Application of backflushing for fouling reduction during microfiltration of yeast suspensions

The results of microfiltration studies using inorganic TiO2/Al2O3 membranes for solutions containing yeast are reported. The aim of the research was experimental as well as mathematical analysis of the permeate flux dependence on time for cross-flow filtration combined with backflushing using permeate. The effect of operating parameters such as cross-flow velocity, pressure difference across the membrane, and filtration time on permeate flux was experimentally investigated. The transmembrane pressure was changed in the range of 120 Á 320 kPa. The feed was pumped to the membrane module at a cross-flow velocity in the range of 2.73 Á 4.55 m/s. Microfiltration processes were performed at constant temperature (208C) using feed with constant content of yeast, i.e. 510 mg/L. The backflushing runs were performed at constant operating parameters, DPR 0/150 kPa; tR 0/60 s with frequency every 10 min. Experimental results were used for analysis in the frame of mathematical equation obtained using semiempirical models of the filtration and steady-state fluxes.

Concentration polarization and fouling phenomena in microfiltration of yeast suspension using ceramic 0.8 mm membrane can be Table 3 The data for calculation of JNET TMP (Pa) 1.15)/105 2.05)/105 3.03)/105 1.25)/105 2.20)/105 3.05)/105 1.20)/105 2.20)/105 3.20)/105 gw (1/s) K (s) n a b 92,815 Re 216.5 150.6 83.6 226.1 191.8 124.7 240.8 192.4 147.5 1.37 0.75 1.23 1.35 1.08 1.10 1.42 0.94 1.11 0.562 0.557 0.325 0.562 0.557 0.325 0.562 0.557 0.325 2.09)/10 (8 2.70)/10 (8 4.95)/10 (7 2.06)/10 (8 1.76)/10 (8 2.07)/10 (7 2.06)/10 (8 1.76)/10 (8 2.07)/10 (7 153,553 226,909 11,857 controlled by backflushing with duration of 60 s and frequency of 600 s. According to expectations, experimental results showed that steadystate permeate fluxes were higher for backflushed systems, compared to non-backflushed ones. Reduction of initial fluxes varied with transmembrane pressure and cross-flow velocity in the range of 45 Á 73% and 37 Á 67% for nonbackflushed and backflushed systems, respectively. The qualitative analysis of the backflushing efficiency for investigated system is to be done by parameters B and RI. An increase of steadystate flux due to backflushing, B , was found in the range of 25 Á 55%. The values of resistance index, RI, for all experiments with backflushing are less than 1.0 and fluctuate in the range of 0.54 Á 0.76; it means that used backflushing technique with constant parameters, DPR 0 150 kPa; tR 060 s and tF 0600 s, reduced partially membrane fouling during microfiltration of yeast suspension. The backflushing is known to remove surface deposits very efficiently but is less effective in removing of internal fouling. The lowest value of RI 00.54 was obtained for low transmembrane pressure (1.15 )105 Pa) and cross-flow velocity (2.73 m/s). It results in decreasing of operating costs and is interesting from a practical point of view. Finally, experimental flux data were analyzed in the frame of mathematical model derived using semiempirical models of the filtration and steady-state fluxes. The net permeate fluxes were well predicted by applied model. / / / / / TMP u yeast concentration in the feed (wt.%) diameter of membrane (m) filtration, feed permeate flux (m3/m2s) initial flux (m3/m2s) flux after backflushing (m3/m2s) pseudo steady-state flux (m3/m2s) steady-state flux (m3/m2s) reverse flux (m3/m2s) net permeate flux (m3/m2s) regression constant, Eq. (2) regression constant, Eq. (2) permeate transmembrane pressure during backflushing runs (Pa) transmembrane pressure during filtration runs (Pa) retentate resistance with backflushing (1/m) resistance without backflushing (1/m) resistance with clean water (1/m) resistance index forward filtration time or backflushing frequency (s) reverse filtration time or backflushing duration (s) transmembrane pressure (Pa) cross-flow velocity (m/s) Greeks gw l m mw r tw wall shear rate (1/s) friction factor feed viscosity (Pas) water viscosity (Pas) feed density (kg/m3) shear stress (Pa) C0 de F JF J0 J0BF Jps Js JR JNET K n P DPR DPF R RF R0 RM RI tF tR Nomenclature a A b B BF regression constant, Eq. (3) reduction of initial flux (%), Eq. (14) regression constant, Eq. (3) increase of steady-state flux (%), Eq. (15) backflushing