• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.66 no.5
• /
• pp.870-876
• /
• 2017

In this paper, we proposed the automatic vision inspection system using multi-layer perceptron to detect the defects occurred on rail surface. The proposed system consists of image acquisition part and analysis part. Rail surface image is acquired as equal interval using line scan camera and lighting. Mean filter and dynamic threshold is used to reduce noise and segment defect area. Various features to characterize the defects are extracted. And they are used to train and distinguish defects by MLP-classifier. The system is installed on HEMU-430X and applied to analyze the rail surface images acquired from Honam-line at high speed up to 300 km/h. Recognition rate is calculated through comparison with manual inspection results.

• Proceedings of the KSR Conference
• /
• 2007.05a
• /
• pp.85-90
• /
• 2007

Aerodynamically generated noise is significant when the train speed exceeds 250km/h. Experiments were performed in order to investigate the characteristics of aerodynamic noise sources generated from exterior of the HSR-350x, especially from the inter-coach spacing. Measurements of both the inside and outside of the cabin are carried out to investigate the characteristics of the noise. Effect of the size of the gap has been investigated through an wind tunnel test and it has been found that the low frequency noise is strongly dependent on the size of gap. Also performed is an array measurement to locate different noise sources from the high-speed train. Spectral characteristics of exterior noise sources have been examined through these experiments.

• Proceedings of the KSR Conference
• /
• 2010.06a
• /
• pp.1011-1016
• /
• 2010

The noise emitted during train operation is generated with various reasons. It is known that the major noise generation is classified according to the ranges of train speed; that is, engine noise at lower speed range, rolling noise at medium speed range, and air-borne noise at higher speed range. These noises are transmitted in combined form with the noises generated from track components and under-carriage, etc. The rolling noise as a major noise at medium speed range is caused by the vibration occurred at wheel/rail interface. The vibration occurred at wheel/rail interface is transmitted to wheel and rail, and this vibration is emitted from wheel and rail as a noise. The object of this study is to investigate the effect of wheel damper of low noise wheel. In this study theoretical and experimental analysis is performed by numerical model calculations and impact test.

• Proceedings of the KSR Conference
• /
• 2007.05a
• /
• pp.1584-1589
• /
• 2007

Plug-in type door is installed in recently designed tube. It is world wide trend to fulfill the customer's demand. But, plug-in door need widely gap between doorway and platform and it is raise an serious accident. Therefore, newly designed train must have safety device like a sliding step to prevent incident. This paper present about development process for sliding step. Trend of sliding step, development plan, course, technical description, and etc. will be included in this paper.

• Proceedings of the KSR Conference
• /
• 2008.06a
• /
• pp.1201-1208
• /
• 2008

Ubiquotuous technology should be adopted in railroad business to provide passenger's security and convenience. In this project, IT and smart sensor technologies are reviewed, benchmarked, designed, and implemented. The target system is the next generation high speed train to be developed and operated in Korea with the maximum speed of 400km/h. Wireless sensor network with smart sensors is implemented around a train car. PC-like IT terminal will be designed and implemented so an individual passenger can use it to do information retrieval through the Internet, personal data processing, the e-learning, shopping on the railroad, and so on. These provision will give comfort, convenience, and safety of a passenger during his/her trip.

• Proceedings of the KSR Conference
• /
• 2009.05a
• /
• pp.119-126
• /
• 2009

The noise emitted during train operation is generated with various reasons. It is known that the major noise generation is classified according to the ranges of train speed; that is, engine noise at lower speed range, rolling noise at medium speed range, and air-borne noise at higher speed range. These noises are transmitted in combined form with the noises generated from track components and under-carriage, etc. The rolling noise as a major noise at medium speed range is caused by the vibration occurred at wheel/rail interface. The vibration occurred at wheel/rail interface is transmitted to wheel and rail, and this vibration is emitted from wheel and rail as a noise. In this paper, a soundproof box which can simply measure the noise emitted from rail and analyze the characteristics of the measured noise is studied.

• Proceedings of the KSR Conference
• /
• 2004.10a
• /
• pp.423-429
• /
• 2004

The concurrent engineering technologies have been broadly used in the field of the design, testing, manufacturing and maintenance works to reduce development time and costs. For this purpose, many design environments with the product data management system, the virtual engineering system and web database system are developed. In this research, we developed the driving simulator of the KTX(Korea Train Express) as a basic study for building the concurrent engineering design environment of rolling stock. The virtual track was developed from the Seoul to the Busan and the Daejeon to Mockpo to generate immersible driving environment. Also. fault generation systems were developed to educate drivers of the KTX. We expect to reduce the time and costs of newly developed rolling stock using the design environment developed in the research.

• Journal of the Korean Society for Railway
• /
• v.7 no.4
• /
• pp.326-331
• /
• 2004

A routing is the path that an actual trainset follows as it moves from one train to another train in a timetable. The number of routings is equivalent to the number of trainsets required to cover the timetable. The primary factors of rolling stock requirement plan include evaluating the minimum number of routings. This can be formulated as a set partitioning problem and solved using enumeration method or column generation method. In this paper we presents an enumeration algorithm which is useful to implement the enumeration method for the rolling stock requirement plan.

• Proceedings of the KSR Conference
• /
• 1998.11a
• /
• pp.445-452
• /
• 1998

The dynamic performance design of catenary-pantograph system which collects current for the next generation Korean high speed train(KHST) was considered. Used was the same dynamic model of the catenary-pantograph system as that of TGV-K which will be introduced for Kyung-bu corridor. Using the model , sensitivity analysis fer design variables were made to improve dynamic performance of KHST system. The results of sensitivity analysis and performance improvement are as follows: (1) It was found that aerodynamic force, tension of contact wire, mass of contact strip, mass of supporting contact strip, mass of clamp, mass of steady arm, and stiffness of plunger were the design variables most influencing the dynamic performance of the system. (2) Pantograph with reductions of 20% aerodynamic force, 34% weight of supporting contact strip, 20% spring constant of plunger, and 34% equivalent mass of steady arm was very possible system for the KHST which will be running at maximum operating speed 350 km/h.

• Smart Media Journal
• /
• v.10 no.1
• /
• pp.79-87
• /
• 2021

In recent years research on automatic font generation with machine learning mainly focus on using transformation-based methods, in comparison, generative model-based methods of font generation have received less attention. Transformation-based methods learn a mapping of the transformations from an existing input to a target. This makes them ambiguous because in some cases a single input reference may correspond to multiple possible outputs. In this work, we focus on font generation using the generative model-based methods which learn the buildup of the characters from noise-to-image. We propose a novel way to train a conditional generative deep neural model so that we can achieve font style control on the generated font images. Our research demonstrates how to generate new font images conditioned on both character class labels and character style labels when using the generative model-based methods. We achieve this by introducing a modified generator network which is given inputs noise, character class, and style, which help us to calculate losses separately for the character class labels and character style labels. We show that adding the character style vector on top of the character class vector separately gives the model rich information about the font and enables us to explicitly specify not only the character class but also the character style that we want the model to generate.