Mimetic Finite Difference Schemes Research Articles

A New Analysis of an Implicit Mimetic Scheme for the Heat Equation

A mimetic finite difference scheme for the transient heat equation under Robin’s conditions is presented. The scheme uses second order gradient and divergence mimetic operators, on a staggered grid, to approximate the space derivatives. The temporal derivative is replaced by a first order backward difference approximation to obtain an implicit formulation. The resulting scheme contains nonstandard finite difference stencils. An original convergence analysis by the matrix’s method shows that the proposed scheme is unconditionally stable. A comparative study against standard finite difference schemes, based on central difference or first order one side approximations, reveals the advantages of our scheme without being its implementation more expensive or difficult to achieve.

Tensorial elastodynamics for coupled acoustic/elastic anisotropic media: incorporating bathymetry

SUMMARYCorrectly implementing the fluid/solid boundary conditions at the seafloor is important for accurate full-wavefield imaging and inversion of marine seismic data. Because bathymetric profiles are rarely flat, the associated undulations influence wave modes interacting with the seafloor and, therefore, the ensuing imaging and inversion results. The conventional method of using single-domain elastic finite-difference (FD) solutions assumes that the fluid/solid contact is welded, which leads to incorrect handling of the boundary conditions and introduces modelling errors. We present a mimetic finite-difference (MFD) approach to solve the equations of anisotropic elastodynamics in a non-orthogonal coordinate system conformal to the bathymetric interface. The vertically deformed coordinate mapping transforms the irregular Cartesian (physical) domain into a regularly sampled generalized computational domain. We partition the medium into the acoustic and elastic subdomains and explicitly satisfy the fluid/solid boundary conditions with a split-node approach involving high-order one-sided MFD operators that achieve uniform spatial accuracy throughout the computational domain. Fully staggered grids (FSGs) are used to solve the velocity–pressure and velocity–stress formulations of the acoustic and anisotropic elastic wave equations, respectively. Numerical examples demonstrate that the proposed MFD+FSG algorithm accurately simulates wavefields even for strongly undulating bathymetric surfaces overlying structurally complex anisotropic media, and produces no spurious numerical artefacts (e.g. staircasing) or unphysical wave modes often caused by improper handling of the strong-contrast bathymetric interface. The wavefields generated by the tensorial MFD scheme closely match those from the more computationally expensive spectral-element method used to validate our implementation. The developed MFD+FSG technique can be effectively employed as the modelling kernel in a variety of coupled acoustic/elastic imaging and inversion applications.

Forward modelling of gravity data on unstructured grids using an adaptive mimetic finite-difference method

One of the techniques to optimize the gradient-based inversion of geophysical gravity data is to deal with the large and dense Jacobian matrices implicitly. This is made possible by solving the normal systems iteratively and to use pseudo-forward problems to calculate the effect of the Jacobian matrices. For this, numerical forward solvers should be used instead of analytical solvers. In this study, we propose mimetic finite-difference schemes for the solution of the forward gravity problem on unstructured grids. While the mimetic finite-difference method shares many traits with the finite-element and finite-volume methods, it has the advantage of naturally accommodating grids with arbitrary polyhedral elements. We use cell-based and vertex-based mimetic finite-difference schemes to solve a diffusion problem in mixed form, and Poisson's problem, respectively. The cell-based scheme is solved for the gravitational potential and attraction at the centroids and the facets of the elements, respectively, and the vertex-based scheme is solved for the potential at the vertices of the elements. We compare these mimetic finite-difference schemes with cell-based and vertex-based finite-volume schemes and a vertex-based finite-element scheme, in terms of accuracy of the potential. We then utilize the versatility of the mimetic finite-difference method, in accommodating arbitrary elements, and develop an adaptive mesh refinement tool. The mesh adaptivity is performed by an iterative h-refinement where goal-oriented error estimates are used to mark the elements for refinement. The marked elements are decomposed into new tetrahedra by regular subdivision. Since arbitrary polyhedra are naturally permitted in the mimetic finite-difference method, the added nodes are not regarded as hanging nodes, and therefore, any modification of the scheme or extra refinement is avoided. This characteristic allows preserving the quality of the initial grid and simplifying the refinement procedure. We demonstrate the practicality of the presented mesh adaptivity for the forward modelling of gravity and gravity gradient data using realistic geological models. We show that the cell-based mimetic scheme is more efficient than the vertex-based scheme in terms of computational resources. We also demonstrate that in the presence of irregular interfaces in the input domain, the regular refinement approach, presented in this study, is more reliable than a standard mesh regeneration technique.

