A domain decomposition approach to finite volume solutions of the Euler equations on unstructured triangular meshes

We report on our recent efforts on the formulation and the evaluation of a domain decomposition algorithm for the parallel solution of two‐dimensional compressible inviscid flows. The starting point is a flow solver for the Euler equations, which is based on a mixed finite element/finite volume formulation on unstructured triangular meshes. Time integration of the resulting semi‐discrete equations is obtained using a linearized backward Euler implicit scheme. As a result, each pseudo‐time step requires the solution of a sparse linear system for the flow variables. In this study, a non‐overlapping domain decomposition algorithm is used for advancing the solution at each implicit time step. First, we formulate an additive Schwarz algorithm using appropriate matching conditions at the subdomain interfaces. In accordance with the hyperbolic nature of the Euler equations, these transmission conditions are Dirichlet conditions for the characteristic variables corresponding to incoming waves. Then, we introduce interface operators that allow us to express the domain decomposition algorithm as a Richardson‐type iteration on the interface unknowns. Algebraically speaking, the Schwarz algorithm is equivalent to a Jacobi iteration applied to a linear system whose matrix has a block structure. A substructuring technique can be applied to this matrix in order to obtain a fully implicit scheme in terms of interface unknowns. In our approach, the interface unknowns are numerical (normal) fluxes. Copyright © 2001 John Wiley & Sons, Ltd.     

A parallel implicit domain decomposition algorithm for the large eddy simulation of incompressible turbulent flows on 3D unstructured meshes

We present a parallel fully implicit algorithm for the large eddy simulation (LES) of incompressible turbulent flows on unstructured meshes in three dimensions. The LES governing equations are discretized by a stabilized Galerkin finite element method in space and an implicit second-order backward differentiation scheme in time. To efficiently solve the resulting large nonlinear systems, we present a highly parallel Newton-Krylov-Schwarz algorithm based on domain decomposition techniques. Analytic Jacobian is applied in order to obtain the best achievable performance. Two benchmark problems of lid-driven cavity and flow passing a square cylinder are employed to validate the proposed algorithm. We then apply the algorithm to the LES of turbulent flows passing a full-size high-speed train with realistic geometry and operating conditions. The numerical results show that the algorithm is both accurate and efficient and exhibits a good scalability and parallel efficiency with tens of millions of degrees of freedom on a computer with up to 4096 processors. To understand the numerical behavior of the proposed fully implicit scheme, we study several important issues, including the choices of linear solvers, the overlapping size of the subdomains, and, especially, the accuracy of the Jacobian matrix. The results show that an exact Jacobian is necessary for the efficiency and the robustness of the proposed LES solver.     

A Hybrid Domain Decomposition and Multigrid Method for the Acceleration of Compressible Viscous Flow Calculations on Unstructured Triangular Meshes

This paper is concerned with the formulation and the evaluation of a hybrid solution method that makes use of domain decomposition and multigrid principles for the calculation of two-dimensional compressible viscous flows on unstructured triangular meshes. More precisely, a non-overlapping additive domain decomposition method is used to coordinate concurrent subdomain solutions with a multigrid method. This hybrid method is developed in the context of a flow solver for the Navier-Stokes equations which is based on a combined finite element/finite volume formulation on unstructured triangular meshes. Time integration of the resulting semi-discrete equations is performed using a linearized backward Euler implicit scheme. As a result, each pseudo time step requires the solution of a sparse linear system. In this study, a non-overlapping domain decomposition algorithm is used for advancing the solution at each implicit time step. Algebraically, the Schwarz algorithm is equivalent to a Jacobi iteration on a linear system whose matrix has a block structure. A substructuring technique can be applied to this matrix in order to obtain a fully implicit scheme in terms of interface unknowns. In the present approach, the interface unknowns are numerical fluxes. The interface system is solved by means of a full GMRES method. Here, the local system solves that are induced by matrix-vector products with the interface operator, are performed using a multigrid by volume agglomeration method. The resulting hybrid domain decomposition and multigrid solver is applied to the computation of several steady flows around a geometry of NACA0012 airfoil.     

