含泥质夹层油藏蒸汽注采过程的自适应网格算法研究

根据泥质夹层的低渗特性及空间分布，本文提出了一种含泥质夹层油藏网格渗透率的粗化计算方法，并在此基础上，将自适应网格算法应用于含泥质夹层油藏的数值模拟，提升其计算效率．在计算过程中，网格的动态划分仅依据流体物理量的变化，泥质夹层区域不全部采用细网格，仅针对流动锋面处的泥质夹层采用细网格，其余泥质夹层处采用不同程度的粗网格．相较于传统算法，网格数大幅下降．数值算例表明，自适应网格算法的计算结果精度与全精细网格一致，能够准确模拟出泥质夹层对于流体的阻碍作用，同时计算效率得到大幅提升，约为全精细网格算法的3~7 倍．     

A high‐order flux reconstruction adaptive mesh refinement method for magnetohydrodynamics on unstructured grids

We report our recent development of the high‐order flux reconstruction adaptive mesh refinement (AMR) method for magnetohydrodynamics (MHD). The resulted framework features a shock‐capturing duo of AMR and artificial resistivity (AR), which can robustly capture shocks and rotational and contact discontinuities with a fraction of the cell counts that are usually required. In our previous paper, 36 we have presented a shock‐capturing framework on hydrodynamic problems with artificial diffusivity and AMR. Our AMR approach features a tree‐free, direct‐addressing approach in retrieving data across multiple levels of refinement. In this article, we report an extension to MHD systems that retains the flexibility of using unstructured grids. The challenges due to complex shock structures and divergence‐free constraint of magnetic field are more difficult to deal with than those of hydrodynamic systems. The accuracy of our solver hinges on 2 properties to achieve high‐order accuracy on MHD systems: removing the divergence error thoroughly and resolving discontinuities accurately. A hyperbolic divergence cleaning method with multiple subiterations is used for the first task. This method drives away the divergence error and preserves conservative forms of the governing equations. The subiteration can be accelerated by absorbing a pseudo time step into the wave speed coefficient, therefore enjoys a relaxed CFL condition. The AMR method rallies multiple levels of refined cells around various shock discontinuities, and it coordinates with the AR method to obtain sharp shock profiles. The physically consistent AR method localizes discontinuities and damps the spurious oscillation arising in the curl of the magnetic field. The effectiveness of the AMR and AR combination is demonstrated to be much more powerful than simply adding AR on finer and finer mesh, since the AMR steeply reduces the required amount of AR and confines the added artificial diffusivity and resistivity to a narrower and narrower region. We are able to verify the designed high‐order accuracy in space by using smooth flow test problems on unstructured grids. The efficiency and robustness of this framework are fully demonstrated through a number of two‐dimensional nonsmooth ideal MHD tests. We also successfully demonstrate that the AMR method can help significantly save computational cost for the Orszag‐Tang vortex problem.     

Cell-vertex Based Parallel and Adaptive Explicit 3D Flow Solution on Unstructured Grids

A parallel adaptive Euler flow solution algorithm is developed for 3D applications on distributed memory computers. Significant contribution of this research is the development and implementation of a parallel grid adaptation scheme together with an explicit cell vertex-based finite volume 3D flow solver on unstructured tetrahedral grids. Parallel adaptation of grids is based on grid-regeneration philosophy by using an existing serial grid generation program. Then, a general partitioner repartitions the grid. An adaptive sensor value, which is a measure to refine or coarsen grids, is calculated considering the pressure gradients in all partitioned blocks of grids. The parallel performance of the present study was tested. Parallel computations were performed on Unix workstations and a Linux cluster using MPI communication library. The present results show that overall adaptation scheme developed in this study is applicable to any pair of a flow solver and grid generator with affordable cost. It is also proved that parallel adaptation is necessary for accurate and efficient flow solutions.     

