Artificial Intelligence Based Optimal Operation of Green Hydrogen Refueling

This study presents an AI-based optimized energy management model for a hydrogen refueling station with a PV system and fuel cell (FC). Using the Time-series Dense Encoder (TiDE) model for day-ahead PV generation and electricity price forecasting, with real data from Spain, the station efficiently manages hydrogen demand and electricity sales. Despite lower-than-expected PV production (11,559.39 kWh vs. 12,991.71 kWh), real-time revenue rose to 1,144.84 Euros, demonstrating operational flexibility and profit.