基于支持向量机的光伏发电功率预测系统设计  

Design of photovoltaic power prediction system based on support vector machine

作　　者：马志春[1] 刘振永[1] 吴淑花[1] MA Zhichun;LIU Zhenyong;WU Shuhua(School of Physics,Shijiazhuang University,Shijiazhuang 050035,China)

机构地区：[1]石家庄学院物理学院,河北石家庄050035

出　　处：《现代电子技术》2021年第10期91-94,共4页Modern Electronics Technique

基　　金：国家自然科学基金项目(41101125);河北省科技计划自筹经费项目(18214320);河北省科技计划自筹经费项目(18210912)。

摘　　要：光伏发电存在显著的变动性,高精度预测其发电功率能优化电网能量管理的效果与调度性。为此,该文设计一种基于支持向量机的光伏发电功率预测系统,主要通过光伏发电功率数据采集工作站获取气象数据与功率数据,并保存在光伏发电功率数据库服务器中。基于光伏发电功率数据库服务器中的数据,光伏发电功率预测模块采用基于支持向量机的光伏发电功率滚动预测模型,实现光伏发电功率预测。实验结果表明,所设计系统对光伏发电功率的预测结果平均误差较小,系统预测结果最为接近于实际功率,且预测耗时短,应用价值较高。There is significant variability in photovoltaic power generation.The high⁃precision prediction of generation power can optimize the effect and scheduling of power grid energy management.Therefore,a photovoltaic power prediction system based on support vector machine(SVM)is designed.The system mainly obtains meteorological data and power data by photovoltaic power generation data acquisition workstation,and stores them in photovoltaic power generation database server.On the basis of the data in photovoltaic power generation database server,the SVM⁃based rolling prediction model of photovoltaic generation power is adopted in the photovoltaic power prediction module to realize the power prediction of photovoltaic generation power.The experimental results show that the designed system has a small average error in the prediction result of photovoltaic generation power,the system prediction result is closest to the actual power,and its prediction time is short,but its application value is high.

关 键 词：光伏发电 功率预测 支持向量机 系统设计 数据采集 数据保存 滚动预测模型 

分 类 号：TN86-34[电子电信—信息与通信工程] TM615[电气工程—电力系统及自动化]