基于数字孪生的风电机组齿轮箱故障诊断方法研究  

Research on Fault Diagnosis Method of Wind Turbine Gearbox Based on Digital Twin

作　　者：孙亚飞 渠叶君 SUN Yafei;QU Yejun(Ningxia New Energy Development Co.,Ltd.of GD Power Development Co.,Ltd.,Yinchuan,Ningxia Hui Autonomous Region,750001 China)

机构地区：[1]国电电力宁夏新能源开发有限公司,宁夏银川750001

出　　处：《科技资讯》2025年第1期102-104,共3页Science & Technology Information

摘　　要：风力发电作为可再生能源的重要支柱,其风电机组齿轮箱受多种因素影响易出现故障,从而影响整个风电机组的稳定性。因此,提出基于数字孪生的风电机组齿轮箱故障诊断方法。首先,构建数字孪生总体架构。其次,采集风电机组齿轮箱数据并过滤,利用集合经验模态分解提取特征。最后,构建风电机组的齿轮箱的数字孪生模型,并采用长短期记忆(Long Short-Term Memory,LSTM)完成故障诊断。实验结果表明,所提方法故障诊断的准确率在0.94以上、漏报率在0.15以下,为风电场的运维管理提供了有力的决策支持。Wind power generation,as an important pillar of renewable energy,the gearbox of its wind turbine is affected by various factors,leading to gearbox failures and affecting the stability and reliability of the entire wind turbine.Therefore,a fault diagnosis method for wind turbine gearbox based on Digital Twin is proposed.Firstly,construct a digital twin overall architecture.Then,collect and filter wind turbine gearbox data,and use set empirical mode decomposition to extract features from wind turbine gearbox data.Finally,construct a digital twin model of the wind turbine gearbox and use Long Short-Term Memory(LSTM)to complete fault diagnosis.The experimental results show that the accuracy of the proposed method for fault diagnosis is above 0.94,and the false alarm rate is below 0.15,providing strong decision support for the operation and maintenance management of wind farms.

关 键 词：数字孪生 风电机组 故障诊断 长短期记忆 齿轮箱 

分 类 号：TH165[机械工程—机械制造及自动化]