基于改进Hilbert变换的车辆变速箱故障特征识别  

Fault Feature Recognition of Vehicle Gearbox Based on Improved Hilbert Transforming

作　　者：高辉 成传明 GAO-Hui;CHENG Chuan-ming(Fujian Industrial School,Fuzhou 350002,China;Hanjiang Normal University,Shiyan 442000,China)

机构地区：[1]福建工业学校,福建福州350002 [2]汉江师范学院,湖北十堰442000

出　　处：《汉江师范学院学报》2022年第3期15-21,共7页Journal of Hanjiang Normal University

基　　金：福建省教育科学“十三五”规划基金项目(项目编号:FJJK19-953)。

摘　　要：针对传统信号时频分析算法消噪效果差、无法应对端点效应及信号模态混叠的不足,提出基于改进的Hilbert变换的车辆变速箱故障特征识别算法.先构造信号的IMF分量,优化Hilbert算法的经验模态分解性能,并基于TVD方法对原始故障信号做消噪处理;利用信号波形曲线变化的斜率来预测处于端点处的信号变化趋势,抑制信号的端点效率和模态混叠现象,最后利用改进的Hilbert变换分解不同故障类型的特征奇异值准确识别故障特征.仿真结果表明提出算法在高、低频状态下都能准确提取变速箱的故障特征,且拥有高达98.38%的故障分类准确率.Aiming at the shortcomings of traditional signal time-frequency analysis algorithm,such as poor diagnosing effect,inability to deal with endpoint effect and signal mode aliasing,a fault feature recognition algorithm for vehicle gearbox based on improved Hilbert transforming is proposed.The IMF component of the signal is constructed to optimize the empirical mode decomposition performance of Hilbert algorithm,and the original fault signal is diagnosed based on TVD method;the change trend of the signal at the end point is predicted by using the change slope of the signal waveform curve,and the endpoint efficiency and mode aliasing phenomenon of the signal are suppressed.The improved Hilbert transforming is used to decompose the characteristic singularity of different fault types value can accurately identify fault features.The simulation results show that the proposed algorithm can accurately extract the fault features of gearbox under high and low frequency conditions,with a fault classification accuracy of 98.38%.

关 键 词：改进Hilbert变换 变速箱 IMF分量 TVD算法 奇异值分解 

分 类 号：U463.2[机械工程—车辆工程]