This project develops a Bayesian physics-informed polynomial chaos expansion (B-PC^2) framework for uncertainty quantification and reliability analysis in high-dimensional, computationally expensive simulation models. Present surrogate approaches either incorporate physics constraints or provide a Bayesian uncertainty treatment, but rarely both. The proposed B-PC^2 approach unifies these aspects to achieve effective uncertainty propagation in realistic engineering applications. The project will provide an open, theoretically consistent foundation for Bayesian, physics- informed polynomial chaos expansions. By combining Bayesian estimation, physics constraints, and active learning, it will advance the state of the art in reliability-oriented uncertainty quantification and provide validated methods applicable to complex, high-dimensional engineering models

Klíčová slova
uncertainty quantification, surrogate modeling, structural reliability

26-24969L

angličtina

Novák Lukáš, doc. Ing., Ph.D. - hlavní řešitel

Ústav stavební mechaniky
- odpovědné pracoviště (21.9.2025 - nezadáno)
Ústav stavební mechaniky
- příjemce (21.9.2025 - nezadáno)

Odpovědnost: Novák Lukáš, doc. Ing., Ph.D.