Page 72 - Membranes for Industrial Wastewater Recovery and Re-Use
P. 72

52  Membranesfor Industria2 Wastewater Recovery and Re-use

          after allowance was made for differing mean particle size of  the cake layer and
          the  bulk  suspension,  which  required  experimental  verification.  A  similar
          limitation appears to apply to the Song model (Song and Elimelech, 1995): an
          excellent fit with theoretically predicted flux was obtained for CFMF of 0.06 pm
          homodispersed colloidal silica once the specific cake resistance R’,  and the cake
          concentration C,  had been surmised (Wang and Song, 1999).
            It  is  generally  the  case  that  the  modified  concentration  polarisation
          models  developed  to  define  dynamic  and  steady-state  behaviour  during
          microfiltration  are  very  sensitive  to  key  parameters  pertaining  to  particle
          transport and the hydraulic resistance of the cake. These parameters are either
          only  calculable  for  highly  idealised  systems  or  else  must  be  determined
          empirically.  For  real  feedwaters,  which  may  be  physically,  chemically  and
          biologically complex, and the useful employment of these models for predictive
          purposes  is  probably  restricted  to  defining  trends.  Even  then,  there  are
          pronounced differences  in  reported  data. For  example  for  a  single matrix  of
          bovine blood, flux dependency on cross-flow appears to vary from U0.47 to direct
          proportionality  (Cheryan, 1998). Moreover, there appears  to be  a  paucity  of
          mechanistic  information  on  fouling  and  dynamic  behaviour  of  key
          heterogeneous matrices,  such  as  natural  organic  matter  (NOM)  in  surface
           waters or biomass in membrane bioreactor liquors, that could be used to form the
           basis of a predictive model.



           2.4 Process design and operation
           Key elements of process design that contribute to costs are:

             e  staging,
             e  power consumption,
             0  fouling and pre-treatment, and
             e  backflushing and cleaning.

             Staging is of critical importance in large-scale dense membrane processes, i.e.
           reverse  osmosis,  nanofiltration  and  electrodialysis.  Fouling  is  ubiquitous
           throughout the entire gamut of  membrane technologies,  but pretreatment to
           suppress or ameliorate fouling is only routinely  practised  in dense membrane
           processes. Backwashing is always carried out when the module design permits
           this, and cleaning is an essential part ofmembrane plant operation.


           2.4.1 Staging
           For most dense membrane processes the conversion of feed into product is limited
           either by  the membrane area or the rate of  extraction  attainable by  passage
           through a  single module.  It is for this reason that most  reverse  osmosis  and
           electrodialysis  technologies  employ  staging, the  use  of  sequential  stages  to
           produce more product than that attainable by a single passage.
   67   68   69   70   71   72   73   74   75   76   77