求解三维流固耦合问题的一种全隐全耦合区域分解并行算法

三维流固耦合问题的非结构网格数值算法在很多工程领域都有重要应用, 目前现有的数值方法主要基于分区算法, 即流体和固体区域分别进行求解, 因此存在收敛速度较慢以及附加质量导致的稳定性问题, 此外, 该类算法的并行可扩展性不高, 在大规模应用计算方面也受到一定限制．本文针对三维非定常流固耦合问题, 提出一种基于区域分解的全隐全耦合可扩展并行算法．首先基于任意拉格朗日?欧拉框架建立流固耦合控制方程, 然后时间方向采用二阶向后差分隐式格式、空间方向采用非结构稳定化有限元方法进行离散．对于大规模非线性离散系统, 构造一种结合非精确Newton法、Krylov子空间迭代法与区域分解Schwarz预条件子的Newton-Krylov-Schwarz (NKS) 并行求解算法, 实现流体、固体和动网格方程的一次性整体求解．采用弹性障碍物绕流的标准测试算例对数值方法的准确性进行了验证, 数值性能测试结果显示本文构造的全隐全耦合算法具有良好的稳定性, 在不同的物理参数下具有良好的鲁棒性, 在“天河二号”超级计算机上, 当并行规模从192增加到3072个处理器核时获得了91%的并行效率．性能测试结果表明本文构造的NKS算法有望应用于复杂区域流固耦合问题的大规模数值模拟研究中．      

????????????????????????????

合隐式和显式时间积分技术,对结构非线性动力反应分析提出一种并行混合时间积分算 法. 该算法采用区域分解技术. 将并发性引入到算法中,即利用显式时间积分技术进行界面 节点积分而利用隐式算法求解局部子区域. 为实现并行混合时间积分算法,设计了灵活的并 行数据信息流. 编写了该算法的程序,在工作站机群实现了数值算例,验证了算法的精度和 性能. 计算结果表明该算法具有良好的并行性能,优于隐式算法.     

Computation of fluid flow with a parallel multigrid solver

A finite volume numerical method for the prediction of fluid flow and heat transfer in simple geometries was parallelized using a domain decomposition approach. The method is implicit, uses a colocated arrangement of variables and is based on the SIMPLE algorithm for pressure-velocity coupling. Discretization is based on second-order central difference approximations. The algebraic equation systems are solved by the ILU method of Stone.¹ To accelerate the convergence, a multigrid technique was used. The efficiency was examined on three different parallel computers for laminar flow in a pipe with an orifice and natural convection in a closed cavity. It is shown that the total efficiency is made up of three major factors: numerical efficiency, parallel efficiency and load-balancing efficiency. The first two factors were thoroughly investigated, and a model for predicting the parallel efficiency on various computers is presented. Test calculations indicate reasonable total efficiency and favourable dependence on grid size and the number of processors.     

大规模并行结构动力分析分层计算方法

多核分布式存储超级计算机的兴起为大规模并行结构动力分析提供了强有力的计算工具。根据多核分布式计算环境的特点,提出了一种大规模并行结构动力分析分层计算方法。该方法在传统隐式动力分析的区域分解法的基础上,利用两级分区和两次缩聚策略进行求解。不但通过进一步缩减求解问题规模有效提高了界面方程的收敛速度,而且通过三层并行计算有效提高了通信效率。该方法并不对有限元模型引入近似,属于精确的动力子区域分层计算方法。典型数值算例表明,该方法计算精度与商业软件ANSYS完全法求解精度相当;同传统区域分解法相比,该方法能够获得较高的并行计算性能。     

Efficient implicit algorithm for the equations of 2-D viscous compressible flow: application to shock-boundary layer interaction

