Filtration Resistance Characteristics of a Membrane Process Coupled with a Biological Granular Activated Carbon (BAC) Process for Enhancing the Performance of a Drinking Water Treatment System

An increase of micropollutant in the surface water being used as water supply source is needed to conventional water treatment processes for improving a drinking water quality. Biological granular activated carbon (BAC) bed have been applied to remove micropollutants in surface water used as a water supply source. In addition, membrane filtration (MF) has also been introduced to water treatment plant to provide higher-quality drinking water by removing bacteria, viruses, microorganisms as well as particulate and organic matter. This study aims to investigate the effect of BAC biofiltration as a pretreatment before MF processes and evaluate the performance of BAC/MF systems for the treatment of lake water. The applied filtration rates during this experiment were <100 m/d for BAC and 0.5 ㎥/㎡.d for MF. Influent concentrations were as follows: humic acid <10 mg/L, turbidity <100 NTU, UV254 < 0.250 cm-1, dissolved organic carbon (DOC) < 9.7 mg/L. The system was found to be considerably effective for reducing foulants. The results indicates that the addition of a BAC bed prior to the MF process is necessary for cost-effective lake water treatment, as if effectively reduces geosmin, 2-MIB, and other micropollutants, while also mitigating membrane fouling by lowering TMP. The total filtration resistance was composed of 35.4 to 44.5% from the intrinsic membrane resistance, 34.7 to 36.7% from cake layer resistance, and 19.8 to 27.9% from internal contamination resistance of the separator, with the contamination of the separator itself playing a significant role in the overall filtration resistance. The Silt density index (SDI) were found to be one of the most effective and convenient variables for evaluating sludge properties and membrane fouling. An empirical relationship between SDI and membrane permeability was proposed for effectively predicting membrane fouling in an MF and combined MF with conventional treatment processes under high turbidity and varying DOC concentrations.