• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2026, Vol. 62 ›› Issue (18): 125-133.doi: 10.3901/JME.260955

• 特邀专栏:大模型驱动的高端装备智能运维 • 上一篇    

扫码分享

基于数据动态增强的滚动轴承全寿命演化过程监测方法

孔德同1,2, 张明睿1, 李乃鹏1, 骆南枝1, 张乐平1,2, 雷亚国1   

  1. 1. 西安交通大学现代设计及转子轴承系统教育部重点实验室 西安 710049;
    2. 华电电力科学研究院有限公司 杭州 310030
  • 收稿日期:2025-09-25 修回日期:2026-06-30 发布日期:2026-11-20
  • 作者简介:孔德同,男,1988年出生,高级工程师。主要从事旋转机械振动信号分析、故障诊断及智能运维研究。E-mail:dt_kong@163.com;张明睿,男,2002年出生,硕士研究生。主要从事风电装备智能运维理论与方法研究。E-mail:zhangmr@stu.xjtu.edu.cn;李乃鹏(通信作者),男,1991年出生,博士,教授,博士研究生导师。主要从事机械装备状态监测、寿命预测与智能运维研究。E-mail:naipengli@mail.xjtu.edu.cn
  • 基金资助:
    国家自然科学基金(52375121,52025056,62233017)和中国科协青年人才托举工程(2022QNRC001)资助项目。

Condition Monitoring Method for the Whole Lifecycle Evolution of Rolling Bearings Based on Data Dynamic Augmentation

KONG Detong1,2, ZHANG Mingrui1, LI Naipeng1, LUO Nanzhi1, ZHANG Leping1,2, LEI Yaguo1   

  1. 1. Key Laboratory of Education Ministry for Modern Design and Rotor-bearing System, Xi'an Jiaotong University, Xi'an 710049;
    2. China HuaDian Electric Power Research Institute Corporation Ltd., Hangzhou 310030
  • Received:2025-09-25 Revised:2026-06-30 Published:2026-11-20

摘要: 滚动轴承作为旋转机械的核心组成部件,其健康状态对旋转机械的安全运行至关重要。为了能够解决监测信号中故障特征被淹没的问题,实现在全寿命周期对滚动轴承演化过程的实时监测,提出了一种基于数据动态增强的滚动轴承全寿命演化过程监测方法。所提出的方法基于阈值函数对监测信号中的有效成分进行自适应识别,实现信噪比的提升;采用脉冲特征提取算法,依据监测信号的差分序列与自相关序列,实现了数据脉冲特征的增强;在特征映射中引入动态激活函数,进一步优化输出机制,依据输入特征自适应调整非线性映射关系,解决了数据特征提取不完整的问题。试验结果表明,所提出方法能够实现有效数据增强,显著提高了滚动轴承全寿命演化过程监测的准确性。

关键词: 滚动轴承, 全寿命周期, 状态监测, 数据动态增强

Abstract: As a core component of rotating machinery, the health condition of rolling bearings is crucial to the safe operation of the entire system. To address the problem that fault features in monitoring signals are obscured and to enable real-time monitoring of the full-life evolution of rolling bearings, a condition monitoring method for the whole lifecycle evolution of rolling bearings based on data dynamic augmentation is proposed. In this method, a threshold function is employed to adaptively identify the informative components in the monitoring signals, so that the signal-to-noise ratio is improved. An impulsive feature extraction algorithm is then applied, in which the difference sequence and autocorrelation sequence of the monitoring signals are exploited to enhance the impulsive characteristics of the data. A dynamic activation function is further introduced into the feature mapping stage to optimize the output mechanism, allowing the nonlinear mapping relationship to be adaptively adjusted according to the input features and thereby alleviating the problem of incomplete feature extraction. Experimental results indicate that effective data enhancement is achieved, and the accuracy of whole lifecycle evolution monitoring for rolling bearings is significantly improved.

Key words: rolling bearing, whole lifecycle, condition monitoring, data dynamic augmentation

中图分类号: