基于改进鸟群算法和极限学习机模型的光伏发电系统输出功率预测研究  

Power prediction of photovoltaic power generation system based on improved bird swarm and extreme learning machine model

作　　者：饶宇飞 刘阳 李玲玲[2] 方舟 曲立楠[3] Rao Yufei;Liu Yang;Li Lingling;Fang Zhou;Qu Linan(State Grid Electic Power Research Institute of Henan Electie Power Coupany,Zhengzhou 450002,China;School of Eletrical Engineering,Hebei University of Technology,Tianjin 300130,China;China Eletric Power Research Instiute,Nanjing 210000,China)

机构地区：[1]国网河南省电力公司电力科学研究院,河南郑州450002 [2]河北工业大学电气工程学院,天津300130 [3]中国电力科学研究院(南京),江苏南京210000

出　　处：《可再生能源》2020年第10期1318-1325,共8页Renewable Energy Resources

基　　金：河北省自然科学基金(E2018202282);国家自然科学基金(51475136)。

摘　　要：准确预测光伏发电系统的输出功率,可以帮助电网调度部门合理安排调度计划,并能够提高光伏发电场的发电效率。为此,文章首先提出了一种改进鸟群(IBSA)算法,并采用IBSA对极限学习机(ELM)进行优化,构建了性能良好的IBSA-ELM预测模型;然后,利用IBSA-ELM模型BSA-ELM模型和SVM模型对光伏发电系统输出功率进行预测,并采用均方根误差(RMSE)和决定系数(R2)对该模型的预测效果进行评估。分析结果表明:IBSA算法的收敛精度优于BSA算法;IBSA-ELM模型的预测精度优于BSA-ELM模型和SVM模型。Accurate prediction of photovoltaic power can effectively help power grid dispatching departments to arrange reasonable dispatching plans and improve the operation efficiency of photovoltaic power plants.In this paper,the improved bird swarm optimization(IBSA)algorithm is proposed.The IBSA algorithm is used to optimize the extreme learning machine(ELM)and the IBSA-ELM prediction model is constructed.Then,the photovoltaic power is predicted by IBSA-ELM model.The root mean square error(RMSE)and the detemination cofficient(R2)are used to evaluate the prediction effect.The test results show that the convergence performance of IBSA is better than BSA,and the prediction performance of IBSA-ELM model is better than that of BSA-ELM and SVM models.The IBSA algorithm with good optimization ability and IBSA-ELM model with accurate prediction ability have been obtained.Their application in the power prediction of photovoltaic power generation systems can effectively improve the stability of the power grid.

关 键 词：光伏发电 预测 评估 改进鸟群算法 极限学习机 

分 类 号：TK519[动力工程及工程热物理—热能工程] TM615[电气工程—电力系统及自动化]