Development of an intelligent real-time capable energy management strategy for a hybrid maritime propulsion system considering component aging

Abstract

Stricter emissions norms, especially on CO2 posed by international organizations encourage the maritime sector to seek new cleaner propulsion technologies. A hybrid propulsion system powered by a fuel cell system and a battery system offers the potential to eliminate exhaust gas emissions and is a promising technology to achieve the complete drivetrain electrification of maritime propulsion systems. In this work, a real-time capable energy management strategy that takes into account the aging of the propulsion components is introduced. The energy management strategy achieves the cost-effective operation of a hybrid drive train powered by a battery and a fuel cell for a large-scale propulsion application of a cruise ship. To achieve this, a Q-learning-based agent has been trained with multiple power demand profiles. In this novel method, a reduction in fuel cell degradation is achieved by decreasing its dynamic operation, while the battery pack degradation is reduced by minimizing its capacity drop and resistance. The aging of both components was performed using parameterized aging models. As a result, intelligent power control rules are obtained which can be directly implemented with comparatively low computational effort for real-time control. The developed energy management strategy improves the fuel economy and reduces the degradation of the propulsion components compared to conventional real-time capable rule-based operation strategies.