Higher order time integration methods for two-phase flow

Time integration methods that adapt in both the order of approximation and time step have been shown to provide efficient solutions to Richards' equation. In this work, we extend the same method of lines approach to solve a set of two-phase flow formulations and address some mass conservation issues from the previous work. We analyze these formulations and the nonlinear systems that result from applying the integration methods, placing particular emphasis on their index, range of applicability, and mass conservation characteristics. We conduct numerical experiments to study the behavior of the numerical models for three test problems. We demonstrate that higher order integration in time is more efficient than standard low-order methods for a variety of practical grids and integration tolerances, that the adaptive scheme successfully varies the step size in response to changing conditions, and that mass balance can be maintained efficiently using variable-order integration and an appropriately chosen numerical model formulation.     

Improving explicit time integration by modal truncation techniques

This paper describes a modal weighting technique that improves the stability characteristics of explicit time-integration schemes used in structural dynamics. The central difference method was chosen as the trial algorithm because of its simplicity, both in terms of formulation and ease of numerical stability and convergence analysis. It is shown how explicit algorithms may be reformulated in order to make them stable for any integration time by attenuating high-frequency oscillation modes that are generated by mesh geometry rather than generic dynamical features. We discuss results from trial calculations obtained from mathematical models that represent hysteretic restoring force elements and an application on a physical, four-degree-of-freedom, base-isolated structure using the pseudodynamic technique. © 1998 John Wiley & Sons, Ltd.     

A two-phase model for fast geomorphic shallow flows

The paper introduces a 2D shallow water model based on a two-phase formulation for the analysis of fast geomorphic transients occurring in the context of river morphodynamics.Mass and momentum conservation principles are separately imposed for both phases.The model naturally accounts for non-equilibrium solid transport,since neither instantaneous adaptation hypothesis nor any lag equation is employed to represent sediment dynamics.The hyperbolic character of the proposed model is shown to be preserved independently on the flow conditions.Results from numerical simulations of both 1D and 2D test-cases are compared with literature experimental data and with available numerical solutions.     

Sand fabric evolution effects on drain design for liquefaction mitigation

This paper revisits the seminal work of Seed and Booker (1977) [21] on the design of infinitely permeable drains for liquefaction mitigation. It is shown that their basic mathematical assumption for the rate of earthquake-induced excess pore pressure generation overlooks sand fabric evolution effects during cyclic loading and eventually leads to underestimation of the drain effectiveness. This is because such effects cause peak excess pore pressures to be attained at the early stages of partially drained shaking, followed by a gradual attenuation even if shaking continues undiminished, a response feature not predicted by the original formulation. In addition, special emphasis is given to the analytical relation describing the excess pore pressure build-up until liquefaction in undrained tests. This relation was considered unique in the original work, for reasons of simplicity, thus neglecting sand fabric evolution effects that may differentiate it for various sands, densities and loading conditions. Hence, a revised analytical formulation is proposed, which takes into account both above effects of sand fabric evolution. The paper provides a quantitative assessment of their influence on drain effectiveness and establishes a new set of charts for drain design. Experimental measurements from shaking table tests, as well as robust numerical simulations are shown, which underline the necessity for the revised solution and design charts.     

Comparative error analysis in finite element formulations of the advection-dispersion equation

The behaviour of numerical solutions of the one-dimensional advection-dispersion equation is investigated and comparisons between the consistent and the lumped formulations of Galerkin finite element schemes are made. Well-known criteria for the control of accuracy in the lumped (finite difference) formulation are reviewed. It is found that, because the numerical error produced by the consistent formulation is generally less than that produced by the lumped formulation, these criteria can also be used for the control of numerical dispersion in the consistent formulation. However, because the error in both types of solutions decreases in time when the discretization is invariant, the criteria can be relaxed with advancing simulation time. For the consistent formulation it is found that beyond some initial time period, the numerical error depends only on the temporal discretization. This suggests that constant accuracy can be maintained throughout the simulation period while allowing the time step length to grow.     

