j c i  c j  c j c k    (8.71)
                                       4AReP r             2     4AReP r              2
                                                b k b i b k b j  b         c k c i c k c j  c
                                                           k                          k
                        and the forcing vector is
                                                                          n
                                                        
                                                         b i T i + b j T j + b k T k
                                                	   1
                                          {f 4 }=        b i T i + b j T j + b k T k    n 1
                                                4A ReP r
                                                                 0
                                                                            n
                                                         
                                                  	   1    c i T i + c j T j + c k T k
                                               +          c i T i + c j T j + c k T k    n 2  (8.72)
                                                 4A ReP r
                                                                  0
                        It should be noted that both the flux and convective heat transfer boundary conditions are
                        treated by using the boundary integral, as discussed in the previous chapter. At higher
                        Reynolds numbers convection stabilization of Equation 8.70 is essential. This can be
                        achieved by introducing characteristic Galerkin method (Chapter 7).
   257   258   259   260   261   262   263   264   265   266   267