Operator splitting methods for degenerate convection–diffusion equations II: numerical examples with emphasis on reservoir simulation and sedimentation

We present an accurate numerical method for a large class of scalar, strongly degenerate convection–diffusion equations. Important subclasses are hyperbolic conservation laws, porous medium type equations, two-phase reservoir flow equations, and strongly degenerate equations coming from the recent theory of sedimentation–consolidation processes. The method is based on splitting the convective and the diffusive terms. The nonlinear, convective part is solved using front tracking and dimensional splitting, while the nonlinear diffusion part is solved by an implicit–explicit finite difference scheme. In addition, one version of the implemented operator splitting method has a mechanism built in for detecting and correcting unphysical entropy loss, which may occur when the time step is large. This mechanism helps us gain a large time step ability for practical computations. A detailed convergence analysis of the operator splitting method was given in Part I. Here we present numerical experiments with the method for examples modelling secondary oil recovery and sedimentation–consolidation processes. We demonstrate that the splitting method resolves sharp gradients accurately, may use large time steps, has first order convergence, exhibits small grid orientation effects, has small mass balance errors, and is rather efficient.     

A computational method for approximating a Darcy–Stokes system governing a vuggy porous medium

We develop and analyze a mixed finite element method for the solution of an elliptic system modeling a porous medium with large cavities, called vugs. It consists of a second-order elliptic (i.e., Darcy) equation on part of the domain coupled to a Stokes equation on the rest of the domain, and a slip boundary condition (due to Beavers–Joseph–Saffman) on the interface between them. The tangential velocity is not continuous on the interface. We consider a 2-D vuggy porous medium with many small cavities throughout its extent, so the interface is not isolated. We use a certain conforming Stokes element on rectangles, slightly modified near the interface to account for the tangential discontinuity. This gives a mixed finite element method for the entire Darcy–Stokes system with a regular sparsity pattern that is easy to implement, independent of the vug geometry, as long as it aligns with the grid. We prove optimal global first-order L ² convergence of the velocity and pressure, as well as the velocity gradient in the Stokes domain. Numerical results verify these rates of convergence and even suggest somewhat better convergence in certain situations. Finally, we present a lower dimensional space that uses Raviart–Thomas elements in the Darcy domain and uses our new modified elements near the interface in transition to the Stokes elements.     

Precise integration method for solving the seepage-stress coupling problem in rock mass with singular possibilities of the coupling matrix

Seepage-stress coupling is a key problem in the field of geotechnical engineering, and the finite element method is one of the main methods to study seepage-stress coupling in rock masses. However, the finite element method has issues of poor stability, low efficiency, and low accuracy in the simulation of the seepage-stress coupling problem. In this paper, the homogeneous saturated rock mass is taken as the object to deduce the control equation based on the Biot's theory. Considering the singularity of the coupling matrix, the discrete equation is converted into a precise integral format, and the equation is solved by the precise integration method to avoid instability and low precision. The precise integration method in this paper has good numerical stability, fast convergence speed, and high simulation accuracy, which effectively facilitates the rapid and stable numerical simulation of the seepage-stress coupling problem using the equivalent continuum medium model. The validity and accuracy of the precise integration method for seepage-stress coupling problems are verified by numerical examples.     

Consistent tangent operators for constitutive rate equations

A general approach for obtaining the consistent tangent operator for constitutive rate equations is presented. The rate equations can be solved numerically by the user's favourite time integrator. In order to obtain reliable results, the substepping in integration should be based on a control of the local error. The main ingredient of the consistent tangent operator, namely the derivative of the stress with respect to the strain increment must be computed simultaneously with the same integrator, applied to a numerical approximation of the variational equations. This information enables finite‐element packages to assemble a consistent tangent operator and thus guarantees quadratic convergence of the equilibrium iterations. Several numerical examples with a hypoplastic constitutive law are given. As numerical integrator we used a second‐order extrapolated Euler method. Quadratic convergence of the equilibrium iteration is shown. Copyright © 2002 John Wiley & Sons, Ltd.     

Direct,partitioned and projected solution to finite element consolidation models

Direct, partitioned, and projected (conjugate gradient‐like) solution approaches are compared on unsymmetric indefinite systems arising from the finite element integration of coupled consolidation equations. The direct method is used in its most recent and computationally efficient implementations of the Harwell Software Library. The partitioned approach designed for coupled problems is especially attractive as it addresses two separate positive definite problems of a smaller size that can be solved by symmetric conjugate gradients. However, it may stagnate and when converging it does not prove competitive with a global projection method such as Bi‐CGSTAB, which may take full advantage of its flexibility in working on scaled and reordered equations, and thus may greatly improve its computational performance in terms of both robustness and convergence rate. The Bi‐CGSTAB superiority to the other approaches is discussed and demonstrated with a few representative examples in two‐dimensional (2‐D) and three‐dimensional (3‐D) coupled consolidation problems. Copyright © 2002 John Wiley & Sons, Ltd.     

