• Journal of Korean Association for Spatial Structures
• /
• v.18 no.4
• /
• pp.69-80
• /
• 2018

In this study, an algorithm applying deep learning to the truss structures was proposed. Deep learning is a method of raising the accuracy of machine learning by creating a neural networks in a computer. Neural networks consist of input layers, hidden layers and output layers. Numerous studies have focused on the introduction of neural networks and performed under limited examples and conditions, but this study focused on two- and three-dimensional truss structures to prove the effectiveness of algorithms. and the training phase was divided into training model based on the dataset size and epochs. At these case, a specific data value was selected and the error rate was shown by comparing the actual data value with the predicted value, and the error rate decreases as the data set and the number of hidden layers increases. In consequence, it showed that it is possible to predict the result quickly and accurately without using a numerical analysis program when applying the deep learning technique to the field of structural analysis.

• Structural Engineering and Mechanics
• /
• v.68 no.5
• /
• pp.537-547
• /
• 2018

This research presents the effect of anisotropy of the hollow disc mode under Brazilian test using PFC3D. The Brazilian tensile strength test was performed on the hollow disc specimens containing the bedding layers and then these specimens were numerically modeled by using the two dimensional discrete element code (PFC3D) to calibrate this computer code for the simulation of the cracks propagation and cracks coalescence in the anisotropic bedded rocks. The thickness of each layer within the specimens varied as 5 mm, 10 mm and 20 mm and the layers angles were changed as $0^{\circ}$, $25^{\circ}$, $50^{\circ}$, $75^{\circ}$ and $90^{\circ}$. The diameter of internal hole was taken as 15 mm and the loading rate during the testing process kept as 0.016 mm/s. It has been shown that for layers angles below $25^{\circ}$ the tensile cracks produce in between the layers and extend toward the model boundary till interact and break the specimen. The failure process of the specimen may enhance as the layer angle increases so that the Brazilian tensile strength reaches to its minimum value when the bedding layers is between $50^{\circ}$ and $75^{\circ}$ but its value reaches to maximum at a layer angle of $90^{\circ}$. The number of tensile cracks decreases as the layers thickness increases and with increasing the layers angle, less layer mobilize in the failure process.

• Journal of the Korean Magnetics Society
• /
• v.5 no.1
• /
• pp.42-47
• /
• 1995

FeN thin films for inductive recording heads were sputter deposited using RF diode sputtering mehtod from a pure iron target onto 7059 glass substrates, and their magnetic properties were measured. The magnetic properties were greatly affected by film thickness, gas pressure, sputter power and flow ratio of $N_{2}$ to Ar. Single layer FeN films with their thickness varied from $1,000\;{\AA}$ to $6,000\;{\AA}$ were doposited. 800 W sputter power, 3 mT gas pressure, $N_{2}$ to Ar flow ratio of 6.6 : 100 were the sputtering conditions. Up to 7 layers of FeN films having total thickness of $6,000\;{\AA}$ were deposited using $SiO_{2}$ of $30\;{\AA}$ thickness as intermediate layers and their coercivity and saturation magnetization were measured. The sputtering conditions were the same as those in the single layer films. Easy axis coercivity of the single layer FeN films gradually decreased as their thickness was increased, but for the films with their thicknesses above $3,000\;{\AA}$, the coercivity changed very little. As the number of the FeN layers were increased, the coercivity decreased We estimated the grain size of FeN films from the FWHM (Full Width at Half Maximum) of X-ray diffraction peaks. The grain size steadily decreased from about $200\;{\AA}$ to $120\;{\AA}$ as the number of layers were increased. Minimum hard axis coercivity of 0.4 Oe was obtained when the number of layers was four. Maximum relative permeability was 2,900 when the number of layers was three. The cut off frequeocy of the multilayer films were above 100 MHz.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.18 no.11
• /
• pp.1001-1006
• /
• 2005

A new fabrication method is proposed to form the stacked polysilicon gate by nitridation in $N_2$ atmosphere using conventional LP-CVD system. Two step stacked layers with an amorphous layer on top of a polycrystalline layer as well as three step stacked layers with polycrystalline films were fabricated using the proposed method. SIMS profile showed that the proposed method would successfully create the nitrogen-rich layers between the stacked polysilicon layers, thus resulting in effective retardation of dopant diffusion. It was observed that the dopants in stacked films were piled-up at the interface. TEM image also showed clear distinction of stacked layers, their plane grain size and grain mismatch at interface layers. Therefore, the number of stacked polysilicon layers with different crystalline structures, interface position and crystal phase can be easily controlled to improve the device performance and reliability without any negative effects in nano-scale CMOSFETs.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.25 no.5
• /
• pp.645-652
• /
• 2001

Heat transfer data are presented which describe characteristics of the transitional thermal boundary layers on the NACA0012 airfoil with upstream wakes. The wakes are generated periodically by circular cylindrical rods which rotate around the airfoil like a squirrel cage. The unsteady wakes simulate those produced by the upstream rotating blade rows in axial turbomachines. The pressure or suction side of the airfoil is also simulated according to the rotating direction of circular rods. As the Reynolds number and the number of rotating rods increase, the boundary layer transition occurs earlier and the Nusselt number increases. The difference of heat transfer coefficient is less on the pressure side than on the suction side. At a constant Reynolds number, the Nusselt number is larger and smaller, respectively, before and after transition as the Strouhal number increases.

