Efficient Global Sensitivity Analysis for Time-Dependent, Multidisciplinary Models

DeCarlo, Erin C.; Mahadevan, Sankaran; Smarslok, Benjamin P.; Sparkman, Daniel

doi:10.2514/6.2017-0594

Cited by 1 publication

(1 citation statement)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…142 Alternatively, Sobol' indices 143 are anoter variance-based GSA technique. GSA based on Sobol' indices is used by DeCarlo et al 144 on an aerothermal problem, where an importance sampling-based kernel method is developed to estimate the indices. It allows for time-dependent multidisciplinary analysis of the sensitivity.…”

Section: B Global Sensitivity Analysismentioning

confidence: 99%

Uncertainty-Based Design Optimization and Technology Evaluation: A Review

Roelofs

Vos

2018

2018 AIAA Aerospace Sciences Meeting

View full text Add to dashboard Cite

Evaluation and assessment of novel technologies for aerospace applications is essential for business strategy and decision making regarding development efforts. Since technology is evaluated in the conceptual design phase and little is known about the technology, large uncertainty is present. This uncertainty needs to be accurately assessed and managed. To investigate the research efforts that have been performed to perform technology evaluation under uncertainty, a literature review was conducted, focusing on methods and modeling approaches to assign and quantify these uncertainties. It is found that probability theory is still the most popular theory for representing uncertainty. Polynomial Chaos Expansions and Stochastic Collocation methods are gaining popularity for propagating uncertainty through a modeling environment, but Monte Carlo Simulations are still widely used. Commonly, surrogate models are used to reduce computational effort. Other efforts focus on the use of multifidelity approaches to reduce computational effort when high-fidelity methods are required. Four issues that may need to be addressed in future research were identified.

show abstract

Section: B Global Sensitivity Analysismentioning

confidence: 99%