Energy Management Optimization for Cellular Networks Under Renewable Energy Generation Uncertainty

The integration of renewable energy (RE) as an alternative power source for cellular networks has been deeply investigated in the literature. However, RE generation is often assumed to be deterministic; an impractical assumption for realistic scenarios. In this paper, an efficient energy procurement strategy for cellular networks powered simultaneously by the smart grid (SG) and locally deployed RE sources characterized by uncertain processes is proposed. For a one-day operation cycle, the mobile operator aims to reduce its total energy cost by optimizing the amounts of energy to be procured from the local RE sources and SG at each time period. Additionally, it aims to determine the amount of extra generated RE to be sold back to SG. A chance constrained optimization is first proposed to deal with the RE generation uncertainty. Then, two convex approximation approaches: 1) Chernoff and 2) Chebyshev methods, characterized by different levels of knowledge about the RE generation, are developed to determine the energy procurement strategy for different risk levels. In addition, their performances are analyzed for various daily scenarios through selected simulation results. It is shown that the higher complex Chernoff method outperforms the Chebyshev one for different risk levels set by the operator.