Green Base Station GCD-BED Modeling And Optimization In Dual Power Supply Mode

The requirements for higher reliability,larger scale,lower latency,and higher frequency bands are proposed in the innovation from the 5th generation mobile communication(5G)to the 6th generation mobile communication(6G).This leads to a significant increase in energy consumption in mobile communication and poses significant challenges to environmental governance.In the energy consumption composition of 5G/6G,the energy consumption of base stations(BS)reaches over 80%and continues to rise rapidly each year.The increase in energy consumption results in increased electricity cost expenditure for telecommunication operators and places significant pressure on the power grid(PG).Renewable energy(RE)possesses characteristics such as clean use and low cost in comparison to traditional energy sources.The RE-PG dual-power supply mode has become a new paradigm due to innovations in energy harvesting(EH)technology,energy storage(ES)technology,and the transformation brought about by artificial intelligence(AI).However,it is still worth exploring how to evaluate the RE output and make reasonable use of "idle" battery resources in this power supply mode.This discussion is focused on not only improving RE integration and battery utilization,but also reducing electricity costs and dependence on the PG.In this study,a hybrid energy supply architecture based on intelligent battery charging and discharging management is initially proposed for the scenario where the RE-PG dual-power supply mode provides power to BS.This architecture involves RE,energy sharing,heterogeneous traffic BS load,and battery status.The evaluation of the RE output in the dual-power supply mode is based on the perspective of energy consumption.Moreover,a battery energy management and grid-connection depth(GCD)model for heterogeneous traffic BS is established.In this model,energy and energy storage information are transmitted between BS through power information lines.Energy sharing is utilized to redistribute energy among BS,maximizing the utilization of renewable energy and reducing dependence on the power grid.Following this,a battery charging and discharging strategy for BS energy storage batteries is proposed based on this model to enable the management of battery energy for BS.Finally,to obtain the optimal GCD value,a GCD optimization platform is established,and the deep Q-learning(DQL)algorithm is employed to learn the time-varying EH information,BS load information,and battery status information for the energy management of BS batteries.Furthermore,considering BS energy storage as a distributed energy storage resource and involving it in PG energy scheduling can enhance BS energy utilization and balance PG stability.Additionally,the battery storage space is divided into backup zones,scheduling zones,and low-cost energy storage zones.It is associated with power load and EH information.Next,a battery energy scheduling scheme and a battery energy scheduling(BED)model are proposed.Moreover,considering the influence of PG timeof-use electricity prices and battery energy status on BED,a BED strategy and a deep learning optimization algorithm are proposed to achieve the optimal BED.It is demonstrated by the simulation that a maximum 25%reduction in GCD value can be achieved by the proposed BS energy management algorithm compared to offline algorithms,under the condition of unknown causal information and real-time energy sharing.Additionally,it is observed that when BS participates in energy scheduling,the amount of electricity purchased during peak electricity prices is effectively reduced,and the backup power during off-peak periods is increased.Furthermore,the number of daily battery charge and discharge cycles remains unchanged.These outcomes showcase not only decrease in electricity expenditure for telecommunication operators but also reduced dependence on the power grid and carbon emissions.