• Journal of The Korean Society of Agricultural Engineers
• /
• v.55 no.6
• /
• pp.37-46
• /
• 2013

Many factors about the stability for the reservoir embankments is determined when the facility is completed. Therefore the initial design of the embankment is important. Many researchers focused the effect of soil parameters although the cross section greatly affects the stability and can be controlled in design step. The objective of this research is to analysis of the effects for the safety factor of slope and seepage according to change cross-section in embankment. As a result, the quantity of seepage decreased as the gradient of downstream slope decreased and was proportional to the height of embankments. There was a linear relationship between the gradient of slope and the safety factor of slope. However the gradient of slope did not affect other side slope. All in a relationship, regressive equations with a high correlation coefficient were calculated and can be applied the simple estimation method of the stability using the cross-section. As results of analyzing the sensitivity, the friction angle and permeability critically effect for the slope stability and the seepage, respectively. The effect of the slope gradient was similar to major soil properties.

• Journal of the Korean Magnetic Resonance Society
• /
• v.10 no.1
• /
• pp.46-58
• /
• 2006

With a simple pulse sequence ($\pi/2$-{gradient, duration T}-acquisition) in solution NMR, detected signal has slowly grown up to percents of the equilibrium magnetization. The source of this unusual resurrection of dephased magnetization after a crushed gradient is cross-correlated effects of radiation damping and the distant dipolar field, which has been demonstrated by a numerical simulation and theoretical analysis.

• 한국가시화정보학회:학술대회논문집
• /
• 2001.12a
• /
• pp.14-22
• /
• 2001

The frequency responses on the sub-pixel evaluation technique were investigated using the Monte-calro Simulation technique. The frequency response by the FFT based cross-correlation gives very good results, however, the gain loss does exist for the small displacement, (less than 0.5 pixel). While, the no gain loss is observed in the Direct Cross-correlation, however, the sub-pixel accuracy was limited to be about 0.1 pixel, i.e., it could not detect the small displacement. To detect the higher accurate sub-pixel displacement, the gradient based technique is the best. For the small interrogation area (e.g., 4x4), only the gradient technique can detect the small displacement correctly.

• Proceedings of the Korean Society For Composite Materials Conference
• /
• 1999.11a
• /
• pp.234-238
• /
• 1999

This study suggests the design of the functionally gradient composite, [0/90/0/core]$_s$ cross-ply laminate, to prevent stress concentration induced from the difference of rigidity between the bone and the artificial hip joint and to reinforce the wear property of the surface and the expectation of their mechanical properties. First, the four-point bending test is done about wet bones and dry bones to know the mechanical properties of the cortical bones. In result, the wet bone shows the viscoelastic behavior and the dry bone shows the elastic behavior. Moreover, we expect the properties of the proposed gradient composites as a function of carbon fiber volume fraction in each layer to apply Halpin-Tsai equation, CLPT(classical laminate plate theory), and Bernoulli beam theory etc. and decide the thickness ratio of each lamina in order to match Young's modulus of the anisotropic cortical bone with the proposed gradient composites.

• Steel and Composite Structures
• /
• v.13 no.1
• /
• pp.71-89
• /
• 2012

The paper investigates beam lateral buckling stability according to linear and non-linear models. Closed form solutions for single-symmetric cross sections are first derived according to a non-linear model considering flexural-torsional coupling and pre-buckling deformation effects. The closed form solutions are compared to a beam finite element developed in large torsion. Effects of pre-buckling deflection and gradient moment on beam stability are not well known in the literature. The strength of singly symmetric I-beams under gradient moments is particularly investigated. Beams with T and I cross-sections are considered in the study. It is concluded that pre-buckling deflections effects are important for I-section with large flanges and analytical solutions are possible. For beams with T-sections, lateral buckling resistance depends not only on pre-buckling deflection but also on cross section shape, load distribution and buckling modes. Effects of pre-buckling deflections are important only when the largest flange is under compressive stresses and positive gradient moments. For negative gradient moments, all available solutions fail and overestimate the beam strength. Numerical solutions are more powerful. Other load cases are investigated as the stability of continuous beams. Under arbitrary loads, all available solutions fail, and recourse to finite element simulation is more efficient.

