Performance Analysis and Resource Allocation for Intelligent Solar Aware Cellular Base Stations

In response to the global climate crisis, solar-powered cellular base stations (BSs) are increasingly attractive to mobile network operators as a green solution to reduce the carbon footprint of networks. However, solar power presents challenges due to its diurnal nature and significant variations in harvestable power due to weather changes. To address these challenges, solar deployments rely on batteries to provide power during the night and periods of low sunlight. Batteries are expensive and have complex environmental considerations, making it desirable to minimize their use at each BS and avoid over-provisioning. Accurate prediction of energy income is essential to minimize the number of Photovoltaic (PV) panels and batteries required while maintaining the quality of service (QoS), achieve a desired energy outage probability, for users. This paper introduces an innovative approach to predict energy harvesting by utilizing a novel conditional Long Short-Term Memory (Cond-LSTM) neural network architecture. Compared with standard LSTM and Transformer models, the Cond-LSTM model reduced the normalized root mean square error (nRMSE) by 69.6% and 42.7%, respectively. The proposed approach facilitates an accurate, cost-optimal PV-battery configuration that meets outage probability requirements and aids in site design for regions lacking historical solar energy data. Additionally, we propose a solar-aware cellular communication scheme and user power allocation to enhance QoS via signal-to-noise ratio (SNR) optimization and minimize the probability of energy outages in the cellular system. Simulation results show the efficiency of our proposed solar aware model in decreasing the overall outage probability of the system and increasing the data throughput of the cellular system. For the same battery configuration, our algorithm achieves 60 percent smaller outage probability and 8 billion bps greater data throughput than a non solar aware approach.