Connecting Space Logistics and Architecture - a Pattern Language for Robust Mission Design
The convergence of new technologies enabling the reuse of in-space transportation, on-orbit servicing, and in-situ resource mining and transformation, coupled with significant interest in ambitious missions such as robotic and crewed Lunar and Mars missions, has increased the focus on missions with complex logistics. Existing mission design tools, such as concurrent engineering, primarily rely on a known or limited set of concepts for investigation, while concept design methods typically yield only a single optimized solution or explore a subset of the trade space based on user assumptions. Since most mission performance is determined during the concept phase, methods to quickly and systematically identify efficient mission concepts are needed. The main challenges identified include the complexity in formalizing architectures, evaluating them, and managing the combinatorial dimensionality of the solution space. To address these challenges, a design support tool is proposed. This tool can, given a set of mission objectives, automatically and rapidly: 1) generate mission architecture concepts, 2) evaluate each concept according to key performance metrics, and 3) illustrate the concepts intuitively, enabling space mission design specialists to quickly understand and compare them, including on qualitative criteria which vary significantly with each mission's context. A method based on a formalized pattern-language system is proposed, using patterns as building blocks to compose solutions. A pattern language specific to space mission design problems, including individual patterns representing possible actions such as launching a payload into Earth orbit or performing electrolysis of water, is developed. The method is illustrated and validated through application and comparison with two real-world cases: the Apollo and Mars Sample Return missions. Additionally, challenges in space technology road-mapping are identified. Requirements for space technologies are defined at the mission concept level, and their development is resource-intensive with long development times. Given the high uncertainty around which mission concepts will be developed in the future, technologies must demonstrate both high performance and versatility for integration into various mission objectives. This relies on the quick generation of diverse concepts for multiple missions and linking them to technology requirements. Therefore, it is demonstrated how statistical analysis of generated concepts using the developed mission concept generation tool can complement technology assessment methods. Overall, it is found that such a tool can indeed capture pivotal mission dynamics and outcomes at the mission concept level and support the identification of technologies likely to be included in high-performing concepts across a range of missions.
Dr Mathias Lerch (président) ; Prof. Jean-Paul Richard Kneib, Prof. Jeffrey Huang (directeurs) ; Prof. Stefana Parascho, Prof. Maarten Bonnema, Prof. Reinhold Bertrand (rapporteurs)
2024-11-15
Lausanne
2024-11-12
10810
148