On magnetic boundary conditions for non-spectral dynamo simulations

We address issues associated with non-local magnetic boundary conditions for non-spectral dynamo simulations. We introduce an integro-differential formulation for a domain bounded by an insulating outer domain. We show how to combine the flexibility of a local discretisation with a rigorous formulation of magnetic boundary conditions in arbitrary geometries. This formulation substantiates from mathematical point of view a new method for numerical solution of magnetohydrodynamic problems with non-local boundary conditions based on coupling finite volumes and boundary elements. Finally, we discuss practical efficiency of this new method.     

Seismic response of an embedded pile in a transversely isotropic half-space under incident P-wave excitations

A rigorous mathematical formulation is presented for the analysis of a thin cylindrical shell embedded in a transversely isotropic half-space under vertically incident P-wave excitation. By virtue of a set of ring-loads Green's functions for the shell and a group of dynamic fundamental solutions for the half-space under arbitrary interfacial dynamic loads, the problem is shown to be reducible to a pair of Fredholm integral equations. By utilizing an adaptive-gradient family capable of capturing regular-to-singular solution transitions smoothly, an accurate numerical procedure is developed. To assess the effect of material anisotropy on the dynamic load-transfer process, a set of comprehensive numerical results presented for various material and geometrical conditions. The accuracy of the proposed numerical scheme is confirmed by its comparison with a benchmark solution for the corresponding isotropic problem.     

Validity limits for the van Genuchten–Mualem model and implications for parameter estimation and numerical simulation

The calculation of the relative hydraulic conductivity function based on water retention data is an attractive and widely used approach, since direct measurements of unsaturated conductivities are difficult. We show theoretically under which conditions an air-entry value for water retention data is definitely required when using the statistical approach of Mualem. Moreover we rigorously specify the conditions for which the classical van Genuchten–Mualem model leads to wrong predictions of relative hydraulic conductivity and, hence, an alternative formulation including an air-entry value should be used. Significant consequences are demonstrated for the inverse parameter estimation based on multistep outflow experiments. Furthermore it is shown that the use of a physically correct formulation of the water retention curve including an air-entry value and the derived hydraulic conductivity function influences not only the stability of numerical simulations but also their final results. This is especially grave as simulations with van Genuchten–Mualem parameters are frequently used to compare experiments and simulations and to draw conclusions on the correctness of Richards’ equation.     

On the combination of global and local data in collocation theory

Due to the successful operation of dedicated satellite gravity missions, nowadays high-accuracy global gravity field models have become available. This triggers the challenge to optimally combine this long to medium wavelength gravity field information derived from space-borne data with high-resolution terrestrial gravity data. In this paper, the least squares collocation concept is revised with the attempt to consistently unify the combination procedure in such a way that the full information contained in both data sets is merged. For example, in local or regional geoid determination the remove-restore method is usually applied only partially taking into account the accuracy of the global model coefficients used for the long-wavelength reduction. The key advantage of the extended formulation is the fact that it automatically accounts for the error covariance of all data types involved. The applicability, feasibility and performance of the proposed method is investigated in the frame of numerical closed-loop simulations. The two main fields of application, i.e., the improvement of a global gravity field model by terrestrial gravity field data, and, vice versa, the support to a regional geoid solution by the incorporation of a global gravity field model, have been analyzed and assessed. Although applied under simplified conditions, it could be shown that the method works and is practically applicable.     

A displacement finite element modelling for convection-diffusion problems

The development of a displacement finite element formulation and its application to convective transport problems is presented. The formulation is based on the introduction of a generalized quantity defined as transport displacement. The governing equation is expressed in terms of this quantity and by using generalized coordinates a variational form of the governing equation is obtained. This equation may be solved by any numerical method, though it is of particular interest for application of the finite element method. Two finite element models are derived for the solution of convection-diffusion boundary value problems. The performance of the two element models is discussed and numerical results are given for different cases of convection and diffusion with two types of boundary conditions. The numerical results obtained show not only the efficiency of the numerical models in handling pure convection, pure diffusion and mixed convection-diffusion problems, but also good stability and accuracy. The applications of the developed numerical models are not limited to diffusion-convection problems but can also be applied to other types of problems such as mass transfer, hydrodynamics and wave propagation.     

