• The Journal of the Korea institute of electronic communication sciences
• /
• v.15 no.3
• /
• pp.513-520
• /
• 2020

Recently, the development of a vision inspection system using machine learning has become more active. This study seeks to develop a defect inspection model using machine learning. Defect detection problems for images correspond to classification problems, which are the method of supervised learning in machine learning. In this study, defect detection models are developed based on algorithms that automatically extract features and algorithms that do not extract features. One-dimensional CNN and two-dimensional CNN are used as algorithms for automatic extraction of features, and MLP and SVM are used as algorithms for non-extracting features. A defect detection model is developed based on four models and their accuracy and AUC compare based on AUC. Although image classification is common in the development of models using CNN, high accuracy and AUC is achieved when developing SVM models by converting pixels from images into RGB values in this study.

• Journal of the Korean Geotechnical Society
• /
• v.19 no.4
• /
• pp.287-298
• /
• 2003

The feasibility of using neural networks to model the complex relationship between piezocone measurements and the undrained shear strength of clays has been investigated. A three layered back propagation neural network model was developed based on actual undrained shear strengths, which were obtained from the isotrpoically and anisotrpoically consolidated triaxial compression test(CIUC and CAUC), and piezocone measurements compiled from various locations around the world. It was validated by comparing model predictions with measured values about new piezocone data, which were not previously employed during development of model. Performance of the neural network model was compared with conventional empirical method, direct correlation method, and theoretical method. It was found that the neural network model is not only capable of inferring a complex relationship between piezocone measurements and the undrained shear strength of clays but also gives a more precise and reliable undrained shear strength than theoretical and empirical approaches. Furthermore, neural network model has a possibility to be a generalized relationship between piezocone measurements and undrained shear strength over the various places and countries, while the present empirical correlations present the site specific relationship.

• Journal of the Korean GEO-environmental Society
• /
• v.6 no.1
• /
• pp.53-62
• /
• 2005

This paper estimates the long-term settlement of Kimpo metropolitan landfill based on measured settlement data from 180 landfill monitors accumulated over a period of 12 years. Comparison of domestic and international settlement records indicate that the domestic compression rate is slightly lower due to greater portion of organic component. Several existing settlement models are used to compare with the observed behavior and also to estimate long-term settlement. The hyperbolic, Gibson & Lo, Bjarngard & Edgers and Power Creep Law models compare well with the measured settlement of the Kimpo metropolitan landfill. The settlement models are further used to estimate long-term settlement. Bjarngard & Edgers and Power Creep Law models result in higher estimates of the long-term settlement compared to the hyperbolic and Gibson & Lo models. Further comparisons indicate that other models, including Sowers and log models, are inapproriate for predicting the long-term settlement of the Kimpo metropolitan landfill.

• Korean Journal of Environment and Ecology
• /
• v.17 no.4
• /
• pp.324-336
• /
• 2004

This study was achieved to present ecological administration plan by analyzing vegetation structure and condition rating class(environmental damage degree) of Sangnim Woods Natural Monument in Hamyang-gun, Korea. In vegetation structure part, actual vegetation was classified by 22 patterns and Quercus serrata Carpinus tschonoskii community(31.8％), Quercus serrata community(14.6％) ranged extensively. Main plant community was 8 types, and is fractionated by 13 plant communities according to stratigraphy development degree it is Quercus serrata community, Quercus serrata Carpinus tschonoskii community, Quercus aliena community, Quercus acutissima community, Carpinus tschonoskii community, Carpinus tschonoskii Quercus serrata community, Zelkova serrata-Quercus serrata community, and Planted area with korean landscape woody plants. Age of old growth trees that diameter of breast height over 38cm was 61∼77years. In condition rating class, area of class 3 was 51,960$m^2$(32.8％), area of class 4 was 6,583$m^2$(3.5％), and area of class 5 was 4,086$m^2$(2.6％) and gross area of class 3∼6 need artificial restoration was 61,619$m^2$(38.9％). Considering actual vegetation, plant community structure, and condition rating class biotope was classified by total 14 types. While distribution area of Queens spp. old growth forest of shrub damaged(51,246$m^2$, 32.4％) and deciduous broad leaved old growth forest of simple-layer structure(19,906$m^2$, 12.6％) is large and that of deciduous broad-leaved old growth forest of multi-layer structure(2,085$m^2$, 1.3％) and Queens spp. old growth forest of multi-layer structure may have to manage with user control by administration plan for stabilization of Sangnim Woods ecosystem for long-term. Also, both vegetation of shrub damaged and simple-layer structure as negative restoration area should be restored for ecological succession and both grassland and planted area with korean woody plants as positive restoration area should be revegetated by using ecological planting model of native vegetation structure in Sangnim Woods.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.39 no.2
• /
• pp.146-156
• /
• 2002

