基于BP网络的减压阀故障诊断

Abstract

提出一种使用bP神经网络检测减压阀故障信号的方法。通过提取减压阀振动时的正常信号和故障信号的均值、标准差、偏度和峰度,作为特征值对建立的bP网络进行训练,再进行故障辨识,取得了令人满意的结果。实验结果证明,利用bP网络进行机械故障检测是可行的。

A method of reducing valve fault diagnosis was proposed based on BP neural network.The BP network was built and trained through eigenvalues,such as mean,standard variance,skewness and kurtosis of normal signals and fault signals extracted from reducing valve's vibration.Then fault identifications were made and the results were satisfactory.The experimental results indicate it is feasible to carry out fault diagnosis using BP neural network.