A fully implicit algorithm has been developed to time integrate the equations of 2-D compressible viscous flow. The algorithm was constructed so as to optimize computational efficiency. The time-consuming block matrix inversions usually associated with implicit algorithms have been reduced to the trivial non-iterative inversion of four sets of scalar bidiagonal matrices. Thus, the algorithm requires virtually no more computer storage than an explicit algorithm. The efficient structure of the implicit algorithm is reflected in comparative timings which slow that it requires only a factor of two more computer time per point per time step than a typical explicit algorithm. Therefore, the algorithm allows more economical solution of given flows than existing explicit methods and also allows more difficult problems to be attempted using available computer resources. Application of the algorithm to the problem of shock-boundary layer interaction produces results consistent with both experimental measurements and other calculations.     

Non-periodic finite-element formulation of Kohn-Sham density functional theory

We present a real-space, non-periodic, finite-element formulation for Kohn-Sham density functional theory (KS-DFT). We transform the original variational problem into a local saddle-point problem, and show its well-posedness by proving the existence of minimizers. Further, we prove the convergence of finite-element approximations including numerical quadratures. Based on domain decomposition, we develop a parallel finite-element implementation of this formulation capable of performing both all-electron and pseudopotential calculations. We assess the accuracy of the formulation through selected test cases and demonstrate good agreement with the literature. We also evaluate the numerical performance of the implementation with regard to its scalability and convergence rates. We view this work as a step towards developing a method that can accurately study defects like vacancies, dislocations and crack tips using density functional theory (DFT) at reasonable computational cost by retaining electronic resolution where it is necessary and seamlessly coarse-graining far away.     

A comparison of coarse and fine grain parallelization strategies for the simple pressure correction algorithm

The primary aim of this work was to determine the simplest and most effective parallelization strategy for control-volume-based codes solving industrial problems. It has been found that for certain classes of problems, the coarse-grain functional decomposition strategy, largely ignored due to its limited scaling capability, offers the potential for significant execution speed-ups while maintaining the inherent structure of traditional serial algorithms. Functional decomposition requires only minor modification of the existing serial code to implement and, hence, code portability across both concurrent and serial computers is maintained. Fine-grain parallelization strategies at the ‘DO loop’ level are also easy to implement and largely preserve code portability. Both coarse-grain functional decomposition and fine-grain loop-level parallelization strategies for the SIMPLE pressure correction algorithm are demonstrated on a Silicon Graphics 4D280S eight CPU shared memory computer system for a highly coupled, transient two-dimensional simulation involving melting of a metal in the presence of thermal-buoyancy-driven laminar convection. Problems requiring the solution of a larger number of transport equations were simulated by including further scalar variables in the calculation. While resulting in slight degradation of the convergence rate, the functional decomposition strategy exhibited higher parallel efficiencies and yielded greater speed-ups relative to the original serial code. Initially, this strategy showed a significant degradation in convergence rate due to an inconsistency in the parallel solution of the pressure correction equation. After correcting for this inconsistency, the maximum speed-up for 16 dependent variables was a factor of 5·28 with eight processors, representing a parallel efficiency of 67%. Peak efficiency of 76% was achieved using five processors to solve for 10 dependent variables.     

A fractional two‐step implicit algorithm using curvilinear collected grids

An efficient fractional two‐step implicit algorithm is reported to simulate incompressible fluid flows in a boundary‐fitted curvilinear collocated grid system. Using the finite volume method, the convection terms are discretized by the high‐accuracy Roe's scheme to minimize numerical diffusion. An implicitness coefficient Π is introduced to accelerate the rate of convergence. It is demonstrated that the proposed algorithm links the fractional step method to the pressure correction procedure, and the SIMPLEC method could be considered as a special case of the fractional two‐step implicit algorithm (when Π=1). The proposed algorithm is applicable to unsteady flows and steady flows. Three benchmark two‐dimensional laminar flows are tested to evaluate the performance of the proposed algorithm. Performance is measured by sensitivity analyses of the efficiency, accuracy, grid density, grid skewness and Reynolds number on the solutions. Results show that the model is efficient and robust. Copyright © 2006 John Wiley & Sons, Ltd.     

