Optimal design of transverse field motor based on hybrid algorithm of harmony search and differential evolution

基于和声差分进化混合算法的横向磁场电机优化设计

Received: 12 March 2020

Abstract

The traditional optimal design of a transverse magnetic field motor mainly uses the electromagnetic relational formulas and the finite element analysis to empirically select its structural parameters, therefore making it difficult to obtain the global optimal solution of each parameter and meeting the requirements for multi-objective optimization. This paper suggests a hybrid algorithm of harmony search and differential evolution and introduces the elite retention strategy to improve the hybrid algorithm. It presents the principles of the hybrid algorithm, and then the algorithm performance test is used to verify the superiority of the hybrid algorithm. The electromagnetic analysis demonstrates the influence of the main dimensional parameters of the transverse magnetic field motor on its maximum torque density, therefore providing the basis for its optimal design. Jointly with the hybrid algorithm, the maximum torque density and the production cost are used as the objective function to optimize the transverse magnetic field motor under two groups of weighted coefficients. The simulation results show that the maximum torque density of the transverse magnetic field motor optimized with the hybrid algorithm is larger compared with the prototype motor but its production cost is lower.

摘要

横向磁场电机传统优化设计大多结合电磁关系公式和有限元分析凭借经验选取结构参数，难以得到各参数的全局最优解并满足多目标优化的需要。提出了一种将和声搜索算法和差分进化算法相结合的和声差分进化混合算法，并引入精英保留策略改进。介绍了该算法的原理，利用算法性能测试验证了该算法的优越性。通过电磁分析揭示了电机各主要尺寸参数对最大转矩密度的影响，为优化设计提供依据。结合该算法，在2组权重系数下以最大转矩密度和生产成本为目标函数对电机进行优化，仿真结果表明，优化后的电机与原型机相比最大转矩密度提高，生产成本降低。所提算法对横向磁场电机的优化设计有一定的参考价值。