Page 108 - Introduction to Computational Fluid Dynamics
P. 108

P1: IWV
                           CB908/Date
                                        0 521 85326 5
            0521853265c04
                                                                                   May 25, 2005
                        4.8 TREATMENT OF TURBULENT FLOWS
                                                                                           ˙    11:7  87
                        u ∂p/∂x represents the pressure–work effect in steady flow, and Q md =

                        ∂/∂y {(  all k  ρ D k ∂ω k /∂y)h k } represents the contribution of species diffusion
                        mass transfer having specific enthalpy h k .If h k equals mixture enthalpy h then
                         ˙
                         Q md = 0.
                           When no chemical reaction is present, R k = 0. However, for a reacting boundary
                        layer, R k will be finite for each species because each may be generated via some
                        reactions and destroyed via some other reactions among the postulated chemical
                        reactions. Very often, for gaseous fuels and for highly volatile solid/liquid fuels,
                        an SCR can be assumed [73]. The SCR is specified as
                                  1kg of fuel + R st kg of oxidant → (1 + R st )kg of product,  (4.86)
                        where R st is the stoichiometric ratio for the fuel under consideration. Thus, there
                        are three species and one must specify R fu , R ox , and R pr . However, in an SCR,
                         R fu = R ox /R st =−R pr /(1 + R st ) so that no net mass is generated or destroyed as a
                        result of chemical reaction. This enables construction of a conserved scalar variable
                        	 = ω fu − ω ox /R st = ω fu + ω pr /(1 + R st ) when mass diffusivities of all species
                        are taken equal. Thus, one may now solve only for ω fu and 	 with R 	 = 0 instead
                                                 ˙
                        of three variables. Further, Q cr =|R fu | H c where  H c is the heat of combustion
                        of the fuel. The value of R fu is obtained from a reaction rate law


                                                                   E
                                                                             n
                                                                          m
                                         R fu = R fu,kin =− A exp −      ω ω ,             (4.87)
                                                                             ox
                                                                          fu
                                                                  R u T
                        where, preexponential constant A and constants E, m, and n are specified for the
                        fuel [82] and R u is the universal gas constant.
                           If turbulent reacting flow is considered then the effective R fu is given by a variant
                        [44] of the eddy-breakup model due to Spalding [74],
                                                             ω ox     ω prod
                                   R fu =−ρ m  min A ω fu , A   , A         , R fu,kin ,   (4.88)
                                              e              R st   (1 + R st )

                        where A = 4 and A = 2. The postulated arguments in favour of this expression
                        are beyond the scope of this book.



                        4.8 Treatment of Turbulent Flows
                        In turbulent flows, 
   in Table 4.1 will assume an effective value. Thus, following
                        Equation 4.63, we have
                                                           µ      µ t
                                                  
  ,eff =   +       ,                    (4.89)
                                                          Pr     Pr t,
                        where suffix t denotes the turbulent contribution. The task now is to represent µ t and
                         Pr t,  via modelled expressions. This exercise, called turbulence modelling, implies
                        validity of the Boussinesq approximation for turbulent viscosity. Although there
   103   104   105   106   107   108   109   110   111   112   113