arXiv subjects

Konstantin Brenner

Publications and source records attributed to Konstantin Brenner.

9 recordsLinked to original sources

arXiv2024

Nonlinear Multilevel Solution Strategies for Diffusive Wave Flood Models in Perforated Domains

This article investigates the numerical solution of the Diffusive Wave equation posed on domains containing a large number of polygonal perforations, motivated by urban flood modeling. Such geometries induce strong multiscale effects driven by geometric complexity, which significantly challenge the robustness of standard nonlinear and linear solvers. The work builds on a multiscale coarse space previously introduced by the authors for linear Poisson problems on perforated domains. This low-dimensional space, constructed on a coarse polygonal partition and spanned by locally discrete harmonic (Trefftz-type) basis functions, is shown to remain effective for the linearized Diffusive Wave problems arising within Newton iterations. This enables the construction of robust two-level preconditioners for the resulting sequence of linear systems. Beyond linearization, the main focus of this work is on the effective solution of the fully nonlinear problem. We assess and combine several Schwarz-based nonlinear preconditioning strategies, including a two-level RASPEN method and a two-step nonlinear method, using the same multiscale coarse space to ensure scalability. While the individual components are drawn from the existing literature, their combination provides a robust and practical solution strategy for a challenging nonlinear problem posed on highly perforated domains. A systematic comparison of the methods and a discussion of algorithmic complexity are presented. The proposed approaches are validated through numerical experiments, including a realistic test case based on topographical data from the city of Nice.

math.NA↗

arXiv2024

Learning local Dirichlet-to-Neumann maps of nonlinear elliptic PDEs with rough coefficients

Partial differential equations (PDEs) involving high contrast and oscillating coefficients are common in scientific and industrial applications. Numerical approximation of these PDEs is a challenging task that can be addressed, for example, by multi-scale finite element analysis. For linear problems, multi-scale finite element method (MsFEM) is well established and some viable extensions to non-linear PDEs are known. However, some features of the method seem to be intrinsically based on linearity-based. In particular, traditional MsFEM rely on the reuse of computations. For example, the stiffness matrix can be calculated just once, while being used for several right-hand sides, or as part of a multi-level iterative algorithm. Roughly speaking, the offline phase of the method amounts to pre-assembling the local linear Dirichlet-to-Neumann (DtN) operators. We present some preliminary results concerning the combination of MsFEM with machine learning tools. The extension of MsFEM to nonlinear problems is achieved by means of learning local nonlinear DtN maps. The resulting learning-based multi-scale method is tested on a set of model nonlinear PDEs involving the $p-$Laplacian and degenerate nonlinear diffusion.

math.NA↗

arXiv2024

Robust Methods for Multiscale Coarse Approximations of Diffusion Models in Perforated Domains

For the Poisson equation posed in a domain containing a large number of polygonal perforations, we propose a low-dimensional coarse approximation space based on a coarse polygonal partitioning of the domain. Similarly to other multiscale numerical methods, this coarse space is spanned by locally discrete harmonic basis functions. Along the subdomain boundaries, the basis functions are piecewise polynomial. The main contribution of this article is an error estimate regarding the H1-projection over the coarse space which depends only on the regularity of the solution over the edges of the coarse partitioning. For a specific edge refinement procedure, the error analysis establishes superconvergence of the method even if the true solution has a low general regularity. Combined with domain decomposition (DD) methods, the coarse space leads to an efficient two-level iterative linear solver which reaches the fine-scale finite element error in few iterations. It also bodes well as a preconditioner for Krylov methods and provides scalability with respect to the number of subdomains. Numerical experiments showcase the increased precision of the coarse approximation as well as the efficiency and scalability of the coarse space as a component of a DD algorithm.

math.NA↗

arXiv2023

A Trefftz-like coarse space for the two-level Schwarz method on perforated domains

We consider a new coarse space for the ASM and RAS preconditioners to solve elliptic partial differential equations on perforated domains, where the numerous polygonal perforations represent structures such as walls and buildings in urban data. With the eventual goal of modelling urban floods by means of the nonlinear Diffusive Wave equation, this contribution focuses on the solution of linear problems on perforated domains. Our coarse space uses a polygonal subdomain partitioning and is spanned by Trefftz-like basis functions that are piecewise linear on the boundary of a subdomain and harmonic inside it. It is based on nodal degrees of freedom that account for the intersection between the perforations and the subdomain boundaries. As a reference, we compare this coarse space to the well-studied Nicolaides coarse space with the same subdomain partitioning. It is known that the Nicolaides space is unable to prevent stagnation in convergence when the subdomains are not connected; we work around this issue by separating each subdomain by disconnected component. Scalability and robustness are tested for multiple data sets based on realistic urban topography. Numerical results show that the new coarse space is very robust and accelerates the number of Krylov iterations when compared to Nicolaides, independent of the complexity of the data.

