Multi-objective optimization of aluminum electrolysis based on functional evolution operator

Abstract:

Aiming at the multi-objective optimization problem that it is difficult to effectively improve current efficiency and reduce DC energy consumption in the aluminum electrolysis manufacturing system (AEMS), a functional evolutionary operator-based NSGA-II (FEONSGA-II) is proposed in this paper. Based on the stable operation of the system, the Pareto non-inferior solution set can be obtained to meet the demands of increasing the efficiency and reducing the consumption of aluminum electrolysis. The crowding entropy is used to update the population, and the distribution of the front solution set at all levels is accurately estimated. To reduce the possibility of destroying the excellent solution set, we introduce arithmetic crossover with a new α-function operator. Then, according to the Gaussian Cauchy variation characteristics, the perturbation related to the number of iterations is generated to expand search ranges and accuracy. Finally, the standard test function is used to detect the performance of the algorithm and three comparison algorithms are applied to solve the aluminum electrolysis example. The experimental results show that the proposed algorithm can obtain the Pareto optimal solution set with uniform distribution, which is conducive to the reference decision of aluminum electrolysis plant to achieve the purpose of improving efficiency and reducing consumption.