改进经验模态分解的电机轴承特征提取方法  

Feature Extraction Method of Motor Bearing Based on Improved Empirical Mode Decomposition

作　　者：张习习 顾幸生 ZHANG Xi-xi;GU Xing-sheng(Key Laboratory of Advanced Control and Optimization for Chemical Process,Ministry of Education,East China University of Science and Technology,Shanghai 200237,China)

机构地区：[1]华东理工大学,化工过程先进控制和优化教育部重点实验室,上海200237

出　　处：《控制工程》2020年第11期1882-1891,共10页Control Engineering of China

基　　金：国家自然科学基金(61573144)。

摘　　要：滚动轴承是电机的重要组成部分,也是出现故障较多的部位,对电机轴承进行特征提取方法研究对于其故障诊断具有重要的意义。首先,介绍了排列熵及其改进算法,并应用于电机轴承故障检测;其次,针对经验模态分解(Empirical Mode Decomposition,EMD)的端点效应和模态混叠问题,提出了极值点对称延拓加聚合经验模态分解以及解相关算法和改进排列熵算法的混合算法(Modified Empirical Mode Decomposition,MEMD),并应用于电机轴承故障特征提取;最后,利用概率神经网络(Probabilistic Neural Network,PNN)对电机轴承故障特征进行了分类。结果表明,利用改进后的经验模态分解算法进行特征提取,可以明显提高故障分类的准确率。Rolling bearing is an important part of the motor,but also the site prone to failure.The research on the feature extraction method of motor bearing is of great importance for fault diagnosis of motor bearing.First,permutation entropy algorithm and its improved algorithm are introduced in this paper,and are used for motor bearing fault detection.Secondly,for endpoint effect and the modal aliasing of empirical mode decomposition(EMD),a hybrid method(Modified Empirical Mode Decomposition,MEMD)combines extreme point symmetric continuation method,convergent empirical mode decomposition,decorrelation algorithm and improved permutation entropy is put forward in this paper,and is applied to the motor bearing fault feature extraction.At the end of the paper the probabilistic neural network(PNN)is applied on the motor bearing fault feature classification.The results show that using the improved empirical mode decomposition algorithm for feature extraction can significantly improve the accuracy of fault classification.

关 键 词：改进排列熵 模态混叠 端点效应 解相关 改进经验模态分解 概率神经网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]