Page 295 - Process Modelling and Simulation With Finite Element Methods
P. 295

282         Process Modelling and Simulation with Finite Element Method5

             In  solid  particle  dynamics,  the  population  balance  equations  also  are
         expressible  in  terms  of  some  convolution  integrals.  Nicmanis  and  Hounslow
         [27] used FEM to describe an integro-differential equation with convolution-type
         integral  terms,  where  the  major  processes  of  this  type  are  aggregation  and
         breakage.  The collision rules for bubbles  and particles  depend  substantially on
         the physicochemical  properties  of  the liquid medium.  Traces of  flocculent and
         coagulent  effect  the  probability  of  bubble-particle  agglomeration  and  floc
         formation.  However,  in  terms  of particle  dynamics,  the  collision  rules  can be
         formulated to match observed kinetic rates.  Thus, a semi-empirical approach to
         population  balance  equations,  fitting  the  coefficients  of  the  aggregation,
         breakage, and growth models, is a successful technique in characterizing particle
         processes.  Randolph  and Larson  [28] cite the change in number density n(v) of
         particles with volume n in the product stream of a continuous mixed-suspension,
         mixed-product removal crystallizer in which these three processes are occurring
         from an inlet stream with feed population ni,(v):

                     ('  ) - nin  ('   + d (v ) ~1 (v )) = b (v ) - d (v )   (7 27)
                                      (G
                         z         dv
         where  T is  the  residence  time  in  the  crystallizer,  G(v)  is  a  volume-dependent
         growth  function  and  the  number  density  of  nuclei  is  incorporated  into  the
         equation  as  a  boundary  condition  n(O)=no.  b(v)  and  d(v)  are  suggestively
         denoted  as  the  birth  and  death  terms  for  the  volume  fraction  of  size  v.  In
         general,  there  are  contributions  to  both  terms  from  both  aggregation  and
         breakage.  Case 1 considered by  [27] is a purely aggregation model, so we will
         cite only the forms derived by Hulburt and Katz [29] for aggregation.

                              %
                       b(v) =  p (v - w, w)n(v - w)n( w)dw
                                   -
                              U                                      (7.28)
                        d(v) = n(v)p  (v,w)n(w)dw
                                   0
         Succinctly, birth by aggregation is due to the probability  of combining particles
         with  volumes  which  sum to v (and  sticking).  Death  is by  the probability  of  a
         particle of volume v participating in a collision (and sticking).
   290   291   292   293   294   295   296   297   298   299   300