Accounting for high-order correlations in probabilistic characterization of environmental variables,and evaluation

Probabilistic characterization of environmental variables or data typically involves distributional fitting. Correlations, when present in variables or data, can considerably complicate the fitting process. In this work, effects of high-order correlations on distributional fitting were examined, and how they are technically accounted for was described using two multi-dimensional formulation methods: maximum entropy (ME) and Koehler–Symanowski (KS). The ME method formulates a least-biased distribution by maximizing its entropy, and the KS method uses a formulation that conserves specified marginal distributions. Two bivariate environmental data sets, ambient particulate matter and water quality, were chosen for illustration and discussion. Three metrics (log-likelihood function, root-mean-square error, and bivariate Kolmogorov–Smirnov statistic) were used to evaluate distributional fit. Bootstrap confidence intervals were also employed to help inspect the degree of agreement between distributional and sample moments. It is shown that both methods are capable of fitting the data well and have the potential for practical use. The KS distributions were found to be of good quality, and using the maximum likelihood method for the parameter estimation of a KS distribution is computationally efficient.     

An internal variable approach applied to the dynamic analysis of elastic–plastic structural systems

Discrete elastic–plastic systems subjected to dynamic load conditions are considered and properties related to an implicit time-integration scheme are discussed on the basis of an internal variable formulation. The material models accounted for are quite general: elastic–perfectly plastic and elastic–plastic (subjected to linear or non-linear, kinematic or isotropic hardening). The internal variable approach adopted in the paper leads easily to extremum theorems from which convergence properties of convenient time-integration schemes immediately follow. Next, taking account of the stabilising effect due to inertial and damping forces, it is shown how the above results can be extended to the case of softening (again linear or non-linear, kinematic or isotropic). Finally, some numerical examples are given and an extension to damage models is envisaged.     

Tube wave modeling by the finite-difference method with varying grid spacing

A finite-difference approach of aP-SV modeling scheme is applied to compute seismic wave propagation in heterogeneous isotropic media, including fluid-filled boreholes. The discrete formulation of the equation of motion requires the definition of the material parameters at the grid points of the numerical mesh. The grid spacing is chosen as coarse as possible with respect to the accurate representation of the shortest wavelength. If we assume frequencies lower than 250 Hz then the grid spacing is usually chosen in the range of a few meters. One encounters difficulties because of the large-scale difference between the grid spacing and the size of the borehole, usually several centimeters.These difficulties can be overcome by a grid refinement technique. This technique provides the construction of grids with varying grid spacing. The grid spacing in the vicinity of the borehole is chosen such that the borehole is properly represented. An example demonstrates the accuracy of this technique by comparisons with other methods. Unlike many analytical methods, the FD method can handle complex subsurface geometries. Further numerical examples of walk-awayVSP configurations show tube wave propagation within fluid-filled boreholes of realistic diameters.     

Influence of Boundary Condition Types on Unstable Density‐Dependent Flow

Boundary conditions are required to close the mathematical formulation of unstable density‐dependent flow systems. Proper implementation of boundary conditions, for both flow and transport equations, in numerical simulation are critical. In this paper, numerical simulations using the FEFLOW model are employed to study the influence of the different boundary conditions for unstable density‐dependent flow systems. A similar set up to the Elder problem is studied. It is well known that the numerical simulation results of the standard Elder problem are strongly dependent on spatial discretization. This work shows that for the cases where a solute mass flux boundary condition is employed instead of a specified concentration boundary condition at the solute source, the numerical simulation results do not vary between different convective solution modes (i.e., plume configurations) due to the spatial discretization. Also, the influence of various boundary condition types for nonsource boundaries was studied. It is shown that in addition to other factors such as spatial and temporal discretization, the forms of the solute transport equation such as divergent and convective forms as well as the type of boundary condition employed in the nonsource boundary conditions influence the convective solution mode in coarser meshes. On basis of the numerical experiments performed here, higher sensitivities regarding the numerical solution stability are observed for the Adams‐Bashford/Backward Trapezoidal time integration approach in comparison to the Euler‐Backward/Euler‐Forward time marching approach. The results of this study emphasize the significant consequences of boundary condition choice in the numerical modeling of unstable density‐dependent flow.     

