Computational morphogenesis method for the lunar shelter based on the principle of inverse hanging simulation and bi-directional evolutionary structural optimization

Lunar exploration has flourished in recent years, with rovers, robots, and eventually buildings expected to be gradually deployed on the Moon to support scientific research. To protect these facilities from extreme environmental conditions, such as large temperature variations, intense radiation, meteoroid impacts, and frequent moonquakes, artificial lunar shelters are required. A computational morphogenesis method for lunar shelters considering the mechanical properties of in-situ sintered lunar regolith is proposed. This workflow integrates the inverse hanging simulation to find forms dominated by compression and bi-directional evolutionary structural optimization to improve structural efficiency. Starting from circular, rectangular, and square plans with a fully fixed perimeter, two fixed opposite edges, and four fixed vertices, respectively, three 3D voxel structural models are generated by this method, and then final prototypes compatible with 3D printing are obtained after post-processing. Finally, their static performance, seismic performance, and impact resistance are evaluated through elastic-plastic finite element analysis. All proposed prototypes satisfy safety requirements under lunar gravity, and can withstand shallow moonquakes with return periods up to 475 years, and meteoroid impacts with a cumulative flux greater than 2.2 × 10⁻¹⁵ m⁻²・s⁻¹. The circular plan prototype performs best overall, remaining undamaged even under shallow moonquakes with return periods of 2475 years. This study contributes to lunar shelter design by developing an automated and objective computational workflow, improving structural performance and reducing material usage, with strong potential for application in both extraterrestrial and terrestrial architectural contexts involving complex load conditions.