Multiscale dimension reduction for flow and transport problems in thin domain with reactive boundaries
Introduction
Mathematical models in thin domains occur in many real-world applications, scientific and engineering problems. Fluid flow and transport in thin tube structures are widely used in biological applications, for example, to simulate blood flow in vessels [1], [2], [3]. In engineering problems, flow simulation is used to study fluid flow in complex pipe structures, for example in pipewise industrial installations, wells in oil and gas industry, heat exchangers, etc. In reservoir simulations, thin domains are related to the fractures that usually have complex geometries with very small thickness compared to typical reservoir sizes [4], [5], [6]. Such problems are often considered with complex interaction processes with surrounding media or walls. In many applications, these problems are transformed to reduced (e.g., one) dimensional problems via some type of apriori postulated models. Our goal is to present an alternative approach to analytical model reduction by using multiscale basis functions.
For the applicability of the convenient numerical methods for simulations of such problems, a very fine grid should be constructed to resolve the geometry's complex structure on the grid level. Moreover, a very small domain thickness provides an additional complexity in the grid construction for thin and long domains. For a fast and accurate solution of the presented problem, a homogenization (upscaling) technique or multiscale models are used, which are based on constructing the lower dimensional coarse grid models. The asymptotic behavior of the solutions in thin domains is intensively studied. In [7], [8], the authors propose a hierarchical model reduction approach, in which a coarse model featuring only the dominant direction dynamics is enriched locally by a fine model that accounts for the transverse variables via an appropriate modal expansion. In [9], the authors consider the problem of complete asymptotic expansion for the time-dependent Poiseuille flow in a thin tube. In [10], the method of asymptotic partial decomposition of a domain (MAPDD) was presented to reduce the computational complexity of the numerical solution of such problems. This method combines the three-dimensional description in some neighborhoods of bifurcations and the one-dimensional description out of these small subdomains, and it prescribes some special junction conditions at the interface between these 3D and 1D submodels. Numerical results were presented in [2] for the method of asymptotic partial domain decomposition for thin tube structures with Newtonian and non-Newtonian flows in large systems of vessels. Our goal is to provide an alternative systematic approach to model reduction for thin domain problems that can add complexity via additional multiscale basis functions.
Model reduction techniques usually depend on a coarse grid approximation, which can be obtained by discretizing the problem on a coarse grid and choosing a suitable coarse-grid formulation of the problem. In the literature, several approaches have been developed to obtain the coarse-grid formulation for the problems in heterogeneous domains, including multiscale finite element method [11], [12], [13], mixed multiscale finite element method [14], [15], generalized multiscale finite element method [16], [17], [18], [19], [20], multiscale mortar mixed finite element method [21], multiscale finite volume method [22], [23], variational multiscale methods [24], [25], [26], and heterogeneous multiscale methods [27] etc. The non - conforming multiscale method is considered for the solution of the Stokes flow problem in a heterogeneous domain in [28], [29]. In [30], [31], we presented the Generalized Multiscale Finite Element Method (GMsFEM) for the solution of the flow problems in perforated domains with continuous multiscale basis functions. GMsFEM shows a good accuracy for solving the nonlinear (non - Newtonian) fluid flow problems [32]. In [33], we presented the Discontinuous Galerkin Generalized Multiscale Finite Element Method (DG-GMsFEM) for the solving the two - dimensional problems in perforated domains.
In this paper, we consider flow and transport processes in thin structures with reactive boundaries. The mathematical model is described using Stokes equations and the unsteady convection-diffusion equation with non - homogeneous boundary conditions. Non-homogeneous boundary conditions occur in many applications. For example, the pore-scale modeling and simulation of reactive flow in porous media have many applications in many branches of science such as biology, physics, chemistry, geomechanics, and geology [34], [35], [36]. In order to handle the complex geometry of the walls and non-homogeneous boundary conditions on them, we present additional spectral multiscale basis functions. In this work, we continue developing the multiscale model reduction techniques for problems with multiscale features and the generalization of the techniques for DG-GMsFEM. In our previous work [37], [38], we considered elliptic problems in perforated domains and constructed additional basis functions to capture non-homogeneous boundary conditions on perforations. In [37], we proposed a non-local multi-continua (NLMC) method for Laplace, elasticity, and parabolic equations with non-homogeneous boundary conditions on perforations. In [38], we considered the Continuous Galerkin Generalized Multiscale Finite Element Method (CG-GMsFEM) for problems in perforated domains with non-homogeneous boundary conditions, where we constructed one additional basis function for local domains with perforations. Recently, we extended this technique for solving unsaturated flow problems in heterogeneous domains with rough boundary [39]. In this work, we consider DG-GMsFEM and present additional spectral basis functions for rough non-homogeneous boundary conditions in transport problems. The concept is based on the separation of the snapshots for each feature and shares a lot of similarities with multiscale methods for fractured media presented in our previous works [40], [41]. Our work is also motivated by a recently developed Edge GMsFEM, where multiscale basis functions are constructed for each coarse grid interface [42], [43].
