基于机群的湍流燃烧数值模拟平台的构建与应用

首先对机群系统的特点和软硬件配置环境进行阐述,然后基于MPI和开源工具OpenFOAM实现一个计算流体力学机群系统的搭建,并对这个机群系统的性能和并行效率进行测试和分析。通过利用此机群系统对湍流燃烧流动反应进行数值模拟及并行计算得到模拟效果的三维流场。实验结果表明,该机群系统针对湍流燃烧数值模拟可以明显提高计算效率并得到较好的仿真效果。     

Modelling of wind load on single and staggered dual buildings

A method of modelling numerically the wind loads on single and staggered dual buildings using Computational Fluid Dynamics is presented in this paper. Simulation of a turbulent boundary layer over test models was carried out at the Supercomputing and Visualization Unit, the National University of Singapore, using FLUENT 6.1.18. Turbulence was introduced at the inlet through a parallel auxiliary simulation and the computation of the flow advanced in time using Large Eddy Simulation with a ReNormalization Group subgrid-scale viscosity model. Wind velocities at different locations and wind pressures on the building faces were recorded. Subsequently the flow characteristics were examined and the force and moment spectra deduced. The results were compared with data from earlier wind tunnel experiments carried out at Virginia Polytechnic Institute and State University. It can be concluded from the study that numerical wind modelling on tall structures is a promising alternative to conventional tests in atmospheric boundary layer wind tunnels.     

Large and Small Eddies Matter: Animating Trees in Wind Using Coarse Fluid Simulation and Synthetic Turbulence

Animating trees in wind has long been a problem in computer graphics. Progress on this problem is important for both visual effects in films and forestry biomechanics. More generally, progress on tree motion in wind may inform future work on two‐way coupling between turbulent flows and deformable objects. Synthetic turbulence added to a coarse fluid simulation produces convincing animations of turbulent flows but two‐way coupling between the enriched flow and objects embedded in the flow has not been investigated. Prior work on two‐way coupling between fluid and deformable models lacks a subgrid resolution turbulence model. We produce realistic animations of tree motion by including motion due to both large and small eddies using synthetic subgrid turbulence and porous proxy geometry. Synthetic turbulence at the subgrid scale is modulated using turbulent kinetic energy (TKE). Adding noise after sampling the mean flow and TKE transfers energy from small eddies directly to the tree geometry. The resulting animations include both global sheltering effects and small scale leaf and branch motion. Viewers, on average, found animations, which included both coarse fluid simulation and TKE‐modulated noise to be more accurate than animations generated using coarse fluid simulation or noise alone.     

Pseudospectral simulation of turbulent viscoelastic channel flow

The methodology and validation of direct numerical simulations of viscoelastic turbulent channel flow are presented here. Using differential constitutive models derived from kinetic and network theories, numerical simulations have demonstrated drag reduction for various values of the parameters, under conditions where there is a substantial increase in the extensional viscosity compared to the shear viscosity (Sureshkumar, Beris, Handler, Direct numerical simulation of turbulent channel flow of a polymer solution, Phys. Fluids 9 (1997) 743–755 and Dimitropoulos, Sureshkumar, Beris, Direct numerical simulation of viscoelastic turbulent channel flow exhibiting drag reduction: effect of the variation of rheological parameters, J. Non-Newtonian Fluid Mech. 79 (1998) 433–468). In this work, new results pertaining to the Reynolds stress and the pressure are presented, and the convergence of the pseudospectral algorithm utilized in the simulations, as well as its parallel implementation, are discussed in detail. It is shown that the lack of mesh refinement, or the use of a larger value for the artificial stress diffusivity used to stabilize the conformation tensor evolution equations, introduce small quantitative errors which qualitatively have the effect of lowering the drag reduction capability of the simulated fluid. However, an insufficient size of the periodic computational domain can also introduce errors in certain cases, which albeit usually small, can qualitatively alter various features of the solution.     

