Incorporating speed forecasting and SOC planning into predictive ECMS for heavy-duty fuel cell vehicles

[EN] This study presents a novel approach specifically designed for real-world driving scenarios of heavy-duty fuel cell vehicles, named P-ECMS. The P-ECMS addresses both charge-sustaining and charge-depleting modes of the battery to optimize the vehicle¿s energy management. To this aim, the P-ECMS...

Descripción completa

Detalles Bibliográficos
Autores: Piras, M., De Bellis, V., Malfi, E., Desantes J.M.|||0000-0002-4124-9393, Novella Rosa, Ricardo|||0000-0002-5123-6924, López-Juárez, Marcos|||0000-0001-9886-4728
Tipo de recurso: artículo
Fecha de publicación:2024
País:España
Institución:Universitat Politècnica de València (UPV)
Repositorio:RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Idioma:inglés
OAI Identifier:oai:riunet.upv.es:10251/222907
Acceso en línea:https://riunet.upv.es/handle/10251/222907
Access Level:acceso abierto
Palabra clave:Hydrogen
Proton exchange membrane fuel cell
Predictive energy management strategy
Fuel cell hybrid electric vehicle
Realistic driving conditions
Battery SOC planning
13.- Tomar medidas urgentes para combatir el cambio climático y sus efectos
Descripción
Sumario:[EN] This study presents a novel approach specifically designed for real-world driving scenarios of heavy-duty fuel cell vehicles, named P-ECMS. The P-ECMS addresses both charge-sustaining and charge-depleting modes of the battery to optimize the vehicle¿s energy management. To this aim, the P-ECMS integrates a velocity prediction layer and a SOC planning layer. The velocity prediction layer utilizes a realistic driving dataset obtained from GT-Real Drive, using information from the TEN-T routes, to accurately predict the speed of the vehicle. The SOC planning layer, leveraging information from a map service provider, plans a target SOC trajectory at the beginning of the driving mission. It employs a neural network trained on the policy obtained from a standard optimized ECMS for various driving cycles and initial SOC values, aiming to achieve a final SOC of 30%. In charge-sustaining operations, the P-ECMS is compared to a conventional Adaptive-ECMS, the reference ECMS (Standard-ECMS), and a rule-based strategy across the HDDT driving cycle. The evaluation focuses on battery SOC sustenance, equivalence factor evolution, and hydrogen consumption. Results show that both the P-ECMS and the A-ECMS outperform the S-ECMS in terms of SOC sustenance, with the P-ECMS achieving a significant 2% reduction in hydrogen consumption compared to the A-ECMS. The study demonstrates that the P-ECMS advantages extend to battery discharge conditions, as it achieves a remarkable reduction in consumption compared to an optimal Charge-Depleting/Charge-Sustaining (up to 5%) when employing a linear battery discharge planning. The integration of the SOC planning layer proves additional benefits, and a comparison between the P-ECMS with linear battery discharge planning and the P-ECMS with the SOC planning layer integrated shows the advantages of SOC trajectory planning for different segment lengths. The study suggests an optimal segment length between 3 km and 8 km, obtained with the necessary data from a map service provider.