Local-Global Model Reduction for Large-Scale Models Integrating Systems-Theoretical Properties
Abstract
The main objective of this proposal is to develop efficient and accurate reduced-order models comprised of multiscale and multiphysics characteristics amenable for fast simulation of large-scale problems of flow and assessment of uncertainty in highly heterogeneous porous media. This effort will incorporate multiscale methods and system theory (reduced-order modeling) for nonlinear systems for a broad spectrum of applications, ranging from single-phase, to multiphase flow and transport phenomena. In our approach, we develop a framework which balances the error from global reduced-order models and local multiscale approximations. Another unique feature of the proposed work involves the development of extensions to nonlinear uncertain parameter-dependent problems in subsurface flow simulation. The project will attempt to achieve the following results:(1) development of a new local-global multiscale model reduction framework} based system theory and multiscale techniques for processes in highly heterogeneous porous media; (2) development of multiscale methods for complex nonlinear systems of two-phase flows; (3) derivation of error estimators for reduced large-scale discretized models} for characterizing model solution accuracy based on system-theoretical properties; (4) extensions of the proposed techniques to nonlinear and stochastic (parameter-dependent) systems
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 29, 2016
- Accession Number
- AD1058557
Entities
People
- Eduardo Gildin
- Yalchin Efendiev
Organizations
- Texas Engineering Experiment Station