Pilot plant study on the performance and optimization of submerged membranes for taste and odor removal

Hybrid sorption-membrane processes are an attractive alternative for meeting a range of water treatment goals in a single process that is compact and cost-effective. This study investigated the performance and optimization of a hybrid sorption-membrane process using powdered activated carbon (PAC) and submerged-style membranes for odor control. Specifically, this study focused on the removal of 2-methylisoborneol (2-MIB) from a Lake Michigan source water and investigated the effects of PAC dose, dosing method, backwash interval, and aeration. Adsorption performance was predicted using a mathematical model, and tested using a pilot-scale, submerged membrane system. Modeling continuously dosed PAC performance agreed well with pilot results, but pulse dosed PAC performance was overestimated by the model. Non-ideal mixing effects were identified as important factors in explaining the pulse dosing results. Pilot results with aeration and pulse dosing were also overestimated by the model. It is hypothesized that aeration disturbs the PAC cake layer that forms on the membrane, and thus eliminates the beneficial effects that this cake layer can have on adsorption. Extending backwash intervals from 30 to 180 minutes with continuous dosing increased the percent removal of 2-MIB by up to 30% in both the model predictions and pilot results. This study highlights some important considerations in the design of full-scale systems and future mathematical models.