U-RANS/PDF方法的并行计算研究

本文对非结构网格中U-RANS/PDF混合方法的并行计算进行了研究,通过对方柱绕流的数值模拟,讨论了并行程序的有效性和计算效率.非结构网格上U-RANS/PDF混合方法的并行计算,提高了PDF方法的计算效率和处理复杂几何结构的能力,使得PDF方法能够方便地应用于实际燃烧问题.     

基于JASMIN的地下水流大规模并行数值模拟

针对具有精细网格剖分、长时间跨度特征的地下水流模拟中计算时间长、存储开销大等瓶颈问题,基于MODFLOW三维非稳定流计算方法,提出基于网格片的核心算法以及基于影像区的通信机制,并在JASMIN框架上研制了大规模地下水流并行数值模拟程序JOGFLOW.通过河南郑州市中牟县雁鸣湖水源地地下水流的模拟,对程序正确性和性能进行了验证;通过建立一个具有精细网格剖分的假想地下水概念模型对可扩展性进行测试.相对于32核的并行程序,在512以及1 024个处理机上的并行效率分别可达77.2%和67.5%.数值模拟结果表明,JOGFLOW具有较好的计算性能与可扩展性,能够有效使用数百上千计算核心,支持千万量级以上网格剖分的地下水流模型的大规模并行计算.     

随机扰动下三维流体界面不稳定性的并行计算

对三维流体界面不稳定性的数值模拟引进了新的数值计算方法,并在MPI并行计算环境下进行了数值模拟.利用LevelSet方法确定界面位置,零水平集对应界面位置.对应离散LevelSet方程和界面两侧的两套Euler方程,借助于Ghost网格方法来完成离散.对最后网格点上的两套状态量的辨认依赖于该点的LevelSet值的符号.并进行了数值计算.     

方柱绕流大涡模拟

采用有限体/有限元混合格式、非结构网格和大涡模拟方法求解可压缩的N-S方程,对Re=22 000的方柱绕流进行数值模拟,并对不同的边界条件进行详细的分析比较.通过对以往研究经验的总结和利用精细的边界条件,使得采用二阶精度的数值格式和较稀疏的网格仍然得到了令人满意的计算结果,甚至优于以往采用密网格的模拟结果.     

基于FDTD的时域混合方法及其在天线前门耦合数值模拟中的应用

强电磁脉冲环境下的平台-机载天线一体化耦合计算属于典型多尺度时域电磁计算问题,采用传统的FDTD方法数值模拟时,由于精细结构的存在导致网格量巨大,计算效率低下。介绍了一种将非均匀FDTD方法与细导线FDTD方法以及多网格集总元件FDTD方法相结合的时域混合方法,能够有效降低计算开销,结合并行计算技术,快速计算得到天线端口上耦合产生的瞬态电压和电流响应,并将该方法成功应用于无人机平台-天线一体化前门耦合数值模拟中。     

基于格子Boltzmann方法的平板射流大涡模拟

应用多GPU技术,将格子Boltzmann方法与大涡模拟相结合(LBM-LES),使用1.12×10⁸网格,对雷诺数Re=4 000,倾斜角α=30°,吹风比M=0.5工况下的平板单孔射流进行了大规模高性能数值模拟研究.合理的定性与定量结果验证了LBM-LES模拟平板射流的有效性与可行性.使用上亿的计算网格捕捉了精细的湍流拟序结构,有利于主流与射流之间的掺混机理研究.此外,使用6个K20M GPU并行计算,模拟了71 680 LBM时间步长,仅耗时15 402秒,计算性能达到521.24MLUPS,即每秒更新5.212 4×10⁸个网格点的数据.     

并行计算环境中的高效气热耦合方法

气热耦合计算有助于高温透平叶片热流和温度分布的准确预估。在气热耦合计算中,流体区域和固体区域可能采用不同的网格形式和不同的求解器;此外,在大涡模拟等大规模计算中,流体区域和固体区域的耦合需要在并行环境中进行。因此,需要发展在不同求解区域支持不同的网格形式和求解器,并且适用于大规模并行计算环境的高效气热耦合方法。本文发展了并行计算环境中的高效气热耦合方法,其流体域和固体域之间的耦合通过高效的并行数据交换来实现,其对求解器类型和网格形式以及负载分配没有限制。数值算例验证了本文气热耦合方法在并行计筛环境中的高效性和灵活性。     

一种可扩展的三维半导体器件并行数值模拟算法

在基于漂移-扩散模型的三维半导体器件数值模拟中,通过有限体积法进行数值离散,采用完全耦合的牛顿迭代求解非线性代数方程组,并使用基于代数多重网格预条件子的GMRES方法求解牛顿迭代中的线性方程组,构造一种稳健且高度可扩展的非结构四面体网格上求解半导体方程的并行算法.基于PHG平台实现该算法的并行计算程序,并对PN结和MOS场效应晶体管等问题进行了最大网格规模达到5亿单元、最大并行规模达到1 024进程的大规模数值模拟实验,结果表明,该算法计算效率高,可扩展性好.     