Assessing the credibility of a series of computational fluid dynamic simulations of open channel flow

Recent developments in numerical algorithms have enabled the construction of three‐dimensional models for the prediction of flows in open channels. These advances encompass improvements in both numerical solutions and the process representation required for an accurate system definition. However, to date, there is still little agreement on how to assess systematically and report the credibility of these simulations. This paper addresses this problem by adopting a Grid Convergence Index approach. The results indicate, for two simple hypothetical cases, a zero‐degree confluence and a meander bend, that the numerical code can be verified to an acceptable numerical standard. However, it is shown that this does not mean that verification is complete, as the literature implies, as whilst the discretization resolution may be sufficient to verify one of the model variables it does not imply that every variable has converged. Furthermore, the scheme may still be insufficient to capture all the processes of interest that are operating within the chosen environment. Copyright © 2003 John Wiley & Sons, Ltd.     

On the primary variable switching technique for simulating unsaturated–saturated flows

Primary variable switching appears as a promising numerical technique for variably saturated flows. While the standard pressure-based form of the Richards equation can suffer from poor mass balance accuracy, the mixed form with its improved conservative properties can possess convergence difficulties for dry initial conditions. On the other hand, variable switching can overcome most of the stated numerical problems. The paper deals with variable switching for finite elements in two and three dimensions. The technique is incorporated in both an adaptive error-controlled predictor–corrector one-step Newton (PCOSN) iteration strategy and a target-based full Newton (TBFN) iteration scheme. Both schemes provide different behaviors with respect to accuracy and solution effort. Additionally, a simplified upstream weighting technique is used. Compared with conventional approaches the primary variable switching technique represents a fast and robust strategy for unsaturated problems with dry initial conditions. The impact of the primary variable switching technique is studied over a wide range of mostly 2D and partly difficult-to-solve problems (infiltration, drainage, perched water table, capillary barrier), where comparable results are available. It is shown that the TBFN iteration is an effective but error-prone procedure. TBFN sacrifices temporal accuracy in favor of accelerated convergence if aggressive time step sizes are chosen.     

The Influence of a Homogeneous Magnetic Field on the Ekman and Stewartson Layers

This contribution is aimed at a comparison of two different methods of how to deal with the solid inner core in geodynamo models. The first method, based on a direct application of the non-slip boundary conditions, was frequently used in the past. The second one, developed by the authors of the present paper, is based on an advanced analytical solution within the boundary layers and consequent formulation of new boundary conditions on the flow in the volume of the outer core. As an example we have used the results obtained by Hollerbach (1997) in the study of the influence of an imposed axial magnetic field on the fluid flow in a differentially rotating spherical shell. In the case of a weak imposed magnetic field, our solutions are very similar to those of Hollerbach. This non-trivial correspondence confirms the correctness of both methods, which are different not only in the formulation of boundary conditions, but also in the numerical methods: whereas Hollerbach used spectral methods, our computer code is based on finite differences. The influence of the conductivity of the inner core on the fluid flow was also studied.     

Mixed finite element methods and higher order temporal approximations for variably saturated groundwater flow

Richards’ equation (RE) is commonly used to model flow in variably saturated porous media. However, its solution continues to be difficult for many conditions of practical interest. Among the various time discretizations applied to RE, the method of lines (MOL) has been used successfully to introduce robust, accurate, and efficient temporal approximations. At the same time, a mixed-hybrid finite element method combined with an adaptive, higher order time discretization has shown benefits over traditional, lower order temporal approximations for modeling single-phase groundwater flow in heterogeneous porous media. Here, we extend earlier work for single-phase flow and consider two mixed finite element methods that have been used previously to solve RE using lower order time discretizations with either fixed time steps or empirically based adaption. We formulate the two spatial discretizations within a MOL context for the pressure head form of RE as well as a fully mass-conservative version. We conduct several numerical experiments for both spatial discretizations with each formulation, and we compare the higher order, adaptive time discretization to a first-order approximation with formal error control and adaptive time step selection. Based on the numerical results, we evaluate the performance of the methods for robustness and efficiency.