重庆理工大学学报2025，Vol.39Issue(7)：106-115,10.DOI:10.3969/j.issn.1674-8425(z).2025.04.014
RNN与MLP融合算法在永磁同步电机谐波抑制中的应用
Application of RNN and MLP fusion algorithm in harmonic suppression of PMSM
摘要
Abstract
This paper proposes a current harmonic suppression algorithm based on recurrent neural network(RNN)and multi-layer perceptron(MLP)for the 5th and 7th harmonic current problems of permanent magnetic synchronous motors(PMSM).It employs two independent RNNs to achieve regression prediction of voltage compensation values and uses an MLP network to perform decision-level fusion on different prediction values.The fused compensation value is injected into the motor winding to effectively suppress harmonic currents.Simulation and experimental results show the algorithm performs better in suppressing the 5th and 7th harmonic currents of PMSM,not only improving the approximation accuracy of the RNN algorithm,but also enhancing the overall harmonic current suppression effect.关键词
永磁同步电机/电流谐波抑制算法/循环神经网络/多层神经网络/决策融合Key words
PMSM/current harmonic suppression algorithm/RNN/MLP/decision fusion分类
动力与电气工程引用本文复制引用
李学成,郭俊杰,徐龙翔..RNN与MLP融合算法在永磁同步电机谐波抑制中的应用[J].重庆理工大学学报,2025,39(7):106-115,10.基金项目
河南省研究生教育改革与质量提升工程项目(YJS2022JD48) （YJS2022JD48）
芜湖市重大科技成果工程化项目(2022ZC07) （2022ZC07）
