基于EMD-SVD模型和SVM滚动轴承故障模式识别

Fault Pattern Recognition of Rolling Bearing Based on EMD-SVD Model and SVM

针对滚动轴承振动信号的非平稳特性和在现实条件下难以获取大量故障样本的实际情况,提出一种经验模态分解、奇异值分解、Renyi熵和支持向量机相结合的故障诊断方法。运用经验模态分解方法对其去噪信号进行分析,利用互相关系数准则对固有模式分量进行筛选,再对所选分量重构相空间得到吸引子轨道矩阵;对矩阵进行奇异值分解求取奇异值,再计算这些奇异值的Renyi熵以组成故障特征向量,并将其作为支持向量机的输入以识别滚动轴承的故障类型。最后,利用实际滚动轴承试验数据的诊断与对比试验验证了该方法的有效性和泛化能力。

According to the non-stationarity characteristics of the vibration signals from rolling bearing and the difficulty for obtaining enough fault samples,a comprehensive fault diagnosis method based on Empirical Mode Decomposition（EMD）,Singularity Value Decomposition（SVD）,Renyi-entropy and Support Sector Machine（SVM） is proposed.Firstly,the denoised vibration signals are decomposed into a finite number of Intrinsic Mode Functions（IMF）.Secondly,some IMF components are selected according to the criterion of mutual correlation coefficient between IMF components and denoised signal.Thirdly,the phase space of the selected IMF components is reconstructed so as to obtain the attractor orbit matrix.Fourthly,with the SVD method,singular value sequences are obtained,and then Renyi-entropies of these sequences are calculated as faulty eigenvector.Finally,the eigenvector serves as input of SVM classifier so that the faults of rolling bearing are recognized.Practical rolling bearing experiment data is used to verify this method,and the diagnosis results and comparative tests fully validate its effectiveness and generalization ability.