Conceptual Moon imaging micro/nano-satellite design optimization under uncertainty

Low-cost Moon imaging micro/nano-satellites have been receiving much attention from the engineering community. The reliability-based robust design optimization (RBRDO) of these weight/size/power constrained satellites is crucial. In this study, the configuration of this conceptual design problem considering uncertainty is analyzed and the interdisciplinary relations are described. Discipline models including orbit, payload, propulsion, and communication are mainly discussed. A new RBRDO approach to lunar micro/nano-satellite system design is presented, involving dimension reduction, dynamic response surface, system decoupling, and multiobjective alliance search. The optimization goal includes payload cost-effective ratio and satellite mass, and five reliability constraints are considered including satellite installation volume and battery cycles. Through identifying a dominant active subspace for high-dimensional inputs, the in-loop uncertainty quantification is greatly accelerated by about two orders of magnitude. A smooth and uniformly distributed Pareto front is therefore obtained to provide beneficial Pareto-optimal solutions, which indicates that the orbit altitude gathers around 200 km and the mission cycle is nearly 2.4 years. An optimal trade-off design has the minimum satellite mass of 8.95 kg and smaller cost-effective ratio. Compared with the deterministic optimization, the RBRDO approach is preferable for designing Moon imaging micro/nano-satellites of high reliability and robustness.