• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.4 no.2
• /
• pp.91-107
• /
• 1992

A numerical formulation is developed to solve the linear three-dimensional hydrodynamic equations which describes wind induced flows in a homogeneous shelf sea. The hydmdynamic equations are at the outset separated into two systems. namely, an equation containing the gradient of sea surface elevation and the mean flow (external mode) and an equation describing the deviation from the mean flow (internal mode). The Galerkin method is then applied to the internal mode equation. The eigenvalues are determined from the eigenvalue problem involving the vertical eddy viscosity subject to a homogeneous boundary condition at the surface and a sheared boundary condition at the sea bed. The model is tested in a one-dimensional channel with uniform depth under a steady, uniform wind. The analytical velocity profile by Cooper and Pearce (1977) using a constant vertical eddy viscosity in channels of infinite and finite length is chosen as a benchmark solution. The model is also tested in a homogeneous, rectangular basin with constant depth under a steady, uniform wind field (the Heaps' Basin of the North Sea scale).

• Journal of Korean Society for Atmospheric Environment
• /
• v.15 no.3
• /
• pp.281-289
• /
• 1999

A new E-$\varepsilon$ turbulence numerical model is proposed for analysing the turbulent air flow field within are urban street canyon. In this model the equations of eddy viscosity and energy dissipation ae reformed by considering the Kolmogorov time scale and streamline curvature effect. Application results of the new E-$\varepsilon$ model have been compared with those of standard E-$\varepsilon$ model and Yang and Shih's one, which are commonly used ones in engineering fields, and with field experiment results of DePaul and Sheih. The new model appears to be generally superior to other both models in the prediction of an air flow field within street canyon.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.22 no.5
• /
• pp.333-343
• /
• 2010

In Mokpo coastal zone, the characteristics showing ebb-dominant tidal flow was confirmed by analysis of observed tide and tidal currents, Physical factors occurring ebb-dominant flow were reviewed. Influence of critical depth for drying, bottom shear stress, coastal reclamation, tidal amplitude, nonlinear tide, and eddy viscosity on the change of ebb-dominant flow was investigated by applying a two-dimensional circulation model. The simulation results for a variety of conditions showed that eddy viscosity and critical depth for drying does little or no impact on the generation of asymmetric flow. Strong bottom friction stress makes ebb-dominant flow clearly. Change of tidal flat into land swells ebb- dominant flow, and change of tidal flat into sea disappears ebb-dominant flow. Nonlinear tides play a decisive role in the generation of asymmetrical tidal flow. Non-linear tides should be included in the open boundary conditions of hydrodynamic modeling in the Mokpo coastal zone.

• Journal of Ocean Engineering and Technology
• /
• v.4 no.1
• /
• pp.49-54
• /
• 1990

This study describes the application of a two dimensional depth integrated numerical model. The explict scheme of finite difference method had been applied to the model of circulation. The nonlinear terms showed a slight difference for the variations of water elevation when calculated grid was small. They were also found to be minor when calculated grid size was increased. For verification of the numerical model, numerical results were compared with predicted values and field data. In the model, the effect of nonlinear advective terms proved not to be significant.

• Magazine of the Korean Society of Agricultural Engineers
• /
• v.36 no.3
• /
• pp.144-154
• /
• 1994

An efficient tidal model, TIDE which is an iterative type, nonlinear finite element model has developed for the analysis of the tidal movement in the coastal area which is characterized by irregular boundaries and bottom topography. Traditional time domain finite element models have been in difficulties with requirement for high eddy viscosity coefficients and small time steps to insure numerical instability. These problems are overcome by operating in the frequency domain with an elaborate grid system by combining the triangular and quadrilateral shape grids. Furthermore, in order to handle non-linearity which will be more significant in the shallow region, an iterative scheme with least square error minimization algorithm has been implemented in the model. The results of TIDE model are agreed with the analytical solutions in a rectangular channel under the condition of tidal waves entering the channel closed at one end.

