Modelling a modular configuration of PEM fuel cell system for vessels applications

Introduction: Increasing restrictions on the vessel emissions both in the ports and at sea has opened to the possibility for the electrification of the vessels using batteries and hydrogen fuel cells thereby reducing the dependency on fossil fuel which emits CO2, NOx, SOx, and other unwanted particle emissions.
Objectives: The hybridization of modularized vessel power trains with fuel cells and batteries is still a novel concept, although it takes advantage of the experience developed in other heavy-duty applications such as cars, buses, forklifts, and trains.
Methods: In this study, we focus on the operation of a modular fuel cell system. Three power allocating strategies across several fuel cell modules were considered: the equal distribution and the sequential distribution (also known as Daisy chain) and the independent distribution. The simulation was run in MATLAB-Simulink where the components models and optimization of the modules power distribution were developed. The system dynamic operation reflects the typical drive cycle and load profile of a ship in a short trip (approx. 30min). The system structure considers fuel cell modules with batteries and ultracapacitors. A real-time Power Management System (PMS) is integrated to decide the optimal fuel splitting between them with the aim of high fuel-efficiency operation. The PMS is developed based on Equivalence Consumption Minimization Strategy (ECMS), an effective energy management technique capable of determining the instantaneous equivalent fuel consumption of energy storage systems and determines the optimal power split with low computational burden and limited calibration of control parameters.
Results: The results point to a modular power splitting strategy which increases the system energy efficiency and reduces the risk of failure. It is finally suggested that a combination of both the sequential and equal distribution of power among the fuel cell modules is in most cases desirable. We find that the sequential distribution provides a better efficiency at lower power demands while equalizing the distribution at higher power demands. The real time optimization suggest that the system can operate at a better efficiency during the load changes by increasing the power supplied by the batteries and ultracapacitors.