滚动轴承性能退化表征与剩余寿命预测方法研究
摘要
滚动轴承是机械设备传动系统中的关键部件，由于其处于重载、高速及高温等极端恶劣的运行环境中，极易发生失效，继而引发系统级故障，因此，掌握滚动轴承的性能退化状态以及剩余寿命是保障机械设备安全可靠运行的关键所在。本文以滚动轴承的性能退化表征和剩余寿命预测为研究主题，开展轴承全寿命周期试验，在此基础上，基于轴承的振动信号，进行了特征提取、性能退化表征以及剩余寿命预测研究，主要内容如下：
（1）以6207深沟球轴承（材料为：GCr15）为试验对象，利用滚动轴承加速寿命试验台，进行全寿命周期性能退化试验，监测、记录并分析全程中的状态监测量（温度、振动）的特征参数及变化规律，为后续性能退化行为表征参数的选取提供参考依据。
（2）针对振动信号的特征提取问题，构建反映轴承退化的特征集。从时域、频域以及时频域提取了与轴承性能退化征兆相关的71个特征参量，构成原始特征集。结果表明，在轴承的全寿命周期中，原始特征表现出不同形式的变化趋势，代表各自特征有关退化过程的独特信息，能够全面有效反映轴承的退化信息。
（3）构建健康指数，表征滚动轴承性能退化状态。以相关性、单调性以及鲁棒性作为特征评价指标，筛选出反映轴承性能退化的敏感特征，并基于PCA方法，对多个敏感特征进行融合，构建出表征轴承性能退化的健康指数。通过试验验证，所构建的健康指数能有效表征轴承正常运行、初始退化以及急剧退化三个
阶段的退化状态。
（4）构建EMD-Kriging模型，实现滚动轴承剩余寿命预测。首先，采用EMD 方法对健康指数进行主趋势提取；其次，基于Kriging模型对轴承剩余寿命进行预测；最后，通过试验以及与典型预测方法进行比较分析，验证了所提模型的可行性和有效性。
本文研究为滚动轴承的性能退化表征与剩余寿命预测提供了方法借鉴，对于提高滚动轴承乃至机械设备传动系统的可靠性和保障性水平具有重要的工程意义。
关键词：滚动轴承；特征提取；特征选择；性能退化表征；剩余寿命预测
分类号：TH133.33; TH17
Abstract
Rolling bearing is the key component in the mechanical equipment transmission system. Because it is in the extremely severe operating environment such as heavy load, high speed and high temperature, it is easy to fail, and then cause system level failure. Therefore, it is the key to ensure the safe and reliable operation of mechanical equipment to master the performance degradation state and remaining useful life of rolling bearing. This paper takes the performance degradation characterization
and remaining useful life prediction of rolling bearing as the research subject, and carries out the bearing life cycle test. On this basis, based on the vibration signal of the bearing, the research on feature extraction, performance degradation characterization and remaining useful life prediction is carried out. The main contents are as follows:
(1) Taking 6207 deep groove ball bearing (material: GCr15) as the test object, the rolling bearing accelerated life test bench was used to conduct a full life cycle performance degradation test to monitor, record and analyze the characteristic parameters of the state monitoring quantity (temperature, vibration) throughout And the change rule provides a reference basis for the selection of subsequent performance degradation behavior characterization parameters.
(2) Aiming at the problem of feature extraction of vibration signals, a feature set reflecting bearing degradation is constructed. 71 feature parameters related to the signs of bearing performance degradation were extracted from the time domain, frequency domain and time-frequency domain to form the original feature set. The results show that in the life cycle of the bearing, the original features show different types of change trends, representing the distinct information about the degradation process, which can fully and effectively reflect the degradation information of the bearing.
(3) Construct a health index to characterize the degradation state of rolling bearing performance. Using correlation, monotonicity and robustness as feature evaluation indexes, the sensitive features reflecting the degradation of bearing performance are selected, and based on the PCA method, multiple sensitive features are fused to construct a health index that characterizes the degradation of bearing performance. It is verified through experiments that the constructed health index can effectively characterize the degradation state of the bearing in three stages of normal operation, slight degradation and severely degradation.
(4) Construct an EMD-Kriging model to predict the remaining useful life of rolling
bearings. First, the EMD method is used to extract the main trend of the health index; secondly, the remaining useful life of the bearing is predicted based on the Kriging model; finally, the feasibility and effectiveness of the proposed model are verified by comparison and analysis with typical prediction methods where the same dataset is used.
The research in this paper provides a method for the characterization of rolling bearing performance degradation and remaining useful life prediction, which has important engineering significance for improving the reliability and security of rolling bearing and even the transmission system of mechanical equipment.
Keywords:rolling bearing; feature extraction; feature selection; performance degradation characterization; remaining useful life prediction
Classification Number: TH133.33; TH17
目录
摘要 ....................................................................................................................... II 目录 .................................................................................................................... I V 1 绪论 .. (1)
1.1 研究背景及意义 (1)
1.2 国内外研究现状 (2)
1.2.1 特征提取 (3)
1.2.2 性能退化表征 (4)
1.2.3 剩余寿命预测 (4)
1.3 现状分析总结 (6)
1.4 本文的研究思路和主要内容 (6)
2 滚动轴承全寿命周期试验 (8)
2.1 滚动轴承结构及失效模式 (8)
2.1.1 滚动轴承结构 (8)
2.1.2 滚动轴承常见失效模式 (9)
2.2 滚动轴承加速寿命试验 (10)
2.2.1 滚动轴承加速寿命试验台 (10)
2.2.2 试验方案及流程 (11)
2.2.3 试验结果分析 (14)
2.3 本章小结 (18)
3 滚动轴承振动信号的特征提取 (19)
3.1 时域特征提取 (19)
3.2 频域特征提取 (23)
3.3 相似相关特征提取 (26)
3.4 时频域特征提取 (28)
3.4.1 小波包分解特征 (28)
3.4.2 经验模态分解特征 (33)
3.5 本章小结 (37)
4 滚动轴承性能退化表征方法研究 (39)
4.1 敏感特征选择 (39)
4.2 基于主成分分析的健康指数构建方法 (41)
4.2.1 主成分分析的基本原理 (41)
4.2.2 健康指数的构建 (43)
4.3 试验验证和结果分析 (44)
4.3.1 PRONOSTIA试验介绍 (44)
4.3.2 方法验证及结果分析 (46)
4.4 本章小结 (51)
5 滚动轴承剩余寿命预测方法研究 (53)
5.1 Kriging模型 (53)
5.2 基于EMD-Kriging的轴承剩余寿命预测模型 (56)
滚动轴承的特点5.2.1 基于EMD-Kriging的预测模型 (56)
5.2.2 轴承剩余寿命预测流程 (57)
5.3 试验验证 (58)
5.4 本章小结 (63)
6 总结与展望 (64)
6.1 全文工作总结 (64)
6.2 研究展望 (64)
参考文献 (66)
作者简历 (71)

版权声明：本站内容均来自互联网，仅供演示用，请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系QQ:729038198，我们将在24小时内删除。