气体动理学统一算法的隐式方法研究

目前的气体动理学统一算法(unified gas kinetic scheme, 简称UGKS) 在求解高速流动问题时的计算效率,难以满足求解复杂工程问题的需求. 为了提高该算法的计算效率, 本文对模型方程的对流项和碰撞项进行了隐式处理, 并针对UGKS 界面通量与演化时间相关的特点, 引入了演化时间平均界面通量, 通过对控制方程矩阵进行近似LU 分解(lower-upper decomposition), 实现了隐式UGKS. 不同来流马赫数的圆柱绕流算例测试表明, 只要演化时间选取得当, 隐式方法可以得到与显式方法完全相同的结果, 且计算效率可以提高1~2 个量级.      

A high-order parallel Eulerian-Lagrangian algorithm for advection-diffusion problems on unstructured meshes

In this paper, we present a high-order discontinuous Galerkin Eulerian-Lagrangian method for the solution of advection-diffusion problems on staggered unstructured meshes in two and three space dimensions. The particle trajectories are tracked backward in time by means of a high-order representation of the velocity field and a linear mapping from the physical to a reference system, hence obtaining a very simple and efficient strategy that permits to follow the Lagrangian trajectories throughout the computational domain. The use of an Eulerian-Lagrangian discretization increases the overall computational efficiency of the scheme because it is the only explicit method for the discretization of convective terms that admits large time steps without imposing a Courant-Friedrichs-Lewy–type stability condition. This property is fully exploited in this work by relying on a semi-implicit discretization of the incompressible Navier-Stokes equations, in which the pressure is discretized implicitly; thus, the sound speed does not play any role in the restriction of the maximum admissible time step. The resulting mild Courant-Friedrichs-Lewy stability condition, which is based only on the fluid velocity, is here overcome by the adoption of the Eulerian-Lagrangian method for the advection terms and an implicit scheme for the diffusive part of the governing equations. As a consequence, the novel algorithm is able to run simulation with a time step that is defined by the user, depending on the desired efficiency and time scale of the physical phenomena under consideration. Finally, a complete Message Passing Interface parallelization of the code is presented, showing that our approach can reach up to 96% of scaling efficiency.     

The hypohydrostatic rescaling and its impacts on modeling of atmospheric convection

The atmospheric circulation spans a wide range of spatial scales, including the planetary scale (~10,000 km), synoptic scale (~2,000 km), mesoscale (~200 km), and convective scales (< 20 km). The wide scale separation between convective motions, responsible for the vertical energy transport, and the planetary circulation, responsible for the meridional energy transport, has prevented explicit representation of convective motions in global atmospheric models. Kuang et al. (Geophys. Res. Lett. 32: L02809, 2005) have suggested a way to circumvent this limitation through a rescaling that they refer to as Diabatic Acceleration and REscaling (DARE). We focus here on a modified version of the procedure that we refer to as hypohydrostatic rescaling. These two strategies are equivalent for inviscid and adiabatic flow in the traditional meteorological setting in which the vertical component of the Coriolis acceleration is ignored, but they differ when atmospheric physics is taken into account. It is argued here that, while the hypohydrostatic rescaling preserves the dynamics of the planetary scale circulation, it increases the horizontal scale of convective motions. This drastically reduces the computational cost for explicit simulation of hypohydrostatic convection in a global atmospheric model. A key question is whether explicit simulations of hypohydrostatic convection could offer a valid alternative to convective parameterization in global models. To do so, radiative–convective equilibrium is simulated with a high-resolution non-hydrostatic model using different model resolutions and values of the rescaling parameter. When the behavior of hypohydrostatic convection is compared with coarse-resolution simulations of convection, the latter set of simulations reproduce more accurately the result from a reference high-resolution simulation. This is particularly true for the convective velocity and cloud ice distributions. Scaling arguments show that hypohydrostatic rescaling increases the convective overturning time. In particular, this convective slowdown associated with the hypohydrostatic rescaling is more significant than the slowdown resulting from under-resolving the convective elements. These results cast doubt on the practical value of the hypohydrostatic rescaling as an alternative to convective parameterization.     

Multidomain implicit numerical scheme

A multidomain method for the solution of elliptic CFD problems with an ADI scheme is described. Two methods of treatment of internal boundary conditions for ADI functions are discussed, namely an explicit and a semi‒implicit method. Stability conditions for the proposed methods are derived theoretically. The semi‒implicit scheme is more stable than the explicit scheme, leading to improved numerical efficiency for multidomain computations. Numerical computations for a linear convection-diffusion equation, for buoyancy‒driven recirculating flow in a square cavity and for turbulent flow in a square duct confirmed the theoretical results. Computer runs of the multidomain code in a distributed memory multiprocessor system were successful and efficient and produced reliable results. © 1997 John Wiley & Sons, Ltd.     

