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 (11): 194-202.DOI: 10.19912/j.0254-0096.tynxb.2021-0513

• • 上一篇    下一篇

基于多层深度互信息变分网络的风电机组轴承故障预警方法

柏林1, 晏锐1, 刘小峰1, 孔德斌2,3, 王明迪2,3   

  1. 1.机械传动国家重点实验室,重庆大学,重庆 400044;
    2.上海空间推进研究所,上海 201112;
    3.上海空间发动机工程技术研究中心,上海 201112
  • 收稿日期:2021-05-13 出版日期:2022-11-28 发布日期:2023-05-28
  • 通讯作者: 柏 林(1972—),男,博士、教授、博士生导师,主要从事设备状态检测与智能诊断方面的研究。Bolin0001@aliyun.com
  • 基金资助:
    国家自然科学基金(52175077; 51975067); 中央高校基本科研业务费(2020CDCGJX022)

WIND TURBINE BEARING FAULT EARLY WARNING BASED ON MULTI-LAYER DEPTH MUTUAL INFORMATION VARIATIONAL NETWORK

Bo Lin1, Yan Rui1, Liu Xiaofeng1, Kong Debin2,3, Wang Mingdi2,3   

  1. 1. State Key Laboratory of Mechanical Transmissions, Chongqing University, Chongqing 400044, China;
    2. Shanghai Institute of Space Propulsion, Shanghai 201112, China;
    3. Shanghai Engineering Research Center of Space Engine, Shanghai 201112, China
  • Received:2021-05-13 Online:2022-11-28 Published:2023-05-28

摘要: 针对变工况复杂环境下风电机组轴承的早期故障潜隐性高且故障阈值设置困难的问题,提出一种基于多层深度互信息变分网络的轴承故障超前预警方法。该网络在变分自编码器的架构上进行多层编码拓展,采用解码信号的二次编码增强了变分网络对输入信号中噪声的鲁棒性,通过隐层变量与输入信号间最大化互信息以及与二次编码特征间的最大化互信息,提高了网络对轴承正常状态空间的建模能力。以二次编码特征与隐层变量的重构误差为基础,构建健康指数,结合三次递推指数加权移动平均模型设置与风电机组轴承工况环境相适应的健康基线。在2个风电机组轴承故障检测试验中的应用结果表明,该方法对比传统的基于模型重构的轴承早期故障检测方法,具有更高故障预警准确率和抗干扰能力。

关键词: 风电机组, 故障检测, 状态监测, 多层互信息, 健康指数构建, 健康基线设置

Abstract: For the wind turbine in variable working conditions and complex environment,the fault monitoring and early warning of its bearing faces the problems of high latent characteristics of and the difficulty in fault threshold setting. Aiming to the problems,an early warning method based on multi-layer deep mutual information variational network(MDMIVN) is proposed. The coding layer of variational auto-coder is extended to multi-layer encoding, and the decoded signal is encoded again to improve the network robustness to the noise in the original fault signal. To improve the ability of modeling the bearing normal state space, the maximum mutual information between the latent variables and the input signal, between the latent variables and the secondary coding features are introduced as the loss functions. The health index is established based on the reconstruction errors between the secondary coding features and the latent variables. And then,the health baseline is adaptively set, by combining the triple exponential weighted moving average model and the updating iteration of the health indexes. The experimental results on two wind turbine bearing vibration data sets show that compared with the traditional early fault detection method based on model reconstruction, the proposed method has high fault warning accuracy, anti-interference ability.

Key words: wind turbines, fault detection, condition monitoring, multi-layer mutual information, health index construction, health baseline setting

中图分类号: