摘要:

This manuscript briefly describes a robust iterative solution methodology used for parallel unstructured finite element simulation of strongly coupled fluid flow, heat and mass transfer. The solution method relies on an inexact Newton scheme and linear system solvers based on preconditioned Krylov subspace methods. The discussion considers computational efficiency and robustness issues related to the proposed schemes. The evaluated preconditioning techniques include simple polynomial expansion and multi-step block iterative methods along with overlapping Schwarz domain decomposition techniques using subdomain solvers based on incomplete factorizations. For this comparison a particular spatial discretization of the governing transport PDEs based on a Galerkin Least Squares (GLS) finite element formulation is used. Results are presented for some standard 2D CFD benchmark problems as well as for a number of 3D problems.

ISSN:1061-8562    

DOI:10.1080/10618569908940825

出版商:Taylor & Francis Group    

年代:1999

数据来源: Taylor

摘要:

This paper presents a new approach to the solution of nonsymmetric linear systems that uses hybrid techniques based on both direct and iterative methods. An implicitly preconditioned modified system is obtained by applying projections onto block rows of the original system. Our technique provides the flexibility of using either direct or iterative methods for the solution of the preconditioned system. The resulting algorithms are robust, and can be implemented with high efficiency on a variety of parallel architectures. The algorithms are used to solve linear systems arising from the discretization of convection-diffusion equations as well as those systems that arise from the simulation of particulate flows. Experiments are presented to illustrate the robustness and parallel efficiency of these methods.

ISSN:1061-8562    

DOI:10.1080/10618569908940826

出版商:Taylor & Francis Group    

年代:1999

数据来源: Taylor

摘要:

We describe massively parallel finite element compulations of fluid dynamics for several crystal growth systems. Examples are presented of how large-scale numerical simulations have been used to gain insights to the workings of several processes, specifically the melt growth of oxide crystals by the Czochralski process and the solution growth of nonlinear optical crystals. These systems are characterized by nonlinear interactions between field and interfacial phenomena-the transport of momentum, heat, and mass coupled with solidification kinetics. Modern finite element methods show great potential to provide the understanding needed to optimize these processes.

ISSN:1061-8562    

DOI:10.1080/10618569908940827

出版商:Taylor & Francis Group    

年代:1999

数据来源: Taylor

摘要:

A space-time scheme is an unconditional stable time-integration scheme and its domain is discretized into space-time slabs, which are employed to weakly enforce the continuity of the solution across the time slabs. It is applied in this study to two-dimensional advection-diffusion problems, and space-mesh adaptation is introduced. Mesh adaptation is a powerful scheme to reduce discretized errors: however, it is found that an error due to the projection between adaptive meshes in successive time slabs is another source of error. To reduce projection errors, the direct projection scheme for space-time method will be used in this study. Galerkin/Least Squares scheme is applied to prevent numerical instability due to the skew-symmetric term in the weak form of the advection-diffusion equation.

ISSN:1061-8562    

DOI:10.1080/10618569908940828

出版商:Taylor & Francis Group    

年代:1999

数据来源: Taylor

摘要:

In this paper, we suggest a numerical simulation method of snowdrift. Two-dimensional numerical simulation method was first examined by comparison with the wind tunnel experiments. We found that the prediction method had the sufficient accuracy. Next, the simulation method was expanded to the three-dimensional simulation method. The simulation of snowdrift for the actual building was conducted to verify the applicability of the method. It became clear that the numerical simulation method was a useful tool to predict the snowdrift around the building.

ISSN:1061-8562    

DOI:10.1080/10618569908940829

出版商:Taylor & Francis Group    

年代:1999

数据来源: Taylor

摘要:

A finite element method for analyzing unsteady incompressible creeping flows is presented. Marker particles are introduced to analyze the flow motions. To determine the marker position in the element, vector products are used. By checking the signs of the product, the marker position during the transient analysis can be determined in a simple manner. A benchmark-type problem for which an analytical solution is available and the filling process of a simple axisymmetrical mould shape are solved to illustrate this method.

ISSN:1061-8562    

DOI:10.1080/10618569908940830

出版商:Taylor & Francis Group    

年代:1999

数据来源: Taylor

摘要:

A finite element method based on ALE formulation has been adopted in order to examine the effect of periodic velocity excitation on the aerodynamic characteristics of an oscillating circular cylinder. Periodic excitation, which was placed on the cylinder surface, stimulated the separated shear layers around the cylinder, and numerical results showed that some excitation can reduce negative damping, which is caused by unsteady lift force, and thereby stabilize the aerodynamics of the cylinder. Furthermore, the change of lift phase caused by periodic excitation seems to be important in stabilizing the aerodynamics of the cylinder. The simulation also confirmed that periodic excitation can suppress the vortex-induced vibration of the cylinder.

ISSN:1061-8562    

DOI:10.1080/10618569908940831

出版商:Taylor & Francis Group    

年代:1999

数据来源: Taylor

摘要:

Numerical aeroelastic simulation of a high-aspect-ratio transport type wing model in transonic region is presented. The aeroelastic responses of the wing are extracted by integrating compressible thin-layer Navier-Stokes equations coupled with the equations of motion of the wing structure, in a time dependent manner. The Yee-Harten implicit TVD scheme and the Wilson's θ method are employed to integrate these equations, respectively. Flutter boundaries were found for Mach number range, 0.7 to 0.85 and the results were compared with experimental flutter boundaries. Futhermore, Limit Cycle Oscillations were found and the characteristics of the flutter and limit cycle oscillations are investigated and discussed.

ISSN:1061-8562    

DOI:10.1080/10618569908940832

出版商:Taylor & Francis Group    

年代:1999

数据来源: Taylor

摘要:

In previous papers, we proposed finite element schemes based on the Petrov-Galerkin weak formulation using exponential weighting functions for solving accurately, and in a stable manner, the flow field of an incompressible viscous fluid. In this paper, we present the Petrov-Galerkin finite element scheme for turbulent flow fields based on large eddy simulation using the standard Smagorinsky model with the Van Driest damping function. The filtered incompressible Navier-Stokes equations are numerically integrated in time by using a fractional step strategy with second-order accurate Adams-Bashforth explicit differencing for both convection an diffusion terms. In order to evaluate more accurately a mass matrix, the well-known multi-pass algorithm was also adopted in this study. Numerical results obtained are compared through flow around a rectangular cylinder at Re = 22,000 with the experimental data and other existing numerical data.

ISSN:1061-8562    

DOI:10.1080/10618569908940833

出版商:Taylor & Francis Group    

年代:1999

数据来源: Taylor

摘要:

The conventional shape function for the finite-element method (FEM) is linear, and it is thus inadequate for analyzing numerically complex flows at high Reynolds numbers. In this study, we propose a new scheme, GSMAC-CIP, using the third-order shape function, which requires continuity of the value of the function and its first space derivative in the whole space and is formulated by a finite element method for the cubic interpolated pseudo-particle (CIP) method. We verified the effectiveness of this new scheme by analyzing the forced-driven convection in a square cavity at Re = 1000, 5000 and 10000. The numerical results obtained by the present scheme are compared with those of GSMAC-FEM using coarser meshes, and it is shown that the present scheme is superior to GSMAC-FEM in terms of space accuracy. Moreover, it is shown that the numerical results obtained by the present scheme using fine meshes were in precise agreement with those obtained by Ghia et al.

ISSN:1061-8562    

DOI:10.1080/10618569908940834

出版商:Taylor & Francis Group    

年代:1999

数据来源: Taylor