• Journal of Biomedical Engineering Research
• /
• v.16 no.2
• /
• pp.149-156
• /
• 1995

This paper suggests a new experimental system and protocol of mechanical analysis for arterial cross-section research. So far, most methods of arterial studies have been focused on the deformation measurement in longitudinal and circumferential direction. The deformation in radial direction has been theoretically assumed by Poisson's ratio and/or the incompressibility of arterial wall. Also, the radial gradient of strains are neglected. In fact, the radial deformation and radial gradient of strains against blood pressure are important to be observed in the pathological point of view of artery. Proposed experimental system and protocol are to measure the deformation of cross-sectional artery. Also, this method enables to measure the deformation of anterior, posterior, and side site of cross-sectional area. It is meaningful to correlate the mechanically experimented data with pathological data of athroscIerotic artery.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2001.07a
• /
• pp.628-631
• /
• 2001

Current lead is one of the first proposed devices for the application of High Temperature-Superconductor(HTSC). The current lead provides high current for electrical machine using superconductor from room temperature. Its characteristics that is zero resistance and low heat transfer rate under critical temperature lead to research for the replacement of existing current lead with HTSC. In this paper, we investigated the temperature distributions of stacking type and rod type current lead with each cross-section area and length using Nastran program and compared each temperature distribution. It is obtained from this paper that stacking type current lead has flat temperature gradient and than rod type one and more stable operation as current lead is closely related with its cross-section area and length.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.2
• /
• pp.189-194
• /
• 2020

This paper analyzes the gradient descent method, which is the one most used for learning neural networks. Learning means updating a parameter so the loss function is at its minimum. The loss function quantifies the difference between actual and predicted values. The gradient descent method uses the slope of the loss function to update the parameter to minimize error, and is currently used in libraries that provide the best deep learning algorithms. However, these algorithms are provided in the form of a black box, making it difficult to identify the advantages and disadvantages of various gradient descent methods. This paper analyzes the characteristics of the stochastic gradient descent method, the momentum method, the AdaGrad method, and the Adadelta method, which are currently used gradient descent methods. The experimental data used a modified National Institute of Standards and Technology (MNIST) data set that is widely used to verify neural networks. The hidden layer consists of two layers: the first with 500 neurons, and the second with 300. The activation function of the output layer is the softmax function, and the rectified linear unit function is used for the remaining input and hidden layers. The loss function uses cross-entropy error.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.22 no.6
• /
• pp.823-831
• /
• 1998

In order to investigate the impulsive noise at the exit of high-speed railway tunnel and the pressure transients inside the tunnel, numerical calculations using a Total Variation Dimishing difference scheme were applied to axisymmetric unsteady compressible flow field. Some compression wave forms were assumed to model the compression wave produced in real high-speed railway tunnel. The numerical data were extensively explored to analyze the peak over-pressure and maximum pressure gradient in the pressure wavefront. The effect of the distance and cross-sectional area ratio between two-continuous ducts on the characteristics of the pressure waves were investigated. The peak over-pressure inside the second duct decreases for the distance and cross-sectional area ratio between two tunnels to increase. The peak over-pressure and maximum pressure gradient of the pressure wavefront inside the second duct increase as the maximum pressure gradient of initial compression wave increases. The present results were qualitatively well agreed with the results of the previous shock tube experiment.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.18 no.4
• /
• pp.1031-1038
• /
• 1994

Characteristics of pulsatile flow and heat transfer have been studied numerically in the constant heat flux curved tube with periodic pressure gradient. As the Womersley number increases, the phase difference between the pressure gradient and the cross section averaged axial velocity becomes larger. In case of the Womersley number $\beta = 2$, when cross section averaged axial velocity reaches periodic state with time, the reverse and the natural flow coexist at phase angle, $\lambda = 1.44\pi$ and $\lambda =1.96\pi$. For all the Womersley numbers of present investigation, the time variation of wall temperature near inner wall is higher than that of near outer wall, independent of phase angle.