基于自适应Morlet小波和WOA方法的牵引齿轮箱故障诊断  

Fault Diagnosis of Traction Gear Box Based on Adaptive Morlet Wavelet and WOA Method

作　　者：杜延鹏 韩得水 吴连军 栾赛 曹朝煜 DU Yanpeng;HAN Deshui;WU Lianjun;LUAN Sai;CAO Chaoyu(CRRC Industrial Research Institute(Qingdao)Co.,Ltd.,Qingdao 266111,China)

机构地区：[1]中车工业研究院(青岛)有限公司,山东青岛266111

出　　处：《机械制造与自动化》2024年第6期160-163,共4页Machine Building & Automation

摘　　要：为提高轨道交通牵引传动系统用齿轮箱早期故障诊断能力,开发一种Morlet小波自适应参数字典算法,可以实现局部分割与整体数据进行全局分析的功能,通过鲸鱼优化算法自主计算小波字典数据。根据正交匹配追踪结果对振动信号开展稀疏分解,以包络谱分析的方法获取齿轮中的早期信号,实现齿轮箱的高效故障诊断。研究结果表明:仿真信号波形内形成了明显的周期故障冲击特征,采用该方法计算的Morlet小波参数极大地缩短了算法所需的时间。与CFA方法相比,该方法对原子小波参数具有更准确的识别性能,具备更强抗噪能力,算法效率也获得明显提升。该研究可以拓展到其他的机械传动系统上,具有很高的推广价值。In order to improve the early fault diagnosis ability of gear box for rail transit traction transmission system,a Morlet wavelet adaptive parameter dictionary algorithm is developed with the aim to realize the function of local segmentation and global analysis of the whole data.The whale optimization algorithm is used to compute the wavelet dictionary data autonomously.According to the results of orthogonal matching pursuit,the sparse decomposision of the vibration signals is carried out,and the envelope spectrum analysis method is applied to obtain the early signals in the gear box,achieving the efficient fault diagnosis of the gear box.The results show that obvious periodic fault impact features are formed in the simulation signal waveform,and the Morlet wavelet parameters calculated by the method greatly shorten the time required by the algorithm.Compared with CFA method,this method has more accurate identification performance for atomic wavelet parameters,stronger anti-noise ability,and remarkablely improved algorithm efficiency.The research can be extended to other mechanical transmission systems with a high value of popularization.

关 键 词：齿轮箱 故障诊断 稀疏表示 MORLET小波 鲸鱼优化算法 

分 类 号：TH132.41[机械工程—机械制造及自动化]