开关磁阻电机小波神经网络无位置传感器控制  

Position Sensorless Control of Switched Reluctance Motor Using Wavelet Neural Networks

作　　者：夏长亮[1] 谢细明[1] 史婷娜[1] 田洋[1] 

机构地区：[1]天津大学电气与自动化工程学院,天津300072

出　　处：《电工技术学报》2008年第7期33-38,共6页Transactions of China Electrotechnical Society

基　　金：天津市应用基础研究计划资助项目(06YFJMJC01900)

摘　　要：提出了一种基于小波神经网络的开关磁阻电机无位置传感器控制新方法。该方法采用两个不同的小波神经网络分别获取相绕组换相逻辑的开通信号和关断信号,经过综合处理得到单相绕组的开关信号。神经网络以相绕组的电流和磁链为输入,以各相的开关信号为输出,从而建立起电流、磁链和开关信号的非线性映射。采用电机在有位置传感器运行条件下的样本对小波神经网络进行训练,训练完成后,用神经网络输出结果取代位置传感器换相信号,实现电机无位置传感器运行。仿真和实验结果表明,由神经网络获得的开关信号和由位置传感器获得的开关信号相比误差小,电机能够准确换相,且输出转矩波动小,转速曲线平滑,电机在无位置传感器下运行良好。This paper presents a new approach to the position sensorless control of the switched reluctance motor（SRM） based on wavelet neural networks（WNNs）. The basic premise of the approach is that two wavelet neural networks with different parameters are constructed to switch on and turn off each phase respectively. The WNNs form a very efficient nonlinear mapping structure from phase current, flux linkage to communication signal with current, flux linkage as input and switching signal as output, therefore the communication signals can be obtained by manipulation of the WNNs＇ outputs. After trained by the data acquired from the system with position sensor, the WNNs replace the position sensor and make the SRM switch to position sensorless operation. The simulation and experimental results show that there is tiny error of switching signals between estimation and reality. The SRM can operate with little torque fluctuation and slight speed vibration.

关 键 词：小波神经网络 开关磁阻电机 无位置传感器控制 梯度下降法 小波函数 

分 类 号：TM352[电气工程—电机]