刘畅,王立勇,吴健鹏,张喜明

• 1:北京信息科技大学机电工程学院
• 2:中国北方车辆研究所

摘要(Abstract)：

针对传统永磁同步电机（permanent magnet synchronous motor, PMSM）损耗预测方法精度低、易陷入局部最优等问题，提出一种基于天牛须搜索（beetle antennae search, BAS）算法优化的反向传播（back propagation, BP）神经网络损耗预测模型。首先，基于有限元方法构建PMSM的电磁场损耗计算仿真模型，并利用最佳空间填充试验设计方法，选取了400组控制参数组合（定子电流、转速、电压和内功率因数角）进行仿真求解，获得用于神经网络训练的数据集。其次，在此基础上，引用BAS算法对BP网络的初始权重和偏置进行全局优化，提升网络对复杂非线性关系的拟合能力，加快训练收敛速度，增强模型预测稳定性。最后，构建多输出预测模型BASBP，该模型可同时预测定子铁损、转子铁损、绕组铜损及永磁体涡流损耗。实验结果表明，BAS-BP模型对各类损耗均具备良好的预测能力，在平均绝对百分比误差（mean absolute percentage error,MAPE）、平均绝对误差（mean absolute error, MAE）与均方根误差（root mean square error, RMSE）等误差指标上明显优于传统BP网络，体现出更高的预测精度。

关键词(KeyWords)： 永磁同步电机;损耗预测;反向传播神经网络;天牛须搜索算法;有限元分析

Abstract:

Keywords:

基金项目(Foundation):

