Wouter Zijl

Numerical homogenization of two-phase flow in porous media

Homogenization has proved its effectiveness as a method of upscaling for linear problems, as they occur in single-phase porous media flow for arbitrary heterogeneous rocks. Here we extend the classical homogenization approach to nonlinear problems by considering incompressible, immiscible two-phase porous media flow. The extensions have been based on the principle of preservation of form, stating that the mathematical form of the fine-scale equations should be preserved as much as possible on the coarse scale. This principle leads to the required extensions, while making the physics underlying homogenization transparent. The method is process-independent in a way that coarse-scale results obtained for a particular reservoir can be used in any simulation, irrespective of the scenario that is simulated. Homogenization is based on steady-state flow equations with periodic boundary conditions for the capillary pressure. The resulting equations are solved numerically by two complementary finite element methods. This makes it possible to assess a posteriori error bounds.

Groundwater flow systems theory: research challenges beyond the specified-head top boundary condition

Groundwater flow systems theory : research challenges beyond the specified-head top boundary condition

Simple Hydraulic Conductivity Estimation by the Kalman Filtered Double Constraint Method

This paper presents the Kalman Filtered Double Constraint Method (DCM-KF) as a technique to estimate the hydraulic conductivities in the grid blocks of a groundwater flow model. The DCM is based on two forward runs with the same initial grid block conductivities, but with alternating flux-head conditions specified on parts of the boundary and the wells. These two runs are defined as: (1) the flux run, with specified fluxes (recharge and well abstractions), and (2) the head run, with specified heads (measured in piezometers). Conductivities are then estimated as the initial conductivities multiplied by the fluxes obtained from the flux run and divided by the fluxes obtained from the head run. The DCM is easy to implement in combination with existing models (e.g., MODFLOW). Sufficiently accurate conductivities are obtained after a few iterations. Because of errors in the specified head-flux couples, repeated estimation under varying hydrological conditions results in different conductivities. A time-independent estimate of the conductivities and their inaccuracy can be obtained by a simple linear KF with modest computational requirements. For the Kleine Nete catchment, Belgium, the DCM-KF yields sufficiently accurate calibrated conductivities. The method also results in distinguishing regions where the head-flux observations influence the calibration from areas where it is not able to influence the hydraulic conductivity.

Face-centered and volume-centered discrete analogs of the exterior differential equations governing porous medium flow II: Examples

This paper presents a ‘physics-oriented’ approach to approximate the continuum equations governing porous media flow by discrete analogs. To that end, the continuity equation and Darcy’s law are reformulated using exterior differential forms. This way the derivation of a system of algebraic equations (the discrete analog) on a finite-volume mesh can be accomplished by simple and elegant ‘translation rules.’ In the discrete analog the information about the conductivities of the porous medium and the metric of the mesh are represented in one matrix: the discrete dual. The discrete dual of the block-centered finite difference method is presented first. Since this method has limited applicability with respect to anisotropy and non-rectangular grid blocks, the finite element dual is introduced as an alternative. Application of a domain decomposition technique yields the face-centered finite element method. Since calculations based on pressures in volume centers are sometimes preferable, a volume-centered approximation of the face-centered approximation is presented too.

Numerical homogenization of the rigidity tensor in Hooke's law using the node-based finite element method

Combining a geological model with a geomechanical model, it generally turns out that the geomechanical model is built from units that are at least a 100 times larger in volume than the units of the geological model. To counter this mismatch in scales, the geological data models heterogeneous fine-scale Youngs moduli and Poissons ratios have to be “upscaled” to one “equivalent homogeneous” coarse-scale rigidity. This coarse-scale rigidity relates the volume-averaged displacement, strain, stress, and energy to each other, in such a way that the equilibrium equation, Hookes law, and the energy equation preserve their fine-scale form on the coarse scale. Under the simplifying assumption of spatial periodicity of the heterogeneous fine-scale rigidity, homogenization theory can be applied. However, even then the spatial variability is generally so complex that exact solutions cannot be found. Therefore, numerical approximation methods have to be applied. Here the node-based finite element method for the displacement as primary variable has been used. Three numerical examples showing the upper bound character of this finite element method are presented.

