计算机工程与应用 ›› 2014, Vol. 50 ›› Issue (11): 267-270.

• 工程与应用 • 上一篇    

SOM和Elman神经网络在整流器故障诊断的应用

康洪铭1,2,李光升2,谢永成2,魏  宁2   

  1. 1.中国空气动力研究与发展中心低速所,四川 绵阳 621000
    2.装甲兵工程学院 控制工程系,北京 100072
  • 出版日期:2014-06-01 发布日期:2015-04-08

Application of SOM and ELMAN neural network on fault diagnosis of bridge rectifier

KANG Hongming1,2, LI Guangsheng2, XIE Yongcheng2, WEI Ning2   

  1. 1.Low Speed Aerodynamic Research and Development Centers in China, Mianyang, Sichuan 621000, China
    2.Department of Mechanical Engineering, Academy of Armored Forced Engineering, Beijing 100072, China
  • Online:2014-06-01 Published:2015-04-08

摘要: 针对装甲车辆电源系统整流器内部二极管的断路和短路故障,提出了一种基于SOM和Elman神经网络相结合的诊断方法。通过建立整流器的仿真模型,利用快速傅里叶变换(FFT)提取各故障模式的谐波次数和幅值,并用SOM网络进行模式分类,由于各模式下具体故障类型存在相位差,通过采样其周期内的电压值,再利用Elman网络可以识别具体故障。从仿真结果来看,实现了整流器的模式分类和故障识别,验证了该方法的可行和正确性。

关键词: 整流器, 自组织映射(SOM), Elman, 故障诊断

Abstract: A new diagnostic method based on SOM and Elman neural network for open or short faults of inner diodes in rectifier of armored vehicle power system is proposed. Through establishing the rectifier mode, FFT is used to get each fault mode’s harmonious number and magnitude, and the modes are classed by SOM network. Considering the phase difference between material faults of some mode, the voltage value is sampled in period, and the faults are identified by Elman network. In results, this method achieves the goal of mode class and fault identify, moreover, it is feasible and correct.

Key words: rectifier, Self-Organizing Map(SOM), Elman, fault diagnosis