Page 63 - The engineering of chemical reactions
P. 63

Variable Density 47

                                             dca                 d x          d x     d x
                                                                             -   + x  -  x2
                                     kf  CA  -  4kb(CAo   -  cAj2 =  ( x  -x1)(x   -  x2)  = x-x1
                               where xi and  x2  are the roots of the polynomial k&A - kb(CA,,  - CJ2.  This giVeS
                                           CA
                                       t =            dC.4        -ji,:,+T&
                                           s kc.4   -  4kb(CAo   -  cAj2
                                          CAo                       CAo        CA”
                                         =  In  CA  --xl  - in  CA  -x2
                                             CAo - xl    CAo   -  x2

                                 It is obvious from this example why we quickly lose interest in solving mass-balance
                            equations when the kinetics become high order and reversible. However, for any single
                            reaction the mass-balance equation is always separable  and soluble as

                                                               CA
                                                                  dC.4
                                                         t=-     -
                                                              s  r(CA>
                                                              CAo

                            Since  CA =  CA,(~   -  X), we can also write this in terms of the fractional conversion X,








                            Since  r(C,J)  or r(X) is usually a polynomial in  CA  or X, one can solve this analytically
                            for  t  (CA) or t(X)  by factoring the denominator of the integrand and solving the resulting
                            partial fractions for CA(t)   or  X(t).  This can always be done, but the solution is frequently
                            neither simple nor instructive.
                                 One can always solve these problems  numerically  for particular values of  CAM  and
                            k,  and we will do this for many situations. However, it is still frequently desirable to
                            have analytical expressions for  CA(t)   because we can then solve this expression by just
                            substituting particular parameters into the equation. In any numerical solution one has to
                            solve the differential equation again for each set of parameters.
                                 Another reason for searching for analytical solutions is that we can only solve numer-
                            ically a problem that is well posed mathematically. We must program a valid mathematical
                            expression of the problem on the computer or the answers may be nonsense. The need for
                            proper descriptions of the equations, initial and boundary conditions, and stoichiometric
                            relations among the variables is the same whether one is interested in an analytical or a
                            numerical  solution.

            VARIABLE DENSITY
                            The previous examples are all simple problems to integrate for CA(~)   or at least for t (CA).
                            We assumed that the density was constant (the volume in the equations). This would be
   58   59   60   61   62   63   64   65   66   67   68