Evaluation of 5G base station energy storage adjustable potential based on MF-LSTM

A major obstacle to the widespread adoption and long-term sustainability of 5G base stations is their high power consumption. Implementing an energy storage system serves as an effective approach to mitigate peak energy demands and operational expenses. However, a reliable approach to accurately evaluate the adjustable energy storage potential of 5G base stations across different communication load scenarios is still insufficient. This study introduces an evaluation technique leveraging the Multi-Factor Long Short-Term Memory Network (MF-LSTM), incorporating key variables such as communication load, meteorological data, and historical power usage to build a predictive model. The experimental findings indicate that, compared to conventional time series forecasting techniques like ARIMA and SVR, the MF-LSTM model enhances the accuracy of predicting energy storage adaptability, thereby optimizing energy dispatching strategies. The findings of this study provide a theoretical basis for the intelligent management of energy storage systems in 5G base stations, laying the groundwork for the development of sustainable and environmentally friendly communication networks.