Patched grid and adaptive mesh refinement strategies for the calculation of the transport of vortices

This paper presents two techniques allowing local grid refinement to calculate the transport of vortices. one is the patched grid (PG) method which allows non‐coincident interfaces between blocks. Treatment of the non‐coincident interfaces is given in detail. The second one is the adaptive mesh refinement (AMR) method which has been developed in order to create embedded sub‐grids. The efficiency of these two methods is demonstrated by some validating tests. Then the PG and AMR strategies are applied in the computation of the transport of vortices. We start with a simple vortex flow in a cubic box. Then, the flowfield around a complex aircraft configuration is calculated using the two refinement techniques. Results are compared with a fine, referenced grid calculation. Copyright © 2002 John Wiley & Sons, Ltd.     

A finite volume scheme for solving anisotropic diffusion on ALE‐AMR grids

In the context of High Energy Density Physics and more precisely in the field of laser plasma interaction, Lagrangian schemes are commonly used. The lack of robustness due to strong grid deformations requires the regularization of the mesh through the use of Arbitrary Lagrangian Eulerian methods. Theses methods usually add some diffusion and a loss of precision is observed. We propose to use Adaptive Mesh Refinement (AMR) techniques to reduce this loss of accuracy. This work focuses on the resolution of the anisotropic diffusion operator on Arbitrary Lagrangian Eulerian‐AMR grids. In this paper, we describe a second‐order accurate cell‐centered finite volume method for solving anisotropic diffusion on AMR type grids. The scheme described here is based on local flux approximation which can be derived through the use of a finite difference approximation, leading to the CCLADNS scheme. We present here the 2D and 3D extension of the CCLADNS scheme to AMR meshes. Copyright © 2017 John Wiley & Sons, Ltd.     

A PARALLEL BLOCK-STRUCTURED MULTIGRID METHOD FOR THE PREDICTION OF INCOMPRESSIBLE FLOWS

In this paper a parallel multigrid finite volume solver for the prediction of steady and unsteady flows in complex geometries is presented. For the handling of the complexity of the geometry and for the parallelization a unified approach connected with the concept of block-structured grids is employed. The parallel implementation is based on grid partitioning with automatic load balancing and follows the message-passing concept, ensuring a high degree of portability. A high numerical efficiency is obtained by a non-linear multigrid method with a pressure correction scheme as smoother. By a number of numerical experiments on various parallel computers the method is investigated with respect to its numerical and parallel efficiency. The results illustrate that the high performance of the underlying sequential multigrid algorithm can largely be retained in the parallel implementation and that the proposed method is well suited for solving complex flow problems on parallel computers with high efficiency.     

Adaptive finite volume methods with well‐balanced Riemann solvers for modeling floods in rugged terrain: Application to the Malpasset dam‐break flood (France, 1959)

The simulation of advancing flood waves over rugged topography, by solving the shallow‐water equations with well‐balanced high‐resolution finite volume methods and block‐structured dynamic adaptive mesh refinement (AMR), is described and validated in this paper. The efficiency of block‐structured AMR makes large‐scale problems tractable, and allows the use of accurate and stable methods developed for solving general hyperbolic problems on quadrilateral grids. Features indicative of flooding in rugged terrain, such as advancing wet–dry fronts and non‐stationary steady states due to balanced source terms from variable topography, present unique challenges and require modifications such as special Riemann solvers. A well‐balanced Riemann solver for inundation and general (non‐stationary) flow over topography is tested in this context. The difficulties of modeling floods in rugged terrain, and the rationale for and efficacy of using AMR and well‐balanced methods, are presented. The algorithms are validated by simulating the Malpasset dam‐break flood (France, 1959), which has served as a benchmark problem previously. Historical field data, laboratory model data and other numerical simulation results (computed on static fitted meshes) are shown for comparison. The methods are implemented in GEO CLAW , a subset of the open‐source CLAWPACK software. All the software is freely available at www.clawpack.org . Published in 2010 by John Wiley & Sons, Ltd.     

