摘要
风功率曲线是考核风电机组发电性能的重要指标,对风电场的运行管理和电力系统的运行调度都具有重要意义。实际运行过程的设备故障及人为控制因素会导致风速-功率曲线中存在大量的异常数据,给风功率曲线的后续应用带来严重影响。本文在分析风电机组风速-功率异常运行数据特征的基础上,根据空间分布位置和形态将异常数据分为曲线底部、中部、上部堆积型异常数据和曲线周围分散型异常数据等四类,提出了基于变点分组法与四分位法组合的异常数据识别清洗方法及流程,与四分位-变点分组法以及局部离群因子算法的对比算例验证结果表明,提出的变点分组-四分位法可有效识别四种类型的异常数据,流程合理,清洗效果好,效率高,并具有较强的通用性。
Wind power curve is an important basis for assessing the power generation performance of wind turbines,and is of great significance to wind farm management and power system scheduling.The equipment failures and control factors will cause a large number of outliers in wind speed-power curve during actual operation,which will directly affect the subsequent applications of the wind power curve.Based on the analysis of the characteristics of outliers,this paper divided the outliers into four categories according to their spatial distribution and shape features,including the bottom,middle and upper stacked outliers as well as scattered outliers around the curve.A combined strategy and its process for eliminating outliers were proposed based on the change point grouping and quartile method.Compared with the quartile-change point grouping method and the local outlier factor(LOF)algorithm,the results show that the proposed method and the combined strategy can eliminate four types of outliers with good effect,high efficiency and strong versatility.
作者
沈小军
付雪姣
周冲成
王伟
Shen Xiaojun;Fu Xuejiao;Zhou Chongcheng;Wang Wei(Department of Electrical Engineering Tongji University Shanghai 200092 China;Global Energy Internet Research Institute Beijing 102209 China)
出处
《电工技术学报》
EI
CSCD
北大核心
2018年第14期3353-3361,共9页
Transactions of China Electrotechnical Society
基金
中央高校基本业务经费资助项目(0800219312)
