基于相似日和主成分分析的光伏发电系统短期出力预测

摘要

In view of the fluctuation and the intermittence of t Photovoltaic power,based on principal component analysis (PCA) and particle swarm optimization (PSO) optimization,a short term forecasting method of BP neural network power is proposed.Firstly,this method analyzes the original input data by principal component analysis,then which uses the analysis results as the input data of the BP neural network.The search speed of particle swarm algorithm is slow,but it has a better overall search capability.The traditional BP neural network search speed is relatively fast,but it is prone to local extreme points.Therefore,the combination of the two can make up for both the disadvantages and the failure of the prediction model,so the prediction accuracy of prediction model is improved.The results show that the prediction model is invariable when the type changes,and the forecast error is less than 20％.%针对光伏发电功率的间歇性和波动性,提出了一种基于主成分分析(PCA)和粒子群优化(PSO)算法的BP神经网络短期发电功率预测方法.该方法先对原始输入数据进行主成分分析,再将分析结果作为BP神经网络的输入数据.由于粒子群算法搜索速度较慢,但全局搜索能力较强,而传统的BP神经网络搜索速度较快,但易陷入局部极值点,因此将两者结合起来,既弥补了各自的劣势,又避免了预测模型的失效,从而提高了预测模型的预测精度.分析结果表明,当天气类型改变时,该预测模型的有效性不变,预测误差均小于20％.