• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 2006.04a
• /
• pp.1001-1008
• /
• 2006

The main objective of this paper is to compare economical effectiveness of typical methods for checking stability in principal components of steel cable-stayed bridges. Elastic and inelastic buckling analyses are carried out for frame-like numerical models of cable-stayed bridges. The axial-flexural interaction equations prescribed in AASHTO Allowable Stress Design (ASD) and AASHTO Load and Resistance Factor Design (LRFD) are used in order to check the stability of principal components. Parametric studies are performed for numerical models which have the center span length of 300m, 600m, 900m and l200m with different girder depths. Peak values of the interaction equations are calculated at the intersection point between girders and towers. These peak values are considered as a major factor to design of principal components of cable-stayed bridges. As a result, more economical design for girders and towers can be feasible using the inelastic buckling analysis. In addition, LRFD codes are more economical about 20% on the average than ASD codes for all numerical models of cable-stayed bridges.

• 한국지구물리탐사학회:학술대회논문집
• /
• 2003.11a
• /
• pp.272-280
• /
• 2003

A definite shut-in pressure in hydraulic fracturing techniques is needed for obtaining the correct information on the in-situ stress regimes in rock masses. The relation between the behaviour of hydraulically induced fractures and the condition of remote stress is considered to be major reasons of an ambiguous shut-in pressure in hydraulic fracturing pressure-time history curves. This paper describes the results of a series of numerical analyses carried out using UDEC(Universal Distinct Element Code, Itasca), which is based on the discrete element method, to compare several methods for determining the shut-in pressure during hydraulic fracturing. The fully coupling of hydraulic and mechanical analysis was applied, and the effects of four different discontinuity geometries in numerical modelling have been investigated for this purpose. The effects of different remote stress regimes and different physical properties on hydraulic fracture propagation have been also analyzed. Several methods for obtaining shut-in pressure from the ambiguous shut-in curves have been applied to all the numerical models. The graphical intersection methods, such as (P vs. t) method, (P vs. log(t)) method, (log(P) vs. log(t)) method, give smaller values of the shut-in pressure than the statistical method, (dP/dt vs. P). Care should be taken in selecting a method for shut-in pressure, because there can be existed a stress anomaly around the wellbore and fracturing from the wellbore by a constant flow rate may have a more complicate mechanism.

• Journal of Korean Society of Transportation
• /
• v.9 no.1
• /
• pp.19-28
• /
• 1991

This study aims to seize the differences in drivers' attitudes and to speculate about the attitudes which may be causally related to traffic accidents based on data collected from group survey questionnaire in the four countries. The questionnaire was composed of 9 problem areas and prepared in Japanese English and Korean languages. The survey was conducted in Japan Canada U.S.A and Korea in 1987 The present study analysed the driver's attitudes toward behaviour of pedestrians traffic signal and causes of traffic accidents. The main results are :(1) The order of high percentage of those who cited the pedestrians as major cause of accidents was Korea Japan, Canada and U,S,A This markes the very high number of accidents involving pedestrians in Korea. (2) The percentage of who answered that most drivers started before the sign changed to Go sign at an intersection were higher in Japan and Korea than in U.S.A (3) Regarding the caues of accidents if the drivers were to meet with an accident the percentage of responses attributing faults to themselves for the accident was very high in Japan. Korea came next. In contrast the percentage of responses attributing faults to others is higher in Canada and U.S.A than Japan and Korea.

• Journal of the Korean Society of Safety
• /
• v.32 no.5
• /
• pp.8-12
• /
• 2017

Bicycles are very popular sporting goods in these days. Thus, the durability of bicycles is very important for the safety of bicyclists. It is well known that a bicycle frame is a major component which is essential to the safety and performance of a complete bicycle. In this study, the durability of bicycle frames were experimentally investigated under the fatigue load. Eighty bicycle frames with different types and materials were prepared and tested according to EN standards. Three kinds of fatigue loads, that is, pedalling, vertical and horizontal fatigue load, which occur constantly during riding a bicycle, were applied to the bicycle frames. The experimental results show that the horizontal fatigue load was the severest mode to pass EN standard. The pass ratio of horizontal fatigue load test was 45.2%, while the pass ratio of vertical fatigue load test was 100%. Most of cracks were found at the right side of bottom bracket shell and at the intersection area between head tube and down tube. It seems that the experimental results can be applied to improve the safety and performance of a bicycle frame.

• Proceedings of the Korean Society of Propulsion Engineers Conference
• /
• 2017.05a
• /
• pp.835-837
• /
• 2017

In this paper, manufacturing processes of cylindrical composite lattice structures using filament winding method was described. Cylindrical composite lattice structures were manufactured in accordance with four major steps. Silicon mold of lattice shape was installed on mandrel and then continuous fiber was wound on silicon mold. After winding process, in order to ensure the same thickness for all regions, compression process was done for its intersection parts. Finally, the composite lattice structure was demoulded after curing in oven. It was found that the manufactured cylindrical composites lattice structure had 2.4% of dimensional error compared to the design requirements.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1993.06a
• /
• pp.1005-1008
• /
• 1993