Parameter investigation with line-implicit lower–upper symmetric Gauss–Seidel on 3D stretched grids

An implicit lower–upper symmetric Gauss–Seidel (LU-SGS) solver has been implemented as a multigrid smoother combined with a line-implicit method as an acceleration technique for Reynolds-averaged Navier–Stokes (RANS) simulation on stretched meshes. The computational fluid dynamics code concerned is Edge, an edge-based finite volume Navier–Stokes flow solver for structured and unstructured grids. The paper focuses on the investigation of the parameters related to our novel line-implicit LU-SGS solver for convergence acceleration on 3D RANS meshes. The LU-SGS parameters are defined as the Courant–Friedrichs–Lewy number, the left-hand side dissipation, and the convergence of iterative solution of the linear problem arising from the linearisation of the implicit scheme. The influence of these parameters on the overall convergence is presented and default values are defined for maximum convergence acceleration. The optimised settings are applied to 3D RANS computations for comparison with explicit and line-implicit Runge–Kutta smoothing. For most of the cases, a computing time acceleration of the order of 2 is found depending on the mesh type, namely the boundary layer and the magnitude of residual reduction.     

Local and Global Transitions to Chaos and Hysteresis in a Porous Layer Heated from Below

The routes to chaos in a fluid saturated porous layer heated from below are investigated by using the weak nonlinear theory as well as Adomian's decomposition method to solve a system of ordinary differential equations which result from a truncated Galerkin representation of the governing equations. This representation is equivalent to the familiar Lorenz equations with different coefficients which correspond to the porous media convection. While the weak nonlinear method of solution provides significant insight to the problem, to its solution and corresponding bifurcations and other transitions, it is limited because of its local domain of validity, which in the present case is in the neighbourhood of any one of the two steady state convective solutions. On the other hand, the Adomian's decomposition method provides an analytical solution to the problem in terms of infinite power series. The practical need to evaluate numerical values from the infinite power series, the consequent series truncation, and the practical procedure to accomplish this task transform the otherwise analytical results into a computational solution achieved up to a finite accuracy. The transition from the steady solution to chaos is analysed by using both methods and their results are compared, showing a very good agreement in the neighbourhood of the convective steady solutions. The analysis explains previously obtained computational results for low Prandtl number convection in porous media suggesting a transition from steady convection to chaos via a Hopf bifurcation, represented by a solitary limit cycle at a sub-critical value of Rayleigh number. A simple explanation of the well known experimental phenomenon of Hysteresis in the transition from steady convection to chaos and backwards from chaos to steady state is provided in terms of the present analysis results.     

Taylor-Galerkin-based spectral element methods for convection-diffusion problems

Several explicit Taylor-Galerkin-based time integration schemes are proposed for the solution of both linear and non-linear convection problems with divergence-free velocity. These schemes are based on second-order Taylor series of the time derivative. The spatial discretization is performed by a high-order Galerkin spectral element method. For convection-diffusion problems an operator-splitting technique is given that decouples the treatment of the convective and diffusive terms. Both problems are then solved using a suitable time scheme. The Taylor-Galerkin methods and the operator-splitting scheme are tested numerically for both convection and convection-diffusion problems.     

A staggered spectral element model with application to the oceanic shallow water equations

A staggered spectral element model for the solution of the oceanic shallow water equations is presented. We introduce and compare both an implicit and an explicit time integration scheme. The former splits the equations with the operator-integration factor method and solves the resulting algebraic system with generalized minimum residual (GMRES) iterations. Comparison of the two schemes shows the performance of the implicit scheme to lag that of the explicit scheme because of the unpreconditioned implementation of GMRES. The explicit code is successfully applied to various geophysical flows in idealized and realistic basins, notably to the wind-driven circulation in the North Atlantic Ocean. The last experiment reveals the geometric versatility of the spectral element method and the effectiveness of the staggering in eliminating sprious pressure modes when the flow is nearly non-divergent.     

