小波局部极大模方法在轴心轨迹辨识中的应用研究  

APPLY WAVELET MODULUS MAXIMA TO IDENTIFY SHAFT CENTER ORBIT

作　　者：彭志科[1] 何永勇[1] 卢 青[1] 陈真勇[1] 褚福磊[1] 

机构地区：[1]清华大学精密仪器与机械学系,北京100084

出　　处：《机械工程学报》2002年第7期6-11,共6页Journal of Mechanical Engineering

基　　金：国家自然科学基金(50105007);国家"九五"攀登计划(PD9521908Z2)资助项目。

摘　　要：提出了一种新的基于小波局部极大模的轴心轨迹辨识方法。该方法通过对轴心轨迹的θ-s曲线进行连续小波变换,提取小波局部极大模线,计算每条局部极大模线对应的Lipschitz指数α。最后用α为正的局部极大模线的条数、正α的平均值、α为负的局部极大模线的条数和负α的平均值等4个量作为轴心轨迹特征,并以这些特征量作为BP网络的输入进行分类识别。采用径向碰摩、联轴器不对中、油膜涡动和转子不平衡等4种故障的试验数据对该方法进行检验。结果表明,该方法有特征量少、神经网络的训练速度快和系统识别正确率高的优点。A novel approach for identification of shaft center orbits is proposed, which bases on wavelet modulus maxima. Inthis approach, continuous wavelet transform is put on Θ-S diagram of shaft center orbits, then, maxima lines are extracted and Lipschitz exponents αof every maxima line are calculated by non-linear least square method. The number of maxima lines with positive α, the mean of the positiveα, the number of maxima lines with negative αand the mean of the negative αare used as features of the shaft center orbits. In succession, the four features are used as the inputs of BP network to classify shaft center orbits. The experimental data of four kinds of faults: rub-impact, oil whirl, coupling misalignment and unbalance are used to test the wavelet modulus maxima based method. The test result indicates that the method can fulfill classification accurately and the method possesses advantages of few features used and of fast training processing of network.

关 键 词：小波变换 局部极大模 Lipschitz指数 轴心轨迹 Θ-s曲线 旋转机械 故障诊断 

分 类 号：TH17[机械工程—机械制造及自动化] TP277[自动化与计算机技术—检测技术与自动化装置]