In order to optimize the local search efficiency of multi-objective parameters of flux switching permanent motor based on traditional NSGA-II algorithm, an improved NSGA-II (iNSGA-II) algorithm is proposed, with an anti-redundant mutation operator and forward comparison operation designed for quick identification of non-dominated individuals. In the initial stage of the iNSGA-II algorithm, half of the individual populations were randomly generated, while the other half was generated according to feature distribution information. Taking the flux switching permanent motor stator/rotor gap, permanent magnets width, stator tooth width, rotor tooth width and other parameters as optimization variables, the flux switching permanent motor maximum output shaft torque and minimum torque ripple are taken as optimization objectives, thus a multi-objective optimization model is established. Real number coding was adopted for obtaining the Pareto optimal solution of flux switching permanent motor structure parameters. The results showed that the iNSGA-II algorithm is better than the traditional NSGA-II on convergence. A 1.8L TOYOTA PRIUS model was selected as the prototype vehicle. By using the optimized parameters, a joint optimization simulation model was established by calling ADVISOR’s back-office function. The simulation results showed that the entire vehicle’s 100-km acceleration time is under 8 s and the battery’s SOC value maintains at 0.5–0.7 in the entire cycle, implying that the iNSGA-II algorithm optimizes the flux switching permanent motor design and is suitable for the initial design and optimizing calculation of the flux switching permanent motor.

中文翻译：

---

基于改进NSGA-II算法的开关磁通电机多目标优化研究

为了优化基于传统NSGA-II算法的磁通切换永磁电机多目标参数的局部搜索效率，提出一种改进的NSGA-II（iNSGA-II）算法，具有抗冗余变异算子和前向比较旨在快速识别非支配个体的操作。在iNSGA-II算法的初始阶段，一半的个体种群是随机生成的，另一半是根据特征分布信息生成的。以磁通切换永磁电机定转子间隙、永磁体宽度、定子齿宽、转子齿宽等参数为优化变量，以磁通切换永磁电机最大输出轴转矩和最小转矩脉动为优化目标，从而建立多目标优化模型。采用实数编码求得磁通切换永磁电机结构参数的帕累托最优解。结果表明，iNSGA-II算法在收敛性上优于传统的NSGA-II算法。1.8L TOYOTA PRIUS 车型被选为原型车。利用优化后的参数，调用ADVISOR的后台功能建立联合优化仿真模型。仿真结果表明，整车百公里加速时间在8 s以下，整个循环中电池SOC值保持在0.5-0.7，说明iNSGA-II算法优化了磁通切换永磁电机设计，适用于磁通切换永磁电机的初步设计和优化计算。