A direct method for the identification of the permeability field based on flux assimilation by a discrete analog of Darcy's law

Inverse models to determine the permeability are generally based on existing forward models for the pressure. The permeabilities are adapted in such a way that the calculated pressures match the specified pressures in a number of points. To assimilate a priori knowledge about the flux, we introduce the ‘flux assimilation method’, which is based on the ‘vector potential–pressure formulation’ of Darcys law. Thanks to an unconventional discretization technique – the ‘edge-based face element method’ – not only the specified pressures, but also specified information about the flux density can easily be assimilated. A relatively simple, but insightful analytical example illustrates the potential of this method.

The symmetry approximation for nonsymmetric permeability tensors and its consequences for mass transport

It is well known that the permeability has a tensor character. In practical applications, this is accounted for by the introduction of three principal permeabilities — three scalars — and three mutually orthogonal principal axes. In this paper, it is investigated whether this is always the exact way of describing anisotropy and, if not, what the consequences of the principal axes approximation are for flow and transport. First, it is shown that spatial upscaling may result in nonsymmetric large-scale permeability tensors, for which principal axes do not exist. However, it is possible to define generalized principal axes: three principal axes for the flux and three for the pressure gradient, with only three principal permeabilities. Since nonsymmetric permeability tensors are undesirable in practical applications, an approximation method making the nonsymmetric permeability symmetric is introduced. The important conclusion is then that the exact large-scale flux and large-scale pressure gradient do not have the same directions as the approximate flux and approximate pressure gradient. A practical consequence is that the principal axes approximation results in a difference between flux and transport direction. When considering miscible displacement or transport of mass dissolved in groundwater, the velocity component normal to the flux direction may be considered as a contribution to the transverse macro dispersion.

Zone-Integrated Double-Constraint Methodology for Calibration of Hydraulic Conductivities in Grid Cell Clusters of Groundwater Flow Models

Abstract This paper deals with the double-constraint methodology for calibration of steady-state groundwater flow models. The methodology is based on updating the hydraulic conductivity of the model domain by comparing the results of two forward groundwater flow models: a model in which known fluxes are specified as boundary conditions and a model in which known heads are specified as boundary conditions. A new zone-integrated double-constraint approach is presented by partitioning the model domain in zones with presumed constant hydraulic conductivity (soft data), and the double-constraint methodology is reformulated accordingly. The feasibility of the method is illustrated by a practical case study involving a numerical steady-state groundwater flow model with about 3 million grid blocks, subdivided into four zones corresponding to the major hydrogeological formations. The results of the zone-integrated double-constraint method for estimating the horizontal and vertical hydraulic conductivities of the zones compare favourably with a classical model calibration based on minimisation of the differences between calculated and measured heads, while the double-constraint method proves to be more robust and computationally less cumbersome.

Introduction: Setting the Scene

This chapter introduces the subject matter: parameter estimation for groundwater flow models by the double constraint methodology (DCM). After a brief introduction of forward and inverse modeling, the difference between imaging and calibration is mentioned. Some relevant literature is reviewed, and the contents of the chapters of this book are presented.

Foundations of Forward and Inverse Groundwater Flow Models

This chapter deals with the basic equation governing groundwater flow. In Sect. 2.1, the equations are presented in their full four-dimensional form (three spatial dimensions + time) with emphasis on the parameters. Section 2.2 introduces Calderon’s approach to determination of the spatially heterogeneous hydraulic conductivity field. To avoid negative hydraulic conductivities in the double constraint methodology (DCM), this approach is based on the square root of the hydraulic conductivity (sqrt-conductivity α). Section 2.3 introduces Stefanescu’s α-center method for parameter estimation, while Sect. 2.4 analyzes Calderon’s approach in more depth using inspiration from Stefanesco’s method. Although this chapter sets the scene for the real subject matter of this book—the double constraint methodology (DCM)—its reading may be skipped by readers who want to go directly to the DCM treated in Chap. 3.