Research on User Side Photovoltaic-Energy Storage-Charging Configuration Strategy Based on Improved Deep Learning Algorithm and Its Engineering Application

At present, there are various types of energy storage on the user side, including the charging piles+energy storage, photovoltaic+energy storage, photovoltaic+charging piles+energy storage, etc. Each charging type has own characteristics. And for charging station applications currently distributed in urban centers, urban-rural integration areas, rural areas, and other application scenarios, there are significant differences in the utilization rate of electric vehicles on charging stations for each application scenario. In this article, the load of typical application scenarios is quantitatively calculated, and a certain capacity of energy storage equipment is configured in each application scenario to achieve the optimal ratio between load and energy storage capacity. This article divides into three types of regional features and effectively calculates the annual charging volume for each hour in each typical 24-hour period. Furthermore, the application scenarios will be extended to different scenarios such as photovoltaic, energy storage, and wind power, simulating the application benefits of integrated wind solar energy storage and charging projects. After obtaining 8760 hours of data on charging stations in each charging station, they were trained and learned through improved deep learning algorithms and RBF neural networks, which can be quantitatively characterized through the non specific function expressions. Specifically, for photovoltaic and photovoltaic power stations equipped with energy storage units, optimal ratio needs to be determined through intelligent algorithms to achieve maximum revenue. Furthermore, optimization strategy for the optical storage and charging configuration based on improved deep learning algorithms proposed in the article will be applied to specific engineering projects, and the profitability of the engineering projects will be quantitatively analyzed. The effectiveness of the proposed theoretical model will be demonstrated through specific data analysis.