Abstract: The present work is focused on the multiobjective optimization of integrated multistage reverse osmosis membrane cascades applied to the ultrapurification of chemicals for the semiconductor industry. The membrane systems were formulated adding product quality metrics to economic criteria to result a bi-criteria nonlinear programming (NLP) problem. The Pareto solutions to the problem were generated via the epsilon constraint method. On the one hand, maximum economic profit solutions corresponded with the configurations applying bypass. The bypass from the feed stream of the last stage was preferred over other proposed options. On the other hand, maximum quality solutions were obtained by low recovery rates (specifically in the last stages of the cascade) and the consequent high flow (and low metallic content) streams recirculated to previous stages.

Authorship

Download reference