• Structural Engineering and Mechanics
• /
• v.11 no.1
• /
• pp.17-34
• /
• 2001

An improved method has been developed for the computation of the section forces and stiffness in nonlinear finite element analysis of RC plane frames. The need for a new approach arises because the conventional technique may have a questionable level of efficiency if a large number of layers is specified and a questionable level of accuracy if a smaller number is used. The proposed technique is based on automatically dividing the section into zones of similar state of stress and tangent modulus and then numerically integrating within each zone to evaluate the sectional stiffness parameters and forces. In the new system, the size, number and location of the layers vary with the state of the strains in the cross section. The proposed method shows a significant improvement in time requirement and accuracy in comparison with the conventional layered approach. The computer program based on the new technique has been used successfully to predict the experimental load-deflection response of a RC frame and good agreement with test and other numerical results have been obtained.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.36 no.3
• /
• pp.226-232
• /
• 2023

As a potential alternative to flash memory, HfO₂/Al₂O₃ stacks appear to be a viable option as charge capture layers in charge trapping memories. The paper undertakes a review of HfO₂/Al₂O₃ stacks as charge trapping layers, with a focus on comparing the number, thickness, and post-deposition heat treatment and γ-ray and white x-ray treatment of such stacks. Compared to a single HfO₂ layer, the memory window of the 5-layered stack increased by 152.4% after O₂ annealing at ±12 V. The memory window enlarged with the increase in number of layers in the stack and the increase in the Al/Hf content in the stack. Furthermore, our comparison of the treatment of HfO₂/Al₂O₃ stacks with varying annealing temperatures revealed that an increased annealing temperature resulted in a wider storage window. The samples treated with O₂ and subjected to various γ radiation intensities displayed superior resistance. and the memory window increased to 12.6 V at ±16 V for 100 kGy radiation intensity compared to the untreated samples. It has also been established that increasing doses of white x-rays induced a greater number of deep defects. The optimization of stacking layers along with post-deposition treatment condition can play significant role in extending the memory window.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.40 no.11
• /
• pp.9-18
• /
• 2003

Two configurations of Film Bulk Acoustic Wave Resonators with acoustic quater-wave bragg reflector layers are theoretically analyzed using equivalent circuits and the difference of their characteristics are discussed. We compare the characteristics of λ/2 mode to those of ideal FBAR with top and bottom electrode contacting air and the characteristics of λ/4 mode to those of ideal FBAR with top electrode contacting air and bottom electrode clamped. We assume that the piezoelectric film is ZnO, the electrode is A1 and the substrate is Si, ABCD parameters are extracted and input impedance is calculated by converting the equivalent circuit from Mason equivalent circuits to the simplified equivalent circuits that ABCD parameters are extracted possible, From the variation of resonance frequency due to the change of thickness of reflector layers and the variation of electrical Q due to the change of mechanical Q of reflector layers, it is confirmed that the reflector layer just under the bottom electrode have the greatest effect on the varation of resonance frequency and electrical Q. It is shown that the number of reflector layers required for the saturation of electrical Q decreases with the increase of the impedance ratio of reflector layers and electrical Q of λ/2 mode is larger than that of λ/4 mode, Electromechanical coupling factor is independent of the number of layers, The impedance ratio of reflector layers becomes larger as the electromechanical coupling factor becomes larger, The electromechanical coupling factor of the two mode are smaller than those of ideal FBARs because of the trapping of acoustic energy in the reflector layers, The insertion loss of the ladder filter decreases with the increase of the number of reflector layers but the bandwidth is not affected much by the number of reflector layers, As the impedance ratio of reflector layers becomes larger the insertion loss becomes smaller and the bandwidth becomes wider, In our analysis of the two mode, characteristics of λ/2 mode appear to be slightly more favorable than that of λ/4 mode

• Applied Microscopy
• /
• v.52
• /
• pp.7.1-7.6
• /
• 2022

The process of encapsulating cobalt nanoparticles using a graphene layer is mainly direct pyrolysis. The encapsulation structure of hybrids prepared in this way improves the catalyst stability, which greatly reduces the leaching of non-metals and prevents metal nanoparticles from growing beyond a certain size. In this study, cobalt particles surrounded by graphene layers were formed by increasing the temperature in a transmission electron microscope, and they were analyzed using scanning transmission electron microscopy (STEM). Synthesized cobalt hydroxide nanosheets were used to obtain cobalt particles using an in-situ heating holder inside a TEM column. The cobalt nanoparticles are surrounded by layers of graphene, and the number of layers increases as the temperature increases. The interlayer spacing of the graphene layers was also investigated using atomic imaging. The success achieved in the encapsulation of metallic nanoparticles in graphene layers paves the way for the design of highly active and reusable heterogeneous catalysts for more challenging molecules.

• Journal of Korea Multimedia Society
• /
• v.23 no.7
• /
• pp.803-811
• /
• 2020

This paper proposes an image classification algorithm that transforms the number of convolution layers in the residual block of ResNet, CNN's representative method. The proposed method modified the structure of 34/50 layer of ResNet structure. First, we analyzed the performance of small and many convolution layers for the structure consisting of only shortcut and 3 × 3 convolution layers for 34 and 50 layers. And then the performance was analyzed in the case of small and many cases of convolutional layers for the bottleneck structure of 50 layers. By applying the results, the best classification method in the residual block was applied to construct a 34-layer simple structure and a 50-layer bottleneck image classification model. To evaluate the performance of the proposed image classification model, the results were analyzed by applying to the cifar10 dataset. The proposed 34-layer simple structure and 50-layer bottleneck showed improved performance over the ResNet-110 and Densnet-40 models.