SPH-GPU并行计算在风沙流中的应用

为了实现小尺度范围风沙运动的真实感模拟,采用基于拉格朗日力学无网格形式的光滑粒子流体动力学（smooth particle hydrodynamics,SPH）方法解决了基于欧拉网格法因网格大变形或者变形边界等引起的各种问题,并克服了不能用固定欧拉网格追踪任意单颗粒子运动轨迹的困难,因此该方法在研究风沙运动方面有着独特的优势。然而,随着风沙流动中SPH粒子数目的增加,该方法计算效率低,计算规模大的缺陷在风沙模拟过程中尤为明显。为了提高其计算效率,在CUDA软硬件平台上,建立SPH-GPU并行加速的二维气沙两相耦合模型,对串行的热点程序进行分析,找出最耗时且适合并行的热点程序;其次对GPU并行计算模型进行验证,宏观上得到了沙粒群运动的时空变化规律,微观上得到了典型沙粒的跃移轨迹和变异的尖角轨迹;最后对比了三种不同粒子数下CPU与GPU的计算效率。模拟结果证明SPH-GPU并行计算方法能够进一步应用在风沙流的数值模拟研究中。     

Recent Developments in Variational Multiscale Methods for Large-Eddy Simulation of Turbulent Flow

The variational multiscale method is reviewed as a framework for developing computational methods for large-eddy simulation of turbulent flow. In contrast to other articles reviewing this topic, which focused on large-eddy simulation of turbulent incompressible flow, this study covers further aspects of numerically simulating turbulent flow as well as applications beyond incompressible single-phase flow. The various concepts for subgrid-scale modeling within the variational multiscale method for large-eddy simulation proposed by researchers in this field to date are illustrated. These conceptions comprise (i) implicit large-eddy simulation, represented by residual-based and stabilized methods, (ii) functional subgrid-scale modeling via small-scale subgrid-viscosity models and (iii) structural subgrid-scale modeling via the introduction of multifractal subgrid scales. An overview on exemplary numerical test cases to which the reviewed methods have been applied in the past years is provided, including explicit computational results obtained from turbulent channel flow. Wall-layer modeling, passive and active scalar transport as well as developments for large-eddy simulation of turbulent two-phase flow and combustion are discussed to complete this exposition.     

风机流场并行数值模拟在集群系统上的应用

介绍集群系统的特点及其软硬件配置环境。利用计算流体力学商用软件Fluent对一种新型垂直轴风力机风轮在曙光天潮超级服务器4000A集群上进行数值模拟及并行计算,得到风轮内外部的三维流场,实现该软件在高性能并行计算机上的应用,为并行有限元分析提供一个基础计算平台。     

槽流拟颗粒模型的并行算法

将流体处理为离散粒子,应用拟颗粒硬球模型来研究槽流中的流动现象,与分子动力学模拟的算法类似,是研究槽流机理的一种行之有效的方法。为了作大规模的模拟,本文采用区域分解算法和消息传递编程模型技术,将该模型串行程序并行化,应用一维划分、单相传递的方法简化了并行算法,采用轮换搜索法来避免硬球碰撞次序对结果的影响。在可扩展的机群系统上用实例计算,通过与串行程序的对比,验证了并行程序的正确性,表明本文设计的并行算法取得了较高的并行计算效率。     

Simulation of flowfields induced by wind blades based on a parallelized low-speed flow solver

In this study, a parallel computing technology is applied on the simulation of a wind turbine flow problem. A third-order Roe type flux limited splitting based on a pre-conditioning matrix with an explicit time marching method is used to solve the Navier–Stokes equations. The original FORTRAN code was parallelized with Message Passing Interface (MPI) language and tested on a 64-CPU IBM SP2 parallel computer. The test results show that a significant reduction of computing time in running the model and a super-linear speed up rate is achieved up to 32 CPUs at IBM SP2 processors. The speed up rate is as high as 49 for using IBM SP2 64 processors. The test shows very promising potential of parallel processing to provide prompt simulation of the current wind turbine problems.     

