Design and optimization of reverse osmosis unit for vinasse purification of Moroccan distillery

Abstract

The aim of this work is to design and optimize a RO unit that can be integrated into a Moroccan distillery, for vinasse purification. The RO membrane was chosen, for its efficiency to remove acetic acid and glycerol. Based on water flux calculations, the RO unit design was carried out, proposing a single-stage configuration with modules grouped in series. Subsequently, Artificial neural network (ANN) and Non-dominated sorting genetic algorithm-II (NSGA-II) are applied to optimize the rejections depending on the operating parameters of the proposed process. The effect of these parameters on the two rejections was conducted by numerical experiments using physical model. The resulting data are exploited to develop a reliable ANN model allowing a satisfactory prediction of the rejections. By coupling this model to the NSGA-II the optimal rejections are 99.8% for acetic acid and 98.8% for glycerol. The cost of vinasse purification is evaluated to 0.48 US$/m3.