流动/运动松耦合与紧耦合计算方法及稳定性分析

为了更加精确地模拟流动/运动耦合问题, 建立了耦合动态混合网格生成非定常流场计算和六自由度运动方程求解的一体化计算方法, 并在统一框架内同时实现了松耦合与紧耦合方法.通过圆柱涡致自激振荡(vortex induced vibration, VIV)的模拟, 对不同时间精度的松耦合和紧耦合算法的优劣及适用范围进行了评估和分析; 通过引入附加质量的概念, 对耦合算法的稳定性进行了理论分析.研究表明:在流体的密度与物体的密度接近时, 松耦合方法是不稳定的, 必须采用紧耦合方法.最后利用耦合算法对二维鱼体的自主游动和钝锥三自由度自由飞过程进行了数值模拟, 证实了理论分析的结论.      

流动和传热问题的分区并行计算

针对交错网格下的SIMPLE数值算法实施了分区并行计算方法,在小型局域网下实现了流动和传热问题的并行数值计算.对两个经典的流动和传热问题的数值模拟实验表明,所建立的并行计算环境和分区并行算法能够得到正确的和收敛的数值结果.但与串行计算结果相比,并行计算误差明显大于串行计算误差.对并行算法做出的性能分析表明,所给出的并行算法得到了明显的加速效率.随着计算规模的增大,加速比和并行效率提高更显著.     

Numerical simulation and application of three-dimensional oil resources migration-accumulation of fluid dynamics in porous media

Numerical simulation of oil migration and accumulation is to describe the history of oil migration and accumulation in basin evolution. It is of great value in evaluation of oil resources and determination of the location and amount of oil deposits. Based on such actual conditions as the effects of fluid mechanics in porous media and 3-dimensional geology characteristics, a kind of modified method of second order upwind finite difference fractional steps implicit interactive scheme was put forward. As for the actual problem of Dongying hollow, Huimin hollow, Tanhai region and Yangxin hollow in Shengli Petroleum Oil Field, a numerical simulation test was carried out, and the result is basically coincident with the actual conditions. For the model problem, optimal order estimates were derived. Thus the well-known problem on oil resources was solved. Supported by the National Basic Research Program of China (Grant No. 1999032803), the National Program for Tackling Key Problems (Grant No. 20050200069), the National Natural Science Foundation of China (Grant Nos. 10771124 and 10372052), and the Doctorate Foundation of the Ministry of Education of China (Grant No. 20030422047)     

流动数值模拟中一种并行自适应有限元算法

给出了一种流动数值模拟中的基于误差估算的并行网格自适应有限元算法.首先,以初网格上获得的当地事后误差估算值为权,应用递归谱对剖分方法划分初网格,使各子域上总体误差近似相等,以解决负载平衡问题.然后以误差值为判据对各子域内网格进行独立的自适应处理.最后应用基于粘接元的区域分裂法在非匹配的网格上求解N-S方程.区域分裂情形下N-S方程有限元解的误差估算则是广义Stokes问题误差估算方法的推广.为验证方法的可靠性,给出了不可压流经典算例的数值结果.     

Numerical simulation of 3D bubbles rising in viscous liquids using a front tracking method

The rise of bubbles in viscous liquids is not only a very common process in many industrial applications, but also an important fundamental problem in fluid physics. An improved numerical algorithm based on the front tracking method, originally proposed by Tryggvason and his co-workers, has been validated against experiments over a wide range of intermediate Reynolds and Bond numbers using an axisymmetric model [J. Hua, J. Lou, Numerical simulation of bubble rising in viscous liquid, J. Comput. Phys. 22 (2007) 769–795]. In the current paper, this numerical algorithm is further extended to simulate 3D bubbles rising in viscous liquids with high Reynolds and Bond numbers and with large density and viscosity ratios representative of the common air–water two-phase flow system. To facilitate the 3D front tracking simulation, mesh adaptation is implemented for both the front mesh on the bubble surface and the background mesh. On the latter mesh, the governing Navier–Stokes equations for incompressible, Newtonian flow are solved in a moving reference frame attached to the rising bubble. Specifically, the equations are solved using a finite volume scheme based on the Semi-Implicit Method for Pressure-Linked Equations (SIMPLE) algorithm, and it appears to be robust even for high Reynolds numbers and high density and viscosity ratios. The 3D bubble surface is tracked explicitly using an adaptive, unstructured triangular mesh. The numerical model is integrated with the software package PARAMESH, a block-based adaptive mesh refinement (AMR) tool developed for parallel computing. PARAMESH allows background mesh adaptation as well as the solution of the governing equations in parallel on a supercomputer. Further, Peskin distribution function is applied to interpolate the variable values between the front and the background meshes. Detailed sensitivity analysis about the numerical modeling algorithm has been performed. The current model has also been applied to simulate a number of cases of 3D gas bubbles rising in viscous liquids, e.g. air bubbles rising in water. Simulation results are compared with experimental observations both in aspect of terminal bubble shapes and terminal bubble velocities. In addition, we applied this model to simulate the interaction between two bubbles rising in a liquid, which illustrated the model’s capability in predicting the interaction dynamics of rising bubbles.     

