The Establishment Of Photovoltaic Cell Model Based On Neural Network

With the development of the economy and the continuous exploitation of energy,the crisis of insufficient fossil fuels has gradually appeared.Solar energy is a new green energy which has the advantages of safety,cleanliness,and large quantities.Traditional energy sources are gradually depleted and environmental problems are increasing year by year.The use of renewable energy,especially solar energy,provides a way for mankind to solve the crisis,and its industrial potential is huge.Solar energy can be converted into electrical energy directly by photovoltaic systems,and the main body of photovoltaic systems is photovoltaic cells,so the research on photovoltaic cell modeling is of great significance for determining the output characteristics of photovoltaic cells under outdoor conditions,especially the maximum power tracking.The main research content of this paper is to use neural network to quickly solve the photovoltaic cell model parameters,and to further improve the accuracy of building the battery model.Secondly,it can solve the photovoltaic cell output current-voltage characteristic curve under a certain condition and predict the maximum power point.This article first introduces the working principle of photovoltaic cells,as well as the commonly used single diode models and shape models of photovoltaic cells,and classifies the parameters of the models.Secondly,the artificial intelligence method commonly used for prediction is explained,the principle of artificial neural network is introduced,and the BP neural network and wavelet neural network are introduced in detail.In order to improve the accuracy of using neural network to obtain model parameters,a method for finding similar days is established to classify weather data,and then based on the prediction method combining wavelet decomposition and neural network to predict the parameters of the photovoltaic cell model under certain conditions,thus Establish the single diode model and shape model of the photovoltaic cell.Finally,the proposed method is verified by the actual measured historical data.The historical weather is divided into three categories:sunny,cloudy,and cloudy using the method based on similar days.The photovoltaic cell models are established for the three types of weather,and NRMSE is used.Two error functions used to evaluate the accuracy of the model with MAPE are used to verify and compare the accuracy of the established model.The verification results show that under three weather conditions,the model parameters predicted by the combination of wavelet decomposition and neural network and the established photovoltaic cell model are faster and more accurate.The method of combining wavelet decomposition and neural network proposed in this paper to predict the parameters of photovoltaic cell model is faster than analytical and numerical methods.The prediction of I-V curve and maximum power point is more accurate than that of BP neural network.