一种改进粒子群算法在横向磁通永磁电机优化中的应用

通过对鸟、鱼等动物群体捕食行为的深入分析,进行了粒子群优化算法的改进研究:在优化过程中,动态调整学习因子和惯性参量,得到了全新的粒子速度与位置更新模式,从而有效防止了粒子的"早熟"问题,并改善了算法的全局寻优能力。使用提出的改进粒子群算法对典型数学函数进行校验,结果表明:改进算法的全局寻优能力比原算法有明显提高。将其引入电磁场逆问题求解过程,对一台聚磁式横向磁通永磁同步电机进行基于磁钢用量和电机整体体积最小的尺寸优化设计,取得了令人较为满意的效果。

Application of An Improved Particle Swarm Optimization for Transverse Flux Permanent Magnet Machine Optimization

Based on a comprehensive simulation of bird flocking or fish schooling behavior,an improved particle swarm optimization(IPSO)algorithm is developed.By the linearly decreasing learning factors and stochasitc adjusting inertia weight,the global search ability of particles is improved which avoiding a stagnation phenomenon as well as achieving a good convergence performance.Benchmark testing of four PSOs is discussed and compared.The application to minimize the permanent magnetic material cost of a transverse flux permanent magnet machine(TFPM)demonstrates the validation of the IPSO.