In this paper, a new virtual-straight line parameter determination methodology and fast time domain simulation technique for non-uniform interconnects are presented and verified. Time domain signal response of interconnects circuit considering the characteristic of non-linear transistor is performed by using model order reduction method. Since model order reduction method is peformed by using per unit length parameters, virtual- straight line parameters for non-uniform interconnects are determined. Its method is integrated into Berkeley SPICE and shown that time domain signal responses using proposed method have a good agreement with the results of conventional circuit simulator HSPICE. The proposed method can be efficiently employed in the high-performance VLSI circuit design since it can provide a fast and accurate time domain signal response of complicated multi - layer interconnects.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.22 no.2
• /
• pp.189-196
• /
• 2009

This paper presents the subparametric finite element model formulated by partial-linear layerwise theory for the analysis of laminate composites. The proposed model is based on refined approximations of two dimensional plane for orthotropic thick laminate plate as well as thin case. Three dimensional problem can be reduced to two dimensional case by assuming piecewise linear variation of in-plane displacement and a constant value of out-of-plane displacement across the thickness. The integrals of Legendre polynomials are chosen to define displacement fields and Gauss-Lobatto numerical integration is implemented in order to directly obtain maximum values occurred at the nodal points of each layer without other extrapolation techniques. The validity and characteristics of the proposed model have been tested by using orthotropic multilayered plate problem as compared to the values available in the published references. In this study, the convergence test has been carried out to determine the optimal layer model in terms of central deflection and stresses. Also, the distribution of displacements and stresses across the thickness has been investigated as the number of layer is increased.

• The Journal of Engineering Geology
• /
• v.28 no.4
• /
• pp.583-592
• /
• 2018

Recently, concern has arisen regarding the lowering of groundwater levels in the hinterland caused by the development of high-capacity radial collector wells in riverbank filtration areas. In this study, groundwater levels are estimated using Modflow software in relation to the water volume pumped by the radial collector well in Anseongcheon Stream. Using the water volume data, an artificial neural network (ANN) model is developed to determine the amount of water that can be withdrawn while minimizing the reduction of groundwater level. We estimate that increasing the pumping rate of the horizontal well HW-6, which is drilled parallel to the stream direction, is necessary to minimize the reduction of groundwater levels in wells OW-7 and OB-11. We also note that the number of input data and the classification of training and test data affect the results of the ANN model. This type of approach, which supplements ANN modeling with observed data, should contribute to the future groundwater management of hinterland areas.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.11
• /
• pp.491-500
• /
• 2021

In this study, we propose a method for detecting fall behavior using MS Kinect v2 RGBD Camera-based Human-Skeleton Keypoints and a 2-Stacked Bi-LSTM model. In previous studies, skeletal information was extracted from RGB images using a deep learning model such as OpenPose, and then recognition was performed using a recurrent neural network model such as LSTM and GRU. The proposed method receives skeletal information directly from the camera, extracts 2 time-series features of acceleration and distance, and then recognizes the fall behavior using the 2-Stacked Bi-LSTM model. The central joint was obtained for the major skeletons such as the shoulder, spine, and pelvis, and the movement acceleration and distance from the floor were proposed as features of the central joint. The extracted features were compared with models such as Stacked LSTM and Bi-LSTM, and improved detection performance compared to existing studies such as GRU and LSTM was demonstrated through experiments.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.22 no.6
• /
• pp.157-167
• /
• 2023

