基于GASA-BRNN的齿槽转矩预测方法与分析  

Prediction Method and Analysis of Cogging Torque Based on GASA-BRNN

作　　者：任鸿昌 王东瑞[1] 姜添惠 张闯 陈文辉 REN Hongchang;WANG Dongrui;JIANG Tianhui;ZHANG Chuang;CHEN Wenhui(Shenyang Institute of Engineering,Shenyang 110036,China)

机构地区：[1]沈阳工程学院,辽宁沈阳110136

出　　处：《机械制造与自动化》2022年第3期196-200,共5页Machine Building & Automation

基　　金：国家级大学生创新创业计划项目(202011632019);沈阳工程学院大学生创新创业计划项目(202011632085)。

摘　　要：为降低现有齿槽转矩分析方法的复杂性,完成对电机结构的快速优化设计,结合贝叶斯正则化与模拟退火遗传算法对BP神经网络进行改进,构建永磁同步电机齿槽转矩网络预测分析模型。选取永磁同步电机的3种结构特征与齿槽转矩的参数化关系作为研究对象,使用有限元方法设计了一款4极24槽的永磁同步电机真模型,通过比较有限元仿真数据以及不同神经网络算法的预测结果,验证了GASA-BRNN预测模型的可行性。In order to reduce the complexity of existing slot torque analysis methods and complete the rapid optimization design of motor structure,the BP neural network is upgraded by combining Bayesian regularization with simulated annealing genetic algorithm,and a PMSM slot torque network prediction analysis model is constructed.With the structural features of permanent magnet synchronous motor and the parametric relation of cogging torque as the reserch subjects,a simulation model of a 4-pole with 24-slot permanent magnet synchronous motor is designed by finite element method.The feasibility of the GASA-BRNN prediction model is verified by comparing the finite element simulation data and the prediction results of different neural network algorithms.

关 键 词：齿槽转矩 神经网络 遗传算法 贝叶斯正则化 

分 类 号：TM351[电气工程—电机]