CUED Publications database

Discovering a one-dimensional active subspace to quantify multidisciplinary uncertainty in satellite system design

Hu, X and Parks, GT and Chen, X and Seshadri, P (2016) Discovering a one-dimensional active subspace to quantify multidisciplinary uncertainty in satellite system design. Advances in Space Research, 57. pp. 1268-1279. ISSN 0273-1177

Full text not available from this repository.

Abstract

© 2016 Published by Elsevier Ltd. Uncertainty quantification has recently been receiving much attention from aerospace engineering community. With ever-increasing requirements for robustness and reliability, it is crucial to quantify multidisciplinary uncertainty in satellite system design which dominates overall design direction and cost. However, coupled multi-disciplines and cross propagation hamper the efficiency and accuracy of high-dimensional uncertainty analysis. In this study, an uncertainty quantification methodology based on active subspaces is established for satellite conceptual design. The active subspace effectively reduces the dimension and measures the contributions of input uncertainties. A comprehensive characterization of associated uncertain factors is made and all subsystem models are built for uncertainty propagation. By integrating a system decoupling strategy, the multidisciplinary uncertainty effect is efficiently represented by a one-dimensional active subspace for each design. The identified active subspace is checked by bootstrap resampling for confidence intervals and verified by Monte Carlo propagation for the accuracy. To show the performance of active subspaces, 18 uncertainty parameters of an Earth observation small satellite are exemplified and then another 5 design uncertainties are incorporated. The uncertainties that contribute the most to satellite mass and total cost are ranked, and the quantification of high-dimensional uncertainty is achieved by a relatively small number of support samples. The methodology with considerably less cost exhibits high accuracy and strong adaptability, which provides a potential template to tackle multidisciplinary uncertainty in practical satellite systems.

Item Type: Article
Subjects: UNSPECIFIED
Divisions: Div A > Energy
Div A > Turbomachinery
Depositing User: Cron Job
Date Deposited: 17 Jul 2017 19:45
Last Modified: 18 Nov 2017 21:54
DOI: