Multi-objective Magnetic Circuit Optimization of Coreless Motors Based on an Improved NSGA-Ⅲ

Abstract

Aiming at the problems that multi-objective optimization algorithms are prone to fall into local optima and yield uneven solution set distributions in the magnetic circuit optimization of permanent magnet DC coreless motors for the aerospace field, an improved NSGA-Ⅲ multi-objective optimization algorithm is proposed. This algorithm incorporates a dynamic adaptive crossover and mutation mechanism, a normal distribution crossover operator, and a dynamic crowding degree operator, which effectively enhances the global search capability in high-dimensional objective spaces and the distribution uniformity of the Pareto solution set, thus solving the problems that traditional algorithms tend to fall into local optima and suffer from insufficient solution set diversity in multi-objective collaborative optimization. Taking the maximization of torque coefficient, minimization of torque ripple, and minimization of magnetic leakage coefficient of the coreless motor as the core optimization objectives, a high-precision surrogate model between the objective functions and decision variables was established based on the response surface methodology, and embedded into the improved NSGA-Ⅲ algorithm to realize the multi-objective optimization of magnetic circuit parameters. Simulation results show that after optimization, the motor torque coefficient is increased by 29%, the torque ripple is reduced by 13%, and the magnetic leakage coefficient is decreased by 37%, with comprehensive performance indicators improved significantly, indicating favorable engineering application value.