Discontinuous finite volume element method for Darcy flows in fractured porous media

https://doi.org/10.1016/j.cam.2020.113025Get rights and content

Abstract

This paper presents a numerical simulation of the single phase Darcy flow model in two-dimensional fractured porous media. Under some physically consistent coupling conditions, the model can be described as a reduced problem by coupling the bulk problem in porous matrix and the fracture problem in fractures. Flows are governed by the primal form of the Darcy’s equations for both the bulk and fractures. The coupled discontinuous finite volume element methods and conforming finite element method are adopted to solve the bulk problem and fracture problem, respectively. We theoretically analyze the well-posedness of the discrete problem, and derive optimal error estimates in standard L2 error and broken H1 error. Numerical experiments include not only the fractures with high permeability as the prior flow conduit, but also the fractures with low permeability as the flow barrier, which demonstrate the accuracy, flexibility and robustness of our discrete formulation for complicated networks of fractures in porous media domain.

Introduction

The modeling of fluid flow and transport in fractured porous media has become a significant problem due to its application in many environmental and energy problems, such as ground water contamination, oil migration, nuclear waste and carbon dioxide storage. Since the fracture is supposed to have a small thickness with respect to the width of the porous medium, the dimension of the fracture is reduced to be d1 in a ddimensional domain. In practical application, the fracture permeability may vary by several orders of magnitude, thus the flow in fractured porous medium is strongly influenced by the presence of the fracture. In general, there are two kinds of fractures. On one hand, the fractures with higher permeability than the surrounding domains act as fast pathways, the fluid tends to flow into fractures and then along them, so the velocity is not supposed to be continuous across fractures. And on the other hand that those with lower permeability act as geological barriers, in this case, the fluid has a tendency to avoid fractures, so the pressure is not identical on both sides of fractures. For both types of fractures, the reduced coupled bulk fracture models have been presented for single phase Darcy flow developed in [1], where an averaging process is applied, and the fracture is treated as an interface that crosses the bulk region. The fracture is additionally assumed to be filled of debris, so that the flow therein can still be modeled by the Darcy law. To close the problem, interface conditions are enforced that to relate the average and jump of the bulk pressure to the normal flux of the fracture pressure.

Various numerical methods have been proposed to solve this reduced coupled bulk fracture model, such as methods to keep robust on polygonal meshes, including mimetic finite difference method [2], [3], hybrid high-order method [4], weak Galerkin method [5], discontinuous Galerkin method [6], mixed finite element method [7], [8], [9], [10], finite volume method[11], [12], [13], [14], [15], two-grid algorithm [16] and methods where the fractures are allowed to arbitrarily cut the bulk grid, including eXtended finite element method [17], [18], [19], [20]. We also assign the following literatures in which the problem of discrete fracture networks have been considered by various novel approaches, for example, virtual element method [21], a PDE-constrained optimization method [22], [23], and so on.

Our focus is here on the discontinuous finite volume element (DFVE) method which is originally introduced in [24] to the second-order elliptic problem, and later applied to solve the Stokes problem [25]. The DFVE method uses discontinuous piecewise polynomial for the trial functions as in discontinuous Galerkin method, and uses piecewise constant for test functions as in finite volume method, so it has the flexibility, high order of accuracy of discontinuous Galerkin method, and the simplicity, conservative property of finite volume element method. The fascinating nature of DFVE method reflects on the smaller conservation control volume which is less than half the size of control volume applied in the existed finite volume method. The localizability of the discontinuous element and its dual partition in DFVE method provides another advantage for parallel computing. There are many works on the discontinuous finite volume element method for both the second order elliptic problem [26], [27], [28], [29] and the Stokes equation [30], [31], [32]. DFVE method has been used to solve elliptic problem with adaptive technique and other applications [33], [34]. Recently, there are a few papers applying DFVE method to nonlinear problem [35], [36] and also for the coupled system [37], [38], [39], [40], [41].

The DFVE method proposed here hinges on a primal formulation in the bulk coupled with a primal formulation inside the fracture. To keep the exposition as simple as possible while retaining all the key difficulties, we focus on the two-dimensional case. More specifically, considering the discontinuity of solutions at the bulk fracture interface, it is quite natural to choose DFVE method to solve the flow problem in fractured porous media. However, this is not the only motivation to use DFVE method. Indeed, the coupling conditions at the interfaces between bulk and fracture flow be formulated using jump and average operators, which can be easily embedded in the variational formulation, so that DFVE method turns out to be a powerful tool for efficiently handling the coupling of bulk fracture model.

In this paper we propose a discretization which combines a DFVE approximation for the problem in the bulk with a conforming finite element approximation in the fracture. Compared with these existing methods, DFVM has the property of local mass conservation at a so-called diamond control volume level. And because of using low-order polynomial as the trial function, DFVM is easy to code, but also, a high order of accuracy can be achieved. Moreover, we obtained the optimal order error estimates corresponding variables in L2-norm and broken H1-norm. Some experiments are presented to illustrate the theory.

The paper is organized as follows. In Section 2, we recall and state the reduced model. In Section 3, the discrete formulation coupled the discontinuous finite volume element method for the bulk problem with the conforming finite element method for the fracture problem is introduced. Section 4 is used to derive some convergence analysis. Section 5 presents numerical experiments on some examples with exact analytical solutions and a series of benchmark problems. Finally we draw a conclusion in Section 6.

