基于LCD与奇异值能量差分谱的齿轮故障诊断方法

Fault Diagnosis Method of Gears based on LCD and Energy Difference Spectrum of Singular Values

齿轮是机械传动中的重要组成部分,其故障的发生已经成为影响设备可靠、稳定运行的主要因素。提出一种基于改进局部特征尺度分解(Local Characteristic-scale Decomposition,LCD)与奇异值能量差分谱(Energy Difference Spectrum of Singular Value,EDSSV)的齿轮故障诊断方法。首先,利用支持向量回归(Support Vector Regression,SVR)对信号进行延拓处理,抑制LCD分解过程中产生的端点效应,分析改进后LCD算法的精确性和可靠性;然后结合奇异值能量差分谱降噪理论,有效剔除各ISC中噪声成分,重构信号频谱,提高信噪比;最后计算分解得到的内禀尺度分量(Intrinsic Scale Component,ISC)模糊熵(Fuzzy Entropy,FE)特征集,利用支持向量机(Support Vector Machine,SVM)进行分类。实验研究表明,提出的基于改进LCD与奇异值能量差分谱的齿轮故障诊断方法能有效诊断出齿轮故障类型。

Gear is an important part of mechanical transmission,its fault has become a major factor affecting the reliability and stability of the equipment.In this paper,a fault diagnosis method of gears based on improved local characteristic-scale decomposition(LCD)and energy difference spectrum of singular values(EDSSV)is proposed.Firstly,support vector regression(SVR)is used for signal continuation processing to restrain the endpoint effects in LCD decomposition.And the accuracy and reliability of the improved LCD algorithm are analyzed.Then,combining with the theory of EDSSV de-noising,the noise components are eliminated and the signal spectrum is reconstructed so that the signalto-noise ratio is raised.Finally,the intrinsic scale component(ISC)fuzzy entropy characteristic sets are computed by decomposing the ISC and classified by the support vector machine(SVM).Experimental results show that the proposed gear diagnosis method based on LCD and EDSSV can effectively identify the fault types of gears.