• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2021, Vol. 57 ›› Issue (9): 78-88.doi: 10.3901/JME.2021.09.078

• 机械动力学 • 上一篇    下一篇

扫码分享

基于自适应增强差分积形态滤波器的滚动轴承故障特征提取方法

苗宝权1,2, 陈长征1,2, 罗园庆1,2, 赵思雨1,2   

  1. 1. 沈阳工业大学机械工程学院 沈阳 110870;
    2. 沈阳工业大学辽宁省振动噪声控制工程技术研究中心 沈阳 110870
  • 收稿日期:2020-05-09 修回日期:2020-12-28 出版日期:2021-05-05 发布日期:2021-06-15
  • 通讯作者: 陈长征(通信作者),男,1964年出生,博士,教授,博士研究生导师。主要研究方向为振动噪声治理及故障诊断。E-mail:czchen@sut.edu.cn
  • 作者简介:苗宝权,男,1994年出生。主要研究方向为机械设备故障诊断与信号处理方向。E-mail:bqmiao512@163.com
  • 基金资助:
    国家自然科学基金资助项目(51675350,51575361)。

Rolling Bearing Fault Feature Extraction Method Based on Adaptive Enhanced Difference Product Morphological Filter

MIAO Baoquan1,2, CHEN Changzheng1,2, LUO Yuanqing1,2, ZHAO Siyu1,2   

  1. 1. School of Mechanical Engineering, Shenyang University of Technology, Shenyang 110870;
    2. Liaoning Vibration and Noise Control Engineering Research Center, Shenyang University of Technology, Shenyang 110870
  • Received:2020-05-09 Revised:2020-12-28 Online:2021-05-05 Published:2021-06-15

摘要: 为了在强背景噪声下提取滚动轴承微弱的故障特征信息。提出一种新的自适应增强差分积形态滤波方法(Adaptive enhanced difference product morphological filter,AEDPO)用于滚动轴承早期的故障诊断。首先,结合已有的四种形态学滤波算子滤波的能力,提出一种改进的增强差分积形态滤波算子(Enhanced difference product morphological filter operation,EDPO),该算子具有在强背景噪声下提取周期性脉冲特征的能力。随后,针对形态滤波过程中最优的结构元素(Structuring element,SE)尺度选择问题,提出一种新的自适应选择策略,名为峭度特征能量积(Kurtosis feature energy product,KF)。最后,EDPO算子凭借最优的SE尺度进行滤波处理,提取滚动轴承早期的故障特征。通过对仿真信号和实测滚动轴承内圈故障信号进行分析,结果表明AEDPO方法能够有效地在强背景噪声中提取滚动轴承微弱的故障特征,对比于传统的形态滤波方法更能体现该方法的准确性和优越性。

关键词: 形态滤波, 滚动轴承, 故障特征提取, 故障诊断

Abstract: In order to extract weak fault feature information of rolling bearings under strong background noise. A new adaptive enhanced different product operation morphological filter (AEDPO) is proposed for early fault diagnosis of rolling bearings. First, combining the existing filtering capabilities of the four morphological filtering operators, an improved enhanced differential product morphological filtering operator (EDPO) is proposed, which has the ability to extract periodic pulse features under strong background noise. Then, a new adaptive selection strategy is proposed for the optimal selection of structuring elements (SE) scale in the morphological filtering process, which is called the kurtosis feature energy product (KF). Finally, the EDPO operator performs filtering based on the optimal SE scale to extract early fault characteristics of rolling bearings. Through the analysis of the simulation signal and the actual fault signal of the inner ring of the rolling bearing, the results show that the AEDPO method can effectively extract the weak fault characteristics of rolling bearings from strong background noise, and it can better reflect its accuracy and superiority than the traditional morphological filtering method.

Key words: morphological filtering, rolling bearing, fault feature extraction, fault diagnosis

中图分类号: