欢迎访问《太阳能学报》官方网站,今天是 分享到:
ISSN 0254-0096 CN 11-2082/K

太阳能学报 ›› 2022, Vol. 43 ›› Issue (5): 343-350.DOI: 10.19912/j.0254-0096.tynxb.2020-0445

• • 上一篇    下一篇

基于图形特征的双输入卷积神经网络风力机轴承剩余寿命预测

余萍, 曹洁1   

  1. 1.兰州理工大学电气工程与信息工程学院,兰州 730050;
    2.甘肃省工业过程控制重点实验室,兰州 730050;
    3.兰州理工大学电气与控制工程国家级实验教学示范中心,兰州 730050
  • 收稿日期:2020-05-18 出版日期:2022-05-28 发布日期:2022-11-28
  • 通讯作者: 余萍(1979—),女,博士研究生、讲师,主要从事模式识别与智能系统、故障诊断与预测等方面的研究。yup@lut.edu.cn
  • 基金资助:
    国家自然科学基金(61763028; 61563032; 61963025); 甘肃省自然科学基金(1506RJZA104)

DUAL-INPUT CONVOLUTIONAL NEURAL NETWORK FOR GRAPHICAL FEATURES BASED REMAINING USEFUL LIFE PROGNOSTICATING OF WIND TURBINE BEARINGS

Yu Ping, Cao Jie1   

  1. 1. College of Electrical and Information Engineering, Lanzhou University of Technology, Lanzhou 730050, China;
    2. Key Laboratory of Gansu Advanced Control for Industrial Processes, Lanzhou 730050, China;
    3. National Demonstration Center for Experimental Electrical and Control Engineering Education, Lanzhou University of Technology, Lanzhou 730050, China
  • Received:2020-05-18 Online:2022-05-28 Published:2022-11-28

摘要: 提出一种基于图形特征的风力机轴承剩余使用寿命(RUL)预测方法。首先,基于连续小波变换(CWT)对时域振动数据样本集进行预处理,得到用于预测的时频图形数据集。然后,采用双输入卷积神经网络(DICNN)从图形数据集中提取特征映射,用于构造高性能健康指数(DICNN-HI)来表征轴承各退化阶段的状态。最后,结合DICNN-HI,采用基于高斯过程回归(GPR)的分析方法进行RUL预测,并用PRONOSTIA滚动轴承数据集进行验证。结果表明,该方法具有较高的健康指数预测精度,能有效反映滚动轴承的劣化状态,有助于实现风力机轴承的RUL预测。同时,也可为其他旋转机械设备的剩余寿命预测提供重要的理论参考,具有一定的实用价值。

关键词: 风力机轴承, 双输入卷积神经网络, 图形特征, 剩余使用寿命, 预测

Abstract: A graphical features based remaining useful life (RUL) prognosticating method for the bearings in wind turbine is proposed in this paper. Firstly, preprocessing the time-domain vibration data sample set based on continuous wavelet transform (CWT) to obtain the time-frequency graphical data set used for prognosticating work. Secondly, dual-input convolutional neural network (DICNN) is employed to extract the feature map from the graphical data set to construct high performance health indicator (DICNN-HI) for representing the state of each degradation stage of the bearing. Finally, according to the predicted DICNN-HI, a Gaussian process regression (GPR)-based analysis is used for RUL prognosticating, which is verified by the PRONOSTIA ball bearing data set. Results illustrate that the proposed method has a high prediction accuracy of health indicator to map the state of degradation of a bearing effectively, which is helpful to realize the RUL prognosticating accurately in this study. It provides an important theoretical reference for RUL prognosticating of bearing and the other rotating machineries, as well as a certain practical value.

Key words: wind turbine bearing, dual-input convolutional neural network, graphical features, remaining useful life, prognosticating

中图分类号: