考虑光伏机理与数据驱动结合的短期功率预测  

Consider Short-term Power Prediction Combining Photovoltaic Mechanism and Data-driven

作　　者：陈凡 李智 丁津津 樊磊 伍骏杰 CHEN Fan;LI Zhi;DING Jin-jin;FAN Lei;WU Jun-jie(State Grid Anhui Electric Power Co.,Ltd.,Hefei 230601,China;State Grid Anhui Electric Power Research Institute,Hefei 230601,China;School of Electrical Engineering and Automation,Anhui University,Hefei 230601,China)

机构地区：[1]国网安徽省电力有限公司,合肥230601 [2]国网安徽省电力有限公司电力科学研究院,合肥230601 [3]安徽大学电气工程与自动化学院,合肥230601

出　　处：《科学技术与工程》2023年第20期8686-8692,共7页Science Technology and Engineering

基　　金：国网公司科技项目(52120522000R)。

摘　　要：精确的光伏功率预测对电网的可靠与稳定运行至关重要。现有研究大多数都是将天气条件直接作为数据驱动的输入,未深入分析天气条件与光伏输出功率直接耦合关系,因此将机理模型与数据驱动方法相结合,提出一种新型的光伏功率预测方法。首先,建立光伏系统物理模型,依据建立的物理模型得到不同的辐照度分量以及光伏电池温度。其次,将这些关键的天气特征重新构建数据驱动的输入,实现光伏机理与数据驱动结合的短期功率预测。最后,进行误差修正然后得到最终的光伏功率预测结果。根据光伏系统实测数据集进行仿真分析,结果表明:因为从物理模型得到了关键天气特征,考虑了天气条件与天气因素的耦合关系,预测精度有了明显提升,验证了所提方法的有效性。Accurate solar power forecasting plays a critical role in ensuring the reliable and economic operation of power grids.Most of existing literature directly uses available weather conditions as input features,which might ignore some key weather factors and the coupling among weather conditions.Therefore,a novel solar power forecasting approach was proposed by exploring key weather factors from photovoltaic(PV)analytical modeling.First,a PV analytical model was established,different irradiance components and PV cell temperatures were derived from PV analytical modeling.Then,these weather features were used to reformulate the input of machine learning methods,a better forecasting performance was achieved.Finally,based on the historical forecasting deviations,a compensation term is presented to adjust the solar power forecast.Based on measured dataset of the PV system,simulation analysis was conducted.The results show that due to obtaining key weather features from the physical model and considering the coupling relationship between weather conditions and weather factors,the prediction accuracy was significantly improved,verifying the effectiveness of the proposed method.

关 键 词：光伏功率预测 数据驱动 物理模型 天气特征 误差修正 

分 类 号：TM615[电气工程—电力系统及自动化]