• 한국분말재료학회지
• /
• 제18권4호
• /
• pp.365-371
• /
• 2011

In this study, gradient porous Al-Cu sintered body was fabricated by powder metallurgy processing. Al-Cu powder mixtures were prepared by low energy ball milling with various milling time. After ball milling for 3h, the shape of powder mixtures changed to spherical type with size of 100~500 ${\mu}m$. Subsequently, Al-Cu powder mixtures were classified (under 150, 150~300 and over 300 ${\mu}m$) and compacted (20, 50 and 100 MPa). Then, they were sintered at $600^{\circ}C$ for various holding time (10, 30, 60 and 120 min) in $N_2$ atmosphere. The sintered bodies had 32~45% of porosity. As a result, the optimum holding time was determined to be 60 min at $600^{\circ}C$ and sintered bodies with various porosity were obtained by controlling the compacting pressure.

• 한국정밀공학회지
• /
• 제18권9호
• /
• pp.110-121
• /
• 2001

During hot pipe bending using induction heating, the wall of bending outside is thinned by tensile stress. In design requirement, the reduction of wall thickness is not allowed to exceed 12.5%. So in this study, two methods of bending, one is loading of reverse moment and the other is loading of temperature gradient, have been investigated to design pipe bending process that satisfy design requirements. For this purpose, finite element analysis with a bending radius 2Do(outer diameter of pipe) has been performed to calculate proper reverse moment and temperature gradient to be applied. Induction heating process has been analyzed to estimate influence of heating process parameters on heating characteristic by finite difference method. Then pipe bending experiments have been performed for verification of finite element and finite difference analysis results. Experimental results are in good agreement with the results of simulations.

• 한국초전도ㆍ저온공학회논문지
• /
• 제18권2호
• /
• pp.21-24
• /
• 2016

In recent years, the iron-dominated high-temperature superconductor (HTS) quadrupole magnets are being developed for heavy ion accelerators. Field analyses for iron-dominated quadrupole magnets were based on the normal-conducting (NC) quadrupole magnet early in the development for accelerators. Some conclusions are still in use today. However, the magnetic field of iron-dominated HTS quadrupole magnets cannot fully follow these conclusions. This study established an HTS quadrupole magnet model and an NC quadrupole magnet model, respectively. The harmonic characteristics of two magnets were analyzed and compared. According to the comparison, the conventional iron-dominated quadrupole magnets can be designed for maximum field gradient; the HTS quadrupole magnet, however, should be considered with varying field gradient. Finally, the HTS quadrupole magnet was designed for the changing field gradient. The field quality of the design was improved comparing with the result of the previous study. The new design for the HTS quadrupole magnet has been suggested.

• 한국지구물리탐사학회:학술대회논문집
• /
• 한국지구물리탐사학회 2007년도 공동학술대회 논문집
• /
• pp.329-336
• /
• 2007

For scaling of the gradient of misfit function, we develop a new pseudo-Hessian matrix constructed by combining amplitude field and pseudo-Hessian matrix. Since pseudo- Hessian matrix neglects the calculation of the zero-lag auto-correlation of impulse responses in the approximate Hessian matrix, the pseudo-Hessian matrix has a limitation to scale the gradient of misfit function compared to the approximate Hessian matrix. To validate the new pseudo- Hessian matrix, we perform frequency-domain elastic full waveform inversion using this Hessian matrix. By synthetic experiments, we show that the new pseudo-Hessian matrix can give better convergence to the true model than the old one does. Furthermore, since the amplitude fields are intrinsically obtained in forward modeling procedure, we do not have to pay any extra cost to compute the new pseudo-Hessian. We think that the new pseudo-Hessian matrix can be used as an alternative of the approximate Hessian matrix of the Gauss-Newton method.

• 대한전자공학회논문지SP
• /
• 제49권2호
• /
• pp.130-136
• /
• 2012

본 논문에서는 등간격 선형 배열 센서에서 초점빔형성을 이용한 근접장 거리 추정 시에 방위각 오차의 영향을 극복하기 위하여, 거리 방향 탐색 시에 일정 범위 안의 방위각 오차를 보정하여 추정 성능을 향상시키는 기법을 제안하였다. 방위각 오차는 탐색 거리가 주어졌을 때 경사법을 기반으로 빔형성기 출력값이 최대가 되도록 보정하게 된다. 모의실험 결과를 통해, 제안된 기법이 주엽의 폭보다 작은 방위각 오차를 보정하여 거리 추정 오차를 방위각 오차가 없는 경우와 같은 수준으로 감소시킬 수 있음을 확인하였다.

