论文部分内容阅读
提出一种综合使用前向选择(forward selection,FS)和K-means聚类以及径向基函数(radical basis function,RBF)神经网络的光伏发电功率预测方法.模型以每小时的气象因素作为输入量,首先采用前向选择法对原始多维输入量进行约减,在降低维数的基础上减小各个变量间的耦合现象.再通过K-means聚类方法对样本进行聚类,继而对各类数据建立不同的RBF预测模型,避免单神经网络的过拟合问题.实验结果表明,相比于传统的神经网络预测模型,该文使用的模型输入变量更少,预测精度更高.“,”This paper presented a method based on forward selection(FS),K-means clustering and RBF neural network to predict the PV output power.Firstly,using forward selection method to reduce input factors (hourly meteorological factors),so that the coupling phenomenon between variables can be reduced.Then using K-means method to clustering the samples and establish different RBF forecasting models to avoid over fitting problem in single neural network.The simulation results showed that the prediction method using forward selection,K-means and RBF has better prediction accuracy and less input factors than general neural network prediction model.