Bayesian inversion for anisotropic hydraulic phase-field fracture

verfasst von
Nima Noii, Amirreza Khodadadian, Thomas Wick
Abstract

In this work, we employ a Bayesian inversion framework to fluid-filled phase-field fracture. We develop a robust and efficient numerical algorithm for hydraulic phase-field fracture toward transversely isotropic and orthotropy anisotropic fracture. In the fluid-driven coupled problem, three primary fields for pressure, displacements, and the crack phase-field are solved while direction-dependent responses (due to the preferred fiber orientation) are enforced via penalty-like parameters. A new crack driving state function is introduced by avoiding the compressible part of anisotropic energy to be degraded. Next, we use a successful extension of the anisotropic hydraulic phase-field fracture as a departure point for Bayesian inversion to estimate material parameters. To this end, we employ the delayed rejection adaptive Metropolis–Hastings (DRAM) algorithm to identify the parameters. The focus is on uncertainties arising from different variables, including elasticity modulus, Biot's coefficient, Biot's modulus, dynamic fluid viscosity and Griffith's energy release rate in the case of the isotopic hydraulic fracture while in the anisotropic setting, we will have additional penalty-like parameters to be identified. Several numerical examples substantiate our algorithmic developments.

Organisationseinheit(en)
Institut für Angewandte Mathematik
Institut für Kontinuumsmechanik
PhoenixD: Simulation, Fabrikation und Anwendung optischer Systeme
Typ
Artikel
Journal
Computer Methods in Applied Mechanics and Engineering
Band
386
ISSN
0045-7825
Publikationsdatum
01.12.2021
Publikationsstatus
Veröffentlicht
Peer-reviewed
Ja
ASJC Scopus Sachgebiete
Numerische Mechanik, Werkstoffmechanik, Maschinenbau, Physik und Astronomie (insg.), Angewandte Informatik
Elektronische Version(en)
https://arxiv.org/abs/2007.16038 (Zugang: Offen)
https://doi.org/10.1016/j.cma.2021.114118 (Zugang: Geschlossen)
 

Details im Forschungsportal „Research@Leibniz University“