基于压缩感知的短期风电功率预测相似数据分析方法  

A method of similar data analysis to the short-term wind power forecasting based on compressive sensing

作　　者：杜凯 王鹤[1] 綦雪松 Du Kai;Wang He;Qi Xuesong(School of Electrical Engineering,Northeast Electric Power University,Jilin 132000,Jilin,China;State Grid Jilin Electric Power Supply Company,Jilin 132000,Jilin,China)

机构地区：[1]东北电力大学电气工程学院,吉林吉林132000 [2]国网吉林供电公司,吉林吉林132000

出　　处：《电测与仪表》2019年第12期98-103,共6页Electrical Measurement & Instrumentation

基　　金：国家电网公司科技发展计划项目(KY-SD-2016-204-JLDKY)

摘　　要：准确及时的短期风电功率预测对包含大规模风电的电力系统运行调度、检修计划、备用安排有着重要意义。引入温度、风速、风向及其各自的最大变化范围等对风电功率影响较大的因素作为风电模式特征,利用其时段周期性,提出一种基于压缩感知的相似数据分析方法,用以为预测模型提取历史基础数据。文中提出的基于压缩感知的相似数据分析方法以风电模式特征为数据类别,预测目标时间段数据为原始信息,利用时段周期性历史样本数据构造冗余字典,通过匹配追踪,求得观测值作为相似数据。实验结果表明,相对于一般的相似数据分析方法,本方法提取的相似数据更为切合实际情况,进而可以提高短期风电功率预测的精度,为风电场运行和调度提供更优的数据参考。Accurate and timely short-term wind power forecasting is of great significance to the operation dispatch, turnaround plan and spare planning of electrical power system with large-scale wind power. This paper introduced temperature, wind speed, wind direction and the maximum range of them that have great influence of wind power as wind power pattern features, and put forward a similar data analysis method based on compressive sensing through using the periodicity, to extract history sample data for forecasting model. In this paper, the similar data analysis method based on compressive sensing takes wind power pattern features as data category, the data of forecast period as the original data, and the periodic history sample data are adopted to redundant dictionary, and matching pursuit to get observation as similar data. Simulation results show that, compared with the general similar data analysis methods, the method proposed in this paper is more accurate, furthermore, it can improve the accuracy of short-term wind power forecasting, and provide optimization data for the wind farm operation dispatch.

关 键 词：风电功率预测 相似数据 压缩感知 

分 类 号：TM933[电气工程—电力电子与电力传动]