In this work, we use the DG-GMsFEM for constructing multiscale basis functions for problems in thin domains. The goals of this paper are in constructing the general approach for problems in complex thin geometries with an accurate approximation of the velocity space and transport processes. We construct local multiscale basis functions to generate a lower-dimensional model on a coarse grid. In DG-GMsFEM for flow problem, we start with constructing the snapshot space that captures possible flows between coarse cell interfaces. After constructing the snapshot space, we perform a dimension reduction by a solution of the local spectral problems. For the pressure approximation, we use piecewise constant functions on the coarse grid. For the transport problem, we construct multiscale basis functions for each interface between coarse grid cells and present additional basis functions to capture non-homogeneous boundary conditions on walls. The presented snapshot spaces can accurately capture processes on the rough wall boundaries with non-homogeneous boundary conditions on them. We present the results of the numerical simulations for three test geometries for two and three-dimensional problems.
As we mentioned earlier, our goal is to investigate the use of multiscale and generalized multiscale methods for dimension reduction for problems in this domains. Many existing approaches propose analytical or semi-analytical reduced-dimensional problems, where the dimensions are determined apriori. Our idea is to use generalized multiscale method and identify the dimension. The proposed approach combined with a posteriori error estimate can further identify the dimension across thin layers and allow obtaining more accurate solution.
The paper is organized as follows. In Section 2, we describe the problem formulation and the fine-scale approximation. In Section 3, we present the multiscale method for flow and transport processes in thin structures with rough reactive boundaries. In Section 4, we present numerical results. The paper ends with a conclusion.
Section snippets
Problem formulation
Let Ω be a thin domain with multiscale features, where the thickness is small compared to the domain length (See Fig. 1 for an illustration). We will consider the following flow and transport equations in the thin domain Ω, where μ is the fluid viscosity, ρ is the fluid density, D is the diffusion coefficient, c is the concentration, u and p are the velocity and pressure.
The system (1) is equipped with following initial conditions
Multiscale method
Let be a coarse-grid partition of the domain Ω with mesh size H where is the number of coarse grid cells (local domains). We use to denote the set of facets in with . In Fig. 3, we present illustration of the fine and coarse grids. Each coarse cell is the subdomain of the global domain Ω. Moreover, the fine grid is conforming with coarse grid facets. Let be a fine-grid partition of the domain , Ω, and are sets of interior and
Numerical results
In this section, we will present some numerical results. We will use the following three computational domains (Fig. 12) to demonstrate the performance of our method:
Geometry 1 with fine grid that contains 17350 cells. Coarse grid contains 10 local domains.
Geometry 2 with fine grid that contains 18021 cells. Coarse grid contains 20 local domains.
Geometry 3 with fine grid that contains 15094 cells. Coarse grid contains 10 local domains.
Conclusions
We developed a multiscale model order reduction technique for the solution of the flow and transport problem in thin domains. Our motivation stems from reducing the problem dimension in thin layer applications. For the fine grid approximation, we apply the discontinuous Galerkin method and use the solution as a reference solution. Our multiscale approach for solving problems in complex thin geometries gives an accurate approximation of the velocity space and transport processes. We presented
CRediT authorship contribution statement
Maria Vasilyeva: Conceptualization, Methodology, Supervision, Validation, Writing – original draft. Valentin Alekseev: Investigation, Methodology, Software, Validation. Eric T. Chung: Conceptualization, Supervision, Writing – review & editing. Yalchin Efendiev: Conceptualization, Supervision, Writing – review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgement
V.A.'s work is supported by the RFBR N19-31-90076 and the mega-grant of the Russian Federation Government (N 14.Y26.31.0013). MV's and YE's works are supported by the mega-grant of the Russian Federation Government (N 14.Y26.31.0013). The research of Eric Chung is partially supported by the Hong Kong RGC General Research Fund (Project numbers 14304719 and 14302018) and CUHK Faculty of Science Direct Grant 2019-20.