math.NA↗

arXiv2021

Gradient discretization of two-phase poro-mechanical models with discontinuous pressures at matrix fracture interfaces

We consider a two-phase Darcy flow in a fractured and deformable porous medium for which the fractures are described as a network of planar surfaces leading to so-called hybrid-dimensional models. The fractures are assumed open and filled by the fluids and small deformations with a linear elastic constitutive law are considered in the matrix. As opposed to [10], the phase pressures are not assumed continuous at matrix fracture interfaces, which raises new challenges in the convergence analysis related to the additional interfacial equations and unknowns for the flow. As shown in [16, 2], unlike single phase flow, discontinuous pressure models for two-phase flows provide a better accuracy than continuous pressure models even for highly permeable fractures. This is due to the fact that fractures fully filled by one phase can act as barriers for the other phase, resulting in a pressure discontinuity at the matrix fracture interface. The model is discretized using the gradient discretization method [22], which covers a large class of conforming and non conforming schemes. This framework allows for a generic convergence analysis of the coupled model using a combination of discrete functional tools. In this work, the gradient discretization of [10] is extended to the discontinuous pressure model and the convergence to a weak solution is proved. Numerical solutions provided by the continuous and discontinuous pressure models are compared on gas injection and suction test cases using a Two-Point Flux Approximation (TPFA) finite volume scheme for the flows and $P_2$ finite elements for the mechanics.

math.NA↗

arXiv2021

Gradient discretization of two-phase flows coupled with mechanical deformation in fractured porous media

We consider a two-phase Darcy flow in a fractured porous medium consisting in a matrix flow coupled with a tangential flow in the fractures, described as a network of planar surfaces. This flow model is also coupled with the mechanical deformation of the matrix assuming that the fractures are open and filled by the fluids, as well as small deformations and a linear elastic constitutive law. The model is discretized using the gradient discretization method [26], which covers a large class of conforming and non conforming schemes. This framework allows for a generic convergence analysis of the coupled model using a combination of discrete functional tools. Here, we describe the model together with its numerical discretization, and we prove a convergence result assuming the non-degeneracy of the phase mobilities and that the discrete solutions remain physical in the sense that, roughly speaking, the porosity does not vanish and the fractures remain open. This is, to our knowledge, the first convergence result for this type of models taking into account two-phase flows in fractured porous media and the non-linear poromechanical coupling. Previous related works consider a linear approximation obtained for a single phase flow by freezing the fracture conductivity [36, 37]. Numerical tests employing the Two-Point Flux Approximation (TPFA) finite volume scheme for the flows and $\mathbb P_2$ finite elements for the mechanical deformation are also provided to illustrate the behavior of the solution to the model.

math.NA↗

arXiv2020

Verification benchmarks for single-phase flow in three-dimensional fractured porous media

Flow in fractured porous media occurs in the earth's subsurface, in biological tissues, and in man-made materials. Fractures have a dominating influence on flow processes, and the last decade has seen an extensive development of models and numerical methods that explicitly account for their presence. To support these developments, we present a portfolio of four benchmark cases for single-phase flow in three-dimensional fractured porous media. The cases are specifically designed to test the methods' capabilities in handling various complexities common to the geometrical structures of fracture networks. Based on an open call for participation, results obtained with 17 numerical methods were collected. This paper presents the underlying mathematical model, an overview of the features of the participating numerical methods, and their performance in solving the benchmark cases.

math.NA↗

arXiv2016

Improving Newton's method performance by parametrization: the case of Richards equation

The nonlinear systems obtained by discretizing degenerate parabolic equations may be hard to solve, especially with Newton's method. In this paper, we apply to Richards equation a strategy that consists in defining a new primary unknown for the continuous equation in order to stabilize Newton's method by parametrizing the graph linking the pressure and the saturation. The resulting form of Richards equation is then discretized thanks to a monotone Finite Volume scheme. We prove the well-posedness of the numerical scheme. Then we show under appropriate non-degeneracy conditions on the parametrization that Newton\^as method converges locally and quadratically. Finally, we provide numerical evidences of the efficiency of our approach.

math.AP↗

arXiv2010

A finite volume method on general meshes for a degenerate parabolic convection-reaction-diffusion equation

We propose a finite volume method on general meshes for the discretization of a degenerate parabolic convection-reaction-diffusion equation. Equations of this type arise in many contexts, such as the modeling of contaminant transport in porous media. We discretize the diffusion term, which can be anisotropic and heterogeneous, via a hybrid finite volume scheme. We construct a partially upwind scheme for the convection term. We consider a wide range of unstructured possibly non-matching polygonal meshes in arbitrary space dimension. The only assumption on the mesh is that the volume elements must be star-shaped. The scheme is fully implicit in time, it is locally conservative and robust with respect to the P\'eclet number. We obtain a convergence result based upon a priori estimates and the Fr\'echet--Kolmogorov compactness theorem.

math.NA↗