Convergence of the mimetic finite difference and fitted mimetic finite difference method for options pricing

We present in this paper two novel numerical spatial discretization techniques based on the mimetic finite difference method for a degenerated partial differential equation (PDE) in one dimension. This PDE is well known as the Black-Scholes PDE which govern option pricing. To handle the degeneracy of the PDE, a novel fitted mimetic finite difference scheme is proposed together with the standard mimetic finite difference method. The temporal discretization is performing using standard implicit scheme. Furthermore rigorous convergence proofs in appropriate normed spaces are proposed. We validate the theoretical results by presenting numerical results and simulations. Those numerical experiments show that our two novel schemes outperform the standard finite difference method and the standard fitted finite volume method in terms of accuracy.

Mimetic finite-difference coupled-domain solver for anisotropic media

Accurately modeling full-wavefield solutions at and near the seafloor is challenging for conventional single-domain elastic finite-difference (FD) methods. Because they treat the fluid layer as a solid with zero shear-wave velocity, the energy partitioning for body and surface waves at the seafloor is distorted. This results in incorrect fluid/solid boundary conditions, which has significant implications for imaging and inversion applications that use amplitude information for model building. To address these issues, here we use mimetic FD (MFD) operators to develop and test a numerical approach for accurately implementing the boundary conditions at a fluid/solid interface. Instead of employing a single “global” model domain, we partition the full grid into two subdomains that represent the acoustic and elastic (possibly anisotropic) media. A novel split-node approach based on one-sided MFD operators is introduced to distribute grid points at the fluid/solid interface and satisfy the wave equation and the boundary conditions. Numerical examples demonstrate that such MFD operators achieve stable implementation of the boundary conditions with the same (fourth) order of spatial accuracy as that inside the split-domain interiors. We compare the wavefields produced by the MFD scheme with those from a more computationally expensive spectral-element method to validate our algorithm. The modeling results help analyze the events associated with the fluid/solid (seafloor) interface and provide valuable insights into the horizontal displacement or velocity components (e.g., recorded in ocean-bottom-node data sets). The developed MFD approach can be efficiently used in elastic anisotropic imaging and inversion applications involving ocean-bottom seismic data.

A fully implicit mimetic finite difference scheme for general purpose subsurface reservoir simulation with full tensor permeability

In the previous article Abushaikha et al. (2017) [1], we presented a fully-implicit mixed hybrid finite element (MHFE) method for general-purpose compositional reservoir simulation. The present work extends the implementation for mimetic finite difference (MFD) discretization method. The new approach admits fully implicit solution on general polyhedral grids. The scheme couples the momentum and mass balance equations to assure conservation and applies a cubic equation-of-state for the fluid system. The flux conservativity is strongly imposed for the fully implicit approach and the Newton-Raphson method is used to linearize the system. We test the method through extensive numerical examples to demonstrate the convergence and accuracy on various shapes of polyhedral. We also compare the method to other discretization schemes for unstructured meshes and tensor permeability. Finally, we apply the method through applied computational cases to illustrate its robustness for full tensor anisotropic, highly heterogeneous and faulted reservoirs using unstructured grids.

High-order mimetic finite-difference operators satisfying the extended Gauss divergence theorem

We present high-order mimetic finite-difference operators that satisfy the extended Gauss theorem. These operators have the same order of accuracy in the interior and at the boundary, no free parameters and optimal bandwidth. They are defined over staggered grids, using weighted inner products with a diagonal norm. We present several examples to demonstrate that mimetic finite-difference schemes using these operators produce excellent results.

High Order Mimetic Finite Difference Operators Satisfying a Gauss Divergence Theorem

High order mimetic finite difference operators that satisfy a discrete extended Gauss Divergence theorem are presented. These operators have the same order of accuracy in the interior as well as the boundary, no free parameters and optimal bandwidth. They are constructed on staggered grids, using weighted inner products with a diagonal norm. We present several examples to demonstrate that mimetic finite difference schemes using these operators produce excellent results.

The Gradient Discretization Method for Optimal Control Problems, with Superconvergence for Nonconforming Finite Elements and Mixed-Hybrid Mimetic Finite Differences

In this paper, optimal control problems governed by diffusion equations with Dirichlet and Neumann boundary conditions are investigated in the framework of the gradient discretisation method. Gradient schemes are defined for the optimality system of the control problem. Error estimates for state, adjoint and control variables are derived. Superconvergence results for gradient schemes under realistic regularity assumptions on the exact solution is discussed. These super-convergence results are shown to apply to non-conforming $\mathbb{P}_1$ finite elements, and to the mixed/hybrid mimetic finite differences. Results of numerical experiments are demonstrated for the conforming, nonconforming and mixed/hybrid mimetic finite difference schemes.

