基于NWP相似性分析的超短期光伏发电功率预测  

POWER FORECASTING OF ULTRA-SHORT-TERM PHOTOVOLTAIC STATION BASED ON NWP SIMILARITY ANALYSIS

作　　者：张姗 冬雷[1] 纪德洋 郝颖[1] 张小凤[2] Zhang Shan;Dong Lei;Ji Deyang;Hao Ying;Zhang Xiaofeng(Schoolof Automation,Beijing Institute of Technology,Beijing 100081,China;Zhuhai College of Beijing Institute of Technology,Zhuhai 519088,China)

机构地区：[1]北京理工大学自动化学院,北京100081 [2]北京理工大学珠海学院,珠海519088

出　　处：《太阳能学报》2022年第4期142-147,共6页Acta Energiae Solaris Sinica

基　　金：国家自然科学基金(51607009)。

摘　　要：针对相似气象情况下,光伏电站发电功率接近的实际情况,提出一种基于数值天气预报(NWP)相似性分析的超短期光伏发电功率预测方法。该方法根据皮尔逊相关系数找到与被预测时刻相似的天气预报数据,根据相似时刻的实际功率估计出被预测时刻的功率。该方法能高效的根据天气预报数据预测发电功率,运算速度快,与神经网络相比预测有着更优的效果,尤其在数据波动大的时段具有更高的可靠性。According to the fact that photovoltaic plants have similar generation power under similar weather conditions,an Ultra-shortterm power forecasting method based on NWP similarity analysis is proposed.The proposed method uses the Pearson correlation coefficient to find weather forecast data similar to the predicted time,and estimates the power in the predicted time based on the actual power of the similar time.The proposed method can efficiently forecast the generated power based on the weather forecast data.Compared with the neural network,the proposed method has a better effect,especially in the period of large data fluctuations,which has higher reliability.

关 键 词：光伏电站 皮尔逊相关系数 功率预测 数值天气预报 相似性分析 

分 类 号：TM615[电气工程—电力系统及自动化]