101. Comparative analysis and test bench validation of energy management methods for a hybrid marine propulsion system powered by batteries and solid oxide fuel cells.
- Author
-
Ünlübayir, Cem, Youssfi, Hiba, Khan, Rehan Ahmad, Ventura, Santiago Salas, Fortunati, Daniele, Rinner, Jonas, Börner, Martin Florian, Quade, Katharina Lilith, Ringbeck, Florian, and Sauer, Dirk Uwe
- Subjects
- *
GREENHOUSE gases , *ELECTRIC power , *HYBRID power systems , *CLIMATE change mitigation , *ELECTRIC propulsion , *SOLID oxide fuel cells - Abstract
Climate protection goals and the transformation of the mobility sector are pushing the shipping industry to develop new propulsion systems emitting fewer or no greenhouse gases. One promising approach to eliminate greenhouse gas emissions from ships is a hybrid propulsion system powered by fuel cells and batteries. A high-temperature solid oxide fuel cell (SOFC) can supply heat and the electrical power demand in combination with a battery. Due to the low dynamic performance of the SOFC when faced with sudden load changes, a battery is responsible for providing the power for the dynamic load components. To ensure the resource-efficient operation of the propulsion components, intelligent energy management methods are required for power distribution control. Implementing a machine-learning-based energy management method based on twin-delayed deep deterministic policy gradient (TD3) improves the overall system efficiency, lifetime, and fuel economy compared to conventional energy management methods. To verify the technical feasibility of the propulsion system including its controls, the system is tested in a hardware-in-the-loop (HIL) environment. By implementing the TD3-based algorithm within the energy management used on the test bench, hydrogen consumption was reduced by approximately 10% and the remaining battery capacity after five years was 6% higher in comparison to conventional energy management methods. [Display omitted] • Hybrid electric propulsion system for a marine application. • Development of machine-learning-based energy management algorithms. • Validation measurements with propulsion components. • Assessment of the economic performance for a large cruise ship application. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF