Water Journal : Water Journal September 2012-1
membrane technology refereed paper technical features 72 SEPTEMBER 2012 water Following removal of the strainers, a significant build-up of fibrous material was observed on the membrane modules (Figure 10). Following concerns raised by the membrane suppliers about the risk of premature membrane failure, maintenance was performed to remove this accumulated material. One of the two membrane trains was taken offline for cleaning while the other train remained in service. Prior to manual cleaning the membranes were soaked in a sodium hypochlorite solution for 24 hours. The manual component involved lifting out the membrane cassettes and removing extraneous material caught in the membrane fibres by hand. After the membranes were cleaned they were bubble tested and any damaged fibres were plugged before re-installation. This operation was both time consuming and labour intensive, with four people required over a week-long operation to clean and repair the membranes on only one of the two membrane trains. This issue of membranes becoming fouled with fibrous matter was also observed on another MBR process commissioned by the Alliance three months after the Paxton upgrade. However, at the time of writing the extent of the material build-up at the other plant had not been clearly identified. The air blowers for the upgraded plant were sized for the ultimate load at design capacity (at year 2030). This factor, combined with the intensive aeration needed for membrane air scour, resulted in excessive aeration and higher total oxidised nitrogen concentrations, making it difficult to comply with the low-effluent total nitrogen concentration limits. In an attempt to remedy the over- aeration problem, both the dissolved oxygen set point and the recirculation rates were reduced until effluent nitrate levels dropped to an acceptable range. Daily sampling was conducted to confirm that the nitrate spikes had been effectively reduced. The feasibility of reducing the pulley ratio on the air blowers to reduce the blower output will also be examined. The installed control philosophy did not include capability to automatically stop the blowers during periods of low oxygen demand. Changes will be implemented to allow this capability. Control changes have also been proposed to allow recirculation rates to be increased when operating without aeration to improve denitrification. It will also be important to ensure that the control philosophy provides operational flexibility for situations such as breakdown of critical equipment and for process optimisation. The residual chemical (sodium hypochlorite) from recovery cleaning was found to adversely affect the biological process. Effluent ammonia spikes were observed the day after each recovery clean was performed. This residual chemical was designed to be returned to the biological process; however, it was concluded that this resulted in a loss of nitrification. When this occurred the wasting rate was temporarily reduced to give the nitrifying biota time to re-establish, which took about one sludge age to occur. The control sequence for the recovery clean was subsequently changed to divert the residual hypochlorite to the aerobic digester. Figure 10. Membrane condition at installation (left) and after six months of operation (right). Figure 11. Membrane bubble testing (top) and repair (bottom). Figure 12. Effluent ammonia concentrations before and after a membrane recovery clean (timing of recovery clean shown by red line). Figure 13. Drive cog on drum screen motor. Figure 14. Drum screen arrangement.
Water Journal November 2012-1
Water Journal August 2012