Unconfined seepage analysis in earth dams using smoothed fixed grid finite element method

One major difficulty in seepage analyses is finding the position of phreatic surface which is unknown at the beginning of solution and must be determined in an iterative process. The objective of the present study is to develop a novel non‐boundary‐fitted mesh finite‐element method capable of solving the unconfined seepage problem in domains with arbitrary geometry and continuously varied permeability. A new non‐boundary‐fitted finite element method named as smoothed fixed grid finite element method (SFGFEM) is used to simplify the solution of variable domain problem of unconfined seepage. The gradient smoothing technique, in which the area integrals are transformed into the line integrals around edges of smoothing cells, is used to obtain the element matrices. The solution process starts with an initial guess for the unknown boundary and SFGFEM is used to approximate the field variable. The boundary shape is then modified to eventually satisfy nonlinear boundary condition in an iterative process. Some numerical examples are solved to evaluate the applicability of the proposed method and the results are compared with those available in the literature. Copyright © 2011 John Wiley & Sons, Ltd.     

Upper bound limit analysis using linear finite elements and non‐linear programming

A new method for computing rigorous upper bounds on the limit loads for one‐, two‐ and three‐dimensional continua is described. The formulation is based on linear finite elements, permits kinematically admissible velocity discontinuities at all interelement boundaries, and furnishes a kinematically admissible velocity field by solving a non‐linear programming problem. In the latter, the objective function corresponds to the dissipated power (which is minimized) and the unknowns are subject to linear equality constraints as well as linear and non‐linear inequality constraints. Provided the yield surface is convex, the optimization problem generated by the upper bound method is also convex and can be solved efficiently by applying a two‐stage, quasi‐Newton scheme to the corresponding Kuhn–Tucker optimality conditions. A key advantage of this strategy is that its iteration count is largely independent of the mesh size. Since the formulation permits non‐linear constraints on the unknowns, no linearization of the yield surface is necessary and the modelling of three‐dimensional geometries presents no special difficulties. The utility of the proposed upper bound method is illustrated by applying it to a number of two‐ and three‐dimensional boundary value problems. For a variety of two‐dimensional cases, the new scheme is up to two orders of magnitude faster than an equivalent linear programming scheme which uses yield surface linearization. Copyright © 2001 John Wiley & Sons, Ltd.     

二维接触问题的互补算法及工程应用

接触问题是一种常见的非线性问题,如何能够很好地模拟接触面的变形及受力特性,以及实现对变形体间的接触问题的真实模拟是该领域研究的难点问题。基于二维接触问题的实际物理意义,分别在法向和切向建立等价的互补模型。用非线性互补函数（NCP）中的Fischer-Burmeister（FB）函数将互补函数模型转化为非光滑方程组表达,用常规的Newton法求解。同时,基于高斯积分法可以用较少的积分点达到较高的精度,为了进一步提高求解精度,改善不连续的通病,对面-面接触模型在高斯点上对接触面进行处理,可通过调节积分点数目对求解精度进行控制,方法易于理解,实现方便。在此基础上建立二维接触有限元模型,通过一系列工程算例验证该方法的可行性与有效性。结果表明,与ABAQUS有限元的计算结果相比,该方法有着较高的精度,更真实地反映问题的实际。     

Semi‐analytical far field model for three‐dimensional finite‐element analysis

A challenging computational problem arises when a discrete structure (e.g. foundation) interacts with an unbounded medium (e.g. deep soil deposit), particularly if general loading conditions and non‐linear material behaviour is assumed. In this paper, a novel method for dealing with such a problem is formulated by combining conventional three‐dimensional finite‐elements with the recently developed scaled boundary finite‐element method. The scaled boundary finite‐element method is a semi‐analytical technique based on finite‐elements that obtains a symmetric stiffness matrix with respect to degrees of freedom on a discretized boundary. The method is particularly well suited to modelling unbounded domains as analytical solutions are found in a radial co‐ordinate direction, but, unlike the boundary‐element method, no complex fundamental solution is required. A technique for coupling the stiffness matrix of bounded three‐dimensional finite‐element domain with the stiffness matrix of the unbounded scaled boundary finite‐element domain, which uses a Fourier series to model the variation of displacement in the circumferential direction of the cylindrical co‐ordinate system, is described. The accuracy and computational efficiency of the new formulation is demonstrated through the linear elastic analysis of rigid circular and square footings. Copyright © 2004 John Wiley & Sons, Ltd.     