Convergence Analysis of the mimetic Finite Difference Method for Elliptic Problems with Staggered Discretizations of Diffusion Coefficients

We study the convergence of the new family of mimetic finite difference schemes for linear diffusion problems recently proposed in [K. Lipnikov, G. Manzini, J. D. Moulton, and M. Shashkov, J. Comput. Phys, 305 (2016), pp. 111--126]. In contrast to the conventional approach, the diffusion coefficient enters both the primary mimetic operator, i.e., the discrete divergence, and the inner product in the space of gradients. The diffusion coefficient is therefore evaluated on different mesh locations, i.e., inside mesh cells and on mesh faces. Such a staggered discretization may provide the flexibility necessary for future development of efficient numerical schemes for nonlinear problems, especially for problems with degenerate coefficients. These new mimetic schemes preserve symmetry and positive-definiteness of the continuum problem, which allow us to use efficient algebraic solvers such as the preconditioned conjugate gradient method. We show that these schemes are inf-sup stable and establish a priori error...

The mimetic finite difference method for the Landau–Lifshitz equation

The Landau–Lifshitz equation describes the dynamics of the magnetization inside ferromagnetic materials. This equation is highly nonlinear and has a non-convex constraint (the magnitude of the magnetization is constant) which poses interesting challenges in developing numerical methods. We develop and analyze explicit and implicit mimetic finite difference schemes for this equation. These schemes work on general polytopal meshes which provide enormous flexibility to model magnetic devices with various shapes. A projection on the unit sphere is used to preserve the magnitude of the magnetization. We also provide a proof that shows the exchange energy is decreasing in certain conditions. The developed schemes are tested on general meshes that include distorted and randomized meshes. The numerical experiments include a test proposed by the National Institute of Standard and Technology and a test showing formation of domain wall structures in a thin film.

A performance analysis of a mimetic finite difference scheme for acoustic wave propagation on GPU platforms

SummaryRealistic applications of numerical modeling of acoustic wave dynamics usually demand high‐performance computing because of the large size of study domains and demanding accuracy requirements on simulation results. Forward modeling of seismic motion on a given subsurface geological structure is by itself a good example of such applications, and when used as a component of seismic inversion tools or as a guide for the design of seismic surveys, its computational cost increases enormously. In the case of finite difference methods (or any other volumen‐discretization scheme), memory and computing demands rise with grid refinement, which may be necessary to reduce errors on numerical wave patterns and better capture target physical devices. In this work, we present several implementations of a mimetic finite difference method for the simulation of acoustic wave propagation on highly dense staggered grids. These implementations evolve as different optimization strategies are employed starting from appropriate setting of compilation flags, code vectorization by using streaming SIMD extensions Advanced Vector Extensions (AVX), CPU parallelization by exploiting the Open Multi‐Processing framework to the final code parallelization on graphics processing unit platforms. We present and discuss the increasing processing speed up of this mimetic scheme achieved by the gradual implementation and testing of all these performance optimizations. In terms of simulation times, the performance of our graphics processing unit parallel implementations is consistently better than the best CPU version. Copyright © 2016 John Wiley & Sons, Ltd.

The mimetic finite difference method for elliptic and parabolic problems with a staggered discretization of diffusion coefficient

Numerical schemes for nonlinear parabolic equations based on the harmonic averaging of cell-centered diffusion coefficients break down when some of these coefficients go to zero or their ratio grows. To tackle this problem, we propose new mimetic finite difference schemes that use a staggered discretization of the diffusion coefficient. The primary mimetic operator approximates div ( k ⋅ ) ; the derived (dual) mimetic operator approximates − ∇ ( ⋅ ) . The new mimetic schemes preserve symmetry and positive-definiteness of the continuum problem which allows us to use algebraic solvers with optimal complexity. We perform detailed numerical analysis of the new schemes for linear elliptic problems and a specially designed linear parabolic problem that has solution dynamics typical for nonlinear problems. We show that the new schemes are competitive with the state-of-the-art schemes for steady-state problems but provide much more accurate solution dynamics for the transient problem.

A new mimetic scheme for the acoustic wave equation

A new mimetic finite difference scheme for solving the acoustic wave equation is presented. It combines a novel second order tensor mimetic discretizations in space and a leapfrog approximation in time to produce an explicit multidimensional scheme. Convergence analysis of the new scheme on a staggered grid shows that it can take larger time steps than standard finite difference schemes based on ghost points formulation. A set of numerical test problems gives evidence of the versatility of the new mimetic scheme for handling general boundary conditions.