Parallel domain decomposition method for finite element approximation of 3D steady state non‐Newtonian fluids

We introduce a stabilized finite element method for the 3D non‐Newtonian Navier–Stokes equations and a parallel domain decomposition method for solving the sparse system of nonlinear equations arising from the discretization. Non‐Newtonian flow problems are, generally speaking, more challenging than Newtonian flows because the nonlinearities are not only in the convection term but also in the viscosity term, which depends on the shear rate. Many good iterative methods and preconditioning techniques that work well for the Newtonian flows do not work well for the non‐Newtonian flows. We employ a Galerkin/least squares finite element method, with stabilization parameters adjusted to count the non‐Newtonian effect, to discretize the equations, and the resulting highly nonlinear system of equations is solved by a Newton–Krylov–Schwarz algorithm. In this study, we apply the proposed method to some inelastic power‐law fluid flows through the eccentric annuli with inner cylinder rotation and investigate the robustness of the method with respect to some physical parameters, including the power‐law index and the Reynolds number ratios. We then report the superlinear speedup achieved by the domain decomposition algorithm on a computer with up to 512 processors. Copyright © 2015 John Wiley & Sons, Ltd.     

Neumann's method for boundary problems of thin elastic shells

The possibility of using Neumann's method to solve the boundary problems for thin elastic shells is studied. The variational statement of the static problems for the shells allows for a problem examination within the distribution space. The convergence of Neumann's method is proven for the shells with holes when the boundary of the domain is not completely fixed. The numerical implementation of Neumann's method normally requires significant time before any reliable results can be achieved. This paper suggests a way to improve the convergence of the process, and allows for parallel computing and evaluation during the calculations.     

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.     

The implementation of an aeronautical CFD flow code onto distributed memory parallel systems

The parallelization of an industrially important in‐house computational fluid dynamics (CFD) code for calculating the airflow over complex aircraft configurations using the Euler or Navier–Stokes equations is presented. The code discussed is the flow solver module of the SAUNA CFD suite. This suite uses a novel grid system that may include block‐structured hexahedral or pyramidal grids, unstructured tetrahedral grids or a hybrid combination of both. To assist in the rapid convergence to a solution, a number of convergence acceleration techniques are employed including implicit residual smoothing and a multigrid full approximation storage scheme (FAS). Key features of the parallelization approach are the use of domain decomposition and encapsulated message passing to enable the execution in parallel using a single programme multiple data (SPMD) paradigm. In the case where a hybrid grid is used, a unified grid partitioning scheme is employed to define the decomposition of the mesh. The parallel code has been tested using both structured and hybrid grids on a number of different distributed memory parallel systems and is now routinely used to perform industrial scale aeronautical simulations. Copyright © 2000 John Wiley & Sons, Ltd.     

二维自适应非结构网格DSMC并行算法研究

研究了二维自适应非结构网格DSMC并行算法实现的过程.首先提出了一类非结构网格自适应策略,有效降低了网格尺度对计算结果的影响,提高了流场的分辨率;然后基于PC-CLUSTER群机并行体系结构与消息传递库MPI并行环境,利用分区并行思想,设计了非结构网格DSMC并行算法,节约了计算时间.利用For-tran90的动态分配内存技术编制了通用计算程序;最后对过渡流域高超声绕流进行了数值模拟,计算结果初步验证了算法的可行性与有效性.     

