基于EMD-1(1/2)维谱熵的滚动轴承故障诊断  

Fault Diagnosis of Rolling Bearings Based on EMD 1(1/2)-Dimensional Spectral Entropy

作　　者：宋平岗[1] 周军[1] 

机构地区：[1]华东交通大学电气与电子工程学院,南昌330013

出　　处：《科学技术与工程》2014年第2期149-153,共5页Science Technology and Engineering

摘　　要：为了准确地诊断出滚动轴承的运行状态,将1(1/2)维谱熵引入滚动轴承故障诊断中。先对滚动轴承原始故障信号进行EMD(empirical mode decomposition)分解得到若干个固有模态函数(intrinsic mode function,IMF),进而再求取各个IMF的1(1/2)维谱熵值,作为表征滚动轴承故障类型的特征向量。将其作为Elman神经网络的输入参数,最后区分滚动轴承故障状态和故障类型。仿真分析和实验研究表明,该方法能够有效地提取出滚动轴承的故障特征,最后通过与小波包分析-BP神经网络故障诊断方法对比,显示出其具有更高的识别率,更加表明其可行性和有效性。A method to diagnose the faults of roiling bearings accurately was presented, into which the 1 （1/2） - dimensional spectral entropy is introduced. First, the original fault signal goes through empirical mode decomposi-tion （EMD） and gets some intrinsic mode functions （IMF）. Second, the 1 （1/2）-dimensional spectral entropy val-ue was calculated by the IMF we get. Then the value is input to the Elman Neural Network as a new eigenvector to characterize the fault type of the rolling bearing. At last, the fault status and fault type of the roiling bearing were distinguished. Simulation analysis and experimental study show that this method can effectively extract the fault fea- tures of rolling bearings. Compared with the Wavelet Packet Analysis-Neural Network fault diagnosis, this method is more feasible and effective with a higher recognition rate.

关 键 词：EMD(empirical MODE decomposition) 形态滤波 ELMAN神经网络 滚动轴承 故障诊断 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]