We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CE

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computational Engineering, Finance, and Science

Title: Multielement polynomial chaos Kriging-based metamodelling for Bayesian inference of non-smooth systems

Abstract: This paper presents a surrogate modelling technique based on domain partitioning for Bayesian parameter inference of highly nonlinear engineering models. In order to alleviate the computational burden typically involved in Bayesian inference applications, a multielement Polynomial Chaos Expansion based Kriging metamodel is proposed. The developed surrogate model combines in a piecewise function an array of local Polynomial Chaos based Kriging metamodels constructed on a finite set of non-overlapping subdomains of the stochastic input space. Therewith, the presence of non-smoothness in the response of the forward model (e.g.~ nonlinearities and sparseness) can be reproduced by the proposed metamodel with minimum computational costs owing to its local adaptation capabilities. The model parameter inference is conducted through a Markov chain Monte Carlo approach comprising adaptive exploration and delayed rejection. The efficiency and accuracy of the proposed approach are validated through two case studies, including an analytical benchmark and a numerical case study. The latter relates the partial differential equation governing the hydrogen diffusion phenomenon of metallic materials in Thermal Desorption Spectroscopy tests.
Subjects: Computational Engineering, Finance, and Science (cs.CE); Artificial Intelligence (cs.AI); Numerical Analysis (math.NA)
Cite as: arXiv:2212.02250 [cs.CE]
  (or arXiv:2212.02250v1 [cs.CE] for this version)

Submission history

From: Emilio Martínez-Pañeda [view email]
[v1] Mon, 5 Dec 2022 13:22:39 GMT (5648kb,D)

Link back to: arXiv, form interface, contact.