References (51)
- et al.
Mathematical modelling and numerical simulation of the cardiovascular system
Handb. Numer. Anal.
(2004) - et al.
Finite element simulation of blood flow in the cerebral artery
Comput. Methods Appl. Mech. Eng.
(2001) - et al.
A multiscale finite element method for elliptic problems in composite materials and porous media
J. Comput. Phys.
(1997) - et al.
Generalized multiscale finite element methods (GMSFEM)
J. Comput. Phys.
(2013) - et al.
Adaptive multiscale model reduction with generalized multiscale finite element methods
J. Comput. Phys.
(2016) - et al.
Iterative multiscale finite-volume method
J. Comput. Phys.
(2008) - et al.
The variational multiscale method?a paradigm for computational mechanics
Comput. Methods Appl. Mech. Eng.
(1998) - et al.
Multiscale model reduction for transport and flow problems in perforated domains
J. Comput. Appl. Math.
(2018) - et al.
A conservative local multiscale model reduction technique for Stokes flows in heterogeneous perforated domains
J. Comput. Appl. Math.
(2017) - et al.
Diffusion, convection, adsorption, and reaction of chemicals in porous media
J. Differ. Equ.
(1991)
Upscaling method for problems in perforated domains with non-homogeneous boundary conditions on perforations using non-local multi-continuum method (nlmc)
J. Comput. Appl. Math.
A generalized multiscale finite element method (gmsfem) for perforated domain flows with Robin boundary conditions
J. Comput. Appl. Math.
Non-local multi-continua upscaling for flows in heterogeneous fractured media
J. Comput. Phys.
Edge multiscale methods for elliptic problems with heterogeneous coefficients
J. Comput. Phys.
Multiscale formulation of two-phase flow at the pore scale
J. Comput. Phys.
Multiscale models of blood flow in the compliant aortic bifurcation
Appl. Math. Lett.
Asymptotic partial domain decomposition in thin tube structures: numerical experiments
Int. J. Multiscale Comput. Eng.
Modeling fractures and barriers as interfaces for flow in porous media
SIAM J. Sci. Comput.
A reduced model for Darcy's problem in networks of fractures
ESAIM: Math. Model. Anal.
On the coupling of 1d and 3d diffusion-reaction equations: application to tissue perfusion problems
Math. Models Methods Appl. Sci.
Hierarchical local model reduction for elliptic problems: a domain decomposition approach
Multiscale Model. Simul.
Himod reduction of advection–diffusion–reaction problems with general boundary conditions
J. Sci. Comput.
Asymptotic analysis of the nonsteady viscous flow with a given flow rate in a thin pipe
Appl. Anal.
Method of asymptotic partial decomposition of domain for multistructures
Appl. Anal.
Multiscale Finite Element Methods: Theory and Applications, vol. 4
Cited by (8)
Multiscale solver for multi-component reaction–diffusion systems in heterogeneous media
2023, Journal of Computational and Applied MathematicsMultiscale model reduction technique for fluid flows with heterogeneous porous inclusions
2023, Journal of Computational and Applied MathematicsCitation Excerpt :The macroscopic equations are formulated on a coarse grid with mesh size independent of the size of perforations. We extend multiscale approach presented in our previous papers [32–34] to solve the nonlinear convective Darcy–Brinkman–Forchheimer equations in heterogeneous domains. The multiscale solver for the coarse grid approximation is constructed using Generalized Finite Element Method (GMsFEM) [20].
Meshfree Generalized Multiscale Finite Element Method
2023, Journal of Computational PhysicsCitation Excerpt :An unstructured GMsFEM for a filtration problem in a fractured media with arbitrary polygonal coarse-scale elements was developed in [19]. Multiscale mathematical modeling for flow and transport problems in a thin domain is considered in [20]. In multiscale modeling, it is necessary to build a particular computational grid.
GENERALIZED MULTISCALE FINITE ELEMENT TREATMENT OF A HETEROGENEOUS NONLINEAR STRAIN-LIMITING ELASTIC MODEL<sup>*</sup>
2024, Multiscale Modeling and SimulationMultiscale Model Reduction Multiscale Finite Element Methods and Their Generalizations
2023, Applied Mathematical Sciences (Switzerland)