Section snippets

Model statement

We consider a bounded, open, convex domain ΩR2, composed of a fractured porous medium, and let ΩD, ΩN denote the Dirichlet boundary and Neumann boundary, respectively, as depicted in Fig. 1. We denote a network formed by the union of NΓ fractures γk, for k=1,2,,NΓ. Each γk is a d1 dimensional C manifold and we have Γ=k=1NΓγk.

In this paper, we suppose that the fractures can intersect only at their endpoints, i.e., γk¯γj¯=γkγj=ikj.

Then we can subdivide the boundary of fractures in

Discontinuous finite volume element method

We consider a family of meshes Rh made of triangular elements in two dimension which are aligned with the fracture Γ, hK is the diameter of element K. Every triangle KRh is divided into three subtriangles T by connecting the barycenter of the triangle K to its corner nodes, then we define the dual partition Th of the primal partition Rh to be the union of the triangles T as shown in Fig. 2. We define the mesh parameter h=maxKRhhK. Let he denote the length of edge e in element K, Eo be the set

Error estimates

In this section, we use the equivalence relationship between the DFVE method and IPDG approximations for the coupled problem (2.1)–(2.12). This relationship plays a key role in the error estimation. We treat the DFVE method as a perturbation of the interior penalty method, as the error estimates of the interior penalty method are developed in [6], we can easily develop the error estimates for the DFVE method by means of the estimation of the difference between the solution of the DFVE method

Numerical experiments

In this section, we present some numerical examples to test the problem (2.1)–(2.12) in two-dimensional. All numerical results demonstrate that the proposed scheme (3.5) is robust, correct and easy to implement. The first three numerical examples are provided to show the convergence and accuracy. Moreover, in the last four examples we investigate the robustness of our method by contrasts between the permeability in the bulk and fracture. In all numerical examples, we take penalty parameters α=

Conclusion

In this paper, we develop a discrete scheme for the reduced model which describes the flow in the fractured porous media. This scheme couples DFVE methods for the bulk flow with a conforming finite element approximation for the fracture flow. The optimal convergence rate is obtained under the suitable norms. Numerical results in last section are not only consistent with our theory, but also demonstrate that our scheme is accurate and robust for fractures with high or low permeability. In future

References (51)

  • FumagalliA. et al.

    Conforming, non-conforming and nonmatching discretization couplings in discrete fracture network simulations

    J. Comput. Phys.

    (2019)
  • MartinV. et al.

    Modeling fractures and barriers as interfaces for flow in porous media

    SIAM J. Sci. Comput.

    (2005)
  • AntoniettiP.F. et al.

    Mimetic finite difference approximation of flows in fractured porous media

    M2AN Math. Model. Numer. Anal.

    (2016)
  • FormaggiaL. et al.

    Analysis of a mimetic finite difference approximation of flows in fractured porous media

    M2AN Math. Model. Numer. Anal.

    (2018)
  • ChaveF. et al.

    A hybrid high-order method for Darcy flows in fractured porous media

    SIAM J. Sci. Comput.

    (2018)
  • WangG. et al.

    Weak Galerkin finite element methods for the simulation of single-phase flow in fractured porous media

    J. Sci. Comput.

    (2018)
  • AntoniettiP.F. et al.

    Discontinuous Galerkin approximation of flows in fractured porous media on polytopic grids

    SIAM J. Sci. Comput.

    (2019)
  • FrihN. et al.

    Modeling fractures as interfaces with nonmatching grids

    Comput. Geosci.

    (2012)
  • ChenH.X. et al.

    Adaptive mixed finite element methods for Darcy flow in fractured porous media

    Water Resour. Res.

    (2016)
  • ChenH.X. et al.

    A residual-based a posteriori error estimator for single-phase darcy flow in fractured porous media

    Numer. Math.

    (2017)
  • AngotP. et al.

    Asymptotic and numerical modeling of flows in fractured porous media

    M2AN Math. Model. Numer. Anal.

    (2009)
  • BrennerK. et al.

    Gradient discretization of hybrid dimensional Darcy flows in fractured porous media with discontinuous pressure at matrix fracture interfaces

    IMA J. Numer. Anal.

    (2017)
  • ChenS. et al.

    A mortar finite volume method for a fractured model in porous media

    J. Math. Anal. Appl.

    (2018)
  • D’AngeloC. et al.

    A mixed finite element method for Darcy flow in fractured porous media with non-matching grids

    M2AN Math. Model. Numer. Anal.

    (2012)
  • PraM.D. et al.

    Well posedness of fully coupled fracture/bulk Darcy flow with XFEM

    SIAM J. Numer. Anal.

    (2006)
  • Cited by (7)

    • Study on the mass and heat transfer of gas-liquid two-phase flow in the porous media based on the 3D reconstruction model

      2021, Powder Technology
      Citation Excerpt :

      These transfer processes are very complex and interact with each other under some conditions. However, they are the basic processes that constitute many natural phenomena and various production and technology, including nuclear reactor cooling [1], storage, transportation and combustion of coal, solar energy collection and heat storage [2–4], chemical reaction, heat and mass transfer of catalytic filling bed [5,6], the underground water exploration, heat and moisture transfer in building and insulation materials, etc. The properties of porous materials are closely related to their microstructure.

    • Mechanism underlying initiation of migration of film-like residual oil

      2022, Journal of Dispersion Science and Technology
    View all citing articles on Scopus

    This work is partially supported by the National Natural Science Foundation of China (Nos. 11901372, 11771262,11771350), and Natural Science Foundation of Shaanxi Province, China (Nos. 2019JQ-077,2020JQ-403), and the Fundamental Research Fund for the Central Universities of China (Nos. GK201903007, GK201901008), and Foundation CMG.

    View full text