基于神经网络的光伏短期出力预测

Short Term Photovoltaic Output Prediction Based on Neural Network

本文分析了光伏发电设备的输出功率与影响天气的因素之间的相关性,影响天气的因素包括辐照度、温度、湿度、天气类型和季节。其次,在此基础上,分析了小波神经网络的相关理论,完成了小波神经网络的预测模型设计,而后,提出了一种基于气象因素寻找相似日的方法,并基于光伏电站的历史气象信息创建了特征向量,并通过计算灰色关联度来确定预测日相似日的样本集。最后,将相似日算法和小波神经网络创建光伏电站的短期电力预测模型。预测结果表明,基于神经网络的预测模型在一定程度上有参考利用价值。

This article analyzes the correlation between the output power of photovoltaic power generation equipment and the factors that affect weather,including irradiance,temperature,humidity,weather type,and season.Secondly,based on this,the relevant theories of BP neural network and wavelet neural network were explained,and the prediction model design of wavelet neural network was completed.Then,a method for finding similar days based on meteorological factors was proposed,and feature vectors were created based on historical meteorological information of photovoltaic power plants.The sample set for predicting similar days was determined by calculating grey correlation degree.Finally,the similar day algorithm and wavelet neural network are used to create a short-term power prediction model for photovoltaic power plants.The prediction results indicate that the prediction model based on neural networks can meet the practical application needs to a certain extent.