(756a) Modeling of Organic Fouling Based on Interaction between Membrane and Foulants for Reverse Osmosis Membrane

Reverse osmosis (RO) technology has progressed steadily over the last few decades and is now the leading technology in the desalination industry. Those gains were achieved through improvements in both RO membrane element performance and energy recovery technologies. For further energy saving and RO plant cost reduction, continuous improvement of membrane element performance and process efficiency are needed. A major problem in RO plant operation is fouling on membranes or spacers that increase specific energy consumption and operating cost. Therefore, to evaluate accurately the effect of developing a new membrane element or a new process, RO plant cost must be estimated with consideration of chronological changes caused by fouling.

Several fouling models have been proposed for membrane separation so far, but most of them rely on parameters that were obtained from filtration experiments or field plant operations, as opposed to fouling mechanisms. These models with experimental or empirical parameters are useful to estimate the chronological changes of water permeate flux or resistance of foulant layers when the parameters of fouling model are known for the specific combination of membrane, element, feed water composition, and other operating conditions. However, for the estimation of gains from new membranes, new elements, or new processes, additional filtration experiments or plant tests are needed. Therefore, it is desirable for the estimation of gains from new membranes, new elements, or new processes to have a fouling model that is less dependent upon experimental or empirical parameters.

This study is focused on the modeling of organic fouling, which is the main cause of RO fouling. In this study, the organic fouling rate was assumed to be sum of the attachment rate of foulant on the membrane surface, the detachment rate from membrane surface, and the deposition rate of foulant due to the presence of obstacles or insufficient cross-flow velocity. For the estimation of the attachment rate, the attachment probability is determined from interactions between the membrane and foulants or among the foulants. The detachment rate is modeled to caused by rolling of attached foulants by cross-flow velocity, and the rolling rate and the detachment probability were estimated. The deposition rate was estimated from the residence time of foulants that were not attached to the membrane surface. The chronological flux transition estimated from a fouling model proposed in this study was compared with the experimental results which use sodium alginate and bovine serum albumin (BSA) as model foulants for polysaccharide and protein, respectively.