Research On Optimal Allocation Of Grid-Connected Solar Photovoltaic Power Generation Based On Power Prediction

Solar power generation is an important aspect of clean renewable energy generation,its abundant resources,and clean pollution-free advantages make it concerned.However,due to the influence of weather and environment in the process of photovoltaic power generation,photovoltaic power generation will have great volatility,resulting in random instability in photovoltaic power generation,easy to cause low power quality,unstable output power,and other problems,thus affecting the safe and stable operation of the distribution network when photovoltaic grid connection.How to increase the development and utilization of solar energy is a hot issue in the research of photovoltaic power generation systems.In this thesis,the optimization configuration of gridconnected solar photovoltaic power generation based on power prediction is studied to improve the accuracy of solar photovoltaic power prediction and optimize the grid-connected system configuration of photovoltaic power generation.Firstly,the prediction technology of photovoltaic power is studied.According to the obtained weather data and other relevant environmental information,the combined neural network algorithm is used to establish the photovoltaic power prediction model based on CNN-VMD-PCA feature fusion.The convolutional neural network and variational mode decomposition are used to analyze the existing weather data and extract the required data features.On this basis,the principal component analysis method is used to analyze the processed data.Finally,the XGBoost prediction method is used to predict photovoltaic power generation.Based on the data of the Taike photovoltaic power station in Wenshan City,the prediction effect of the CNN-VMD-PCA feature fusion prediction model was verified by an example,and the prediction results were compared with the traditional regression model,XBGoost prediction model,and CNN-VMD-KPCA model.The results show that the prediction effect of the proposed model is better,the goodness of fit is up to 0.994,and the prediction accuracy is higher than other models.Secondly,the optimal configuration of the photovoltaic power generation system in the process of grid connection is analyzed.Aiming at the optimization of the location and capacity of the photovoltaic power generation system in the distribution network,an appropriate mathematical model is established to minimize the loss of the distribution network and stabilize the voltage of the distribution network.The simulation platform of the photovoltaic power generation system was built by MATLAB software,and the optimization calculation of the simulation system was carried out by power flow calculation and particle swarm optimization algorithm.The simulation results show that the network loss reduction rate of the established site-location optimization model reaches62.52%,which can effectively reduce the grid loss of the photovoltaic power generation system when it is connected to the grid,greatly improve the voltage stability of the distribution network,and effectively improve the voltage quality of the distribution network.The research of grid-connected optimal allocation of solar photovoltaic power based on power prediction starts from two aspects: improving the accuracy of photovoltaic power prediction and optimizing the location and capacity allocation of grid-connected photovoltaic power generation system,which can provide practical reference for improving the comprehensive strength of photovoltaic power stations to a great extent,and provide more effective solutions for practical engineering applications.