基于CEEMDAN–MPE算法的地铁短波长钢轨波磨识别  

Identification Method for Metro Rail Corrugation Based on CEEMDAN–MPE Algorithm

作　　者：马添翼 李伟[1] 陈建政[1] 赖伊雯 杨吉忠[2] MA Tianyi;LI Wei;CHEN Jianzheng;LAI Yiwen;YANG Jizhong(State Key Laboratory of Traction Power,Southwest Jiaotong University,Chengdu 610031,China;Research Institute of Science and Technology,China Railway Eryuan Engineering Group Co.,Ltd.,Chengdu 610031,China)

机构地区：[1]西南交通大学牵引动力国家重点实验室,成都610031 [2]中铁二院工程集团有限责任公司科学技术研究院,成都610031

出　　处：《噪声与振动控制》2023年第2期120-126,153,共8页Noise and Vibration Control

基　　金：国家自然科学基金资助项目(52002343)。

摘　　要：针对地铁钢轨波磨检测效率低、准确率低等问题,提出一种基于自适应噪声完备集合经验模态分解–多尺度排列熵的地铁短波长钢轨波磨识别方法。首先采用自适应噪声完备集合经验模态分解(CEEMDAN)方法将轴箱振动加速度信号逐级分解为多个不同特征时间尺度相对平稳的固有模态函数(IMF);引入多尺度排列熵(MPE)分析各IMF在不同尺度上的熵均值,计算各IMF的相关系数,然后将随机程度较大或信噪比较低的分量剔除,计算剩余IMF的能量值及总能量值,通过设定阈值判断是否存在钢轨波磨,最后对能量值超过阈值的分量进行小波包时频分析,提取振动中心频率确定钢轨波磨波长。结果表明:(1)基于CEEMDAN–MPE算法可通过轴箱振动加速度准确识别短波长钢轨波磨及发生区域;(2)该方法在识别短波长波磨的准确性上优于EEMD能量值和小波包能量熵方法,并且能够较为准确地提取短波长波磨不平顺导致的振动响应,便于针对波磨的波深做进一步分析。To solve the problem of low efficiency and low accuracy in current subway rail corrugation detection,this paper proposes a short-wavelength subway rail corrugation recognition method based on adaptive noise complete ensemble empirical mode decomposition(CEEMDAN)and multi-scale permutation entropy(MPE).First of all,the CEEMDAN method is used to decompose the axle box vibration acceleration signal into several relatively stable intrinsic mode functions(IMF)with different characteristic time scales,and the MPE is introduced to analyze the average entropy of each IMF on different scales,so that the correlation coefficients of each IMF are calculated.Then,the components with greater randomness or low signal-to-noise ratio are removed and the energy value and total energy value of the remaining IMFs are calculated.The threshold value is set in order to judge whether the rail corrugation exists.Finally,the wavelet packet timefrequency analysis is used to the components whose energy value exceeds the threshold value,and the vibration central frequency is extracted to determine the rail corrugation wavelength.The results show that(1)the CEEMDAN-MPE algorithm can accurately identify the short-wavelength rail corrugation and its location through the acceleration of the axle box vibration;(2)the method has better accuracy than EEMD energy value and wavelet packet energy entropy methods in identifying short-wavelength rail corrugation,and is able to more accurately extract the vibration response caused by irregularities of the short-wavelength rail corrugation,which facilitates further analysis of the wave depth of the rail corrugation.

关 键 词：故障诊断 地铁钢轨波磨 自适应噪声完备集合经验模态分解 多尺度排列熵 轴箱振动加速度 

分 类 号：U211.3[交通运输工程—道路与铁道工程]