摘要
为了提高永磁同步电机(PMSM)的性能,本文以一台72槽60极永磁同步电机为例,针对永磁同步电机多目标优化过程中多次有限元迭代导致的计算时间长和优化效率低的问题,提出了一种基于果蝇优化算法(FOA)的多目标优化方法。选取磁钢尺寸作为优化变量,以电机平均转矩、转矩波动和齿槽转矩作为优化目标,采用权重系数的多目标优化函数。首先通过有限元仿真获得各变量的样本空间,其次采用广义回归神经网络(GRNN)对仿真数据集进行拟合训练,得到非线性模型,然后运用FOA进行优化。最后,通过有限元仿真分析,结果表明FOA能有效抑制转矩波动以及增大平均转矩,且具有参数设置少、收敛速度快等优点,具有较好的应用价值。
To improve the performance of Permanent Magnet Synchronous Motors(PMSM),this paper used a 72-slot 60-pole PMSM as an example.It addressed the issue of extended computational time and low optimization efficiency during multi-objective optimization by proposing a multi-objective optimization method based on the Fruit Fly Optimization Algorithm(FOA).The optimization variables selected are the dimensions of the magnetic steel,the optimization objectives include the motor's average torque,torque fluctuation,and slot torque.These objectives were incorporated into a multi-objective optimization function with weighting coefficients.The approach begined by obtaining the sample space of each variable through finite element simulations.Subsequently,a Generalized Regression Neural Network(GRNN)was employed to fit and train the simulation dataset,creating nonlinear models.Finally,the Fruit Fly Optimization Algorithm(FOA)was applied for optimization.The finite element simulation analysis shows that the FOA algorithm effectively reduces torque fluctuation and increases the average torque.It also offers advantages such as minimal parameter configuration,fast convergence,and resistance to getting trapped in local optima.This method holds significant practical value.
作者
柳洪
陈玮
王定龙
吴顺海
LIU Hong;CHEN Wei;WANG Dinglong;WU Shunhai(Hunan CRRC Shangqu Electric Co.,LTD.,Zhuzhou Hunan412001,China)
出处
《微电机》
2024年第7期18-23,共6页
Micromotors
关键词
永磁同步电机
果蝇优化算法
广义回归神经网络
多目标优化
有限元分析
permanent magnet synchronous motor
fruit fly optimization algorithm
generalized regression neural network
multi-objective optimization
finite element analysis
