The Ice2Thrust project advances well beyond existing research by experimentally analysing the efficiency of lunar water extraction and recovery under varying conditions, supported by computational simulations. Its work goes beyond earlier feasibility studies by examining the composition of icy regolith and optimising the cold-trap process for ice retrieval, paving the way for future in-situ analyses on the Moon and a more mature extraction technology.
A space-capable PEM electrolyser could be systematically characterised for varying environmental conditions. Additionally, for the first time worldwide, additively manufactured porous metal structures were used for transpiration cooling of a rocket engine thrust chamber. The results show a strong potential for a significantly increase in fuel economy due to a decrease in amount of coolant needed. Furthermore, a docking adapter design could be investigated that is more compact and simplistic than currently available designs for the transfer of water between spacecraft.
At the same time, the project demonstrates the successful use of reinforcement learning for autonomous docking, achieving robust simulation-to-real transfer through large-scale training and validation in the Zero-G laboratory. Continued work on constrained and model-based reinforcement learning is expected to enhance safety, adaptability, and autonomy for future in-space servicing operations.
Finally, the development of a dedicated Remote Interface Unit for a satellite platform marks a technological breakthrough in managing the complex operations of a water electrolysis propulsion system. This architecture validates water as a safe, sustainable propellant and lays the groundwork for a new European ecosystem in in-orbit servicing, refuelling, and assembly, with future progress depending on hardware maturation, in-orbit demonstration, and the adoption of standardised interfaces to enable commercial interoperability.