Hazardous events in membrane bioreactors – Part 3: Impacts on microorganism log removal efficiencies

• Impact of hazardous events on LRV of microbial indicators was quantified.
• Monte Carlo simulation was used to benchmark impact against normal variability.
• Chemical shock reduced activated sludge removal but fouling mitigated breakthrough.
• Overall removal of virus indicators was impacted by changes in solution chemistry.
• Physical defects on the membrane were shielded by the activated sludge flocs.

Under normal operation, membrane bioreactors (MBRs) produce high quality effluent leading to common application in water recycling schemes. When hazardous events, outside of normal operation, occur strategies must be in place to ensure that microbially safe water is consistently produced. Limited literature is available to quantify the consequence of hazardous events on microorganism log removal value (LRV) by MBRs. As a result, current MBR risk management practices cannot be based on sufficient quantitative evidence. In this study, MBRs were subjected to hazardous events designed to reflect a range of potential severe feed water variations and process failures. During these challenge events, removal of four microbial indicator organisms representing viruses, bacteria and protozoa were quantified. Hazardous event impacts were benchmarked against the 5th percentile of normal process LRV, determined via probabilistic techniques. Severe feed water variation decreased the efficacy of bio-predation due to adversely affected activated sludge performance. However, overall LRV was preserved due to the onset of fouling, aiding rejection by the membrane. Overall FRNA bacteriophage LRV significantly decreased to 0.3 log units below the 5th percentile (LRV=4.1) after the NaCl shock, most likely due to inhibition of adsorption mechanisms. Fibre breakage resulted in a significant reduction in LRV, but was mitigated by suspended solids plugging the membrane lumen within 15 min. The findings presented here can inform risk management strategies for MBRs by permitting control strategy prioritisation according to quantified event consequences.