• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.40 no.9
• /
• pp.597-604
• /
• 2016

A parallel algorithm of bi-conjugate gradient method was developed based on CUDA for parallel computation of the incompressible Navier-Stokes equations. The governing equations were discretized using splitting P2P1 finite element method. Asymmetric stenotic flow problem was solved to validate the proposed algorithm, and then the parallel performance of the GPU was examined by measuring the elapsed times. Further, the GPU performance for sparse matrix-vector multiplication was also investigated with a matrix of fluid-structure interaction problem. A kernel was generated to simultaneously compute the inner product of each row of sparse matrix and a vector. In addition, the kernel was optimized to improve the performance by using both parallel reduction and memory coalescing. In the kernel construction, the effect of warp on the parallel performance of the present CUDA was also examined. The present GPU computation was more than 7 times faster than the single CPU by double precision.

• Proceedings of the KSME Conference
• /
• 2001.06d
• /
• pp.377-381
• /
• 2001

Film cooling characteristics has been investigated numerically with the aid of FLUENT software for the sunk or the lifted upstream wall from the slot injection exit. In this study, with the fixed blowing ratio of 1 and the fixed coolant injection angle of $30^{\circ}$, the downstream flow field and the downstream temperature field were examined in terms of velocity vector, turbulent kinetic energy, temperature contours, and downstream wall temperature. Upstream wall was sunk or lifted from 1d to 5d(d=slot width). The result shows that the up-Id upstream wall has the best film cooling performance. This is due to the fact that the up-1d upstream wall configuration reduces velocity gradient just enough to minimize the turbulent mixing between the mainstream and the coolant just off the slot exit.

• Proceedings of the KIEE Conference
• /
• 2005.10b
• /
• pp.3-5
• /
• 2005

Vessel boundary detection and modeling is a difficult but a necessary task in analyzing the mechanics of inflammation and the structure of the microvasculature. In this paper we present a method of analyzing the structure by means of an active contour model(using GVF Snake) for vessel boundary detection and 3D reconstruction. For this purpose we used a virtual vessel model and produced a phantom model. From these phantom images we obtained the contours of the vessel by GVF Snake and then reconstructed a 3D structure by using the coordinates of snakes.

• 제어로봇시스템학회:학술대회논문집
• /
• 2002.10a
• /
• pp.57.4-57
• /
• 2002

$\textbullet$ Image segmentation is an essential technique of image analysis. In spite of the traditional issues in contour initialization and boundary concavities, active contour models(snakes) are popular and known as successful methods for segmentation. $\textbullet$ We could find in experiment that snake using Gaussian External Force is fast in time but low in accuracy and snake using Gradient Vector Flow by Chenyang Xu and Jerry L. Prince is high in accuracy but slow in time. $\textbullet$ In this paper, we presented a new active contour model, GGF snake, for segmentation of endoscopic image. Proposed GGF snake made up for the defects of the traditional snakes in contour initialization and boundary...

• Journal of Power System Engineering
• /
• v.21 no.2
• /
• pp.48-56
• /
• 2017

The steady-state, incompressible and three-dimensional numerical analysis was carried out to evaluate the velocity fields around the seabed tiller used for the improvement of the seabed soil and the pulling force and buoyancy generated by driving the seabed tiller. The turbulence model used in this study is a realizable $k-{\varepsilon}$ well known to be excellent for predicting the performance of the flow separation and recirculation flow as well as the boundary layer with rotation and strong back pressure gradient. As a results, a typical vortex pair appears near the adjacent rotor vane tip. When the current is stopped, there is no force when pulling the seabed tiller, but when the current flows at 1.2 knots, the force acts on the downstream side and the pulling force is much greater. In stationary currents, the buoyancy of the seabed tiller acts more strongly towards the seabed as the number of rotations of the rotor increases, but acts more strongly toward the sea surface at 1.2 knots of current.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.18 no.10
• /
• pp.2585-2596
• /
• 1994

Many particle filled materials like Poweder/Binder mixtures for poweder injection moldings, have complicated rheological behaviors such as an yield stress and slip phenomena. In the present study, numerical simulation programs via a finite element method and a finite difference method were developed for the quasi-three-dimensional flows and the two-dimensional flow models, respectively, with the slip phenomena taken into account in terms of a slip velocity. In order to qualitatively understand the slip effects, typical numerical results such as vector plots, pressure contours in the cross-channel plane, and isovelocity controus for the down-channel direction were discussed with respect to various slip coefficients. Slip velocities along the boudary surfaces were also investigated to find the effects of the slip coefficient and processing conditions on the overall flow behavior. Based on extensive numerical calculations varying the slip coefficients, pressure gradient, aspect ratio, and power law index, the screw characteristics of the extrusion process were studied in particular with comparisons between the slip model and non-slip model.