Parallel preconditioned WENO scheme for three-dimensional flow simulation of NREL Phase VI Rotor

The preconditioned weighted essentially non-oscillatory (P-WENO) solver for viscous flows (Huang et al. (2009) [9]) is extended to non-inertial reference frames. In the present scheme, patched multi-block grid system is employed and parallel computing is adopted as well. With the present parallel P-WENO solver, three-dimensional flows of the Phase VI Rotor from National Renewable Energy Laboratory (NREL) can be simulated and analyzed for different wind speeds. Our simulation results show good agreement with the numerical predictions based on incompressible Navier–Stokes (N–S) equations as well as the available wind tunnel data from NREL. The flow phenomena, including separation and attachment line, can be captured by the present scheme. The parallel strategy adopted is a block-domain decomposition method for the patched multi-block grid system. To balance the load among different computing nodes, a Tabu search algorithm is adopted for the parallelization. The parallel efficiency of the parallel P-WENO scheme is examined for node numbers ranging from 1 to 64. It is found that the parallel efficiency is monotonically decreased as the node number adopted is increased; the parallel efficiency is retained over 90% for all cases of different node numbers. Due to the high parallel efficiency, our parallel P-WENO solver is validated for applying to practical fluid problems from compressible to incompressible limits.     

Lattice Boltzmann large eddy simulation of subcritical flows around a sphere on non-uniform grids

In this work, the suitability of the lattice Boltzmann method is evaluated for the simulation of subcritical turbulent flows around a sphere. Special measures are taken to reduce the computational cost without sacrificing the accuracy of the method. A large eddy simulation turbulence model is employed to allow efficient simulation of resolved flow structures on non-uniform computational meshes. In the vicinity of solid walls, where the flow is governed by the presence of a thin boundary layer, local grid-refinement is employed in order to capture the fine structures of the flow. In the test case considered, reference values for the drag force in the Reynolds number range from 2000 to 10 000 and for the surface pressure distribution and the angle of separation at a Reynolds number of 10 000 could be quantitatively reproduced. A parallel efficiency of 80% was obtained on an Opteron cluster.     

On a mechanism of stabilizing turbulent free shear layers in cavity flows

Turbulent free shear flows are subject to the well-known Kelvin–Helmholtz type [Panton RL. Incompressible flow. John Wiley and Sons; 1984. p. 675] instability, and it is well-known that any free shear flow which approximates a thin vorticity layer will be unstable to a wide range of amplitudes and frequencies of disturbance. In fact, much of what constitutes flow control in turbulent free shear layers consists of feeding a prescribed destabilizing disturbance to these layers. The question in the control of free shear flows is not whether the shear layer will be stable, but whether you can influence how the layer becomes unstable. In most cases, since these flows are so receptive to forcing input, and naturally tend toward instability, large changes in flow conditions can be achieved with very small amplitude periodic inputs.

Recently, it has been discovered that turbulent free shear flows can also be stabilized using periodic forcing. This is, at first glance, counter-intuitive, considering our long history of considering these flows to be very unstable to forcing input. It is a phenomenon not described in modern fluid dynamic text books. The forcing required to achieve this effect (which we will call turbulent shear layer stabilization) is of a much higher amplitude and frequency than the more traditional type of shear layer flow control effect seen in the literature (which we will call turbulent shear layer destabilization).

A numerical study is undertaken to investigate the effect of frequency of pulsed mass injection on the nature of stabilization, destabilization and acoustic suppression in high speed cavity flows. An implicit, 2nd-order in space and time flow solver, coupled with a recently developed hybrid RANS-LES (Reynolds Averaged Navier Stokes-Large Eddy Simulation) turbulence model by Nichols and Nelson [Nichols RH, Nelson CC. Weapons bay acoustic predictions using a multi-scale turbulence model. In: Proceedings of the ITEA 2001 aircraft-stores compatibility symposium, March 2001], is utilized in a Chimera-based parallel format. This tool is used to numerically simulate both an unsuppressed cavity in resonance, as well as the effect of mass-addition pulsed jet flow control on cavity flow physics and ultimately, cavity acoustic levels.

