Systematic Development of a Multi-Objective Design Optimization Process Based on a Surrogate-Assisted Evolutionary Algorithm for Electric Machine Applications

Surrogate model (SM)-based optimization approaches have gained significant attention in recent years due to their ability find optimal solutions faster than finite element (FE)-based methods. However, there is limited previous literature available on the detailed process of constructing SM-based for multi-parameter, multi-objective design electric machines. This paper aims present a systematic an interior permanent magnet synchronous machine (IPMSM), including thorough examination construction SM and adjustment its parameters, which are crucial reducing computation time. The performances candidates such as Kriging, artificial neural networks (ANNs), support vector regression (SVR) analyzed, it found that Kriging exhibits relatively better performance. hyperparameters each fine-tuned using Bayesian avoid manual empirical tuning. In addition, convergence criteria determining number FE computations needed construct discussed detail. Finally, validity proposed verified by comparing Pareto fronts obtained from conventional FE-based results show procedure can significantly reduce total time approximately 93% without sacrificing accuracy compared method.