摘要
利用状态监测与故障诊断能够确保风电机组运行的可靠性与安全性。为了解决风电机组故障诊断适用性差、精度低的问题,针对风电机组早期故障预警及定位工作,在传动链故障诊断中引入深度学习算法,基于监控数据采集系统提供的数据基础,结合稀疏字典对对抗变分自动编码器(AVAE)进行改进,引入非线性深度表示,以降低数据维数,进而实现对原始数据内在特征的多样化、有效提取。同时,提出了一种AVAE-SDL故障诊断模型,可以有效排除训练过程中随机噪声造成的影响,有利于从高维数据中进一步提取内在特征。案例分析结果证明,AVAE-SDL故障诊断模型能够准确检测出机组故障,不存在误报情况,可作为风电机组传动链故障诊断的可靠工具。
The use of condition monitoring and fault diagnosis can ensure the reliability and safety of wind turbine operation.In order to solve the problems of poor applicability and low accuracy of wind turbine fault diagnosis,aiming at the early fault warning and positioning of wind turbines,a deep learning algorithm is introduced into the drive chain fault diagnosis,based on the data provided by the monitoring data acquisition system,combined with a sparse dictionary,the adversarial variational autoencoder(AVAE)is improved,and the nonlinear depth representation is introduced to reduce the data dimension,so as to realize the diversified and effective extraction of the intrinsic characteristics of the original data.At the same time,an AVAE-SDL fault diagnosis model is proposed,which can effectively eliminate the influence of random noise in the training process,which is conducive to further extracting intrinsic features from high-dimensional data.Case analysis results show that the AVAE-SDL fault diagnosis model can accurately detect the unit fault without false alarms,and can be used as a reliable tool for the fault diagnosis of the wind turbine drive train.
作者
颜毅斌
陈清化
管俊杰
范刚
谭香玲
Yan Yibin;Chen Qinghua;Guan Junjie;Fan Gang;Tan Xiangling(Hunan Railway Technology Vocational and Technical College,Zhuzhou,Hunan 412006,China)
出处
《机电工程技术》
2024年第3期78-80,135,共4页
Mechanical & Electrical Engineering Technology
基金
湖南省自然科学基金资助项目(2022JJ60074)
湖南省教育厅资助科研项目(20C1226)。
关键词
风电机组
对抗变分自动编码器
稀疏字典学习
传动链故障
wind turbine
anti variational automatic encoder
sparse dictionary learning
drive chain failure
