噪声参数最优ELMD与LS-SVM在轴承故障诊断中的应用与研究

Application of noise parametric optimization with ELMD and LS-SVM in bearing fault diagnosis

针对轴承振动信号的非平稳特征和现实中难以获得大量典型故障样本,提出基于噪声参数最优的总体局部均值分解(Ensemble Local Mean Decomposition,ELMD)与最小二乘支持向量机(Least Squares Support Vector Machine,LS-SVM)相结合的轴承故障诊断方法。首先对轴承振动信号进行噪声参数最优ELMD分解并得到一系列窄带乘积函数(Product Function,PF),然后计算各PF分量能量以构造能量特征向量,最后将高维能量特征向量作为最小二乘支持向量机的输入来识别轴承故障类型。通过对轴承故障振动信号分析,结果表明噪声参数最优ELMD方法能有效地抑制模态混叠,与LS-SVM结合可以准确地识别轴承的工作状态和故障类型。

Aiming at non-stationary features of bearing vibration signals and difficulties to obtain a large number of fault samples in reality, a method of bearing fault diagnosis based on noise parametric optimization with ensemble local mean decomposition （ ELMD ） and least squares support vector machine （ LS-SVM ） was proposed. Firstly, a bearing vibration signal was decomposed into a series of narrow band product functions （ PFs） using the optimal noise parameters ELMD method. Then, the energy of each PF was calculated to construct energy feature vectors. Finally, the high-dimensional energy feature vectors were taken as inputs of LS-SVM to identify bearing fault types. The results of bearing fault vibration signals analysis indicated that the optimal noise parameters ELMD method can suppress mode-mixing effectively and this approach combined with LS-SVM can identify operating conditions and fault types of bearings correctly.