• Journal of computational fluids engineering
• /
• v.16 no.2
• /
• pp.49-55
• /
• 2011

Turbulent flows with wavy wall are simulated using Residual-based Variational Multiscale Method (RB-VMS) which is proposed by Bazilves et al(2007) as new Large Eddy Simulation methodology. Incompressible Navier-Stokes equations are integrated using Isogeometric analysis which adopt the basis function as NURBS. The Reynolds number is 6760 based on the bulk velocity and averaged channel height. And the amplitude (${\alpha}/{\lambda}$) of wavy wall is 0.05. The computational domain is $2{\lambda}{\times}1.05{\lambda}{\times}{\lambda}$ in the streamwise, wall normal and spanwise direction. Mean quantities and turbulent statistics near wavy wall are compared with DNS results of Cherukat et al.(1998). The predicted results show good agreement with reference data.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.26 no.4
• /
• pp.223-231
• /
• 2013

This paper presents a time-domain Gauss-Newton full-waveform inversion method for the material profile reconstruction in heterogeneous semi-infinite solid media. To implement the inverse problem in a finite computational domain, perfectly-matchedlayers( PMLs) are introduced as wave-absorbing boundaries within which the domain's wave velocity profile is to be reconstructed. The inverse problem is formulated in a partial-differential-equations(PDE)-constrained optimization framework, where a least-squares misfit between measured and calculated surface responses is minimized under the constraint of PML-endowed wave equations. A Gauss-Newton-Krylov optimization algorithm is utilized to iteratively update the unknown wave velocity profile with the aid of a specialized regularization scheme. Through a series of one-dimensional examples, the solution of the Gauss-Newton inversion was close enough to the target profile, and showed superior convergence behavior with reduced wall-clock time of implementation compared to a conventional inversion using Fletcher-Reeves optimization algorithm.

• Proceedings of the KSME Conference
• /
• 2003.11a
• /
• pp.85-89
• /
• 2003

In a closed square cavity filled with a liquid, a cooled the upper horizontal wall and a heated the lower horizontal wall, the flow isn't generated under the ground-based condition when Rayleigh number is lower than 1700. In such case the flow phenomena near an air bubble under a cooled horizontal wall were investigated. The temperature and the flow fields were studied by using the Thermo-sensitive Liquid-Crystal and the image processing. The qualitative analysis for the temperature and the flow fields were carried out by applying the image processing technique to the original data. Injecting bubble at the center point of upper cooled wall, the symmetry shape of two vortexes near an air bubble was observed. The bubble size increased, the size of velocity and the magnitude of velocity increased. In spite of elapsed time, a pair of two vortexes was the unique and steady-state flow in a square cavity and wasn't induce to the other flow in the surround region.

• Journal of the Korean Crystal Growth and Crystal Technology
• /
• v.16 no.2
• /
• pp.53-58
• /
• 2006

When external field was applied to congruent $LiNbO_3$, it was investgated for domain formation and expansion of $LiNbO_3$. The domain wall velocities of 0.5 mm thickness $LiNbO_3$ were 28.70, 16.02 and $5.75{\mu}m/sec$ under poling field of 23.5, 22.0 and 21.0 kV/mm, respectively. As $1 M{\Omega}$ resistor was used in domain inversion system, harmonic domain inversion was not achieved by rapid domain expansion. And 50% duty cycle periodically poled $LiNbO_3$ have been fabricated by charge control using $10 M{\Omega}$ resistor.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.32 no.8
• /
• pp.621-628
• /
• 2008

A numerical study for the flow, heat and mass transfer characteristics of the evaporating tube with the films flowing down on both the inside and outside tube walls has been carried out. The condensation occurs along the outside wall while the evaporation occurs at the free surface of the inside film. The transport equations for momentum and energy are parabolized by the boundary-layer approximation and solved by using the marching technique. The calculation domain of 2 film flow regions (evaporating and condensation films at the inside and outside tube wall respectively) and tube wall is solved simultaneously. The coupling technique for the problem with the 3 different regions and the 2 interfaces of them has been developed to calculate the temperature field. The velocity and temperature fields and the amount of the condensed and evaporated mass as well as the position where the evaporating film is completely dried out are successfully predicted for various inside pressures and inside film inlet flow rates.

