• KIPS Transactions on Computer and Communication Systems
• /
• v.8 no.11
• /
• pp.251-262
• /
• 2019

In this paper, we propose a simulator that provides various algorithms of RBF networks on neuromorphic chips. To develop algorithms based on neuromorphic chips, the disadvantages of using simulators are that it is difficult to test various types of algorithms, although time is fast. This proposed simulator can simulate four times more types of network architecture than existing simulators, and it provides an additional a two-layer structure algorithm in particular, unlike RBF networks provided by existing simulators. This two-layer architecture algorithm is configured to be utilized for multiple input data and compared to the existing RBF for performance analysis and validation of utilization. The analysis showed that the two-layer structure algorithm was more accurate than the existing RBF networks.

• Korean Journal of Crystallography
• /
• v.14 no.2
• /
• pp.93-104
• /
• 2003

The crystals of group-Ⅲ nitride semiconductors with wurtzite structure exhibit a strong polarity. Especially, GaN has characteristics of different growth rate, anisotropic electrical and optical properties due to the polarity. In this work, GaN epilayer was grown and the polarities of the crystals were observed by the chemical wet etching and SP-EFM. GaN thin films were deposited on c-plane A1₂O₃ substrate under the variations of growth conditions by HVPE such as the deposition temperature of the buffer layer, the deposition time, the ratio of Group-V and Ⅲ and the deposition temperature of the film. The adquate results were obtained under the conditions of 500℃, 90 seconds, 1333 and 1080℃, respectively. It is observed that the GaN layer grown without the buffer layer has N-polarity and the GaN layer grown on the buffer layer has Ga-polarity. Fine crystal single particles were grown on c-plane A1₂O₃ and SiO₂, layer. The external shape of the crystal shows ｛10-11｝｛10-10｝(000-1) planes as expected in the PBC theory and anisotropic behavior along c-axis is obvious. As a result of etching on each plane, (000-1) and ｛10-11｝planes were etched strongly due to the N-polarity and ｛10-10｝ plane was not affected due to the non-polarity. In the case of the crystal grown on c-plane A1₂O₃, two types of crystals were grown. They were hexagonal pyramidal-shape with ｛10-11｝plane and hexagonal prism with basal plane. The latter might be grown by twin plane reentrant edge (TPRE) growth.

• Korean Journal of Environmental Agriculture
• /
• v.27 no.4
• /
• pp.421-427
• /
• 2008

This study was carried out to offer a basis data for application of effective management through species composition and community structure of Castanopsis cuspidata var. thunbergii stands (Bijindo and Yokjido). The upper, middle and lower layer in Bijindo and Yokjido consisted of 10, 11 and 27 species, and 9, 6 and 28 species, respectively. The importance value of Castanopsis cuspidata var. thunbergii were highest in upper and middle layer both two studied site. The herbaceous species were 33 and 47 species in Bijindo and Yokjido, respectively. The species diversity ranged from 0.523 to 1.280 in Bijindo and 0.699 to 1.364 in Yokjido. Dominance in Bijindo and Yokjido were 0.477 and 0.242 at upper layer, 0.231 and 0.102 at middle layer, and 0.106 and 0.057 at lower layer, which showed that Castanopsis cuspidata var. thunbergii stands are dominated by a large number of species at all layer. The distribution pattern by Morisita's index showed that Castanopsis cuspidata var. thunbergii was distributed randomly in upper layer both the studied sites.

• Proceedings of the KIEE Conference
• /
• 1991.11a
• /
• pp.39-42
• /
• 1991

The Application of Artificial Neural Network(ANN) to forecast a load in a power system is investigated. The load forecasting is important in the electric utility industry. This technique, methodology based on the fact that parallel structure can process very fast much information is a promising approach to a load forecasting. ANN that is highly interconnected processing element in a hierachy activated by the each input. The load pattern can be divided distinctively into two patterns, that is, weekday and weekend. ANN is composed of a input layer, several hidden layers, and a output layer and the past data is used to activate input layer. The output of ANN is the load forecast for a given day. The result of this simulation can be used as a reference to a electric utility operation.

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 1997.10a
• /
• pp.42-49
• /
• 1997

In this paper, the method for soil-structure interaction analyses in the time domain is proposed. The far field soil region which is the outside of the artificial boundary is modeled by using explicit frequency-dependent two dimensional infinite elements which can include multiple wave components propagating into the unbounded medium. Since the dynamic stiffness matrix of the far field soil region using the proposed infinite elements is obtained explicitly in terms of exciting frequencies and constants in the frequency domain, the matrix can be easily transformed into the displacement unit-impulse response matrix, which corresponds to a convolution integral of it in the time domain. To verify the proposed method for soil-structure interaction analyses in the time domain, the displacement responses due to an impulse load on the surface of a soil layer with the rigid bed rock are compared with those obtained by the method in the frequency domain and those by models with extend finite element meshes. Good agreements have been found between them.

