结合相似日GA-BP神经网络的光伏发电预测

PV Generation Forecasting Combined with Similar Days and GA-BP Neural Network

为了高精度预测光伏发电,减小并网光伏对电力系统运行的不利影响,本文引入相似日的概念,对预测日的天气信息进行分析,根据天气信息、季节等数据,通过聚类的方法,在历史数据中筛选出与预测日特征相似的历史发电数据和天气数据,作为预测模型的训练样本,并采用GA-BP神经网络对系统进行建模以及光伏发电预测。通过对某光伏系统数据验证,计算了预测误差。分析结果表明该方法具有较高的预测精度。

In order to forecast the photovoltaic （PV） power generation with high accuracy and reduce the impact of gridconnected PV system on power system, the concept of similar day is introduced in this paper to analyze the weather information on the prediction day. According to the data such as weather and seasonal information, the historical generation data and weather data which have similar characteristics to those on the prediction day are chosen from historical data through the clustering method, and they are further used as training samples for the prediction model. Moreover, genetic algorithm-back propagation （GA-BP）neural network is used to establish the prediction model and forecast the PV power generation. The proposed method is verified based on the actual data of a PV system, and the forecasting error is calculated. The analysis results show that the forecasting method has higher prediction accuracy.