摘要
将大数据分析应用到风机轴承故障预警中,使用栈式自动编码器(SAE)为基本结构,通过逐层提取风机轴承监视控制与数据采集系统(SCADA)数据深层特征,将散乱的SCADA大数据转化成能够深度刻画风机轴承运行状态的内在特征。利用预训练、微调的方法并结合误差反向传播算法(BP)构建SAE故障预警模型,通过SAE模型对大数据处理得到反映风机轴承运行状态的重构误差平均值,以均值漂移聚类算法动态地计算出风机轴承稳定运行状态重构误差基准值为预警的标准。最后利用某风电场机组的SCADA数据进行工程实例仿真分析,验证了基于大数据分析用于风机轴承故障预警的可行性。
Big data analysis is applied to the fault early warning of wind turbine bearing.Stacked autoencoder(SAE)is used as the basic structure,and the data deep features of the supervisory control and data acquisition system(SCADA)for the wind turbine bearing are extracted layer by layer.The scattered big data from SCADA is transformed into intrinsic features,revealing the operating state of the wind turbine bearing.Pre-training and fine-tuning is combined with error back propagation(BP)algorithm to construct SAE fault early warning model,and the SAE model is used to process the big data and obtain the average value of the reconstruction error indicating the operating state of the wind turbine bearing.In addition,mean shift cluster algorithm is used to dynamically calculate the reference value of the stable operation state of the wind turbine bearing,using as the early warning standard.Finally,the simulation is done with the data from SCADA for a wind farm.and the feasibility of applying the big data analysis to the fault early warning of the wind turbine hearing is verified.
作者
李俊卿
王焕仲
季刚
马阳硕
LI Junqing;WANG Huanzhong;JI Gang;MA Yangshuo(School of Electrical and Electronic Engineering,North China Electric Power University,Baoding 071000,China)
出处
《智慧电力》
北大核心
2020年第2期25-30,共6页
Smart Power
基金
国家自然科学基金资助项目（51777074）~~
关键词
大数据
故障预警
风机轴承
SAE
big data
fault early warning
wind turbine bearing
SAE
