Research On Short-term Output Forecast Of Photovoltaic Power Plants In Weining, Guizhou

In recent years,with more and more serious problems such as energy shortage and environmental pollution,energy transformation is imperative.As a clean energy,the installed capacity of photovoltaic is increasing day by day,but the ensuing potential problems are also gradually highlighted.As photovoltaic power generation is affected by various uncertain weather factors and has strong randomness and volatility,large-scale photovoltaic grid-connection will pose a threat to the safe and stable operation of the power grid.Therefore,accurate and rapid photovoltaic output prediction is conducive to the scheduling department to formulate scheduling plans,expand the proportion of photovoltaic grid-connection,promote photovoltaic consumption,and reduce the adverse effects of photovoltaic grid-connection.Firstly,the paper introduces the development of photovoltaic power generation at home and abroad and the research status of short-term prediction of photovoltaic output Based on the historical data of Weining meteorological station,the solar energy resources and meteorological elements of a photovoltaic power station in Weining area of Guizhou were analyzed by using the data obtained from meteorological method and remote sensing inversion method.Secondly,in order to improve the accuracy of short-term output prediction of photovoltaic power stations,this paper briefly introduces the principle of photovoltaic cell power generation and grid-connected photovoltaic system,and then analyzes the main factors affecting the output of photovoltaic power stations one by one.Through the Pearson correlation coefficient method,the influencing factors with high correlation were selected as the input data of the prediction model,so as to improve the accuracy and calculation efficiency of the prediction model.Then,aiming at the problem that the traditional neural network prediction method has insufficient ability to mine the time correlation of data,a short-term output prediction model of photovoltaic power station based on Long Short-Term Memory(LSTM)is proposed.Based on the historical measured data of a photovoltaic power station in Weining,Guizhou Province,BP,RNN and LSTM prediction simulation models are established to compare the prediction results under different weather types.The results show that the LSTM model is effective and has better prediction performance.Finally,considering the mechanism characteristics of photovoltaic output prediction data,a short-term output prediction model based on CNN-LSTM is proposed by introducing a Convolutional Neural Network(CNN)on the basis of the LSTM prediction model.The model combines the feature extraction ability of CNN with the historical memory ability of LSTM,so as to realize the fusion of spatial and temporal information.Compared with the traditional prediction model,the results verify the effectiveness of the proposed model,and the model has stronger fitting ability and generalization.