• 설비공학논문집
• /
• 제15권7호
• /
• pp.557-562
• /
• 2003

Flue gas from apartment heating gas boiler is exhausted outside through an exhaust duct mounted horizontally in a vertical row on the wall. The flue gas includes nitrogen-oxides (NOx) and carbon monoxide. To investigate the possible entrainment of the flue gas into the apartments through the windows, a large eddy simulation (LES) based numerical method is utilized. Distribution of the velocity intensity and temperature around the exhaust ducts is presented for three numerical parameters: exhaust velocity, temperature of the flue gas, and exhaust duct length. The flow field visualized with particles inserted at the ends of the ducts is also presented. The results clearly show that the exhausted flue gas may flow into the apartments when the windows are open.

• 한국안전학회지
• /
• 제18권2호
• /
• pp.40-45
• /
• 2003

To evaluate the smoke control systems, the large eddy simulation turbulence model based Fire Dynamics Simulate was applied to a 2m $\times$ 2m $\times$ 2.4m room with an opening. The smoke removal rate was investigated for three different smoke control systems: ventilation, extraction and pressurization. When the opening was closed, the smoke removal rates of the smoke control systems were almost the same as expected. The pressurization system showed a lower smoke removal rate compared with the other two smoke control systems for the room with the opening, and hence the pressurization system might not be efficient for a place with large openings. It was shown that the lower extraction flowrate is, the longer time the ventilation system requires to remove smoke. From these results, the ventilation system is recommended for subway stations where several large openings exist.

• 한국안전학회지
• /
• 제18권3호
• /
• pp.34-38
• /
• 2003

The large eddy simulation based Fire Dynamics Simulator was utilized to investigate the effects of the size of an opening on smoke removal performance for the three smoke control systems-ventilation purge, and extraction. Three different opening sizes, $r_A$=1, 2, and 3 were investigated while the flow rate remained 0.75 $m^3/s$ at the inlet or outlet depending on the systems. Increase of the opening size did not give a significant difference in the smoke removal rate for the three smoke control systems, though the increasing opening size slightly improved smoke removal. The extraction system was shown the best smoke control system, and the purge system yielded low performance compared to the other two systems for all the different opening sizes.

• 한국전산유체공학회지
• /
• 제17권2호
• /
• pp.93-99
• /
• 2012

The present study focuses on the case of developing flow with in a channel containing a long array of sinusoidal waves (2a/${\lambda}$=0.1, ${\lambda}$=h, ${\lambda}$ is the wavelength, 2a is the wave height, h is the mean channel depth) at the bottom wall. The Reynolds number defined with channel height, h and the mean velocity, U, is Re=6,700. The channel is sufficiently long such that transition is completed and the flow is fully developed over the downstream half of the channel. For the case of an incoming steady flow with no resolved turbulence, the instantaneous flow fields in the transition region are characterized by the formation of arrays of highly-organized large-scale hairpin vortices whose dimensions scale with that of the roughness elements. The paper explains the mechanism for the formation of these arrays of hairpin vortices and shows these eddies play the primary role in the formation of the large-scale streaks of high and low velocity over the wavy wall region. The presence of resolved turbulence in the incoming flow, reduces the streamwise distance needed for the streaks to develop over the wavy region, but does not affect qualitatively the transition process. In the fully-developed region, isolated and trains of large-scale hairpins play an important role in the dynamics of the streaks over the wavy wall.

• 한국철도학회논문집
• /
• 제10권2호
• /
• pp.224-230
• /
• 2007

The performance and applicability of FDS code is analyzed for flow simulation in railway tunnel. FDS has been built in NIST(USA) for simulation of fire-driven flow. RANS and DNS's results are compared with FDS's. AJL non-linear ${\kappa}-{\epsilon}$[7,8] model is employed to calculate the turbulent flow for RANS. DNS data by Moser et al.[9] are used to prove the FDS's applicability in the near wall region. Parallel plate is used for simplified model of railway tunnel. Geometrical variables are non-dimensionalized by the height (H) of parallel plate. The length of streamwise direction is 50H and the length of spanwise direction is 5H. Selected Re numbers are 10,667 for turbulent flow and 133 for laminar low. The characteristics of turbulent boundary layer are introduced. AJL model's predictions of turbulent boundary layer are well agreed with DNS data. However, the near wall turbulent boundary layer is not well resolved by FDS code. Slip conditions are imposed on the wall but wall functions based on log-law are not employed by FDS. The heavily dense grid distribution in the near wall region is necessary to get correct flow behavior in this region for FDS.