Uncertainty quantification of two-phase flow problems via measure theory and the generalized multiscale finite element method

The problem of multiphase phase flow in heterogeneous subsurface porous media is one involving many uncertainties. In particular, the permeability of the medium is an important aspect of the model that is inherently uncertain. Properly quantifying these uncertainties is essential in order to make reliable probabilistic-based predictions and future decisions. In this work, a measure-theoretic framework is employed to quantify uncertainties in a two-phase subsurface flow model in high-contrast media. Given uncertain saturation data from observation wells, the stochastic inverse problem is solved numerically in order to obtain a probability measure on the space of unknown permeability parameters characterizing the two-phase flow. As solving the stochastic inverse problem requires a number of forward model solves, we also incorporate the use of a conservative version of the generalized multiscale finite element method for added efficiency. The parameter-space probability measure is used in order to make predictions of saturation values where measurements are not available, and to validate the effectiveness of the proposed approach in the context of fine and coarse model solves. A number of numerical examples are offered to illustrate the measure-theoretic methodology for solving the stochastic inverse problem using both fine and coarse solution schemes.     

Coupled formulation and algorithms for the simulation of non‐planar three‐dimensional hydraulic fractures using the generalized finite element method

This paper presents a coupled hydro‐mechanical formulation for the simulation of non‐planar three‐dimensional hydraulic fractures. Deformation in the rock is modeled using linear elasticity, and the lubrication theory is adopted for the fluid flow in the fracture. The governing equations of the fluid flow and elasticity and the subsequent discretization are fully coupled. A Generalized/eXtended Finite Element Method (G/XFEM) is adopted for the discretization of the coupled system of equations. A Newton–Raphson method is used to solve the resulting system of nonlinear equations. A discretization strategy for the fluid flow problem on non‐planar three‐dimensional surfaces and a computationally efficient strategy for handling time integration combined with mesh adaptivity are also presented. Several three‐dimensional numerical verification examples are solved. The examples illustrate the generality and accuracy of the proposed coupled formulation and discretization strategies. Copyright © 2015 John Wiley & Sons, Ltd.     

Solution of quasi‐static large‐strain problems by the material point method

A procedure for solving quasi‐static large‐strain problems by the material point method is presented. Owing to the Lagrangian–Eulerian features of the method, problems associated with excessive mesh distortions that develop in the Lagrangian formulations of the finite element method are avoided. Three‐dimensional problems are solved utilizing 15‐noded prismatic and 10‐noded tetrahedral elements with quadratic interpolation functions as well as an implicit integration scheme. An algorithm for exploiting the numerical integration procedure on the computational mesh is proposed. Several numerical examples are shown. Copyright © 2010 John Wiley & Sons, Ltd.     

Hybrid FE‐DQ for dynamic analysis of multilayered half‐space subjected to concentrated point impulsive loading

A hybrid finite element method and differential quadrature method (DQM) is developed to estimate the dynamic response of two‐dimensional multilayered half‐spaces subjected to impulsive point loading. Nonreflecting absorbing boundary conditions consist of appropriate springs, and dampers are considered. The capabilities of the finite element method for solving boundary value problems with general domain, loading and systematic boundary treatment are combined with accurate and stable time marching capabilities of the DQM to develop an accurate and efficient numerical technique. The capability, efficiency, robustness and convergence of the DQM for solving the dynamic problem are demonstrated through numerical simulations of various half‐spaces with different time increments and layer arrangement. Also, comparison study when using Newmark's time integration scheme for the same problem is done. It can be concluded that the DQM as an unconditionally stable method is suitable for solving such a problem. Also, parametric study is performed to show the effect of the absorbing boundary conditions on the dynamic response. Copyright © 2012 John Wiley & Sons, Ltd.     

强度折减有限元法中的单元阶次影响分析

强度折减有限元法是当前较为有效的边坡稳定性评价方法,且应用越来越广泛。但影响强度折减有限元法的因素有很多,单元阶次是其中比较重要的一个。通过3个经典算例,这些算例分别是二维地基承载力问题、二维边坡和三维边坡问题,分析了单元阶次的选择对强度折减法的影响。计算结果表明：随着单元的增多,线性单元和二次单元都从大于真实解的一侧来逼近真解;相对于二次单元,由于线性单元过“刚”,因此,会过高地估计安全系数,对于实际工程会偏于危险,且误差大,二次单元的误差是线性单元误差的1/8左右。在采用系统最大位移收敛与否的评判标准的基础上,利用二次单元来进行强度折减分析,则可以弥补这种线性单元的不足,得到更加合理的安全系数。二次单元比线性单元更适合于强度折减有限元法。     

A novel finite element method for seepage analysis

A novel finite element method has been proposed in this paper for the solution of seepage problems economically and accurately. In this method the governing equation and the prescribed boundary conditions are transformed so that they refer to a suitable logarithmically condensed ‘image’ space; the physical problem domain is also mapped into the image space. The transformed equation is then solved in the image space using standard finite elements, subject to the transformed boundary conditions. Because physical space is logarithmically condensed in the image space, the proposed method is capable of dealing with large or very large aspect ratio seepage problems economically and accurately. The validity of the method has been demonstrated by means of a number of examples including anisotropy and non-linearity. In all cases an excellent degree of accuracy was achieved, efficiently and economically.     

