Process robustness and impacts of hazardous events on membrane bioreactor performance

Abstract

Implementation of water treatment processes for recycling requires validation to demonstrate that such processes are capable of achieving required water quality objectives. In order to fully validate the performance of treatment systems, it is necessary to consider both normal (expected) operational conditions, as well as a range of potential “hazardous event” scenarios. The performance of membrane bioreactors (MBRs), including the removals of 48 trace organic chemical contaminants and key bulk water quality and operational parameters, was investigated under normal and hazardous event conditions. Full mass balance was achieved by monitoring both aqueous dissolved chemicals and those adsorbed to biomass. Hazardous events investigated included organic, salinity, 2,4-dinitrophenol (DNP) and, ammonia shock loads, feed starvation, loss of power and physical membrane damage. Under organic, salinity, DNP, and ammonia shock conditions, removals of moderately and very hydrophobic chemicals were not affected. Since these chemicals are largely adsorbed to biomass, these results imply that biotransformation within the biomass structure itself was maintained. However, removals of hydrophilic chemicals were commonly observed to be impeded under hazardous event conditions, indicating loss of the bioactivity in the aqueous phase. This was observed primarily for chemicals of low or moderate ready biotransformability, while easily biotransformable chemicals were still largely removed. The removals of all chemicals were unaffected by the feed starvation (absence of new assimilable substances for 6 days) while the removals of some less readily biotransformable hydrophilic chemicals were measurably affected by the loss of power conditions (2 hours). Impacts from physical membrane damage were investigated by sequentially cutting two hollow-fibre membranes within the MBR. Turbidity and chemical oxygen demand (COD) analyses revealed that these impacts were ‘self-repaired’ by blocking of the breach within approximately 15 minutes. Accordingly, these hazardous events were shown to have an insignificant impact to overall trace chemical removals. The variability of trace chemical removals during hazardous event scenarios has enabled the identification of sensitive chemical indicators for the validation of MBR system performance. The application of these indicators for future risk assessment and management is described.