A parallel gas-kinetic Bhatnagar–Gross–Krook method for the solution of viscous flows on two-dimensional hybrid grids

In this study, a parallel implementation of gas-kinetic Bhatnagar–Gross–Krook method on two-dimensional hybrid grids is presented. Boundary layer regions in wall bounded viscous flows are discretised with quadrilateral grid cells stretched in the direction normal to the solid surface while the rest of the flow domain is discretised by triangular cells. The parallel solution algorithm on hybrid grids is based on the domain decomposition using METIS, a graph partitioning software. The flow solutions obtained in parallel significantly improve the computation time, a significant deficiency of gas-kinetic methods. Several validation test cases presented show the accuracy and robustness of the method developed.     

CartGen: Robust,efficient and easy to implement uniform/octree/embedded boundary Cartesian grid generator

An efficient and easy to implement method to generate Cartesian grids is presented. The presented method generates various kinds of Cartesian grids such as uniform, octree and embedded boundary grids. It supports the variation of grid size along each spatial direction as well as anisotropic and non‐graded refinements. The efficiency and ease of implementation are the main benefits of the presented method in contrast to the alternative methods. Regarding octree grid generation, applying a simple and efficient data compression method permits to store all grid levels without considerable memory overhead. The presented method generates octree grids up to a 13‐level refinement (8192³ grids on the finest level) from a complicated geometry in a few minutes on the traditional desktop computers. The FORTRAN 90 implementation of the presented method is freely available under the terms of the GNU general public license. Copyright © 2007 John Wiley & Sons, Ltd.     

Adaptive thermo-fluid moving boundary computations for interfacial dynamics

In this study,we present adaptive moving boundary computation technique with parallel implementation on a distributed memory multi-processor system for large scale thermo-fluid and interfacial flow computations.The solver utilizes Eulerian-Lagrangian method to track moving(Lagrangian) interfaces explicitly on the stationary(Eulerian) Cartesian grid where the flow fields are computed.We address the domain decomposition strategies of Eulerian-Lagrangian method by illustrating its intricate complexity of the computation involved on two different spaces interactively and consequently,and then propose a trade-off approach aiming for parallel scalability.Spatial domain decomposition is adopted for both Eulerian and Lagrangian domain due to easy load balancing and data locality for minimum communication between processors.In addition,parallel cell-based unstructured adaptive mesh refinement(AMR) technique is implemented for the flexible local refinement and even-distributed computational workload among processors.Selected cases are presented to highlight the computational capabilities,including Faraday type interfacial waves with capillary and gravitational forcing,flows around varied geometric configurations and induced by boundary conditions and/or body forces,and thermo-fluid dynamics with phase change.With the aid of the present techniques,large scale challenging moving boundary problems can be effectively addressed.     

Implementation of some higher-order convection schemes on non-uniform grids

A generalized formulation is applied to implement the quadratic upstream interpolation (QUICK) scheme, the second-order upwind (SOU) scheme and the second-order hybrid scheme (SHYBRID) on non-uniform grids. The implementation method is simple. The accuracy and efficiency of these higher-order schemes on non-uniform grids are assessed. Three well-known bench mark convection-diffusion problems and a fluid flow problem are revisited using non-uniform grids. These are: (1) transport of a scalar tracer by a uniform velocity field; (2) heat transport in a recirculating flow; (3) two-dimensional non-linear Burgers equations; and (4) a two-dimensional incompressible Navier-Stokes flow which is similar to the classical lid-driven cavity flow. The known exact solutions of the last three problems make it possible to thoroughly evaluate accuracies of various uniform and non-uniform grids. Higher accuracy is obtained for fewer grid points on non-uniform grids. The order of accuracy of the examined schemes is maintained for some tested problems if the distribution of non-uniform grid points is properly chosen.     

