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 (4): 288-295.DOI: 10.19912/j.0254-0096.tynxb.2020-0457

• 电化学储能安全性与退役动力电池梯次利用关键技术专题 • 上一篇    下一篇

基于优化VMD复合多尺度散布熵及LSTM的风力发电机齿轮箱故障诊断方法研究

王宏伟1, 孙文磊1, 张小栋2, 何丽1   

  1. 1.新疆大学机械工程学院,乌鲁木齐 830046;
    2.西安交通大学机械工程学院,西安 710049
  • 收稿日期:2020-05-20 出版日期:2022-04-28 发布日期:2022-10-28
  • 通讯作者: 孙文磊(1962—),男,博士、教授,主要从事能源装备技术、数字化设计与制造等方面的研究。sunwenxj@163.com
  • 基金资助:
    新疆维吾尔自治区科技支疆项目(2017E0276); 国家自然科学基金(51565055); 新疆维吾尔自治区研究生创新项目(XJ2019G030)

FAULT DIAGNOSIS METHOD OF WIND TURBINE’S GEARBOX BASED ON COMPOSITE MULTISCALE DISPERSION ENTROPY OF OPTIMISED VMD AND LSTM

Wang Hongwei1, Sun Wenlei1, Zhang Xiaodong2, He Li1   

  1. 1. School of Mechanical Engineering, Xinjiang University, Urumqi 830046, China;
    2. School of Mechanical Engineering, Xi’an Jiaotong University, Xi’an 710049, China
  • Received:2020-05-20 Online:2022-04-28 Published:2022-10-28

摘要: 以风力发电机齿轮箱加速度信号为研究对象,提出一种数据驱动的风力发电机齿轮箱故障诊断方法,该方法以灰狼优化的变分模态分解方法(AGWO-VMD)、复合多尺度规范化散布熵(NCMDE)及长短期记忆网络(LSTM)为基础,实现齿轮箱故障的快速诊断。首先将时域信号转换至角域;然后通过AGWO-VMD方法对角域信号进行自适应分解,并采用NCMDE算法提取分解后及原始信号中的故障特征构成特征向量;最后利用LSTM模型对特征向量进行智能识别与分类。对实际采集的6种故障齿轮信号进行测试与验证,试验结果表明该方法能快速有效区分齿轮故障类型。

关键词: 风力机, 齿轮箱, 故障检测, 灰狼优化算法, 变分模态分解, 复合多尺度规范化散布熵, 长短期记忆网络

Abstract: A data driven diagnosis method based on acceleration signals for the gearbox in wind turbine is proposed, which on the basis of the grey wolves optimised variational modal decomposition (AGWO-VMD), normalized composite multiscale dispersion entropy (NCMDE) and long short-term memeory (LSTM), the gearbox faults diagnosis is realized rapidly. Firstly, the discrete signal in time domain is converted to angular domain. Secondly, AGWO-VMD algorithm is used to decompose the signal adaptively, and NCMDE algorithm is used to extract fault features as feature vectors from both original and decomposed signals. At last, the LSTM model is used for intelligentive classification of feature vectors. The proposed method is validated by 100 groups of data under 6 types of faults collected from WTDS, and the result shows that , it can recognize the right type of gearbox's fault rapidly and effectively.

Key words: wind turbines, gearbox, fault detection, grey wolf optimizer, variational modal decomposition, normalized composite multiscale dispersion entropy, long short-term memeory network

中图分类号: