基于WPT-ESCDE的电气设备运输车轮对轴承故障特征提取方法  

Fault feature extraction method of wheel set bearings of rail transit vehicles of electric equipment based on WPT-ESCDE method

作　　者：张敏 万书亭[2] 王萱 蔡伟 张雄[2] ZHANG Min;WAN Shuting;WANG Xuan;CAI Wei;ZHANG Xiong(Super High Voltage Branch of State Grid Zhejiang Electric Power Co.,Ltd.,Hangzhou 310000,Zhejiang,China;Department of Mechanical Engineering,North China Electric Power University,Baoding 071003,Hebei,China;School of Mechanical Engineering,Yanshan University,Qinhuangdao 066004,Hebei,China)

机构地区：[1]国网浙江省电力有限公司超高压分公司,浙江杭州310000 [2]华北电力大学机械工程系,河北保定071003 [3]燕山大学机械工程学院,河北秦皇岛066004

出　　处：《中国工程机械学报》2023年第2期183-188,共6页Chinese Journal of Construction Machinery

基　　金：国网浙江省电力有限公司科技项目(B311MR210003)。

摘　　要：变压器等电气设备的吊装、转运环节是疏于监控的薄弱环节,极易发生机械冲击引起的二次损伤,而轨道运输车的轮对承载特性及轮对轴承运行状态关乎运输安全。综合考虑轨道运输车轮对轴承运输环境,分析振动信号中存在的主要成分及特征,提出一种基于小波包-包络谱相关散布熵(WPT-ESCDE)的故障特征提取方法。首先,对振动信号的离散时间序列进行小波包分解,并对小波包子带系数进行重构;其次,对每个小波包子带计算平方包络谱,得到离散频率序列,将得到的小波包子带包络谱离散序列看作广义时间序列进行相关分析,得到包络谱相关函数;最后,计算包络谱相关函数的散布熵,筛选最优小波包子带序列进行特征提取。通过仿真分析和QPZZ-Ⅱ旋转机械故障模拟实验台实测信号验证了所提方法的有效性。The hoisting and transportation of electrical equipment such as transformers are weak links that are neglected in monitoring,and secondary damage caused by mechanical impact is very easy to occur.The bearing characteristics of wheel sets and the running state of wheel set bearings of rail transit vehicles are related to the transportation state.In this paper,the main components and characteristics of vibration signals are analyzed,and a fault feature extraction method based on wavelet packet transform envelope spectrum correlation dispersion entropy(WPT-ESCDE)is proposed.Firstly,the discrete time series of vibration signals are decomposed by wavelet packet,and the wavelet packet subband coefficients are reconstructed.Then,the square envelope spectrum is calculated for each wavelet packet subband to obtain the discrete frequency series.The obtained wavelet packet subband envelope spectrum discrete series is regarded as a generalized time series for correlation analysis,and the envelope spectrum correlation function is obtained.Finally,the dispersion entropy of the envelope spectrum correlation function is calculated,and the most wavelet packet subband sequence is selected for feature extraction.The effectiveness of the proposed method is verified by simulation analysis and experimental signals of QPZZ-II rotating machinery simulation test-bed.

关 键 词：电气设备轨道运输车 轮对轴承 振动信号分析 小波包 包络谱相关散布熵 

分 类 号：TP206[自动化与计算机技术—检测技术与自动化装置]