Publication detail

ON PHYSICALLY CONSTRAINED NON-INTRUSIVE POLYNOMIAL CHAOS EXPANSION

NOVÁK, L. SHARMA, H. SHIELDS, M.

Original Title

ON PHYSICALLY CONSTRAINED NON-INTRUSIVE POLYNOMIAL CHAOS EXPANSION

Type

conference paper

Language

English

Original Abstract

Surrogate modeling of costly mathematical models representing physical systems represents a challenging task, since it is typically not possible to create a large experimental design and thus it is beneficial to restrict the approximation by additional information reflecting the physics of the original mathematical model. This paper presents a novel methodology for the construction of polynomial chaos expansion combining experimental design and additional information arising from the physics of an original model – physical constraints. Physical constraints investigated in this paper are represented by boundary conditions and given partial differential equations. A computationally efficient numerical algorithm based on least angle regression and Lagrange multipliers is proposed and compared to standard least angle regression. It is shown that the proposed algorithm leads to superior accuracy of the approximation and does not bring any significant computational burden.

Keywords

Adaptive Sampling; Polynomial Chaos Expansion; Sequential Sampling

Authors

NOVÁK, L.; SHARMA, H.; SHIELDS, M.

Released

12. 6. 2023

Publisher

National Technical University of Athens

Location

Athens, Greece

ISBN

9786185827021

Book

UNCECOMP 2023 Proceedings

Pages from

151

Pages to

157

Pages count

7

BibTex

@inproceedings{BUT187084,
  author="NOVÁK, L. and SHARMA, H. and SHIELDS, M.",
  title="ON PHYSICALLY CONSTRAINED NON-INTRUSIVE POLYNOMIAL CHAOS EXPANSION",
  booktitle="UNCECOMP 2023 Proceedings",
  year="2023",
  pages="151--157",
  publisher="National Technical University of Athens",
  address="Athens, Greece",
  isbn="9786185827021"
}