北京航空航天大学学报 ›› 2018, Vol. 44 ›› Issue (2): 241-249.doi: 10.13700/j.bh.1001-5965.2017.0114

• 论文 • 上一篇    下一篇

基于ITD和改进形态滤波的滚动轴承故障诊断

余建波1, 吕靖香1, 程辉2, 孙习武2, 吴昊3   

  1. 1. 同济大学 机械与能源工程学院, 上海 201804;
    2. 上海航天设备制造总厂, 上海 201100;
    3. 山东省特种设备检验研究院, 济南 250101
  • 收稿日期:2017-03-02 修回日期:2017-04-24 出版日期:2018-02-20 发布日期:2017-09-18
  • 通讯作者: 余建波,E-mail:jbyu@tongji.edu.cn E-mail:jbyu@tongji.edu.cn
  • 作者简介:余建波,男,博士,教授,博士生导师。主要研究方向:设备智能预诊维护与可靠性、模式识别与机器学习;吕靖香,女,硕士研究生。主要研究方向:故障诊断和信号处理。
  • 基金资助:
    国家自然科学基金(51375290,71777173);上海航天科技创新基金(SAST2015054);中央高校基本科研业务费专项资金(22120180068)

Fault diagnosis for rolling bearing based on ITD and improved morphological filter

YU Jianbo1, LYU Jingxiang1, CHENG Hui2, SUN Xiwu2, WU Hao3   

  1. 1. School of Mechanical Engineering, Tongji University, Shanghai 201804, China;
    2. Shanghai Aerospace Equipment Manufacturing Factory, Shanghai 201100, China;
    3. Special Equipment Inspection and Research Institute of Shandong Province, Jinan 250101, China
  • Received:2017-03-02 Revised:2017-04-24 Online:2018-02-20 Published:2017-09-18
  • Supported by:
    National Natural Science Foundation of China (51375290, 71777173);Shanghai Aerospace Science and Technology Innovation Fund (SAST2015054);the Fundamental Research Funds for the Central Universities (22120180068)

摘要: 为从受谐波和随机噪声干扰的振动信号中提取出故障冲击成分,融合四大基本形态学算子提出了改进形态滤波方法——平均组合差值形态滤波(ACDIF)方法,同时与固有时间尺度分解(ITD)相结合,并将ITD-ACDIF方法应用到滚动轴承的故障诊断中。首先,对轴承振动信号进行ITD分解得到一系列旋转分量(PRC);然后,以峭度为准则筛选出含故障信息丰富的有效PRC,对每个有效分量进行ACDIF滤波提取冲击成分进行信号重构;最后,利用频谱分析提取重构信号中的故障特征。数值仿真和轴承故障振动信号的试验结果表明,本文方法可有效滤除谐波干扰,提取强背景噪声下的冲击故障特征,实现设备的故障诊断。

关键词: 轴承故障, 固有时间尺度分解(ITD), 旋转分量(PRC), 改进形态滤波, 故障诊断

Abstract: For extracting fault impulse components embedded in the vibration signal with much noise and harmonics, an improved morphological filter algorithm called average combination difference morphological filter (ACDIF) is proposed on the basis of the four basic morphological operators. Then ACDIF is combined with the intrinsic time scale decomposition (ITD) and the ITD-ACDIF method is employed in the fault diagnosis for rolling bearing. In the ITD-ACDIF fault diagnosis method, ITD is applied to the original vibration signal and a series of proper rotation components (PRC) are obtained, and then the kurtosis is regarded as criterion to select effective PR components which contain much fault-related information. After that, ACDIF filtering is performed on each effective PR in order to pick up bidirectional impulses, and filtered PRs are combined into a signal. Finally, the fault feature is extracted from reconstructed signal by amplitude spectrum. The experimental results on simulated signal and actual bearing vibration signal demonstrate that the proposed method can effectively suppress noise interference and extracting the characteristics of impact fault under the strong background noise to realize fault diagnosis of equipment.

Key words: bearing fault, intrinsic time scale decomposition (ITD), proper rotation component (PRC), improved morphological filter, fault diagnosis

中图分类号: 


版权所有 © 《北京航空航天大学学报》编辑部
通讯地址:北京市海淀区学院路37号 北京航空航天大学学报编辑部 邮编:100191 E-mail:jbuaa@buaa.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发