Multi-Strategy Ensemble Non-Dominated Sorting Genetic Algorithm and Application in Energy-Enviro-Economic Multi-Objective Optimization of Isopropyl Alcohol/Isopropyl Ether Separation Process

Abstract

The separation of isopropyl alcohol (IPA)/isopropyl ether (IPE) is crucial for the recovery of valuable solvents from alcohol ether waste liquids, but involves high energy consumption and emissions. Therefore, the multi-objective optimization of energy, economy and environment in the IPA/IPE separation process is conducive to the sustainable development of the organic solvent industry. Considering the slow convergence and the tendency to fall into local optimum for the most commonly used multi-objective optimization method, namely non-dominant sorting genetic algorithm-II (NSGA-II), a novel multi-strategy ensemble non-dominated sorting genetic algorithm (MENSGA-II) is proposed. In this algorithm, two evolution strategies based on individual neighborhood guidance and random walk are developed, aiming at accelerating convergence speed and enhancing search ability, and a selection strategy based on double space is proposed to improve the distribution of the Pareto front obtained. The MENSGA-II is tested on the benchmark functions and the actual separation process of IPA/IPE. The results prove that the proposed MENSGA-II algorithm has superiority in robustness, convergence speed and distribution of Pareto front. Compared with the actual operating condition, the annual exergy destruction and carbon emission can be reduced by 18.49% and 15.47% when the annual gross profit increases by 3.24% via multi-objective optimization.