• 한국해양공학회지
• /
• 제34권6호
• /
• pp.442-453
• /
• 2020

Underwater color images suffer from low visibility and color cast effects caused by light attenuation by water and floating particles. This study applied single image enhancement techniques to enhance the quality of underwater images and compared their performance with real underwater images taken in Korean waters. Dark channel prior (DCP), gradient transform, image fusion, and generative adversarial networks (GAN), such as cycleGAN and underwater GAN (UGAN), were considered for single image enhancement. Their performance was evaluated in terms of underwater image quality measure, underwater color image quality evaluation, gray-world assumption, and blur metric. The DCP saturated the underwater images to a specific greenish or bluish color tone and reduced the brightness of the background signal. The gradient transform method with two transmission maps were sensitive to the light source and highlighted the region exposed to light. Although image fusion enabled reasonable color correction, the object details were lost due to the last fusion step. CycleGAN corrected overall color tone relatively well but generated artifacts in the background. UGAN showed good visual quality and obtained the highest scores against all figures of merit (FOMs) by compensating for the colors and visibility compared to the other single enhancement methods.

• 한국환경복원기술학회지
• /
• 제3권2호
• /
• pp.47-52
• /
• 2000

This study was carried out to analyze the effects of elapsed years and physical properties on fill-slope alignment of forest road. For the study, 21 forest roads in Puy$\check{o}$-gun and Asan-shi of Chungchungnam-do were selected and fifteen factors that might influence on fill-slope alignment were analyzed. The major forms of alignment of the fill-slope in the time of construction were concave and compound. But, with the elapsed years, the concave and compound forms have decreased and linear and convex forms have increased. And fill-slope alignment was strongly related with fill-slope distance, fill-slope gradient and ground gradient among the physical properties.

• Journal of applied mathematics & informatics
• /
• 제32권5_6호
• /
• pp.707-719
• /
• 2014

In this paper, we investigate a variational discretization approximation of elliptic optimal control problems with control constraints. The state and the co-state are approximated by piecewise linear functions, while the control is not directly discretized. By using some proper intermediate variables, we derive a second-order convergence in $L^2$-norm and superconvergence between the numerical solution and elliptic projection of the exact solution in $H^1$-norm or the gradient of the exact solution and recovery gradient in $L^2$-norm. Then we construct a posteriori error estimates by using the superconvergence results and do some numerical experiments to confirm our theoretical results.

• 한국신재생에너지학회:학술대회논문집
• /
• 한국신재생에너지학회 2005년도 제17회 워크샵 및 추계학술대회
• /
• pp.587-590
• /
• 2005

A detailed geothermal characteristics survey with numerical simulations of the heat transfer in a site for ground source heat pump system is necessary for deploying a shallow geothermal utilization system. Density, specific heat, thermal diffusivity, and thermal conductivity are measured on 91 core samples from a 300 m deep borehole in KIGAM(Korea Institute of Geoscience and Mineral Resources). The heat flow is estimated from the thermal gradient and average thermal conductivity and the correlation between fracture system and hydraulic conductivity is analyzed. From the obtained ground information of the study site the performance of the ground heat pump system can be analyzed with some detailed numerical simulations for seasonal heat pump operation skill and optimal system design techniques.

• 한국멀티미디어학회논문지
• /
• 제21권10호
• /
• pp.1142-1149
• /
• 2018

Recently, computer vision application is increasing by using CNN which is one of the deep learning algorithms. However, CNN does not provide perfect classification performance due to gradient vanishing problem. Most of CNN algorithms use an activation function called ReLU to mitigate the gradient vanishing problem. In this study, four activation functions that can replace ReLU were applied to four different structural networks. Experimental results show that ReLU has the lowest performance in accuracy, loss rate, and speed of initial learning convergence from 20 experiments. It is concluded that the optimal activation function varied from network to network but the four activation functions were higher than ReLU.