一种基于随机森林的太阳能辐射预测模型  

A Solar Radiation Prediction Model Based on Radom Forest

作　　者：刘剑[1] 曹美燕 高治军[1] 许可[1] 

机构地区：[1]沈阳建筑大学信息与控制工程学院,沈阳110168

出　　处：《控制工程》2017年第12期2472-2477,共6页Control Engineering of China

基　　金：国家自然科学基金(61272253);辽宁省教育厅科学研究项(L2015443)

摘　　要：为了高效利用太阳能,准确的预测太阳能辐射情况极其重要。针对太阳能辐射的预测问题,研究了基于随机森林的太阳能辐射预测方法,根据影响太阳能辐射的因子建立了随机森林树型分类器,构建了一种基于随机森林的太阳能辐射量预测模型;将该模型的预测结果与BP神经网络、SVM辐射模型的预测结果进行对比。并对具体地区的太阳辐射进行预测,检验了该模型有效性。结果表明:采用随机森林模型预测效果较好,减少了均方根误差、提高了模型的预估精度,对复杂环境下的太阳能辐射量预测、光伏发电有效利用具有重要指导意义和应用前景。Aiming at the low utilization rate of solar energy, the solar radiation is needed to be predicted accurately. The proposed method in this paper conducts a research on the solar radiation prediction method based on random forest, which aims at the solar radiation prediction problem. According to the influence of solar radiation factors, the random forest tree classifier is established and a solar radiation prediction model is built up as well. Prediction results of this model are compared with BP neural network and SVM radiation model predictions. Then the model which is tested with the real data is used to predict the solar radiation of the specific area and its effectiveness is tested. Experimental results show that the prediction effect of the random forest is better than the other two models, the root mean square error is reduced and the prediction accuracy of this model is improved. This method has important research values and application prospect for the prediction of solar radiation quantity in complex environments.

关 键 词：太阳能 辐射 随机森林 预测模型 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]