基于小波特征提取和概率神经网络的机械故障诊断  

Machine Fault Diagnosis Based on Wavelet Feature Extraction and Probabilistic Neural Networks

作　　者：王云松[1] 

机构地区：[1]江苏技术师范学院电气信息工程学院,江苏常州213001

出　　处：《江苏技术师范学院学报》2011年第6期20-24,共5页Journal of Jiangsu Teachers University of Technology

摘　　要：论述了基于离散小波变换系数的特征提取和概率神经网络在机械故障诊断中的应用。该方法利用离散小波获取振动信号各有效频带的能量作为故障参数,用概率神经网络构建设备运行状态模型,根据历史数据确定故障值并设置故障参数。实验结果从应用程序对轴承故障诊断表明,相比传统方法,该方法能够有效地提取测试信号内在的重要信息内容,并增加机械整体故障诊断的准确性,在机械设备故障处理系统中有良好的应用前景。In this paper,a feature extraction from the coefficients of a discrete wavelet transform and probabilistic neural networks are proposed for machine fault diagnosis.This proposed method is that the energy information in every frequency band obtained by the discrete wavelet decomposition is used as the fault diagnosis parameters,and the improved probabilistic neural network is used to construct the model of real running condition,the fault parameters is determined according to history data from the running machine.The experimental results from the application on bearing fault diagnosis have shown that the proposed approach is able to effectively extract important intrinsic information content of the test signals and increase the overall fault diagnostic accuracy,as compared to conventional methods.It has very good application prospect in the alarm processing system of machinery.

关 键 词：故障诊断 小波特征提取 概率神经网络 

分 类 号：TB533.1[理学—物理] TP206.3[理学—声学]