Frequency (and in a limited number of cases, amplitude) of pulse is varied, from 0 Hz (steady) up to 5000 Hz. The change in the character of the flow control effect as pulsing frequency is changed is described, and linked to changes in acoustic levels. Limited comparison to 1/10th scale experiments is presented.

The observed local stabilization of the cavity turbulent shear layer, when subjected to high frequency pulsed blowing, is shown in simulation to be the result of a violent instability and breakdown of a pair of opposite sign vortical structures created with each high frequency “pulse”. This unique shear layer stabilization behavior is only observed in simulation above a certain critical pulsing frequency. Below this critical frequency, pulsing is shown in simulation to provide little benefit with respect to suppression of high cavity acoustic levels.     

Parallelization and scalability of a spectral element channel flow solver for incompressible Navier–Stokes equations

Direct numerical simulation (DNS) of turbulent flows is widely recognized to demand fine spatial meshes, small timesteps, and very long runtimes to properly resolve the flow field. To overcome these limitations, most DNS is performed on supercomputing machines. With the rapid development of terascale (and, eventually, petascale) computing on thousands of processors, it has become imperative to consider the development of DNS algorithms and parallelization methods that are capable of fully exploiting these massively parallel machines. A highly parallelizable algorithm for the simulation of turbulent channel flow that allows for efficient scaling on several thousand processors is presented. A model that accurately predicts the performance of the algorithm is developed and compared with experimental data. The results demonstrate that the proposed numerical algorithm is capable of scaling well on petascale computing machines and thus will allow for the development and analysis of high Reynolds number channel flows. Copyright © 2007 John Wiley & Sons, Ltd.     

Large-eddy simulation in a complex hill terrain enabled by a compact fractional step OpenFOAM® solver

We report computational fluid dynamics (CFD) code developments using the high-level programming syntax of the open source C++ library OpenFOAM®. CFD simulations utilizing the large-eddy simulation (LES) approach are carried out using the developed code in a real-world application. We investigate wind flowing over the Bolund hill, Denmark. In the present configuration a west–east wind meets the steep west side of the hill. Such conditions lead to flow separation at the location of a sharp cliff. A full scale simulation, with a simulation duration of over one month, is carried out on a supercomputer. Physically, about 45 min of real time is simulated in the LES enabling the statistical averaging of the results. The novelty of the paper consists of the following features: (1) we report validation results of the newly developed LES code for the Bolund hill case, (2) we show the high-level LES solver code in its entirety in a few tens of code lines which promotes transparency in CFD-code development in the OpenFOAM® environment, (3) the study is the first study to use LES in pointing out the complex 3d characteristics of the Bolund hill case with the computationally challenging west–east (270°) wind direction, and (4) based on the comparison with previous experimental data, and Reynolds averaged Navier–Stokes (RANS) simulations, the present LES gives so far the best match for the turbulent kinetic energy increase at the considered measurement positions.     

A parallel, finite-volume algorithm for large-eddy simulation of turbulent flows

A parallel, finite-volume algorithm has been developed for large-eddy simulation (LES) of compressible turbulent flows. This algorithm includes piecewise linear least-square reconstruction, trilinear finite-element interpolation, Roe flux-difference splitting (FDS), and second-order MacCormack time marching. A systematic and consistent means of evaluating the surface and volume integrals of the control volume is described. Parallel implementation is done using the message-passing programming model. To validate the numerical method for turbulence simulation, LES of fully developed turbulent flow in a square duct is performed for a Reynolds number of 320 based on the average friction velocity and the hydraulic diameter of the duct. Direct numerical simulation (DNS) results are available for this test case, and the accuracy of this algorithm for turbulence simulations can be ascertained by comparing the LES solutions with the DNS results. For the first time, a finite volume method with Roe FDS was used for LES of turbulent flow in a square duct, and the effects of grid resolution, upwind numerical dissipation, and subgrid-scale dissipation on the accuracy of the LES are examined. Comparison with DNS results shows that the standard Roe FDS adversely affects the accuracy of the turbulence simulation. For accurate turbulence simulations, only 3–5% of the standard Roe FDS dissipation is needed.     

