Analysis of reverse osmosis membrane behaviors in a long-term verification test

Almost all of reverse osmosis (RO) desalination data have been reported in terms of flu: and salt rejection. Fouling phenomena have also been reported in raw data. Such a way of presenting data is misleading to know the real causes of membrane fouling, and thus the reliable design of RO unit becomes impossible. The correct presentation of RO data should be in the form of transport coefficients of a membrane based on the correct transport equations. In this paper a couple of solution-diffusion type equations is compared carefully with those proposed by Spiegler and Kedern. The former uses two parameters, A and B, while the latter uses three, namely, L(p), sigma, and P. In long run experiments, however, the only data available are flux and rejection, and it is impossible to determine three coefficients from two parameters. Since RO membranes used for seawater desalination show their sigma is very close to 1, it can be shown that value B is very close to P, and it is concluded the former type of equation can be effectively used for a long-run data analysis. An example of such analysis is presented based on the data taken by WRPC, Japan, at Chigasaki, with 12 '' Hollosep hollow fiber modules made by Toyobo. Results showed an interesting trend of variation of A and B values and their relation with time. Although real causes of membrane fouling phenomena are not dear at this stage, the method proposed here is valid and should be extended further to compile such data, which will be related to the physical and chemical changes of membranes and their materials.