三维湍流Rayleigh-Bénard热对流的高效并行直接求解方法

高Ra数Rayleigh-Bénard热对流的湍流特性研究是当前国际上的一个热门研究课题, DNS模拟计算是研究该课题的重要手段之一. 当计算规模增大而网格数巨大时计算工作难以实现, 高Ra湍流热对流的数值模拟研究面临重大挑战. 本文创建了大规模高效并行计算的三维湍流热对流直接求解方法. 采用FFT变换解耦压力泊松方程, 将其变换成沿z方向上的块三对角方程组, 并利用块三对角方程的MPI与OpenMP联立的大规模高效并行近似解求解方案, 创建了可以高效并行计算的热对流直接求解方法. 通过对该方法并行效率的验证计算, 证明新的直接求解并行计算方法具有很好的并行效率和计算时效. 三维窄方腔热对流的计算结果表明, 本文方法计算的三维热对流特性是合理的. 本文创建的可大规模高效并行计算的三维湍流热对流直接求解方法, 也很可能是关于计算流体力学不可压NS方程大规模高效并行计算在特殊情况中计算技术上的一个突破.     

湍流热对流的高效并行直接求解方法

结合泊松方程直接求解的PDD算法,创建二维湍流热对流DNS模拟的并行直接求解方法（PDM-DNS）.在"天河二号"超级计算机上完成高Ra数湍流热对流大规模DNS计算,计算效果令人惊喜,实现了一年前难以达到的计算规模.高分辨率的湍流热对流计算结果显示,不同高Ra数的湍流热对流流动状态完全不同.为高Ra数湍流热对流大规模并行计算和数值模拟研究提供了有价值的计算方法.     

滑移边界对二维Rayleigh-Bénard热对流流态和传热特性的影响

采用并行直接数值模拟（PDM-DNS）计算无滑移和滑移边界二维Rayleigh-Bénard热对流.与无滑移边界形成的随机羽流运动的湍流热对流不同,滑移边界热对流最终形成湍流特征消失,且温度仅分布于四壁的一个大尺度环流的流动形态.平均场近底板的温度分布特性,无滑移边界逐渐变化而滑移边界出现过冲现象.宽高比Γ=1时,Nusselt数（Nu）随Rayleigh数（Ra）的变化具有相同标度指数,Nu~Ra^0.3.滑移边界热对流具有传热增强作用.滑移边界热对流Nu随Γ变化明显,并分为两个阶段,在Γ=0.5时出现Nu_max≈250,是无滑移边界热流Nu的5倍.     

自然对流换热的高精度非结构网格有限体积法

用非结构网格有限体积法求解自然对流换热时,传统的对流项离散格式难以兼顾数值精度与计算效率,我们发展了一种耦合高精度格式的延迟修正方法,用于对流项的离散.高Re数下方腔驱动流数值计算验证了该方法具有较高的计算精度和较好的稳定性.Boussinesq流体的自然对流换热数值模拟,表明该方法能有效克服高Ra数时数值计算发散,可准确捕捉自然对流换热问题中不同偏心率下的等温线和流线分布特征.     

利用格子Boltzmann方法模拟矩形腔内纳米流体Raleigh-Benard对流

利用格 子 Boltzmann方法模拟矩形腔内纳米流体Rayleigh-Benard对流, 得到温度场和流线分布, 比较分析不同Ra数、体积分数、粒径下纳米流体对流换热的变化情况. 结果表明: 在相同的Ra 数和体积分数下, 纳米流体的对流换热随着粒径的增大而减弱; 在相同的Ra数和粒径下, 纳米流体的对流换热随着体积分数增大而增强. 关键词： 纳米流体 Raleigh-Benard 多相流 格子Boltzmann方法     

二维湍流热对流羽流运动路径对传热特性的影响

本文采用并行直接求解方法计算了0.7和4.3两个Pr数的二维湍流热对流,最高Ra=10~(13).高和极高Ra数的二维湍流热对流传热Nu数与反映羽流运动的大尺度环流路径随Ra数的变化具有很好的相关性,并具有两个Ra数转折点.第一转折点出现在大尺度环流由椭圆变为圆形,其周长C_(LSC)随Ra数变化突然减小,第二转折点出现在大尺度环流周长C_(LSC)最小值处,随后羽流变为旋涡团状而周长C_(LSC)随Ra数增高而增大.Pr数较小对应的转折Ra数较低.传热Nu数的变化规律用Ra~(0.3)补偿后发现, Nu的局部标度律在大尺度环流周长C_(LSC)较小的情况下减小,出现偏离GL理论预测倍数线的现象.当Ra数大于第二转折点后,极高Ra数二维湍流热对流Nu数随Ra数的变化与GL理论预测倍数线符合良好,走向保持一致.     

