Research On Short-term Prediction Of Photovoltaic Power Based On ARIMA And LM-BP Neural Network

Short-term photovoltaic power prediction based on ARIMA and LM-BP neural networkWith the intensification of the energy crisis and environmental pollution,the advantages of photo voltaic power generation such as environmental protection and renewable energy have been favored by people.However,the volatility and randomness of photo voltaic power generation will have an impact on the stable operation of the power grid,and photo voltaic output prediction has played a crucial role in solving this problem.The prediction of photo voltaic output will help the stable operation of the power grid and improve the effective use of solar energy,which is of great significance to the development of the photo voltaic industry.In this paper,the differential autoregressive moving average-neural network model(ARIMA and LM-BP)is used to predict photovoltaic output.The purpose is to solve the problem of large fluctuations and instability in the prediction results of a single neural network.By analyzing the factors that affect the output of photo voltaics,it divides the direct and indirect factors.The direct factor is used as the input of the autoregressive integral moving average model to predict the trend of photo voltaic output and extract the corresponding error sequence.The indirect factors are constructed as similar daily sets,together with the error sequence as the input of the subsequent LM-BP neural network model to predict the fluctuation of photo voltaic output..The specific work content of this article is as follows:(1)Conduct research on the working principle of photo voltaic panels and the equivalent circuit of photo voltaic power generation.Through the derivation of mathematical models and simple experiments,it is concluded that the irradiation intensity and the photo voltaic output in an ideal state have a nearly linear relationship,which affects the photovoltaic output Analyze the factors one by one to lay the foundation for subsequent similar days.(2)The theoretical basis for predicting the photo voltaic output by the autoregressive integral moving average model is expounded,and the irradiation intensity is predicted and corrected.In view of the difference in the irradiation intensity in different seasons,a four-day sunny power generation prediction model is constructed.(3)The selection and processing methods of similar days are introduced in detail,and a set of similar days is established.Introduce the principle of neural network,build two neural network models,analyze the advantages of the LM-BP algorithm through the comparison of examples,and finally combine the LM-BP neural network model with the autoregressive integral moving average model to complete the ARIMA and LM-BP neural network model Building.(4)Compare the experimental results obtained by the LM-BP neural network,the traditional BP neural network and the ARIMA and LM-BP neural network model.The results show that the ARIMA and LM-BP algorithm has good performance in terms of prediction accuracy and stability,showing this article The effectiveness and practicability of the proposed method.