Energy Scheduling Model for Photovoltaic 5G Base Station Based on Deep Constrained Deterministic Policy Gradient

With the development of energy internet technology, the configuration of distributed photovoltaic and energy storage batteries in 5G base stations will become a potential solution for the high energy consumption and large electricity cost of communication base stations. In this paper, a day-ahead optimal scheduling strategy based on a deep constrained deterministic policy gradient (DCDPG) algorithm is proposed for photovoltaic 5G base station microgrid. The daily operation and maintenance cost of the base station system and the aging of energy storage battery are considered comprehensively, making the energy dispatching plan for each power unit. Aiming to enhance the accuracy of action selection and accelerate convergence of algorithm, further action constraints are embedded in DCDPG. The simulation results show that the proposed model and algorithm can satisfy the requirements of 5G base station economic operation with high-performance.