A computationally efficient 3D finite‐volume scheme for violent liquid–gas sloshing

We describe a semi‐implicit volume‐of‐fluid free‐surface‐modelling methodology for flow problems involving violent free‐surface motion. For efficient computation, a hybrid‐unstructured edge‐based vertex‐centred finite volume discretisation is employed, while the solution methodology is entirely matrix free. Pressures are solved using a matrix‐free preconditioned generalised minimum residual algorithm and explicit time‐stepping is employed for the momentum and interface‐tracking equations. The high resolution artificial compressive (HiRAC) volume‐of‐fluid method is used for accurate capturing of the free surface in violent flow regimes while allowing natural applicability to hybrid‐unstructured meshes. The code is parallelised for solution on distributed‐memory architectures and evaluated against 2D and 3D benchmark problems. Good parallel scaling is demonstrated, with almost linear speed‐up down to 6000 cells per core. Finally, the code is applied to an industrial‐type problem involving resonant excitation of a fuel tank, and a comparison with experimental results is made in this violent sloshing regime. Copyright © 2015 John Wiley & Sons, Ltd.     

A fast method for solving fluid–structure interaction problems numerically

The paper presents a semi‐implicit algorithm for solving an unsteady fluid–structure interaction problem. The algorithm for solving numerically the fluid–structure interaction problems was obtained by combining the backward Euler scheme with a semi‐implicit treatment of the convection term for the Navier–Stokes equations and an implicit centered scheme for the structure equations. The structure is governed either by the linear elasticity or by the non‐linear St Venant–Kirchhoff elasticity models. At each time step, the position of the interface is predicted in an explicit way. Then, an optimization problem must be solved, such that the continuity of the velocity as well as the continuity of the stress hold at the interface. During the Broyden, Fletcher, Goldforb, Shano (BFGS) iterations for solving the optimization problem, the fluid mesh does not move, which reduces the computational effort. The term ‘semi‐implicit’ used for the fully algorithm means that the interface position is computed explicitly, while the displacement of the structure, velocity and the pressure of the fluid are computed implicitly. Numerical results are presented. Copyright © 2008 John Wiley & Sons, Ltd.     

A Jacobian‐free Newton–Krylov method for thermalhydraulics simulations

The current paper is focused on investigating a Jacobian‐free Newton–Krylov (JFNK) method to obtain a fully implicit solution for two‐phase flows. In the JFNK formulation, the Jacobian matrix is not directly evaluated, potentially leading to major computational savings compared with a simple Newton's solver. The objectives of the present paper are as follows: (i) application of the JFNK method to two‐fluid models; (ii) investigation of the advantages and disadvantages of the fully implicit JFNK method compared with commonly used explicit formulations and implicit Newton–Krylov calculations using the determination of the Jacobian matrix; and (iii) comparison of the numerical predictions with those obtained by the Canadian Algorithm for Thermaulhydraulics Network Analysis 4. Two well‐known benchmarks are considered, the water faucet and the oscillating manometer. An isentropic two‐fluid model is selected. Time discretization is performed using a backward Euler scheme. A Crank–Nicolson scheme is also implemented to check the effect of temporal discretization on the predictions. Advection Upstream Splitting Method+ is applied to the convective fluxes. The source terms are discretized using a central differencing scheme. One explicit and two implicit formulations, one with Newton's solver with the Jacobian matrix and one with JFNK, are implemented. A detailed grid and model parameter sensitivity analysis is performed. For both cases, the JFNK predictions are in good agreement with the analytical solutions and explicit profiles. Further, stable results can be achieved using high CFL numbers up to 200 with a suitable choice of JFNK parameters. The computational time is significantly reduced by JFNK compared with the calculations requiring the determination of the Jacobian matrix. Copyright © 2015 John Wiley & Sons, Ltd.     

Integrated numerical model for vegetated surface and saturated subsurface flow interaction

The construction of an integrated numerical model is presented in this paper to deal with the interactions between vegetated surface and saturated subsurface flows. A numerical model is built by integrating the previously developed quasi-three-dimensional (Q3D) vegetated surface flow model with a two-dimensional (2D) saturated groundwater flow model. The vegetated surface flow model is constructed by coupling the explicit finite volume solution of 2D shallow water equations (SWEs) with the implicit finite difference solution of Navier-Stokes equations (NSEs) for vertical velocity distribution. The subsurface model is based on the explicit finite volume solution of 2D saturated groundwater flow equations (SGFEs). The ground and vegetated surface water interaction is achieved by introducing source-sink terms into the continuity equations. Two solutions are tightly coupled in a single code. The integrated model is applied to four test cases, and the results are satisfactory.     

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.     

Convective heat transfer in a rotating horizontal cylindrical fluid layer

This paper describes the thermal convection and heat transfer in a cylindrical fluid layer rotating around a horizontal axis, with various constant temperatures set at the layer boundaries. The influence of the rotational speed of the cylindrical fluid layer on the convective heat transfer in this layer is studied. The study results are presented as functions of dimensionless parameters that characterize the action of two convective mechanisms: centrifugal and thermal-oscillatory. It is shown that, with low rotational speed, the heat transfer is determined by quasistationary gravitational convection.     

