The increasing development of 5G technology has  focused attention on the energy consumption of its base stations. As a result, it is crucial to establish energy-efficient 5G networks  and reduce the operating costs associated with 5G base stations. In this paper, a multi-time-scale energy management strategy  based on model predictive control (MPC) is proposed to achieve  this aim. Firstly, a 5G base station model that takes into account  several factors is established, including backup energy storage,  inverter air conditioning scheduling potential, photovoltaic  output fluctuations, load, and temperature. Secondly, a day ahead optimal economic dispatch model for minimizing  operational costs is developed. Thirdly, an intraday rolling  optimization strategy based on MPC to dynamically adjust the  day-ahead operation scheme is proposed. Finally,  comprehensive case studies are carried out, which indicate that  the proposed strategy can effectively improve the robustness  and economy of the system. 

5G base station;model predictive control;rolling optimization;optimal scheduling