作者(Author): 刘畅,王立勇,吴健鹏,张喜明

DOI: 10.16508/j.cnki.11-5866/n.2025.05.006

参考文献(References)：

• [1] SUN T F, WANG J B,JIA C L,et al. Integration of FOC with DFVC for interior permanent magnet synchronous machine drives[J]. IEEE Access,2020,8:97935-97945.
• [2] CAI W,WU X G,ZHOU M H,et al. Review and development of electric motor systems and electric powertrains for new energy vehicles[J]. Automotive Innovation,2021,4(1):3-22.
• [3]陈彬，王书恒，景弋洋，等.车用驱动电机损耗温升分析与多目标优化[J].微特电机，2024,52(1):36-42.CHEN B,WANG S H,JING Y Y,et al. Analysis and multiobjective optimization of the loss and temperature rise of drive motor for microvan[J]. Small&Special Electrical Machines,2024,52(1):36-42.(in Chinese)
• [4]陈楷翼.纯电动商用车驱动电机损耗计算及温度场分析[D].重庆：重庆理工大学，2024.CHEN K Y. Loss calculation and temperature field analysis of drive motor for pure electric commercial vehicle[D]. Chongqing:Chongqing University of Technology,2024.(in Chinese)
• [5]王为术，亓月欣，彭岩，等.高速永磁电机损耗特性研究[J].电工技术，2022(21):234-238.WANG W S,QI Y X,PENG Y,et al. Research on loss characteristics of high speed permanent magnet motor[J].Electric Engineering,2022(21):234-238.(in Chinese)
• [6]刘强.永磁同步驱动电机电磁损耗及温度场研究[D].上海：上海交通大学，2020.LIU Q. Study on electromagnetic loss and thermal model of permanent magnet synchronous motor used in drive system[D].Shanghai:Shanghai Jiaotong University,2020.(in Chinese)
• [7]李良辉，李乐，王茜，等.基于DBO-BP的永磁同步电机损耗预测模型[J].现代制造工程，2025(2):130-137.LI L H,LI L,WANG Q,et al. A loss prediction model for PMSM based on BP neural network optimized by dung beetle optimizer[J]. Modern Manufacturing Engineering,2025(2):130-137.(in Chinese)
• [8]金逸珲，吴滨，顾晓峰，等.基于BAS算法优化的感应电机参数估计[J].现代电子技术，2020,43(3):112-115.JIN Y H,WU B,GU X F,et al. Induction motor parameter estimation based on BAS algorithm optimization[J]. Modern Electronics Technique,2020,43(3):112-115.(in Chinese)
• [9]刘霞，王鑫宇，路敬祎，等.基于BAS优化PNN网络的电机轴承故障诊断方法[J].吉林大学学报（信息科学版），2021,39(4):439-444.LIU X,WANG X Y,LU J Y,et al. Fault diagnosis method of motor bearing based on PNN optimized by BAS algorithm[J].Journal of Jilin University(Information Science Edition),2021,39(4):439-444.(in Chinese)
• [10]彭倩，吴浪，徐笑，等. BAS-PSO算法优化的无传感器轮毂电机转速控制[J].江苏理工学院学报，2024,30(4):16-27.PENG Q,WU L,XU X,et al. Speed control of sensorless hub motor optimized by BAS-PSO algorithm[J]. Journal of Jiangsu University of Technology,2024,30(4):16-27.(in Chinese)
• [11]任钊月.基于神经网络的永磁同步电机模型预测控制研究[D].十堰：湖北汽车工业学院，2023.REN Z Y. Research on permanent magnet synchronous motor model predictive control based on neural network[D]. Shiyan:Hubei University of Automotive Technology,2023.(in Chinese)
• [12]胡陈杉.永磁同步电机及驱动系统损耗建模与效率优化控制策略[D].长沙：湖南大学，2023.HU C S. Loss modeling and efficiency optimization control strategy of permanent magnet synchronous motor and drive system[D]. Changsha:Hunan University,2023.(in Chinese)
• [13] MISIR O,AKAR M. Efficiency and core loss map estimation with machine learning based multivariate polynomial regression model[J]. Mathematics,2022,10(19):3691.
• [14]王涛.基于时步有限元的高速电机损耗精细化分析[D].北京：华北电力大学，2019.WANG T. Fine analysis of high speed motor loss based on timestep finite element method[D]. Beijing:North China Electric Power University,2019.(in Chinese)
• [15]安智洋.车用永磁同步电机损耗分析及测试方法[J].汽车零部件，2025(1):55-60.AN Z Y. Loss analysis and measurement method of automotive permanent magnet synchronous motor[J]. Automobile Parts,2025(1):55-60.(in Chinese)
• [16]朱洒，曾峰，陆剑波，等.考虑PWM谐波损耗的车用扁线内嵌式永磁同步电机效率图简化工程计算[J].电工技术学报，2022,37(22):5687-5703.ZHU S,ZENG F,LU J B,et al. Simplified engineering calculation of efficiency map of interior permanent magnet synchronous machines with hairpin windings considering PWMinduced harmonic losses[J]. Transactions of China Electrotechnical Society,2022,37(22):5687-5703.(in Chinese)
• [17]毕成江.基于代理模型的高速永磁电机磁热耦合优化方法研究[D].沈阳：沈阳工业大学，2022.BI C J. Research on magneto-thermal coupling optimization method of high speed permanent magnet motor based on surrogate model[D]. Shenyang:Shenyang University of Technology,2022.(in Chinese)
• [18]王琰琰，任俊玲.基于GA-ACO-BP神经网络的日用消费品物流需求预测[J].北京信息科技大学学报（自然科学版），2024,39(1):91-98.WANG Y Y,REN J L. Logistics demand prediction of daily consumer goods based on GA-ACO-BP neural network[J].Journal of Beijing Information Science and Technology University(Science and Technology Edition),2024,39(1):91-98.(in Chinese)
• [19]王占，孙晶，王芳，等.基于优化BP神经网络的铁路物资需求预测模型研究[J].铁路采购与物流，2025,20(4):17-20.WANG Z,SUN J,WANG F,et al. Research on railway material demand forecasting model based on optimized BP neural network[J]. Railway Purchase and Logistics,2025,20(4):17-20.(in Chinese)
• [20] ZHANG H M. Financial crisis warning analysis for companies based on BP neural network[J]. Applied Mechanics and Materials,2014,519/520:1529-1533.
• [21]廖列法，欧阳宗英.基于二次插值的天牛须搜索算法[J].计算机应用研究，2021,38(3):745-750.LIAO L F,OUYANG Z Y. Beetle antennae search based on quadratic interpolation[J]. Application Research of Computers,2021,38(3):745-750.(in Chinese)
• [22]刘娣，孙佳倩，余钟波.基于改进BP神经网络的多层土壤湿度反演[J].节水灌溉，2023(11):19-27.LIU D,SUN J Q,YU Z B. Multi-layer soil moisture inversion based on improved back propagation neural network technique[J]. Water Saving Irrigation,2023(11):19-27.(in Chinese)
• [23]王杰.基于天牛须搜索算法的神经网络并行优化研究及预测应用[D].杭州：杭州电子科技大学，2020.WANG J. Research and forecasting applications of neural network parallel optimization based on beetle antennae search algorithm[D]. Hangzhou:Hangzhou Dianzi University,2020.(in Chinese)
• [24]姚启达，平鹏，朱心怡，等.基于机器学习的磁性元件磁芯损耗预测方法[J].南通大学学报（自然科学版），2025,24(2):29-38.YAO Q D,PING P,ZHU X Y,et al. Core loss prediction method for magnetic components based on machine learning[J]. Journal of Nantong University(Natural Science Edition),2025,24(2):29-38.(in Chinese)