Domain decomposition for the incompressible Navier–Stokes equations: solving subdomain problems accurately and inaccurately

For the solution of practical flow problems in arbitrarily shaped domains, simple Schwarz domain decomposition methods with minimal overlap are quite efficient, provided Krylov subspace methods, e.g. the GMRES method, are used to accelerate convergence. With an accurate subdomain solution, the amount of time spent solving these problems may be quite large. To reduce computing time, an inaccurate solution of subdomain problems is considered, which requires a GCR-based acceleration technique. Much emphasis is put on the multiplicative domain decomposition algorithm since we also want an algorithm which is fast on a single processor. Nevertheless, the prospects for parallel implementation are also investigated. © 1998 John Wiley & Sons, Ltd.     

An Adaptive Cartesian Mesh Based Method to Simulate Turbulent Flows of Multiple Rotating Surfaces

An efficient Cartesian cut-cell/level-set method based on a multiple grid approach to simulate turbulent turbomachinery flows is presented. The finite-volume approach in an unstructured hierarchical Cartesian setup with a sharp representation of the complex moving boundaries embedded into the computational domain, which are described by multiple level-sets, ensures a strict conservation of mass, momentum, and energy. Furthermore, an efficient kinematic motion level-set interface method for the rotation of embedded boundaries described by multiple level-set fields on a computational domain distributed over several processors is introduced. This method allows the simulation of multiple boundaries rotating relatively to each other in a fixed frame of reference. To demonstrate the efficiency of the numerical method and the quality of the computed findings the generic test problem of a rotating cylinder surrounded by a stationary hull and the flow over a ducted rotating axial fan with a stationary turbulence generating grid at the inflow are simulated. The computational results of the axial fan show a good agreement with the experimental data.     

A conservative unstructured scheme for rapidly varied flows

This article introduces a new semi‐implicit, staggered finite volume scheme on unstructured meshes for the modelling of rapidly varied shallow water flows. Rapidly varied flows occur in the inundation of dry land during flooding situations. They typically involve bores and hydraulic jumps after obstacles such as road banks. Near such sudden flow transitions, the grid resolution is often low compared with the gradients of the bathymetry. Locally the hydrostatic pressure assumption may become invalid. In these situations, it is crucial to apply the correct conservation properties to obtain accurate results. An important feature of this scheme is therefore its ability to conserve momentum locally or, by choice, preserve constant energy head along a streamline. This is achieved using a special interpolation method and control volumes for momentum. The efficiency of inundation calculations with locally very high velocities, and in the case of unstructured meshes locally very small grid distances, is severely hampered by the Courant condition. This article provides a solution in the form of a locally implicit time integration for the advective terms that allows for an explicit calculation in most of the domain, while maintaining unconditional stability by implicit calculations only where necessary. The complex geometry of flooded urban areas asks for the flexibility of unstructured meshes. The efficient calculation of the pressure gradient in this, and other semi‐implicit staggered schemes, requires, however, an orthogonality condition to be put on the grid. In this article a simple method is introduced to generate unstructured hybrid meshes that fulfil this requirement. Copyright © 2008 John Wiley & Sons, Ltd.     

Parallel implementation of a non‐hydrostatic model for free surface flows with semi‐Lagrangian advection treatment

The parallel implementation of an unstructured‐grid, three‐dimensional, semi‐implicit finite difference and finite volume model for the free surface Navier–Stokes equations (UnTRIM ) is presented and discussed. The new developments are aimed to make the code available for high‐performance computing in order to address larger, complex problems in environmental free surface flows. The parallelization is based on the mesh partitioning method and message passing and has been achieved without negatively affecting any of the advantageous properties of the serial code, such as its robustness, accuracy and efficiency. The key issue is a new, autonomous parallel streamline backtracking algorithm, which allows using semi‐Lagrangian methods in decomposed meshes without compromising the scalability of the code. The implementation has been carefully verified not only with simple, abstract test cases illustrating the application domain of the code but also with advanced, high‐resolution models presently applied for research and engineering projects. The scheme performance and accuracy aspects are researched and discussed. Copyright © 2008 John Wiley & Sons, Ltd.