微特电机 ›› 2020, Vol. 48 ›› Issue (7): 57-60.

• 驱动控制 • 上一篇    下一篇

开关磁阻电机神经网络无位置传感器控制

程昭竣, 李姗, 陈茂才, 曹阳   

  1. 中国船舶重工集团公司第七〇四研究所,上海 200031
  • 收稿日期:2020-04-28 出版日期:2020-07-28 发布日期:2020-07-17

Neural Network Sensorless Control for Switched Reluctance Motor

CHENG Zhao-jun, LI Shan, CHEN Mao-cai, CAO Yang   

  1. 704 Research Institute of China Shipbuilding Industry Corporation,Shanghai 200031,China
  • Received:2020-04-28 Online:2020-07-28 Published:2020-07-17

摘要: 建立了开关磁阻电机(SRM)无位置传感器控制的神经网络模型,以样机数据为样本对神经网络进行训练。经训练的神经网络模型能够在导通区间内检测出SRM样机转子位置,但在绕组非导通时无法检测。为得到完整转子位置,采用了单神经网络加积分的优化方法,并在适当的位置进行校正以消除积分误差。样机进行了神经网络无位置传感器控制实验,实验结果表明,该方法适用于不同控制方式下的无位置传感器控制。

关键词: 开关磁阻电机, 无位置传感器控制, 神经网络

Abstract: A neural network model for SRM sensorless control was established,and then trained by using the prototype data. The rotor position of SRM prototype can be detected by the trained neural network model in the conduction range, but it can't be detected when the winding is not conducted. In order to get the complete rotor position, single neural network plus integration method was applied,and the integration error was eliminated at the appropriate position. The neural network sensorless control method was carried out on the prototype. The experimental results showed that the proposed method can be used for SRM sensorless control under different control modes.

Key words: switched reluctance motor (SRM), sensorless control, neutral network

中图分类号: