HYDROGEN PRODUCTION WITH OFFSHORE WIND ENERGY TO FUEL THE DESTINATION GOTLAND FERRIES: A COMPARISON BETWEEN CENTRALIZED AND DECENTRALIZED SYSTEMS

Sammanfattning

The Swedish Island Gotland gets connected to the mainland through a ferry operation by Destination Gotland. In the long term, this route is aimed to be conducted emission free which is why a new generation of ferries is planned. Multifuel gas turbines will propel those ships with the goal to run entirely on green hydrogen. With an assumed energy demand of 16 tons of hydrogen per departure, a solution for a large-scale hydrogen production is needed. The not yet build offshore windfarm Aurora, south-west of Gotland, could provide the electricity for an electrolysis process. As the literature suggests, such a hydrogen production could either be onshore with a cable connection or centralized offshore on a large platform. A third option could be to implement the required equipment at each turbine and therefore reduce the need for extra platforms. This thesis looks at a technical comparison of the two offshore solutions. To achieve comparable results, a simulation tool is developed with Python and the implemented toolbox PyPSA, Python for Power System Analysis. Including simulated wind data for Aurora and the exact demand data for the Destination Gotland ferries, both over a period of one year, allows a direct application to the conducted Gotland Case study. Start-stop cycles, turned-off times and low-load operation are considered; however, the main differences can be found in the storage demand for hydrogen. The decentralized system exceeds feasible capacities, while the centralized system can reach significantly lower capacities. Generally, the centralized system comes with more advantages, even though the economic analyses found in the literature, grant the decentralized system a higher feasibility. Further research needs to be done by combining a technical comparison with more economic factors and by increasing the resolution of the looked at time frame and the level of detail of the simulation.