Large Eddy Simulation of an Internal Combustion Engine Using an Efficient Immersed Boundary Technique

This paper presents highly resolved large eddy simulations (LES) of an internal combustion engine (ICE) using an immersed boundary method (IBM), which can describe moving and stationary boundaries in a simple and efficient manner. In this novel approach, the motion of the valves and the piston is modeled by Lagrangian particles, whilst the stationary parts of the engine are described by a computationally efficient IBM. The proposed mesh-free technique of boundary representation is simple for parallelization and suitable for high performance computing (HPC). To demonstrate the method, LES results are presented for the flow and the combustion in an internal combustion engine. The Favre-filtered Navier-Stokes equations are solved for a compressible flow employing a finite volume method on Cartesian grids. Non-reflecting boundary conditions are applied at the intake and the exhaust ports. Combustion is described using a flame surface density (FSD) model with an algebraic reaction rate closure. A simplified engine with a fixed axisymmetric valve (see Appendix A) is employed to show the correctness of the method while avoiding the uncertainties which may be induced by the complex engine geometry. Three test-cases using a real engine geometry are investigated on different grids to evaluate the impact of the cell size and the filter width. The simulation results are compared against the experimental data. A good overall agreement was found between the measurements and the simulation data. The presented method has particular advantages in the efficient generation of the grid, high resolution and low numerical dissipation throughout the domain and an excellent suitability for massively parallel simulations.     

Reordering and incomplete preconditioning in serial and parallel adaptive mesh refinement and coarsening flow solutions

The effects of reordering the unknowns on the convergence of incomplete factorization preconditioned Krylov subspace methods are investigated. Of particular interest is the resulting preconditioned iterative solver behavior when adaptive mesh refinement and coarsening (AMR/C) are utilized for serial or distributed parallel simulations. As representative schemes, we consider the familiar reverse Cuthill–McKee and quotient minimum degree algorithms applied with incomplete factorization preconditioners to CG and GMRES solvers. In the parallel distributed case, reordering is applied to local subdomains for block ILU preconditioning, and subdomains are repartitioned dynamically as mesh adaptation proceeds. Numerical studies for representative applications are conducted using the object‐oriented AMR/C software system libMesh linked to the PETSc solver library. Serial tests demonstrate that global unknown reordering and incomplete factorization preconditioning can reduce the number of iterations and improve serial CPU time in AMR/C computations. Parallel experiments indicate that local reordering for subdomain block preconditioning associated with dynamic repartitioning because of AMR/C leads to an overall reduction in processing time. Copyright © 2011 John Wiley & Sons, Ltd.     

Boiling flow simulations on adaptive octree grids

Boiling flow simulations are conducted on adaptive octree grids. A phase change model consistent with the mixture formulation, in conjunction with the Volume-of-Fluid (VOF) model, is used to track the liquid–vapor interface. Test cases including Rayleigh Taylor instability and bubble growth in a uniform superheat are conducted to validate the phase change model on adaptive grids. The validated model is then used to conduct film boiling simulations on both two-dimensional and three-dimensional adaptive grids. The average wall Nusselt number agrees well with the widely accepted correlations of Berenson (1961) and Klimenko (1981) and Klimenko and Shelepen (1982) for film boiling on a horizontal surface. For the test cases presented, the efficiency of the adaptive technique, as measured by the adaptive mesh refinement (AMR) efficiency, is mostly in the range of 50–80%. Although this efficiency is a function of the nature and dimensionality of the problem, this range of efficiency is comparable to those obtained in the simulations of primary jet atomization conducted by Fuster et al. (2009). This work opens the prospect of conducting more realistic (three-dimensional) multi-modal boiling flow simulations, and problems of similar complexity, in an efficient manner.     

Parallel DNS with Local Grid Refinement

A parallel finite volume method for unstructured grids is used for a direct numerical simulation of the flow around a sphere at Re = 5000 (based on the sphere diameter and undisturbed velocity). The observed flow structures are confirmed by visualization experiments. A quantitative analysis of the Reynolds averaged flow provides a data base for future model evaluations.     

Parallel Implementation of DSMC Using Unstructured Mesh

The parallel implementation of the Direct Simulation Monte Carlo (DSMC) method on memory-distributed machines using unstructured mesh is reported. Physical domain decomposition is used to distribute the workload among multiple processors. A high-speed driven cavity flow is used as the benchmark problem for the validation of the parallel implementation. Three static partitioning techniques including simple coordinate partitioning, two-step partitioning (JOSTLE) and multi-level partitioning (METIS) are used for static domain decomposition, respectively. A cell renumbering technique is used to improve the memory management efficiency. Results of parallel efficiency show that two-step partitioning using JOSTLE performs the best, with 63% up to 25 processors, due to better load balancing among the processors. The powerful computational capability of the parallel implementation is demonstrated by computing a 2-D, near-continuum, hypersonic flow over a cylinder as well as a 3-D hypersonic flow over a sphere, respectively, both using 25 processors. Results compare reasonably well with previous simulated and experimental studies.     