地球外核热对流运动的区域分解多重网格并行数值模拟

旋转球层中热对流运动的数值模拟是地球发电机模型的重要组成部分,对研究地球发电机作用机理具有重要意义.本文设计一个基于国产超级计算平台并行性能良好的地球外核热对流运动并行数值模型.时间积分方案采用与Crank-Nicolson格式和二阶Adams-Bashford公式相结合的近似分解分步法,空间离散基于立方球网格的二阶精度有限体积格式.所得到的两个大规模稀疏线性代数方程组采用带预处理的Krylov子空间迭代法进行求解.为加速迭代求解过程及提高并行性能,迭代过程采用区域分解多重网格的多层限制型加法Schwarz预处理子,减少了求解程序的计算时间,提高了数值模型的并行性能,模型被很好地扩展到上万处理器核数.数值模拟结果与基准模型算例0的参考值吻合得很好.     

标量湍流的能量传输特性

应用直接数值模拟数据,从标量湍流传输的三波关系出发,进行湍流及标量湍流传输谱的多尺度分析,研究不同尺度间的能量传输性质,证实标量能量的传输与湍动能传输具有不同性质,大尺度速度脉动对标量传输有较大贡献,尤其是与标量小尺度脉动的相互作用,使标量模拟需要有比速度场更高的网格分辨率;并发现标量湍流的能量传输具有明显的非局部性;另外,定义了能量传输系数,发现在相同的Re数和Pe数条件下,标量湍流的对流惯性较速度脉动的惯性子区宽.     

不连续冷源布局多孔介质内热流耦合LBM数值模拟

为研究不连续冷源边界对内置发热体多孔介质方腔内传热及流动的影响,采用格子Boltzmann方法对REV尺度下多孔介质方腔内的自然对流进行计算,并研究瑞利数（Ra）、达西数（Da）、孔隙度对多孔介质方腔内传热流动的影响.发现Da对方腔内的流体流型影响很大,Da为10^-4时,多孔介质方腔内只有一个涡流,而Da为10^-2时,方腔内有两个涡流.增大Ra、Da、孔隙度可以提高冷源壁面的平均努赛尔数（Nu）,增强散热效果,孔隙率对平均Nu影响程度和Da的大小有关.当冷源布置在壁面上方,壁面的平均Nu随Ra的增加剧烈变化,方腔处于高Ra条件下时,将冷源布置在边界的上方可以提高散热效果.6种布置方案中Case 6的散热效果最好.     

基于LBM的部分热活跃边界下多孔腔体内自然对流换热

采用Boltzmann方法模拟部分热活跃边界下的多孔腔体内自然对流,探讨不同热边界布置方案、孔隙度、Da数及Ra数对其流动传热的影响.数值计算表明：Da=10^-4时,腔体内中央出现一个循环流模式,只在Ra数很大时孔隙度才对传热有影响; Da=10^-2时,腔体内出现两个循环流,在Ra数很小时孔隙度对传热产生强烈的的影响.热活跃边界位置影响腔体内流体对流传热的强度,加热边界布置在底部、而冷却边界布置在顶部（Bottom-Top布置方式）,对多孔腔体内对流传热最有利,优于全热边界布置方式的传热效果.     

Heat Convection Between Two Confocal Elliptic Tubes Placed at Different Orientations

In this paper, transient and steady natural convection heat transfer in an elliptical annulus has been investigated. The annulus occupies the space between two horizontal concentric tubes of elliptic cross-section. The resulting velocity and thermal fields are predicted at different annulus orientations assuming isothermal surfaces. The full governing equations of mass, momentum and energy are solved numerically using the Fourier Spectral method. The heat convection process between the two tubes depends on Rayleigh number, Prandtl number, angle of inclination of tube axes and the geometry and dimensions of both tubes. The Prandtl number and inner tube axis ratio are fixed at 0.7 and 0.5, respectively. The problem is solved for the two Rayleigh numbers of $10^4$ and $10^5$ considering a ratio between the two major axes up to 3 while the angle of orientation of the minor axes varies from $0^\circ$ to $90^\circ$. The results for local and average Nusselt numbers are obtained and discussed together with the details of both flow and thermal fields. For isothermal heating conditions, the study has shown an optimum value for major axes ratio that minimizes the rate of heat transfer between the two tubes. Another important aspect of this paper is to prove the successful use of the Fourier Spectral Method in solving confined flow and heat convection problems.     

The ultimate state of thermal convection in Rayleigh-Taylor turbulence

This paper discusses the so-called ultimate state of thermal convection, first proposed by Kraichnan almost 50 years ago and recently observed in numerical simulations of turbulent convection in the absence of boundaries. We focus on numerical simulations of turbulence generated by the Rayleigh-Taylor instability in a wide range of Rayleigh and Prandtl numbers. Our results point out to the conclusion that RT turbulence provides a natural realization of the ultimate state of thermal convection thus highlighting the relationship between the absence of boundaries and the emergence of the ultimate state scaling for global statistical quantities.     