• International Journal of Aeronautical and Space Sciences
• /
• v.3 no.1
• /
• pp.74-85
• /
• 2002

The numerical experiment has been conducted to investigate the unsteady shock wave reflecting phenomena. The cell-vertex finite-volume, Roe's upwind flux difference splitting method with unstructured grid is implemented to solve unsteady Euler equations. The $4^{th}$-order Runge-Kutta method is applied for time integration. A linear reconstruction of the flux vector using the least-square method is applied to obtain the $2^{nd}$-order accuracy for the spatial derivatives. For a better resolution of the shock wave and slipline, the dynamic grid adaptation technique is adopted. The new concept of grid adaptation technique, which is much simpler than that of conventional techniques, is introduced for the current study. Three error indicators (divergence and curl of velocity, and gradient of density) are used for the grid adaptation procedure. Considering the quality of the solution and the numerical efficiency, the grid adaptation procedure was updated up to $2^{nd}$ level at every 20 time steps. For the convenience of comparison with other experimental and analytical results, the case of interaction between the straight incoming shock wave and a sharp wedge is simulated for various flow conditions. The numerical results show good agreement with other experimental and analytical results, in the shock wave reflecting structure, slipline, and the trajectory of the triple points. Some critical cases show disagreement with the analytical results, but these cases also have been proven to show hysteresis phenomena.

• International journal of advanced smart convergence
• /
• v.13 no.2
• /
• pp.48-60
• /
• 2024

With the exponential growth of satellite data utilization, machine learning has become pivotal in enhancing innovation and cybersecurity in satellite systems. This paper investigates the role of machine learning techniques in identifying and mitigating vulnerabilities and code smells within satellite software. We explore satellite system architecture and survey applications like vulnerability analysis, source code refactoring, and security flaw detection, emphasizing feature extraction methodologies such as Abstract Syntax Trees (AST) and Control Flow Graphs (CFG). We present practical examples of feature extraction and training models using machine learning techniques like Random Forests, Support Vector Machines, and Gradient Boosting. Additionally, we review open-access satellite datasets and address prevalent code smells through systematic refactoring solutions. By integrating continuous code review and refactoring into satellite software development, this research aims to improve maintainability, scalability, and cybersecurity, providing novel insights for the advancement of satellite software development and security. The value of this paper lies in its focus on addressing the identification of vulnerabilities and resolution of code smells in satellite software. In terms of the authors' contributions, we detail methods for applying machine learning to identify potential vulnerabilities and code smells in satellite software. Furthermore, the study presents techniques for feature extraction and model training, utilizing Abstract Syntax Trees (AST) and Control Flow Graphs (CFG) to extract relevant features for machine learning training. Regarding the results, we discuss the analysis of vulnerabilities, the identification of code smells, maintenance, and security enhancement through practical examples. This underscores the significant improvement in the maintainability and scalability of satellite software through continuous code review and refactoring.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.29 no.11 s.242
• /
• pp.1265-1276
• /
• 2005

The elliptic conceptual second moment models for turbulent heat fluxes, which are proposed on the basis of elliptic-blending and elliptic-relaxation equations, are applied to calculate the combined forced and natural turbulent convection in a vertical plane channel. The models satisfy the near-wall balance between viscous diffusion, viscous dissipation and temperature-pressure gradient correlation, and also have the characteristics of approaching its respective conventional high Reynolds number model far away from the wall. Also the models are closely linked to the elliptic blending model which is used for the prediction of Reynolds stress. In order to calibrate the heat flux models, firstly, the distributions of mean temperature and scala flux in fully developed channel flow with constant wall difference temperature are solved by the present models. The buoyancy effect on the turbulent characteristics including the mean velocity and temperature, the Reynolds stress tensor, and the turbulent heat flux vector are examined. In the opposing flow, the turbulent transport is greatly enhanced with both the Reynolds stresses and the turbulent heat fluxes being remarkably increased; whereas, in the aiding flow, the opposite change is observed. The results of prediction are directly compared to the DNS to assess the performance of the model predictions and show that the behaviors of the turbulent heat transfer in the whole flow region are well captured by the present models.

• Journal of the Korean Society of Visualization
• /
• v.20 no.3
• /
• pp.94-101
• /
• 2022

The evaluation of the drag and lift as the aerodynamic performance of airfoils is essential. In addition, the analysis of the velocity and pressure fields is needed to support the physical mechanism of the force coefficients of the airfoil. Thus, the present study aims at establishing two different deep learning models to predict force coefficients and flow fields of the airfoil. One is the convolutional neural network (CNN) model to predict drag and lift coefficients of airfoil. Another is the Encoder-Decoder (ED) model to predict pressure distribution and velocity vector field. The images of airfoil section are applied as the input data of both models. Thus, the computational fluid dynamics (CFD) is adopted to form the dataset to training and test of both CNN models. The models are established by the convergence performance for the various hyperparameters. The prediction capability of the established CNN model and ED model is evaluated for the various NACA sections by comparing the true results obtained by the CFD, resulting in the high accurate prediction. It is noted that the predicted results near the leading edge, where the velocity has sharp gradient, reveal relatively lower accuracies. Therefore, the more and high resolved dataset are required to improve the highly nonlinear flow fields.