Shakedown-driven Design of Reusable Spacecraft Structures via Topology Optimization and Physics-Informed Machine Learning
Please login to view abstract download link
To meet the stringent requirements of safety, lightweight, and reusability for future crewed spacecraft, conventional elastic limit-based design approaches are becoming obsolete. To this end, this study proposes employing elastic shakedown as a more advanced design criterion and develops a comprehensive framework for designing complex aerospace components. Specifically, we focus on two representative structures: the integrally stiffened cabin body and lightweight brackets. To evaluate these structures, a shakedown analysis algorithm for shell elements with drilling degree of freedom is developed, providing the basis for a subsequent shakedown-based optimization of stiffened panel. Furthermore, a shakedown-oriented topology optimization algorithm is established to allow for the synergistic design of stiffener layouts and material distribution, ensuring structural integrity under cyclic loading while minimizing mass. The proposed methodology is validated through both numerical simulations and shakedown experiments, and in addition to the conventional numerical optimization techniques for solving the shakedown problem, we also discuss how to combine Graph Neural Networks (GNN) with Physics-Informed Neural Networks (PINN) to achieve a more efficient shakedown analysis. This research provides a robust and efficient path for the lightweight design of next-generation reusable space structures, shifting the design paradigm from conservative elastic limits to more realistic cyclic plasticity boundaries.