New perspectives in turbulent Rayleigh-Bénard convection

Recent experimental, numerical and theoretical advances in turbulent Rayleigh-Bénard convection are presented. Particular emphasis is given to the physics and structure of the thermal and velocity boundary layers which play a key role for the better understanding of the turbulent transport of heat and momentum in convection at high and very high Rayleigh numbers. We also discuss important extensions of Rayleigh-Bénard convection such as non-Oberbeck-Boussinesq effects and convection with phase changes.     

New perspectives in turbulent Rayleigh-Bénard convection

Recent experimental, numerical and theoretical advances in turbulent Rayleigh-Bénard convection are presented. Particular emphasis is given to the physics and structure of the thermal and velocity boundary layers which play a key role for the better understanding of the turbulent transport of heat and momentum in convection at high and very high Rayleigh numbers. We also discuss important extensions of Rayleigh-Bénard convection such as non-Oberbeck-Boussinesq effects and convection with phase changes.     

Simulation of COPO II Lo experiments on natural convection of heat generating liquid at high Rayleigh numbers

The methodology is described and the results are presented concerning numerical modeling of COPO II Lo experiments on heat transfer in liquid with internal heat generation at very high internal Rayleigh numbers typical for natural convection in a core melt that can appear during progress of severe accident at a nuclear power plant (NPP). The work is keeping in the course of development of CFD-based tool for quantitative analysis of heat transfer in a stratified molten pool of different configurations possible in severe accident scenarios with melt retention in the reactor vessel or in the VVER core catcher. Such CFD methodology would be used for testing of simplified correlation models for simulation of the core melt interaction with NPP structures in system code SOCRAT. During verification the available experimental data on the core melt thermohydraulics were analyzed, and it was concluded that they are insufficient to measures of CFD quality. The data uncertainties, along with the complexities of convective flow, uncertainties of the reactor core melt conditions, limitations of experimental possibilities and of turbulence modeling, actually constrain the multivariate CFD simulations of natural convection at very high Rayleigh numbers. RANS turbulence models only can be efficiently applied here, and they are to be checked for such purposes. In a series of numerical modeling of COPO II Lo experiments and some others, availability of a k-? realizable model with included buoyancy effects was estimated, and the optimal set of CFD options was formed for minimizing numerical artifacts. It was demonstrated that in the investigated range of Rayleigh numbers the k-? model works qualitatively correctly, but is inclined to systematical deformation of the melt boundary heat transfer distribution. This allows one to use this model for qualitative multivariate CFD estimations but requires improvement of the model or finding of its efficient and more exact equivalent.     

Analysis of coherent structures in Rayleigh–Bénard convection

Using direct numerical simulation, we investigate characteristics of coherent structures in Rayleigh–Bénard convection in a soft turbulence regime. The role of thermal plumes, essential structures in Rayleigh–Bénard convection, is studied by splitting flow regimes into thermal plume and background by investigating joint probability density function (PDF) of invariants of velocity gradient tensor. The contribution to thermal dissipation rate by these two regions is analysed separately. Through the joint PDF of invariants, we also examine the thermal effect on velocity structures.     

A HOLOGRAPHIC INTERFEROMETRY STUDY OF THE DOUBLE-DIFFUSIVE LAYERED SYSTEM

This study investigates double-diffusive convection in a two-layer, salt-stratified solution destablized by lateral Keating and cooling. The two-wavelength holographic interferometry method was used to measure the transient temperature, concentration, and density distributions. The evolution of the two-layered system can be divided into three stages. In the first stage, the thermally driven convective layers form rapidly, and the existing diffusion layer adjusts itself into a thin interface by convection motion. In the second stage, a quasi-steady state is attained. The temperature distribution is S-type and the temperature difference across the diffusion region does not change much. The concentration distribution is uniform in the two fluid layers, but the concentration and density differences decrease linearly with time. When the interface becomes very thin, unstable finger-type convection appears. Finally, the interface is destroyed by the boundary layers at the side walls in the third stage. The interfacial Nusselt number and Sherwood number art found to increase with the thermal Rayleigh number, and the effect of the solutal Rayleigh number seems to be less significant. The dimemumless mixing time is found to correlate well with thermal and solutal Rayleigh numbers. Results from numerical simulation are demonstrated and compared with the experiments.     

Multiscale SOC in turbulent convection

Using data obtained in a laboratory thermal convection experiment at high Rayleigh numbers, it is shown that the multiscaling properties of the observed mean wind reversals are quantitatively consistent with analogous multiscaling properties of the Bak–Tang–Wiesenfeld prototype model of self-organized criticality in two dimensions.