基于小波熵和ANFIS的轴承故障诊断方法  

Fault Diagnosis of Rolling Bearing Based on Wavelet Entropy and ANFIS

作　　者：隋文涛[1] 张丹[2] 张宇[1] 

机构地区：[1]山东理工大学机械工程学院,山东淄博255049 [2]山东理工大学电气与电子工程学院,山东淄博255049

出　　处：《机床与液压》2010年第11期137-139,共3页Machine Tool & Hydraulics

基　　金：山东省自然科学基金资助项目(Y2005F12)

摘　　要：提出一种基于小波熵和自适应神经模糊推理系统(ANFIS)的滚动轴承故障诊断方法。对原始振动信号进行小波包分解,提取小波熵特征,然后用ANFIS作为分类器进行故障模式识别。比较人工神经网络(ANN)与ANFIS的故障诊断效果。结果表明,该算法诊断准确率高于神经网络。A fault diagnosis method of rolling bearings based on wavelet entropy and adaptive neuro-fuzzy inference system（ANFIS）was presented.The vibration signal was decomposed using wavelet packet,and wavelet entropy was calculated and selected as feature.ANFIS model was used as classifier to identify different faults.For comparison purposes,artifical neural network method was investigated.The results indicate that the rate of accuracy of fault diagnosis of ANFIS model is higher than that of artifical neural network.

关 键 词：故障诊断 自适应神经模糊推理系统 小波熵 神经网络 

分 类 号：TP306[自动化与计算机技术—计算机系统结构]