利用蝙蝠算法优化SVR的太阳辐照度预测方法研究  

Solar radiation prediction method using bat algorithm optimized SVR

作　　者：姚海成[1] 周剑[1] 林琳[2] 邢恩恺 黄南天[3] 陈艳伟 Yao Haicheng;Zhou Jianl;Lin Lin;Xing Enkai;Huang Nantian;Chen Yanwei(Power Dispatching and Communication Center of CSG,Guangzhou 510623,China;College Information and Control Engineering Jilin Institute of Chemical Technology,Jilin 132022,China;College of Electrical Engineering Northeast Dianli University,Jilin 132012,China;Bei Jing Tsingsoft Technology Co.,Ltd.,Beijing 100085,China)

机构地区：[1]中国南方电网电力调度控制中心,广东广州510700 [2]吉林化工学院信息与控制工程学院,吉林吉林132022 [3]东北电力大学电气工程学院,吉林吉林132012 [4]北京清软创新科技股份有限公司,北京100085

出　　处：《可再生能源》2018年第11期1612-1617,共6页Renewable Energy Resources

基　　金：国家自然科学基金项目(51307020);吉林省科技发展计划项目(20160411003XH);吉林省教育厅"十三五"科技项目(JJKH20170219KJ)

摘　　要：为了提高太阳辐照度的预测精度,提出一种利用蝙蝠算法(BA)优化支持向量回归(SVR)的太阳辐照度预测方法。首先,确定SVR预测器的基本结构,选取环境温度、云量、风速、风向、环境湿度以及年积日等与太阳辐照度关系较为紧密的气象监测数据,构成SVR的输入特征向量,将待预测时段小时平均太阳辐照度作为SVR的输出;然后,以预测精度为判断依据,利用蝙蝠算法对SVR的惩罚因子和RBF核函数方差进行寻优;最后,利用最优参数建立SVR预测模型,并对太阳辐照度进行预测。分析结果表明,相比于无参数优化SVR预测模型和利用粒子群算法优化SVR模型的太阳辐照度预测方法,文章所提出的预测方法具有更高的预测精度。In order to improve the prediction accuracy of solar radiation, a new solar radiation prediction method is proposed based on Bat Algorithm optimized Support Vector Regression（SVR）by. First of all, the basic SVR-based predictor is designed. The characters with high relationship with solar radiation such as historical radiation, temperature, cloud amount, wind speed, wind direction, humidity and annual product date are used to construct the input feature vector of SVR.And the hourly average irradiance is used as the output of SVR predictor. Then, the bat algorithm is used to optimize the parameters of the SVR include penalty factor and the variance of RBF kernel function based on the forecasting accuracy of predictor. Finally, the SVR prediction model with optimized parameters is constructed and used to solar radiation forecasting. Experimental results show that the new method has better prediction accuracy than the SVR without optimized parameters and the SVR optimized by Particle Swarm Algorithm.

关 键 词：太阳辐照度预测 支持向量回归 蝙蝠算法 惩罚因子 核函数 

分 类 号：TK511[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]