Uncertainty quantification in structural dynamic analysis using two-level Gaussian processes and Bayesian inference

• Published in: Journal of sound and vibration 2018-01, Vol.412, p.95-115
• Main Authors: Zhou, K., Tang, J.
• Format: Article
• Language: English
• Subjects: Bayesian analysis; Bayesian inference; Component mode synthesis; Component mode synthesis (CMS); Components; Finite element method; Fluid dynamics; Gaussian distribution; Gaussian process; Mathematical models; Model updating; Normal distribution; Order-reduced modeling; Probabilistic analysis; Probabilistic inference; Statistical inference; Structural engineering; Two-level Gaussian processes emulator; Uncertainty; Uncertainty quantification
• ISSN: 0022-460X; 1095-8568
• DOI: 10.1016/j.jsv.2017.09.034

A probabilistic framework for efficient uncertainty quantification in structural dynamic analysis is presented. This framework is built upon the combination of two-level Gaussian processes emulator and Bayesian inference technique. The underlying idea is to employ the two-level Gaussian processes emulator to integrate together small amount of high-fidelity data from full-scale finite element analysis and large amount of low-fidelity data from order-reduced analysis to improve the response variation prediction. As component mode synthesis (CMS) is adopted in order-reduced modeling, we then utilize the improved response variation prediction on modal characteristics to update the CMS model to facilitate the efficient probabilistic analysis of any responses of concern. The effectiveness of this framework is demonstrated through systematic case studies.