• 한국전산유체공학회지
• /
• 제17권3호
• /
• pp.87-92
• /
• 2012

The presence of a layer of vegetation which is relevant in river engineering or coastal engineering can modify the overall flow resistance, turbulent characteristics of flow. The patch of vegetation can be modelled and studied in a simple porous cylinder by previous researchers. Fully three dimensional Large Eddy Simulation is conducted in flow past a porous cylinder with a solid volume fraction (SVF) 0f 20%. The porous cylinder of diameter D contains 89 smaller cylinders which diameter is 0.048D in a regular staggered way. Reynolds number based on porous cylinder diameter D and the bulk velocity is 10,000. The large scale shedding is qualitatively similar to the one observed in the non-porous case (SVF=100%). The difference in the dynamics of the separated shear layer and the streamwise flow penetrating through the porous cylinder are compared with those in the non-porous cylinder. In particular, the wake billows form a larger distance from the back of the porous cylinder.

• 한국전산유체공학회지
• /
• 제17권3호
• /
• pp.1-10
• /
• 2012

Large Eddy Simulation(LES) of turbulent mass transfer in fully developed turbulent pipe flow has been performed to study the effect of Reynolds number on the concentration fields at $Re_{\tau}=180$, 395, 590 based on friction velocity and pipe radius. Dynamic subgrid-scale models for the turbulent subgrid-scale stresses and mass fluxes were employed to close the governing equations. Fully developed turbulent pipe flows with constant mass flux imposed at the wall are studied for Sc=0.71. The mean concentration profiles and turbulent intensities obtained from the present LES are in good agreement with the previous numerical and experimental results currently available. To show the effects of Reynolds number on the turbulent mass transfer, the mean concentration profile, root-mean-square of concentration fluctuations, turbulent mass fluxes, cross-correlation coefficient, turbulent diffusivity and turbulent Schmidt number are presented.

• 한국전산유체공학회지
• /
• 제17권3호
• /
• pp.59-67
• /
• 2012

Large Eddy Simulation(LES) of turbulent mass transfer in fully developed turbulent pipe flow has been performed to study the effect of Reynolds number on the concentration fields at $Re_{\tau}=180$, 395, 590 based on friction velocity and pipe radius. Dynamic subgrid-scale models for the turbulent subgrid-scale stresses and mass fluxes were employed to close the governing equations. Fully developed turbulent pipe flows with constant mass flux imposed at the wall are studied for Sc=0.71. The mean concentration profiles and turbulent intensities obtained from the present LES are in good agreement with the previous numerical and experimental results currently available. The effects of Reynolds number on the turbulent mass transfer are identified in the higher-order statistics(Skewness and Flatness factor) and instantaneous concentration fields. The budgets of turbulent mass fluxes and concentration variance were computed and analyzed to elucidate the effect of Reynolds number on turbulent mass transfer. Furthermore, to understand the correlation between near-wall turbulence structure and concentration fluctuation, we present an octant analysis in the vicinity of the pipe wall.

• 한국전산유체공학회지
• /
• 제18권1호
• /
• pp.49-57
• /
• 2013

Analysis of fluid-structure interaction for two nearby underwater vehicles immersed in the sea is quite challenging because simulation of flow around them is very difficult due to the complexity of underwater vehicle shapes. The conventional approach using body-fitted or unstructured grids demands much time in dynamic grid generation, and yields slow convergence of solution. Since an analysis of fluid-structure interaction must be based on accurate simulation results, a more efficient way of simulating flow around underwater vehicles, without sacrificing accuracy, is desirable. An immersed boundary method facilitates implementation of complicated underwater-vehicle shapes on a Cartesian grid system. An LES modeling is also incorporated to resolve turbulent eddies. In this paper, we will demonstrate the effectiveness of the immersed boundary method we adopted, by presenting the simulation results on the flow around a modeled high-speed underwater vehicle interacting with a modeled low-speed one.

• 대한조선학회논문집
• /
• 제41권4호
• /
• pp.1-8
• /
• 2004

As a computer has been continuously progressed to reduce R&D time and cost, the study of the flow physics has been significantly relied on the numerical method. Recently, Large Eddy Simulation(LES) has been widely used in CFD community to accurately capture the turbulent flows. The LES code requires high accuracy in time, as well as in space. Also, it should have strong robustness to ensure the convergence in various complicated flows. The objective of the present work is to develop a base code for LES simulation, having 2$^{nd}$ order accuracy in time and 4$^{th}$ order accuracy in space. To achieve the present objective, the four-step fractional step method was enhanced by adopting compact Pade'scheme. The standard Smagorinsky model was implemented for the first stage of the present code development. The flows over a cylinder and an airfoil were successfully simulated. and an airfoil were successfully simulated.