Producing hydrogen from renewable energy sources will play a pivotal role in the energy transition. It can reduce emissions in hard-to-abate sectors and be used for grid balancing services. The energy industry is looking into multiple different system configurations to produce green hydrogen. One such system is an offshore stand-alone wind turbine with the sole purpose of producing hydrogen. The reason for this interest is due to decreasing offshore wind costs, electrolyzer capital expenditure requirements are dropping, and such a system will have high electrolyzer operating hours. Further cost reductions can be achieved by substituting the electrical infrastructure in offshore wind farms as well as on the wind turbine, for hydrogen infrastructure.

When directly connecting an electrolyzer to a non-grid connected fluctuating power source, it will be hard to prevent frequently turning the electrolyzers on and off. That has been found to accelerate component degradation, thereby reducing the system lifetime.

This report aims to address how the system design and the advisory control strategy can increase hydrogen production while decreasing the number of electrolyzer turn-offs. The investigated system comprises a 20 MW wind turbine directly connected to PEM electrolyzers, a desalination system, a
water tank, and a battery. Two PEM electrolyzer configurations were compared, namely 4x5 MW and 2x10 MW. Furthermore, two interesting system additions were analyzed. Firstly, the inclusion of a weather forecast to aid the advisory control with making decisions on when to turn electrolyzers on.
Secondly, an additional storage element that could keep the last electrolyzer idling during a period of low wind speeds.

The individual components of the system were firstly modeled and then connected to get a representative system model. Two different advisory control strategies, which operate the components, were then designed to simulate the system behavior with regard to hydrogen production and the number of
turn-offs, among other key performance indicators.

This thesis project was done in collaboration with Vattenfall, who provided access to 5-second resolution power output data from a single offshore wind turbine. Using that data, different system designs and versions of the two advisory control strategies were simulated with the aim of selecting the best system design and advisory control strategy.

The results showed that the system design and advisory control could significantly reduce the number of electrolyzer turn-offs while producing close to the highest potential of hydrogen. The weather forecast and additional storage element brought substantial benefits regarding the electrolyzer turn-offs but did
not largely impact the hydrogen production.