A finite element framework for some mimetic finite difference discretizations

In this work we derive equivalence relations between mimetic finite difference schemes on simplicial grids and modified Nédélec–Raviart–Thomas finite element methods for model problems in H(curl) and H(div). This provides a simple and transparent way to analyze such mimetic finite difference discretizations using the well-known results from finite element theory. The finite element framework that we develop is also crucial for the design of efficient multigrid methods for mimetic finite difference discretizations, since it allows us to use canonical inter-grid transfer operators arising from the finite element framework. We provide special Local Fourier Analysis and numerical results to demonstrate the efficiency of such multigrid methods.

Local Fourier Analysis for Edge-Based Discretizations on Triangular Grids

AbstractIn this paper, we present a local Fourier analysis framework for analyzing the different components within multigrid solvers for edge-based discretizations on triangular grids. The different stencils associated with edges of different orientation in a triangular mesh make this analysis special. The resulting tool is demonstrated for the vector Laplace problem discretized by mimetic finite difference schemes. Results from the local Fourier analysis, as well as experimentally obtained results, are presented to validate the proposed analysis.

New perspectives on polygonal and polyhedral finite element methods

Generalized barycentric coordinates such as Wachspress and mean value coordinates have been used in polygonal and polyhedral finite element methods. Recently, mimetic finite difference schemes were cast within a variational framework, and a consistent and stable finite element method on arbitrary polygonal meshes was devised. The method was coined as the virtual element method (VEM), since it did not require the explicit construction of basis functions. This advance provides a more in-depth understanding of mimetic schemes, and also endows polygonal-based Galerkin methods with greater flexibility than three-node and four-node finite element methods. In the VEM, a projection operator is used to realize the decomposition of the stiffness matrix into two terms: a consistent matrix that is known, and a stability matrix that must be positive semi-definite and which is only required to scale like the consistent matrix. In this paper, we first present an overview of previous developments on conforming polygonal and polyhedral finite elements, and then appeal to the exact decomposition in the VEM to obtain a robust and efficient generalized barycentric coordinate-based Galerkin method on polygonal and polyhedral elements. The consistent matrix of the VEM is adopted, and numerical quadrature with generalized barycentric coordinates is used to compute the stability matrix. This facilitates post-processing of field variables and visualization in the VEM, and on the other hand, provides a means to exactly satisfy the patch test with efficient numerical integration in polygonal and polyhedral finite elements. We present numerical examples that demonstrate the sound accuracy and performance of the proposed method. For Poisson problems in ℝ2 and ℝ3, we establish that linearly complete generalized barycentric interpolants deliver optimal rates of convergence in the L2-norm and the H1-seminorm.

A high-order mimetic method on unstructured polyhedral meshes for the diffusion equation

We present a new family of mimetic finite difference schemes for solving elliptic partial differential equations in the primal form on unstructured polyhedral meshes. These mimetic discretizations are built to satisfy local consistency and stability conditions. The consistency condition is an exactness property, i.e., the mimetic schemes are exact when the solution is a polynomial of an assigned degree. The stability condition ensures the well-posedness of the method. The degrees of freedom are the solution moments on mesh faces and inside mesh cells. Higher order schemes are built using higher order moments. The developed schemes are verified numerically on diffusion problems with constant and spatially variable (possibly, discontinuous) tensorial coefficients.

GRADIENT SCHEMES: A GENERIC FRAMEWORK FOR THE DISCRETISATION OF LINEAR, NONLINEAR AND NONLOCAL ELLIPTIC AND PARABOLIC EQUATIONS

Gradient schemes are nonconforming methods written in discrete variational formulation and based on independent approximations of functions and gradients, using the same degrees of freedom. Previous works showed that several well-known methods fall in the framework of gradient schemes. Four properties, namely coercivity, consistency, limit-conformity and compactness, are shown in this paper to be sufficient to prove the convergence of gradient schemes for linear and nonlinear elliptic and parabolic problems, including the case of nonlocal operators arising for example in image processing. We also show that the schemes of the Hybrid Mimetic Mixed family, which include in particular the Mimetic Finite Difference schemes, may be seen as gradient schemes meeting these four properties, and therefore converges for the class of above-mentioned problems.

Staggered grids for three-dimensional convection of a multicomponent fluid in a porous medium

Convection in a porous medium may produce strong nonuniqueness of patterns. We study this phenomena for the case of a multicomponent fluid and develop a mimetic finite-difference scheme for the three-dimensional problem. Discretization of the Darcy equations in the primitive variables is based on staggered grids with five types of nodes and on a special approximation of nonlinear terms. This scheme is applied to the computer study of flows in a porous parallelepiped filled by a two-component fluid and with two adiabatic lateral planes. We found that the continuous family of steady stable states exists in the case of a rather thin enclosure. When the depth is increased, only isolated convective regimes may be stable. We demonstrate that the non-mimetic approximation of nonlinear terms leads to the destruction of the continuous family of steady states.