Fouling mechanisms of an end-free submerged membrane (Yonsei End Free;YEF) module under different filtration modes

This study investigated effects of the concentration of ball clay suspensions and different filtration modes on the fouling of a submerged module (YEF; Yonsei End Free). In a crossflow filtration mode, short filtration time could reduce the membrane fouling at the one-minute stop duration. In order to see the effect of the stop duration on the increase in transmembrane pressure (TMP), the stop duration was increased to 0.75, 1.5 and 3 minutes while the suction time was set at 15 minutes in a dead-end filtration mode. The overall TMP increasing rate of 1.5-minute duration mode was twice less than that of 0.75-minute stop duration. Whereas, the TMP increasing rates of 1.5-minute stop duration and 3-minute stop duration were quite similar. Thus, the stop duration of 1.5 minutes was chosen for further experiments. To account for different individual fouling mechanisms, three fouling models were tested. The blocking model could fit the experimental data well in whole cycle period for both crossflow and dead-end filtration modes. The gel resistance of the crossflow filtration was 1.7 times higher than that of dead-end filtration suggesting that the gel formed on the surface of membrane tends to be taken off along with cake when the aeration restarted after the pause. On the other hand, the cake resistance of dead-end filtration was 5.9 times higher than that of crossflow filtration. Irreversible fouling was mainly caused by blockage of fine fraction of the ball clay particles.

The different operation conditions affected the filtration characteristics of a ball clay suspension. The fouling models were evaluated based on pilot-scale experiments with YEF module in this study. The increasing rate of TMP of 30-minute filtration was 3 times less than that of the 60-minute filtration with the crossflow mode. Shortening filtration time could improve the reduction of the membrane fouling at the 1-minute stop duration. It was obvious that the increasing rate of TMP depended not only on stop period but also the filtration time. Under various stop duration conditions (0.75, 1.5 and 3 minutes) increasing patterns of TMP were investigated in dead-end filtration mode. The TMP increasing rate of 1.5-minute stop period was 2 times less than that of 0.75-minute stop period. But, in stop periods of 1.5 and 3 minutes the increasing rate of TMP was quite similar. Thus, from the economic point of view, the stop period of 1.5 minutes was favorable. The blocking model could fit the experimental data well in whole cycle period for the crossflow and the dead-end filtration modes. In the crossflow mode, the fouling model transition from the complete blocking to the standard blocking was observed as the filtration time decreased. On the other hand, in the dead-end filtration, the fouling model transition from the standard blocking to the complete blocking was observed as the stop period decreased. The gel resistance of the crossflow filtration was 1.7 times higher than that of the dead-end filtration. Further, the cake resistance of the dead-end filtration was 5.9 times higher than that of crossflow filtration. Irreversible fouling was caused mainly by the blockage of fine particle fraction of the ball clay. The filtration of ball clay suspension gave an irreversible fouling at the surface of the membrane where fine particles appeared to block the membrane pores. Therefore, the validity of the result is probably limited to the cases of filtration of inorganic suspensions of fine particles. To confirm this, further experiments using large particles alone and binary particles are needed.