Reliability-based robust design optimization of modern small satellites has recently been receiving much attention. The role of multidisciplinary design optimization considering system uncertainty is increasingly being recognized in improving satellite performance, safety, and reliability. However, high-dimensional uncertainty hampers the efficiency and accuracy of reliability-based robust design optimization. In this study, a novel multidisciplinary optimization methodology suitable for high-dimensional uncertainty is established and verified. The in-loop uncertainty quantification employing active subspace identification makes the uncertainty propagation and management much easier. Based on the discovered active subspace and a system decoupling strategy, a multiobjective alliance algorithm is presented for the complex constrained design optimization. Typical uncertain factors in small satellites are characterized and interdisciplinary relations are analyzed. A standard test example of speed reducer design first proves the methodology, and then an Earth observation satellite fully illustrates the efficacy. Through 23-dimensional uncertainty represented by one dimension, the multiobjective alliance algorithm gives well-distributed Pareto-optimal (PO) solutions, providing a fruitful reference for decision makers compromising different design targets, like satellite mass, total cost, and ground resolution. The proposed approach is proved to be fairly accurate and adaptable, which can be widely applied to uncertainty-based multidisciplinary design optimization (UMDO) of satellite systems.
Multidisciplinary Optimization Under High-Dimensional Uncertainty for Small Satellite System Design
AIAA Journal ; 54 , 5 ; 1732-1741
2016-02-22
10 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Multidisciplinary Optimization Under High-Dimensional Uncertainty for Small Satellite System Design
Online Contents | 2016
|Uncertainty-Based Multidisciplinary Design Optimization
British Library Conference Proceedings | 1998
|Multidisciplinary Optimization under Uncertainty Using Bayesian Network
SAE Technical Papers | 2016
|