Multi-objective optimization with multiple Pareto frontiers applied to a MED desalination system

Abstract

This article presents the simulation results of a Multi Effect Desalination system coupled to a Mechanical Vapor Compression system. The system is optimized with the Non-dominated Sorting Genetic Algorithm II genetic algorithm by obtaining the optimal results for the desalination plant operating with 8–12 effects, with Forward Feeding and Parallel Cross Feeding. The best results are obtained for two cases: a) For a Multi Effect Desalination system with Parallel Cross Feeding operating with 12 effects, a Top Brine Temperature of 64 °C and a feed water flow rate of 293 kg/s. This configuration presents a Unit Cost of Distilled Water of 1.94 USD/m3 and 128.1 kg/s of fresh water. b) And for a Multi Effect Desalination system with Forward Feeding with 9 effects, a feed water flow rate of 380 kg/s, a Top Brine Temperature of 64 °C, which allows obtaining a Unit Cost of Distilled Water of 1.79 USD/m3. The same system but with 12 effects, a Top Brine Temperature of 72 °C, a feed water flow rate of 380 kg/s allows obtaining a higher water production of 152.3 kg/s, which shows that there is more than one optimal solution for this kind of systems, which depends on the figure to be optimized.