Theoretical and experimental flux maximization by optimization of backpulsing

Abstract

A model of rapid backpulsing for fouling control in cross-flow microfiltration is developed for solid–liquid separations where only a portion of the foulant is reversibly deposited on the membrane. It predicts the duration of each backpulse and time between backpulses which maximize the net flux. The foulant removed during each backpulse of liquid flow in the reverse direction through the membrane (i.e. from the permeate side to the feed side) is modeled as an exponential rise, with time constant τ b and maximum fraction of foulant removed β max which must be determined experimentally for each system. The net flux increases with decreasing τ b and increasing β max due to more effective cleaning by short backpulses. Experiments to test the model were performed using washed bacterial cells and a flat-sheet membrane module. From experiments with a long period of forward filtration followed by individual pulses of reverse flow, the proposed foulant-removal model provides a good fit to the data, with τ b = 0.1–0.2 s and β max = 0.4–0.6 for reverse trans-membrane pressures of 5 psi (35 kPa) or greater. For rapid-backpulsing experiments with repeated cycles of forward and reverse flow, a similar value of τ b was obtained. However, a smaller value of β max = 0.1–0.2 provides the best fit of the rapid-backpulsing data, indicating greater irreversible membrane fouling. Nevertheless, the net flux with backpulsing for the bacterial system studied is as much as six-fold greater than the long-term flux without backpulsing.