• Proceedings of the KSME Conference
• /
• 2003.04a
• /
• pp.2162-2167
• /
• 2003

It is well-known that high anisotropic characteristic of turbulent flow field is dominant inside tip leakage vortex. This anisotropic nature of turbulence invalidates the use of the conventional isotropic eddy viscosity turbulence model based on the Boussinesq assumption. In this study, to check whether an anisotropic turbulence model is superior to the isotropic ones or not, the results obtained from steady-state Reynolds averaged Navier-Stokes simulations based on the RNG ${\kappa}-{\varepsilon}$ and the Reynolds stress model in two test cases, such as a linear compressor cascade and a forward-swept axial-flow fan, are compared with experimental data. Through the comparative study of turbulence models, it is clearly shown that the Reynolds stress model, which can express the production term and body-force term induced by system rotation without any modeling, should be used to predict the complex tip leakage flow, including the locus of tip leakage vortex center, quantitatively.

• Proceedings of the KSME Conference
• /
• 2002.08a
• /
• pp.23-26
• /
• 2002

An evaluation of one algebraic and two one-equation eddy viscosity-transport turbulence closure models as implemented to the CFDS(Characteristic Flux Difference Splitting) scheme is presented for the efficient computation of the turbulent flow. Comparisons of Baldwin-Lomax model as algebraic turbulence model and Baldwin-Barth and Spalart-Allmaras model as one-equation turbulence model are presented for three test cases for 3-dimensional flow. The numerical result of the CFDS schem is examined through comparison with the experimental data.

• Korean Journal of Air-Conditioning and Refrigeration Engineering
• /
• v.11 no.2
• /
• pp.157-166
• /
• 1999

Numerical calculations of turbulent natural convection in an enclosure of the 20 kYA oil-immersed transformer model are presented. The transformer is modelled as two concentric cylinders with different heights and diameters. The thermal boundary layers are well represented in the temperature distributions along the wall of the transformer model. The flow stratification between the hot and cold walls can not be seen in the transformer model. The turbulence eddy viscosity has its maximum at the center of the core and its maximum values at the top of the core are larger than those at the bottom of the core.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.14 no.2
• /
• pp.355-365
• /
• 1994

An analytical model of return flow is presented outside the surf zone. The governing equation is derived from the Navier-Stokes equation and the continuity. Each term of the governing equation is evaluated by the ordering analysis. Then the infinitesimal terms, i.e. the turbulent normal stress, the squared vertical velocity of water particle and the streaming velocity, are neglected. The driving forces of return flow are calculated using the linear wave theory for the shallow water approximation. Especially, the space derivative of local wave heights is described considering a shoaling coefficient. The vertical distribution of eddy viscosity is discussed to the customary types which are the constant, the linear function and the exponential function. Each coefficient of the eddy viscosities which sensitively affect the precision of solutions is uniquely decided from the additional boundary condition which the velocity becomes zero at the wave trough level. Also the boundary conditions at the bottom and the continuity relation are used in the integration of the governing equation. The theoretical solutions of present model are compared with the various experimental results. The solutions show a good agreement with the experimental results in the case of constant or exponential function type eddy viscosity.

• Journal of computational fluids engineering
• /
• v.19 no.4
• /
• pp.29-36
• /
• 2014

A computational study of a thermal striping in the upper plenum of PGSFR(Prototype Generation-IV Sodium-cooled Fast Reactor) being developed at the KAERI(Korea Atomic Energy Research Institute) is presented. The LES(Large Eddy Simulation) approach is employed for the simulation of thermal striping in the upper plenum of the PGSFR. The LES is performed using the WALE (Wall-Adapting Local Eddy-viscosity) model. More than 19.7 million unstructured elements are generated in upper plenum region of the PGSFR using the CFX-Mesh commercial code. The time-averaged velocity components and temperature field in the complicated upper plenum of the PGSFR are presented. The time history of temperature fluctuation at the eight locations of solid walls of UIS(Upper Internal Structure) and IHX(Intermediate Heat eXchanger) are additionally stored. It has been confirmed that the most vulnerable regions to thermal striping are the first plate of UIS. From the temporal variation of temperature at the solid walls, it was possible to find the locations where the thermal stress is large and need to assess whether the solid structures can endure the thermal stress during the reactor life time.