• Structural Engineering and Mechanics
• /
• v.26 no.3
• /
• pp.251-262
• /
• 2007

This paper examines the application of artificial neural networks (ANN) to the response prediction of geometrically nonlinear truss structures. Two types of analysis (deterministic and probabilistic analyses) are considered. A three-layer feed-forward backpropagation network with three input nodes, five hidden layer nodes and two output nodes is firstly developed for the deterministic response analysis. Then a back propagation training algorithm with Bayesian regularization is used to train the network. The trained network is then successfully combined with a direct Monte Carlo Simulation (MCS) to perform a probabilistic response analysis of geometrically nonlinear truss structures. Finally, the proposed ANN is applied to predict the response of a geometrically nonlinear truss structure. It is found that the proposed ANN is very efficient and reasonable in predicting the response of geometrically nonlinear truss structures.

• The Journal of the Acoustical Society of Korea
• /
• v.34 no.6
• /
• pp.470-478
• /
• 2015

A heuristic method treating a layered ocean bottom in a ray modeling is to use the plane wave reflection coefficient for multiple-layered structure, named an one-layer assumption in this paper. We examine the validity of one-layer assumption in the case of two-layered ocean bottom, and obtain a simple inequality condition depending on the sound speed ratio, the ratio of layer thickness to source-receiver range, and the grazing angle of first reflected ray. From this inequality condition, it is shown that an one-layer assumption can be applicable to ray propagation problems at mid frequencies. Finally, numerical experiments are performed in the ocean environment similar to the East Sea in Korea. Incoherent transmission loss is calculated by the geometrical beam model with the plane wave reflection coefficient for multiple-layered ocean bottom and compared with the result of SNUPE 2.0, which is a parabolic equation package developed in Seoul National University.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.19 no.3
• /
• pp.273-279
• /
• 2006

High-$T_c$ superconducting (HTSC) power cable is one of the interesting parts in power application using HTSC wire. However, its stacked structure makes the current distribution between conducting layers non-uniform due to difference between self inductances of conducting layers and mutual inductances between two conducting layers, which results in lower current transmission capacity of HTSC power cable. In this paper, the transport current distribution between conducting layers was investigated through the numerical analysis for the equivalent circuit of HTSC power cable with a shield layer, and compared with the case of without a shield layer. The transport current distribution due to the increase of the contact resistance in each layer was improved. However, its magnetization loss increased as the contact resistance increased. It was confirmed from the analysis that the shield layer was contributed to the improvement of the current distribution between conducting layers if the winding direction and the pitch length were properly chosen.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.9 no.4
• /
• pp.403-413
• /
• 1985

Hot-wire measurements of second and third-order mean products of velocity fluctuations have been made in the separated, reattached, and redeveloping boundary layer behind a vertical fence. Mean velocity, wall static pressure distributions have also been measured in the whole flow field. Upstream of the reattachment point, the separated shear layer developes as a free mixing layer, but the gradient of the maximum slope thickness, turbulent intensities and the Reynolds shear stress are higher than that of the mixing layer due to initial streamline curvature and the effects of highly turbulent recirculating flow region. In the reattachment region, Reynolds shear stress and triple products near the surface is far more rapid than the decrease of the shear stress; that is the presence of the solid wall has a marked effect on the apparent gradient diffusivity of intensity or shear stress and throws doubts upon the usefulness of the simple gradient diffusivity model in this region.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.6 no.4
• /
• pp.124-131
• /
• 2003

When a secondary gaseous flow is injected vertically into a supersonic flow through circular nozzle, a complicated structure of flow field is produced around the injection area. The interaction between the two streams produces a strong bow shock wane on the upstream side of the side-jet. The results show that bow shock wave and turbulent boundary layer interaction induces the boundary layer separation in front of the side-jet. This study is to analyze the structure of flow fields and distribution of surface pressure on the flat plate according to total pressure ratio using a supersonic cold-flow system and also to study the control force of affected side-jet. The nozzle of main flow was designed to have Mach 2.88 at the exit. The injector has a sonic nozzle with 4mm diameter at the exit of the side-jet. In experiments, The oil flow visualization using a silicone oil and ink was conducted in order to analyze the structure of flow fields around the side-jet. The flow fields are visualized using the schlieren method. In this study, a computational fluid dynamic solution is also compared with experimental results.