| Accurate prediction of photovoltaic power generation capacity plays an important role in the safe and stable operation of photovoltaic power generation systems.However,due to the instability,intermittentness and randomness of solar energy,the existing photovoltaic power generation forecast models have problems such as large forecast errors and low generalization ability.If there are too many power stations,training a single photovoltaic power station separately for the power generation prediction model will increase the training cost.Therefore,this article will study the application of multi-power station photovoltaic power generation prediction based on hybrid neural network.The specific work is as follows:First of all,This paper summarizes the current research status of photovoltaic power generation forecasting and analyzes related technologies.It focuses on the application of hybrid neural network models in the field of photovoltaic power generation forecasting.The hybrid model based on neural networks is selected for multi-station photovoltaic power generation forecasting.Secondly,According to the characteristics of daily power generation data and hourly power generation data of multiple power stations and corresponding meteorological data,data preprocessing operations such as missing value supplementation,abnormal value detection,weather digitization,and normalization are performed respectively.Then,this article will study the ultra-short-term forecast and short-term forecast of photovoltaic power generation from multiple power stations..For the correlation between the hourly power generation used by the ultra-short-term forecasting model and the corresponding meteorological data,a dual-channel forecasting model was constructed.It uses an attention mechanism based long short-term memory neural network to learn the time characteristics of the data and uses an dilated causal convolutional neural network to learn the spatial characteristics of the data.After fusion,single-step prediction and multi-step prediction were performed,and the prediction accuracy under different weather conditions was compared from different angles,which verified the effectiveness of the proposed model.For the short-term prediction model,the daily power generation and corresponding meteorological data are used,and the short-term prediction model of photovoltaic power generation based on genetic algorithm and multiple deep learning algorithms is adopted.The effectiveness of the model is verified by analyzing the prediction effects of different seasonal models.Finally,designed a photovoltaic power generation forecasting system based on the B/S structure,using the django framework for back-end development,and vue.js for front-end development.The system realized the display of multi-power station daily power generation forecast curve and hourly power generation forecast curve display. |