The use of big data for transportation often involves using data that includes personal information, such as the driver's driving routes and coordinates. This study explores the creation of a route choice prediction model using a large dataset from mobile navigation apps using federated learning. This privacy-focused method used distributed computing and individual device usage. This study established preprocessing and analysis methods for driver data that can be used in route choice modeling and compared the performance and characteristics of widely used learning methods with federated learning methods. The performance of the model through federated learning did not show significantly superior results compared to previous models, but there was no substantial difference in the prediction accuracy. In conclusion, federated learning-based prediction models can be utilized appropriately in areas sensitive to privacy without requiring relatively high predictive accuracy, such as a driver's preferred route choice.

• Proceedings of the Korean Vacuum Society Conference
• /
• 1999.07a
• /
• pp.38-38
• /
• 1999

KSTAR(Korea Superconducting Tokamak Advanced Research) 핵융합 실험 장치의 진공용기 및 진공용기 내부의 플라즈마 대향 부품들은 초고진공 (5$\times$10-9 Torr)의 달성을 위해 진공용기 내부의 이물질(H2, H2O, CO, CO2, CH4 등) 제거를 목적으로 SS316LN인 진공용기는 25$0^{\circ}C$, 탄소 물질인 플라즈마 대향부품은 35$0^{\circ}C$ 정도까지 가열(이하 베이킹)할 필요성이 있다. 이 가열방법으로 고온 질소가스를 진공용기 이중벽 사이로 흘려주는 방식과 코일에 저주파 교류전류를 흘려 진공용기를 유도가열하는 방식이 고려되고 있는데, 유도가열방식은 최대 유도 전력이 70kW 정도로 실제 베이킹에 필요한 열량을 공급하는데 있어 적잖이 부족하며 또 국부적인 가열 특성으로 인하여 KSTSR의 베이킹 방식은 전자의 가열방식을 우선적으로 채택하고 있다. 본 논문에서는 0-차원 해석을 통하여 진공용기와 플라즈마 대향 부품들에 대한 베이킹 계획을 결정하고 이를 만족시키기 위해 투입해야 할 열량을 직선적으로 증가하는 온도 곡선에서 각 부분의 온도 상승률을 다르게 설정한 세 경우와 F-자 형태로 변화하는 온도 곡선의 경우에 대해 각각 적용하여 시간에 따른 필요열량을 비교.검토하였으며, 이를 근거로 안정적인 베이킹 계획을 선정하였고 이 베이킹 계획의 실현을 위해 투입해야 할 고온 질소가스의 유량과 온도 도달시간까지 매 시간에서의 가스온도를 산출하였다. 토러스 형상의 토카막 진공용기와 플라즈마 대향 부품 및 다층단열재에 대한 해석 모델은 길이가 유한한 0-차원 실린더 모델로 가정하였고, 이에 대한 기하학적 성질 및 열역학적 성질은 유효계수를 고려하여 산출하였다. 진공용기 이중 벽 내부로 흐르는 질소가스의 유량과 온도의 계산은 진공용기 내벽과 외벽을 각각 독립적인 열전달 요소로 가정하여 구성한 모델을 이용하였다. 전체 해석에서 각 열전달 요소의 비열 값은 온도에 따라 변화하는 비열의 특성을 반영하였으며. 진공용기와 플라즈마 대향 부품의 방사율(emissivity)은 앞서 가정했던 각 온도 상승 곡선에 대해서 각각 0.1, 0.2, 1.3의 경우를 가정하여 계산하였다. 직선적으로 증가하는 온도 상승 곡선중 2$0^{\circ}C$/hr의 온도상승율을 갖는 경우가 다른 베이킹 시나리오 모델에 비해 효과적이라 생각되며 초대 필요 공급열량은 200kW 정도로 산출되었다. 실질적인 수치를 얻기 위해 보다 고차원 모델로의 해석이 필요하리라 생각된다. 끝으로 장기적인 관점에서 KSTAR 장치의 베이킹 계획도 살펴본다.