Hybrid Surrogate Model Based Multiobjective Optimization Design of Flux-Modulated Permanent Magnet Machine for Shaftless Pump-Jet Propulsor

This article proposes a novel hybrid surrogate model (SM) based multiobjective optimization method to optimize flux-modulated permanent magnet (PM) machine (FPMM) for the shaftless pump-jet propulsor. The proposed optimization method combining radial basis function neural network (RBFNN), support vector regression (SVR), and nondominated sorting genetic algorithm-III (NSGA-III) to achieve high torque performance and low harmonics of the back electromotive force (EMF). The topology and operating principle of the FPMM are addressed. The design variables are divided into different levels based on sensitivity analysis and the optimization objectives are selected. A hybrid SM is established based on the data space at different levels. NSGA-III is applied to obtain the nondominance solutions and the final design point is selected. The electromagnetic performance of the initial and optimized schemes is compared by finite element analysis (FEA), which verifies the effectiveness and superiority of the proposed multiobjective optimization method.