Research on the integration and optimization of the PV-storage-DC-flexibility system in highway tunnel applications

This study proposes an optimization scheme for the PV-storage-DC-flexibility system based on the combination of Particle Swarm Optimization (PSO) and Q-learning reinforcement learning. The aim is to improve the energy efficiency, stability, and economic performance of photovoltaic (PV) generation and energy storage systems in highway tunnels. A comprehensive model including PV generation, energy storage systems, and DC flexible grids is established, and simulation verification is conducted using MATLAB/Simulink. The PSO algorithm is employed to optimize the system's energy distribution, enhancing the self-consumption rate of PV generation and energy storage, while maximizing economic benefits under varying lighting and load conditions. The Q-learning algorithm implements dynamic scheduling strategies by learning and adjusting real-time states of energy storage charging and discharging, as well as PV power supply methods, thereby optimizing system operation. Experimental results show that the optimized system achieves an energy self-sufficiency rate of about 90% under high lighting conditions and over 50% under low lighting conditions. In cases of significant load fluctuations, the system’s power supply reliability increases to 0.85, which significantly reduces dependence on the external grid compared to traditional grid systems. Hardware-in-the-loop experiments further validate the system’s stability and robustness. Finally, a life cycle cost analysis reveals that the PV-storage-DC-flexibility system demonstrates clear economic advantages over a 10-year operation period, with saved electricity costs increasing year by year and equipment life being significantly extended.