• Journal of computational fluids engineering
• /
• v.15 no.2
• /
• pp.35-40
• /
• 2010

In the present work, LES with new variational multiscale method is conducted on the fully developed channel flow with Reynolds number, 180 based on the friction velocity and the channel half width. Incompressible Navier-Stokes equations are integrated using finite element method with the basis function of NURBS. To solve space-time equations, Newton's method with two stage predictor multicorrector algorithm is employed. The code is parallelized using MPI. The computational domain is a rectangular box of size $2{\pi}{\times}2{\times}4/3{\pi}$ in the streamwise, wall normal and spanwise direction. Mean velocity profiles and velocity fluctuations are compared with the data of DNS. The results agree well with those of DNS and other traditional LES.

• 한국전산유체공학회:학술대회논문집
• /
• 2009.11a
• /
• pp.56-59
• /
• 2009

In the present work, LES with new variational multiscale method is conducted on the fully developed channel flow with Reynolds number is 180 based on the friction velocity and the channel half width. Incompressible Navier-Stokes equations are integrated using finite element method with the basis function of NURBS. To solve space-time equations, Newton's method with two stage predictor multicorretor algorithm is employed. The code is parallelized using MPI. The computational domain is a rectangular box of size $2{\pi}{\times}2{\times}4/3{\pi}$ in the streamwise, wall normal and spanwise direction. Mean velocity profiles and velocity fluctuations are compared with the data of DNS. The results agree well with those of DNS and other traditional LES.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2001.07a
• /
• pp.94-94
• /
• 2001

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.18 no.2
• /
• pp.479-491
• /
• 1994

Discrete vortex method was applied for simulating an active control of turbulent leading- edge separation bubble. The leading-edge separation zone was perturbed by a time-dependent sinusoidal perturbation of different frequencies and levels. In order to describe the local sinusoidal perturbation at the separation point, a source pulsation vortex technique was proposed. The present two-dimensional vortex simulations were qualitatively compared with the experimental results for a blunt circular cylinder, where perturbation was introduced along the square-cut leading edge of the cylinder $(Kiya et al.^{(6,7)}).$ It was found that the reattachment length attained a minimum point at low levels of perturbation and two minima at a moderate higher perturbation frequency. The effects of local perturbation on the evolution of leading-edge separation bubble were scrutinized by comparing the perturbed flow with the natural flow. These comparisons were made for the distributions of mean velocity and its velocity fluctuations, intermittency and wall velocity. The motions of instantaneous reattachment in the space-time domain were demonstrated, which were also compared with the experimental findings. In order to investigate the reduction mehanism of reattachment length in the separation bubble, various cross-correlations for velocity and pressure and the relevant convection velocities were evaluated. It was observed that the convection velocity was closely associated with its corresponding pulsationg frequency.

• Wind and Structures
• /
• v.37 no.4
• /
• pp.255-274
• /
• 2023

The aim of this paper is to enhance the accuracy of predicting time-averaged external surface pressures on low-rise buildings by utilizing Computational Fluid Dynamics (CFD) simulations. To achieve this, benchmark studies of the Silsoe cube and the Texas Tech University (TTU) experimental building are employed for comparison with simulation results. The paper is structured into three main sections. In the initial part, an appropriate domain size is selected based on the precision of mean pressure coefficients on the windward face of the cube, utilizing Reynolds Averaged Navier-Stokes (RANS) turbulence models. Subsequently, recommendations regarding the optimal computational domain size for an isolated building are provided based on revised findings. Moving on to the second part, the Silsoe cube model is examined within a horizontally homogeneous computational domain using more accurate turbulence models, such as Large Eddy Simulation (LES) and hybrid RANS-LES models. For computational efficiency, transient simulation settings are employed, building upon previous studies by the authors at the Windstorm Impact, Science, and Engineering (WISE) Lab, Louisiana State University (LSU). An optimal meshing strategy is determined for LES based on a grid convergence study. Three hybrid RANS-LES cases are investigated to achieve desired enhancements in the distribution of mean pressure coefficients on the Silsoe cube. In the final part, a 1:10 scale model of the TTU building is studied, incorporating the insights gained from the second part. The generated flow characteristics, including vertical profiles of mean velocity, turbulence intensity, and velocity spectra (small and large eddies), exhibit good agreement with full-scale (TTU) measurements. The results indicate promising roof pressures achieved through the careful consideration of meshing strategy, time step, domain size, inflow turbulence, near-wall treatment, and turbulence models. Moreover, this paper demonstrates an improvement in mean roof pressures compared to other state-of-the-art studies, thus highlighting the significance of CFD simulations in building aerodynamics.