Supervised learnmg 01 recurrent neural networks (RNNs) is discussed. First, we review the present state of art, featuring their major properties in contrast of those of the multilayer neural networks. Then, we concisely describe one of the most practical learning algorithms, i.e. backpropagation through time. Revising the basic formulation of the learning algorithms, we derive a general formula to solve for the exact solution(s) of the whole connection weights w of RNNs. On this basis we introduce a novel interpretation of the supervised learning. Namely, we define a multidimensional Euclidean space, by assigning the cost function E(w) and every component of w to each coordinate axis. Since E=E(w) turns up as a hyper surface in this space, we refer to the surface as learning surface. We see that topological features of the learning surface are valleys and hills. Finally, after explicating that the numerical procedures of learning are equivalent to descending slopes of the learning surface along the steepest gradient, we show that a minimal value of E(w) is the intersection of curved valleys.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• v.1
• /
• pp.479-483
• /
• 2006

Indoor positioning is recently highlighted and various kinds of indoor positioning systems are under developments. Since positioning systems have their own characteristics, proper positioning scheme should be chosen according to the required specifications. Positioning methods are often classified into time-based and angle-based one, and this paper presents the error analysis of these location methods. Because measurement equations of these methods are nonlinear, linearization is usually needed to get the position estimate. In this paper, Gauss-Newton method is used in the linearization. To analyze the position error, we investigate the error ellipse parameters that include eccentricity, rotation angle, and the size of ellipse. Simulation results show that the major axes of error ellipses of TOA and AOA method lie in different quadrants at most region of workspace, especially where the geometry is poor. When the TOA/AOA hybrid scheme is employed, it is found that the error ellipse is reduced to the intersection of ellipses of TOA and AOA method.

• Smart Structures and Systems
• /
• v.31 no.4
• /
• pp.383-392
• /
• 2023

Post-earthquake building condition assessment is crucial for subsequent rescue and remediation and can be automated by emerging computer vision and deep learning technologies. This study is based on an endeavour for the 2nd International Competition of Structural Health Monitoring (IC-SHM 2021). The task package includes five image segmentation objectives - defects (crack/spall/rebar exposure), structural component, and damage state. The structural component and damage state tasks are identified as the priority that can form actionable decisions. A multi-task Convolutional Neural Network (CNN) is proposed to conduct the two major tasks simultaneously. The rest 3 sub-tasks (spall/crack/rebar exposure) were incorporated as auxiliary tasks. By synchronously learning defect information (spall/crack/rebar exposure), the multi-task CNN model outperforms the counterpart single-task models in recognizing structural components and estimating damage states. Particularly, the pixel-level damage state estimation witnesses a mIoU (mean intersection over union) improvement from 0.5855 to 0.6374. For the defect detection tasks, rebar exposure is omitted due to the extremely biased sample distribution. The segmentations of crack and spall are automated by single-task U-Net but with extra efforts to resample the provided data. The segmentation of small objects (spall and crack) benefits from the resampling method, with a substantial IoU increment of nearly 10%.

• Journal of Biomedical Engineering Research
• /
• v.28 no.4
• /
• pp.520-529
• /
• 2007

An efficient method for implementing image reconstruction algorithms for Compton cameras is presented. Since Compton scattering formula establishes a cone surface from which the incident photon must have originated, it is crucial to implement a computationally efficient cone-surface integration method for image reconstruction. In this paper we assume that a cone is made up of a series of ellipses (or circles) stacked up one on top of the other. In order to reduce computational burden for tracing ellipses formed by the intersection of a cone and an image plane, we propose a new method using a series of imaginary planes perpendicular to the cone axis so that each plane contains a circle, not an ellipse. In this case the cone surface integral can be performed by simply accumulating the circles along the cone axis. To reduce the computational cost of tracing circles, only one of the circles in the cone is traced and the rest are determined by using simple trigonometric ratios. For our experiments, we used the three different schemes for tracing ellipses; (i) using the samples generated by the ellipse equation, (ii) using the fixed number of samples along a circle on the imaginary plane, and (iii) using the fixed sampling interval along a circle on the imaginary plane. We then compared performance of the above three methods by applying them to the two reconstruction algorithms - the simple back-projection method and the expectation-maximization algorithm. The experimental results demonstrate that our proposed methods (ii) and (iii) using imaginary planes significantly improve reconstruction accuracy as well as computational efficiency.

• Journal of Korean Society of Transportation
• /
• v.34 no.1
• /
• pp.29-42
• /
• 2016

Speeding is an important factor in traffic safety. Speed not only affects crash severity, but is also related to the possibility of crash occurrence. This study presents results from an analysis of 27,968 speed violation cases collected from 36 red light and speed enforcement cameras at signalized intersections in the city of Cheongju. Data included details of their violation history such as speeding tickets within a recent 3-year span and their demographic characteristics. The goal of this analysis is to understand the correlation between speed violations and various factors in terms of humans, vehicles and road environments. This study used descriptive statistics and Binary Logistics Regression(BLR) analysis with SPSS 20.0 software. The major results of this study are as follows. First, speed violations occurred at rural and suburban area. Second, about 25.6% of the violators committed to more than 20km/h over a speed limit. Third, the difference between speed violators and normal drivers clearly appeared in location of intersection(urban/rural/suburban area), gender and age. Finally, a statistically significant model(Hosmer and Lemeshow test: 11.586, p-value: 0.171) was developed through the BLR.