基于网格片的氧碘化学激光器多块并行数值模拟

为实现氧碘化学激光器(COIL)喷管流场大规模数值模拟,采用VICON程序中的基本方程与数值算法,应用JASMIN框架中的基本数据结构——网格片,以及多块结构网格拼接并行算法,发展了三维多块COIL并行模拟程序。数值实验结果展示了该并行模拟程序的正确性及可扩展性,并在2048个处理器核上模拟450万网格单元算例,加速比超过420。     

风电场地形绕流的CFD结果确认研究

掌握复杂地形区域的风能分布对风电场微观选址至关重要.为把CFD技术应用于风场中实际地形的风流动模拟,本文选择三种具有实验数据的典型地形进行数值计算,并对数值结果进行确认研究.文中简述了不同网格和不同湍流模型对旋涡位置和速度分布的影响,并探讨了为增加风能利用率,如何确定风力机的最佳安装位置.     

Free Mesh Method: fundamental conception, algorithms and accuracy study

The finite element method (FEM) has been commonly employed in a variety of fields as a computer simulation method to solve such problems as solid, fluid, electro-magnetic phenomena and so on. However, creation of a quality mesh for the problem domain is a prerequisite when using FEM, which becomes a major part of the cost of a simulation. It is natural that the concept of meshless method has evolved. The free mesh method (FMM) is among the typical meshless methods intended for particle-like finite element analysis of problems that are difficult to handle using global mesh generation, especially on parallel processors. FMM is an efficient node-based finite element method that employs a local mesh generation technique and a node-by-node algorithm for the finite element calculations. In this paper, FMM and its variation are reviewed focusing on their fundamental conception, algorithms and accuracy.     

惯性约束聚变内爆中基于多块结构网格的高效辐射扩散并行算法

针对三维球形靶丸内爆高效模拟需求和传统笛卡尔正交网格上辐射加源困难的问题, 发展一种多块结构非正交网格生成方法, 并基于此种计算网格提出高效的三维扩散格式并行算法, 将其应用于辐射流体方程组的求解和三维内爆不对称性的数值模拟, 数值结果显示了算法的有效性。并行性能测试显示该算法可扩展到5400个核上, 并行效率达到69%。     

Bubble motion in inclined pipes

Highly nonlinear free-surface flows in vertical, inclined, and horizontal pipes are analyzed. The problem of bubble motion in a vertical pipe is closely related to the Rayleigh-Taylor instability problem. Inclined pipe flows are intensively studied as related to gas and oil transportation. A new theory of motion of large bubbles in pipes is developed. As distinct from previous approaches, which relied on semiempirical methods or numerical fitting, analytical methods of potential theory and complex analysis are used. A careful comparison of 2D and 3D solutions is presented. It is shown that a higher dimensionality may not correspond to a higher bubble velocity. For the first time, free-surface flows in inclined pipes are analyzed by means of direct numerical simulation, which makes it possible to develop a new approach to the Rayleigh-Taylor instability problem (bubbles with wedge-and cone-shaped noses).     

强爆炸火球辐射流体自适应网格高精度数值模拟

提出一种自适应网格方法,应用于基于Euler方法的强爆炸辐射流体高精度数值求解。通过与Zinn数值结果对比,验证该方法的正确性。研究自适应网格对冲击波和光辐射输出模拟精度的影响,对比不同网格尺度下的计算耗时。在相同的条件下,使用自适应网格与均匀网格加密3倍得到的冲击波超压分布、光辐射输出演化接近,计算效率提升约8.5倍。由此可知,提出的自适应网格方法可用于强爆炸问题的高精度数值模拟,显著地提升模拟精度和模拟效率。