Development of a convection–diffusion‐reaction magnetohydrodynamic solver on non‐staggered grids

This paper presents a convection–diffusion‐reaction (CDR) model for solving magnetic induction equations and incompressible Navier–Stokes equations. For purposes of increasing the prediction accuracy, the general solution to the one‐dimensional constant‐coefficient CDR equation is employed. For purposes of extending this discrete formulation to two‐dimensional analysis, the alternating direction implicit solution algorithm is applied. Numerical tests that are amenable to analytic solutions were performed in order to validate the proposed scheme. Results show good agreement with the analytic solutions and high rate of convergence. Like many magnetohydrodynamic studies, the Hartmann–Poiseuille problem is considered as a benchmark test to validate the code. Copyright © 2004 John Wiley & Sons, Ltd.     

On the Stefan problem with volumetric energy generation

This paper presents results of solid–liquid phase change, driven by volumetric energy generation (VEG), in a vertical cylinder. We show excellent agreement between a quasi-static, approximate analytical solution valid for Stefan numbers less than one, and a computational model solved using the computational fluid dynamics code FLUENT®. A computational study also shows the effect that the VEG has on both the mushy zone thickness and convection in the melt during phase change.     

Digital Filtering Techniques for Parallel Computation of Explicit Schemes

A computational technique is presented for designing a filter to improve the computational efficiency of a numerical scheme. For an explicit scheme, the integration time step is increased, causing several waves to become unstable. These waves are filtered without disturbing the accuracy of the solution and the accuracy of the remaining waves are controlled. The scheme is applied to the solution of the Euler equations by using the NPARC code.     

Fluid flow investigations within a 37 element CANDU fuel bundle supported by magnetic resonance velocimetry and computational fluid dynamics

The current work presents experimental and computational investigations of fluid flow through a 37 element CANDU nuclear fuel bundle. Experiments based on Magnetic Resonance Velocimetry (MRV) permit three-dimensional, three-component fluid velocity measurements to be made within the bundle with sub-millimeter resolution that are non-intrusive, do not require tracer particles or optical access of the flow field. Computational fluid dynamic (CFD) simulations of the foregoing experiments were performed with the hydra-th code using implicit large eddy simulation, which were in good agreement with experimental measurements of the fluid velocity. Greater understanding has been gained in the evolution of geometry-induced inter-subchannel mixing, the local effects of obstructed debris on the local flow field, and various turbulent effects, such as recirculation, swirl and separation. These capabilities are not available with conventional experimental techniques or thermal-hydraulic codes. The overall goal of this work is to continue developing experimental and computational capabilities for further investigations that reliably support nuclear reactor performance and safety.     

A generic framework for time-stepping partial differential equations (PDEs): general linear methods,object-oriented implementation and application to fluid problems

Time-stepping algorithms and their implementations are a critical component within the solution of time-dependent partial differential equations (PDEs). In this article, we present a generic framework – both in terms of algorithms and implementations – that allows an almost seamless switch between various explicit, implicit and implicit–explicit (IMEX) time-stepping methods. We put particular emphasis on how to incorporate time-dependent boundary conditions, an issue that goes beyond classical ODE theory but which plays an important role in the time-stepping of the PDEs arising in computational fluid dynamics. Our algorithm is based upon J.C. Butcher's unifying concept of general linear methods that we have extended to accommodate the family of IMEX schemes that are often used in engineering practice. In the article, we discuss design considerations and present an object-oriented implementation. Finally, we illustrate the use of the framework by applications to a model problem as well as to more complex fluid problems.     

全机绕流Euler方程多重网格分区计算方法

全机三维复杂形状绕流数值求解只能采用分区求解的方法,本文采用可压缩Euler方程有限体积方法以及多重网格分区方法对流场进行分区计算。数值方法采用改进的van Leer迎风型矢通量分裂格式和MUSCL方法,基于有限体积方法和迎风型矢通量分裂方法,建立一套处理子区域内分界面的耦合条件。各个子区域之间采用显式耦合条件,区域内部采用隐式格式和局部时间步长等,以加快收敛速度。计算结果飞机表面压力分布等气动力特性与实验值进行了比较,二者基本吻合。计算结果表明采用分析“V”型多重网格方法,能提高计算效率,加快收敛速度达到接近一个量级。根据全机数值计算结果和可视化结果讨论了流场背风区域旋涡的形成过程。