AI-driven hydrogen fuel cell and battery energy management system for off-road construction heavy equipment

To comply with zero-carbon regulations, proton exchange membrane fuel cells (PEMFC) are considered as power sources for off-road heavy equipment. Owing to their high energy density and highly dynamic driving conditions, PEMFC-battery (PEMFC-BAT) hybrid power systems have attracted considerable attention. In this study, energy management for PEMFC-BAT hybrid systems was optimized using real vehicle driving test data. AI-based load forecasting was performed to stabilize the output power of the PEMFC. Long short-term memory (LSTM) networks were used for time-series forecasting of the required load. By learning from various real-vehicle driving data, the energy management system (EMS) stabilized the output power of the PEMFC and reduced hydrogen consumption. The final system is expected to increase the lifetime of PEMFCs.