Finite difference techniques for modelling geothermal reservoirs

Two finite difference algorithms suitable for long-time simulation of the exploitation of a two-phase geothermal reservoir are presented. One is based on the hopscotch method proposed by Saul'yev¹ and analysed further by Gordon² and Gourlay.³ The other is based on the well-known ADI method. Both methods use a Newton–Raphson iterative technique in order to obtain accurate solutions of the non-linear difference equations involved. Rapid convergence of the iterative schemes occurs both for single-phase and two-phase reservoir problems. One- and two-dimensional model problems are presented.     

Ground response curves for rock masses exhibiting strain‐softening behaviour

A literature review has shown that there exist adequate techniques to obtain ground reaction curves for tunnels excavated in elastic‐brittle and perfectly plastic materials. However, for strain‐softening materials it seems that the problem has not been sufficiently analysed. In this paper, a one‐dimensional numerical solution to obtain the ground reaction curve (GRC) for circular tunnels excavated in strain‐softening materials is presented. The problem is formulated in a very general form and leads to a system of ordinary differential equations. By adequately defining a fictitious ‘time’ variable and re‐scaling some variables the problem is converted into an initial value one, which can be solved numerically by a Runge–Kutta–Fehlberg method, which is implemented in MATLAB environment. The method has been developed for various common particular behaviour models including Tresca, Mohr–Coulomb and Hoek–Brown failure criteria, in all cases with non‐associative flow rules and two‐segment piecewise linear functions related to a principal strain‐dependent plastic parameter to model the transition between peak and residual failure criteria. Some particular examples for the different failure criteria have been run, which agree well with closed‐form solutions—if existing—or with FDM‐based code results. Parametric studies and specific charts are created to highlight the influence of different parameters. The proposed methodology intends to be a wider and general numerical basis where standard and newly featured behaviour modes focusing on obtaining GRC for tunnels excavated in strain‐softening materials can be implemented. This way of solving such problems has proved to be more efficient and less time consuming than using FEM‐ or FDM‐based numerical 2D codes. Copyright © 2003 John Wiley & Sons, Ltd.     

A finite difference method for measuring soil thermal diffusivity in situ

A method is devised for measuring soil thermal diffusivity in situ. It is based on direct experimental simulation of the finite difference approximation to the one-dimensional heat conduction equation. The method does not require the soil to be homogeneous except between the three thermometers that are used, at depths z + d, z and z ? d. Nor need the energy input curve be sinusoidal. However, it must be fairly smooth for the finite difference approximation to be accurate. Experimental results for London Clay are presented, obtained using thermometers at depths of 1, 6 and 11 cm to give a mean thermal diffusivity of 0.0074 cm²/s at a depth of 6 cm. This value is consistent with other estimates of diffusivity for clay soils. The method is capable of automation, and should be suitable for use on engineering sites, at low cost. The method is capable of generalization to other linear diffusion equations containing one independent parameter. The same limitation also applies to its application to constitutive or geometrical non-linear one dimensional diffusion equations, and each equation requires individual study to assess feasibility of use of the method. The method in effect uses the usual finite difference approximation, not to prepare a numerical solution, but to design an experiment carried out essentially within the finite difference ‘molecule’. The measured parameter of the diffusion equation is the usable product of the method.     

各向异性介质中大地电磁场的异常特征

给出各向异性介质二维地电断面大地电磁场的边值问题以及等价的变分问题。对计算区域采用矩形网格中进一步三角细化的剖分方式并在三角单元内进行线性插值,解出有限单元法数值解。通过典型模型的正演模拟,得到大地电磁测深曲线,并与前人的研究工作对比,验证了该方法的有效性。     

A finite element solution of a fully coupled implicit formulation for reservoir simulation

An iterative method is presented for solving a fully coupled and implicit formulation of fluid flow in a porous medium. The mathematical model describes a set of fully coupled three-phase flow of compressible and immiscible fluids in a saturated oil reservoir. The finite element method is applied to obtain the simultaneous solution (SS) for the resulting highly non-linear partial differential equations where fluid pressures are the primary unknowns. The final discretized equations are solved iteratively by using a fully implicit numerical scheme. Several examples, illustrating the use of the present model, are described. The increased stability achieved with this scheme has permitted the use of larger time steps with smaller material balance errors.