基于粗糙集的光伏输出功率组合预测模型  

A Combined Photovoltaic Output Forecasting Method Based on Rough Set Theory

作　　者：杨锡运[1] 任杰[1] 肖运启[1] YANG Xiyun REN Jie XIAO Yunqi(Department of Control and Computer Engineering, North China Electric Power University, Beijing 102206, China)

机构地区：[1]华北电力大学控制与计算机工程学院,北京102206

出　　处：《中国电力》2016年第12期133-138,共6页Electric Power

基　　金：国家自然科学基金资助项目(51677067);中央高校基本科研业务费专项基金(2015MS32)~~

摘　　要：光伏输出功率预报是减小光伏发电系统并网对电网造成不良影响的有效手段,提高预测的精度是保障光伏功率预报的重要基础。提出了基于粗糙集的组合预测模型。首先建立基于相似日、支持向量机和持续法的3种单一预测模型,然后根据粗糙集理论中确定属性重要度的方法确定单一预测模型的组合权重,建立了基于粗糙集的组合预测模型。仿真结果表明,采用粗糙集的相关理论能得到合理的组合权重,建立的光伏功率组合预测模型具有更高的预测精度。Photovoltaic output prediction is an effective way to decrease impacts for photovoltaic power connecting to grid. Improving accuracy is the key to photovoltaic output prediction. A combined prediction method is proposed based on rough set theory. Firstly, three different models, namely, similar day method, support vector machine and persistence method, are constructed. Then rough set theory is used to obtain relative weights by calculating degree of importance. Finally, the combined model is established by using the weights obtained. A case study shows that rough set theory is effective to obtain appropriate weights, which lead to high prediction accuracy of proposed method.

关 键 词：光伏发电 光伏并网 光伏系统 功率预测 粗糙集 组合预测 相似日 

分 类 号：TM615[电气工程—电力系统及自动化]