电子学报 ›› 2021, Vol. 49 ›› Issue (11): 2217-2224.DOI: 10.12263/DZXB.20201086

• 学术论文 • 上一篇    下一篇

强噪声背景下动车组轴承微弱故障信号检测

孙鑫威1, 纪爱敏1, 陈曦晖1, 林新海2, 许行2   

  1. 1.河海大学机电工程学院,江苏 常州 213000
    2.中车戚墅堰机车车辆工艺研究所有限公司,江苏 常州 213000
  • 收稿日期:2020-09-30 修回日期:2021-04-01 出版日期:2021-11-25 发布日期:2021-11-25
  • 作者简介:孙鑫威 男,1997年3月出生于江苏省无锡市,现为河海大学硕士研究生,主要研究方向为机械故障诊断.E-mail:sxw727290457@163.com
    纪爱敏(通信作者) 男,1965年3月出生于安徽安庆,2001于中国科学技术大学获得博士学位,现为河海大学机电工程学院教授,博士生导师.主要研究方向为机械故障诊断理论与技术、优化设计理论等.E-mail:jiam@hhuc.edu.cn

Detection of Weak Fault Signals for EMU Bearings Under Strong Noise

Xin-wei SUN1, Ai-min JI1, Xi-hui CHEN1, Xin-hai LIN2, Xing XU2   

  1. 1.School of Mechanical and Electrical Engineering,Hohai University,Changzhou,Jiangsu 213000,China
    2.CSR Qishuyan Locomotive and Vehicle Technology Research Institute Co. ,Ltd. ,Changzhou,Jiangsu 213000,China
  • Received:2020-09-30 Revised:2021-04-01 Online:2021-11-25 Published:2021-11-25

摘要:

动车在高速行驶中,齿轮箱轴承易发生裂纹、点蚀等故障.为了在故障发生的初期检测出微弱的故障频率成分,本文提出了一种基于小波降噪预处理的周期势振动共振的轴承故障诊断方法.利用小波包提取轴承的固有共振频带,重构提取出的信号,滤除其中的强噪声干扰,随后将信号输入周期势振动共振系统,增强了故障特征.同时,本文建立了考虑振动共振系统中高频激励信号幅值的优化模型,并采用蚁群算法实现了其参数的自适应优化,得到输出信号后将其转化到频域分析,从而检测出轴承早期故障.实例分析表明,所提方法的数据处理结果相比单独采用随机共振的结果更精确,误差缩减至0.3%.

关键词: 齿轮箱轴承, 微弱故障信号, 小波降噪, 周期势振动共振, 蚁群算法

Abstract:

In the high-speed motion of the motor car, the bearing of the gearbox is prone to cracks, pitting and other failures. In order to detect the weak fault frequency component in the early stage of fault occurrence, a bearing fault diagnosis method based on wavelet de-noising preprocessing and periodic potential vibration resonance feature-enhancing is proposed. Wavelet packet is used to extract the natural resonance frequency band of the bearing, reconstruct the extracted signal, filter out the interference of the strong noise, and then input the signal into the periodic potential vibration resonance system for enhancing the fault characteristics. Meanwhile, an optimization model considering the amplitude of the high-frequency excitation signal in the vibration excitation system is established, and the ant colony algorithm is used to adaptively optimize the parameters. After obtaining the output signal, it is converted into frequency domain analysis to detect early failure of the bearing. The example analysis shows that the error of the proposed method is reduced to 0.3% compared with the result of stochastic resonance.

Key words: gearbox bearings, weak fault signal, wavelet de-noising, periodic potential vibration resonance, ant colony algorithm

中图分类号: