计及信息不确定性的风电机组健康状态实时评估方法

发布时间：2018-05-04 19:12

  本文选题：风电机组 + 预测与健康管理　； 参考：《电力系统自动化》2017年18期

【摘要】：运行工况识别作为风电机组状态监测与健康管理领域的重要环节,往往受到不确定信息以及高速实时数据流的影响,造成健康状态评估难以有效实施。在此背景下,文中提出一种基于Spark流式处理的健康状态实时评估方法。首先,采用大数据分析技术实现风电机组运行工况的空间划分;然后,在充分考虑风电机组监测信息不确定性的情况下,结合数据采集与监控(SCADA)历史运行数据,对基于高斯云模型和高斯云变换的健康状态评估模型进行训练,并以健康指数作为风电机组健康状态评估的指标。最后,将该评估方法应用在中国北方某风电场1.5 MW风电机组故障前的健康状态评估中。算例分析结果表明,该方法可监测到风电机组健康状态的变化趋势,初步实现了故障的早期预警。
[Abstract]:As an important link in the field of wind turbine condition monitoring and health management, operating condition identification is often affected by uncertain information and high speed real-time data flow, which results in the difficulty of effective implementation of health status assessment. In this context, a real-time assessment method of health status based on Spark flow processing is proposed. First of all, big data analysis technology is used to realize the space division of wind turbine operating conditions, and then, with full consideration of the uncertainty of monitoring information of wind turbines, the historical operation data of SCADAs are combined with data acquisition and monitoring. The health state assessment model based on Gao Si cloud model and Gao Si cloud transformation is trained, and the health index is used as the index of wind turbine health assessment. Finally, the method is applied to evaluate the health status of a 1.5 MW wind turbine unit in northern China before failure. The result of example analysis shows that the method can monitor the change trend of wind turbine's health state and realize the early warning of failure.
【作者单位】： 华北电力大学控制与计算机工程学院;浙江大学电气工程学院;文莱科技大学电机与电子工程系;
【基金】：国家自然科学基金资助项目(51407076) 河北省自然科学基金资助项目(F2014502050) 中央高校基本科研业务费专项资金资助项目(2015ZD28)~~
【分类号】：TM315
，

本文编号：1844283