An evaluation of two classes of shock hydrodynamic codes

The accuracy of four industrial shock hydrodynamics codes for blast environments in baffled systems is evaluated based on the shadowgraph data of Reichenbach and Kuhl (1992,3). Both problems involve a planar shock passing through a baffled channel. The numerical methods employed in these codes are representative of two classes, namely, the set of high-resolution schemes advanced in the 1980's, and the classical finite-difference schemes from the late 1960's. The four codes are: (1) the AMR code based on the higher-order Godunov scheme with adaptive grids, (2) the FEM-FCT code based on the flux-corrected transport scheme with unstructured grids, (3) and (4) the finite-difference based HULL and SHARC codes with fixed grids. From the comparisons of these calculations it is concluded that the high-resolution schemes: (1) calculate sharper shocks and sharper density profiles across vortices, (2) predict shear layer rollup forming coherent structures in the spiral vortices immediately downstream of every baffle, and (3) predict development of inviscid instabilities from these shear layers that, upon interaction with the reverberating shocks in the system, quickly become ‘turbulent’. The finite-difference codes predict essentially laminar behavior for the shear layers. Comparisons with shadowgraph data suggest that both classes of codes are able to predict shock reflections and diffractions in the baffled systems. The high-resolution codes give better agreement in the spiral vortices and the shear layers. As expected, turbulent flow features involving highly dissipative flow fields are not predicted by the high-resolution codes. Received March 5, 1995 / Accepted June 20, 1995     

Execution of a parallel edge-based Navier–Stokes solver on commodity graphics processor units

The implementation of an edge-based three-dimensional Reynolds Average Navier–Stokes solver for unstructured grids able to run on multiple graphics processing units (GPUs) is presented. Loops over edges, which are the most time-consuming part of the solver, have been written to exploit the massively parallel capabilities of GPUs. Non-blocking communications between parallel processes and between the GPU and the central processor unit (CPU) have been used to enhance code scalability. The code is written using a mixture of C++ and OpenCL, to allow the execution of the source code on GPUs. The Message Passage Interface (MPI) library is used to allow the parallel execution of the solver on multiple GPUs. A comparative study of the solver parallel performance is carried out using a cluster of CPUs and another of GPUs. It is shown that a single GPU is up to 64 times faster than a single CPU core. The parallel scalability of the solver is mainly degraded due to the loss of computing efficiency of the GPU when the size of the case decreases. However, for large enough grid sizes, the scalability is strongly improved. A cluster featuring commodity GPUs and a high bandwidth network is ten times less costly and consumes 33% less energy than a CPU-based cluster with an equivalent computational power.     

Parallel SOR methods with a parabolic-diffusion acceleration technique for solving an unstructured-grid Poisson equation on 3D arbitrary geometries

This paper presents a parallel algorithm for the finite-volume discretisation of the Poisson equation on three-dimensional arbitrary geometries. The proposed method is formulated by using a 2D horizontal block domain decomposition and interprocessor data communication techniques with message passing interface. The horizontal unstructured-grid cells are reordered according to the neighbouring relations and decomposed into blocks using a load-balanced distribution to give all processors an equal amount of elements. In this algorithm, two parallel successive over-relaxation methods are presented: a multi-colour ordering technique for unstructured grids based on distributed memory and a block method using reordering index following similar ideas of the partitioning for structured grids. In all cases, the parallel algorithms are implemented with a combination of an acceleration iterative solver. This solver is based on a parabolic-diffusion equation introduced to obtain faster solutions of the linear systems arising from the discretisation. Numerical results are given to evaluate the performances of the methods showing speedups better than linear.