A parallel implementation of an MHD code for the simulation of mechanically driven, turbulent dynamos in spherical geometry

A parallel implementation of a nonlinear pseudo-spectral MHD code for the simulation of turbulent dynamos in spherical geometry is reported. It employs a dual domain decomposition technique in both real and spectral space. It is shown that this method shows nearly ideal scaling going up to 128 CPUs on Beowulf-type clusters with fast interconnect. Furthermore, the potential of exploiting single precision arithmetic on standard x86 processors is examined. It is pointed out that the MHD code thereby achieves a maximum speedup of 1.7, whereas the validity of the computations is still granted. The combination of both measures will allow for the direct numerical simulation of highly turbulent cases (1500<Re<5000), which have been previously impractical, given today's computational speed.     

脉动风场中的飘雪模拟

飘雪场景的模拟可以大大提高虚拟场景的逼真效果,其中风场的建模是重要的环节.为更真实的表现风的随机性,本文基于格子BGK建立风场,用D3Q7模型改进三维风场的建模方法,引入湍流强度的概念描述风场的湍流程度,根据地表面的粗糙程度和离地高度确定其大小,实验表明该方法能更好的增强模拟风场中飘雪场景的真实性.     

Numerical modeling of three-dimensional flow and pollutant dispersion around structures

This paper describes a numerical modeling approach that can be used to provide estimates of emissions at industrial sites. In particular the models presented are capable of simulating the wind flow and dispersion of airborne pollutants around surface-mounted structures such as buildings or building complexes. The calculational procedure in this approach consists of two sequential steps, namely: (i) prediction of the mean flow via a turbulent flow model; and (ii) employment of the calculated flow field to drive a particle-in-cell transport and diffusion model. A benchmark simulation is performed in which numerical results from the flow model are compared with other numerical models and with experimental data for flow over a backward-facing step. Results from three-dimensional simulations of flow and dispersion over a two-building complex are also presented.     

A Parallel Overset Grid High-Order Flow Solver for Large Eddy Simulation

This work describes the development and validation of a parallel high-order compact finite difference Navier–Stokes solver for application to large-eddy simulation (LES) and direct numerical simulation. The implicit solver can employ up to sixth-order spatial formulations and tenth-order filtering. The parallelization of the solver is founded on the overset grid technique. LES were then performed for turbulent channel flow with Reynolds numbers ranging from Re _τ=180 to 590, and flow past a circular cylinder with a transitional wake at Re _D =3900. The channel flow solutions were obtained using both an implicit LES (ILES) approach and a dynamic sub-grid scale model. The ILES method obtained virtually identical solutions at half the computational cost. The original vector and new parallel solvers produce indistinguishable mean flow solutions for the circular cylinder. Repeating the cylinder simulation on a much finer mesh resulted in significantly better agreement with experimental data in the near wake than the coarse grid solution and other previous numerical studies.     

“天河一号”大规模并行应用程序测试

目前安装在国家超级计算天津中心的"天河一号"是我国首台千万亿次超级计算机,在2010年11月世界超级计算机Top500排名中,位列世界第一。"天河一号"采用了CPU与GPU相结合的异构融合计算体系结构,自主设计实现了高速互连通信系统,在多个高性能计算应用领域具有应用适应性强、系统运行稳定可靠、性能可扩展性好等特点,为科学研究和应用提供了重要的高性能计算平台。采用石油地震数据处理、飞行器流场模拟、生物分子动力学模拟、磁约束聚变数值模拟、湍流数值模拟、晶体硅分子动力学模拟、全球大气浅水波全隐式数值模拟、地球外核热流动数值模拟等应用的典型算例对"天河一号"进行了大规模并行程序测试,结果表明,"天河一号"在上述应用领域具有良好的可扩展性和并行效率。