摘要
风电场SCADA系统中存在大量异常监测数据,不利于风功率曲线的准确建模和风能预测等后续研究的开展。为此,根据负值点以及分散型、堆积型异常数据的分布特征,提出一种基于边缘检测与方差变点的风功率数据清洗方法。首先进行数据预清洗,以识别负值点;接着基于边缘检测识别曲线主体,以清洗分散型异常数据;然后通过方差变点分区间获得风速功率点中的方差突变点,以清洗堆积型异常数据;最后得到分类清洗后的风功率数据。算例验证结果表明,所提方法可有效地分类识别异常数据,通用性较好,且有利于风功率曲线的准确建模。
The large amount of abnormal monitoring data in the SCADA system of the wind farm are not conducive to the development of accurate modeling of wind power curves and follow-up studies such as wind energy prediction.Therefore,according to the distribution characteristics of three types of data including the negative points,scattered outliers and stacked outliers,this paper presents a cleaning method of wind power data based on the edge detection and variable-point variance.It firstly carries out data pre-cleaning to identify the negative points,and then by means of edge detection,it identifies the main curve as as to clean the scattered abnormal data.Afterwards,based on the variable-point variance,it obtains the the mutation point of variance in each wind speed sub-interval so as to clean the stacked outliers,and finally it obtains the normal wind power data.Validation examples show that the proposed method can effectively classify and identify abnormal data with a good universality and it is conducive to accurate modeling of wind power curves.
作者
苏荣
张斌
沈晨
陈俊生
SU Rong;ZHANG Bin;SHEN Chen;CHEN Junsheng(Southern Offshore Wind Power Joint Development Co.,Ltd.,Zhuhai,Guangdong 519080,China;State Key Lab of Power Transmission Equipment&System Security,Chongqing University,Chongqing 400044,China;College of Automation,Chongqing University of Posts and Telecommunications,Chongqing 400065,China)
出处
《广东电力》
2021年第5期48-56,共9页
Guangdong Electric Power
基金
重庆市自然科学基金面上项目(cstc2020jcyj-msxmX0687)。
关键词
风功率
异常数据清洗
边缘检测
方差突变点
wind power
abnormal data cleaning
edge detection
variable-point variance