微特电机 ›› 2018, Vol. 46 ›› Issue (5): 44-47.doi: 1004-7018-46-5-44-47

• 设计分析 • 上一篇    下一篇

基于RBF神经网络的无刷直流电动机故障诊断

王炜,钟书辉   

  1. 中国空空导弹研究院,洛阳 471009
  • 收稿日期:2018-01-30 出版日期:2018-05-28 发布日期:2018-05-28
  • 作者简介:王炜(1978—),男,博士,研究方向为伺服系统控制。
  • 基金资助:
    航空基金项目(2016ZD12028)

The Fault-Detection System with Brushless DC Motor Based on the Wavelet

WANG Wei, ZHONG Shu-hui   

  1. China Airborne Missile Academy, Luoyang 471009,China
  • Received:2018-01-30 Online:2018-05-28 Published:2018-05-28

摘要:

为尽可能检测空空导弹舵机多种故障,提出采用导弹舵机三相工作电流作为舵机故障诊断的信号源,针对舵机无刷直流电动机驱动器的开路、短路等故障,在MATLAB中构建无刷直流电动机及驱动器的模型,并进行故障仿真。选取Haar小波基函数对故障信号进行提取、分析和处理,利用提取的故障特征值对起到故障识别作用的RBF神经网络进行训练。仿真结果证明,训练得到的RBF神经网络能够有效识别出舵机中无刷直流电动机驱动器的故障,表明该方法的正确性。

关键词: 故障诊断, 无刷直流电动机, 小波分析, RBF神经网络, MATLAB

Abstract:

In order to detect more servo system's faults of air-to-air missile, a method was proposed that the three-phase current of motor in the servo system of missile was chosen to be the fault-detection signal. A fault model of the brushless DC motor and driver were built under MATLAB. The fault signals were analyzed by the wavelet and transferred to the RBF network which was used to train the analyzed signals. The test results of RBF showed the method could effectively distinguish the motor driver's fault.

Key words: fault-detection, brushless DC motor, wavelet, RBF network, MATLAB

中图分类号: