• Proceedings of the Korean Information Science Society Conference
• /
• 2010.06b
• /
• pp.389-392
• /
• 2010

실시간 임베디드 시스템에서 멀티미디어 시스템 같은 빠른 응답성을 요구하는 시스템이 존재 할 수 있다. 현재 EDF, RM 알고리즘의 경우 각각 마감시간, 수행 주기 시간에 의해서 우선순위가 선정되고 수행된다. 그러나 시스템 상에 실시간 프로세스 외에 다양한 프로세스가 존재할 수 있다. 본 논문에서는 마감 시간과 프로세스의 기본 우선순위를 고려하여 새로운 우선순위를 이용하며, 전체 프로세스들이 마감 시간내에 스케줄링 가능하고 문맥교환에 의한 오버헤드에 의해 마감시간을 넘어가지 않을 경우 각 프로세스들에게 타임 슬라이스를 할당하여 수행할 수 있도록 하여 기존 알고리즘들 보다 응답성을 향상시키는 기법을 제시한다.

• Proceedings of the KIPE Conference
• /
• 2016.11a
• /
• pp.51-52
• /
• 2016

본 논문은 속도 운전 영역 확장을 위한 자동 진상각 제어 알고리즘을 제안한다. 속도 운전 영역 확장을 위하여 상전류의 위상을 역기전력의 위상에 대하여 진상으로 제어한다. 기존의 진상각 제어 방식은 속도 - 진상각 참조표를 이용하거나 추가적인 하드웨어를 구성하여 제어한다. 참조표를 이용한 제어 방식의 경우 전동기의 특성에 매우 민감하며 운전점이 변경될 경우 제어 특성이 현저히 나빠지는 단점이 있으며, 추가적인 하드웨어를 이용하는 경우 전동기의 제정수에 의존적임과 동시에 제어 시스템의 부피 및 비용 증가와 같은 문제를 야기시킨다. 본 논문은 유효 전압의 크기와 전류 제어기의 궤환 전압을 이용하여 필요한 진상각을 계산한다. 본 논문에서 제안하는 방식의 유용함은 컴퓨터 시뮬레이션을 통하여 타당함을 검증하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.92-94
• /
• 2023

사물인터넷이 발전함에 따라 사용자에게 여러 서비스를 제공하기 위한 사물인터넷 플랫폼이 활발하게 개발되고 있다. 그러나 각각의 플랫폼은 독자적인 환경에서 개발되고 있어 상호운용성에 어려움이 있다. 특히, 이기종 사물인터넷 플랫폼이 리소스 보호를 위해 동일한 속성기반 접근제어를 사용하더라도 속성을 표현하는 표현법이 상이하여 유효한 속성을 소유하고 있음에도 타 플랫폼의 리소스에 접근할 수 없는 문제가 있다. 이에, 본 논문에서는 속성기반 접근제어를 활용하는 이기종 사물인터넷플랫폼 간의 서로 다른 속성 표현법을 이해하여 상호운용할 수 있는 프레임워크를 제안한다. 본 프레임워크는 자가적응 모듈을 통해 동일한 속성이더라도 표현법의 차이로 발생하는 상호운용의 문제를 해결하여 이기종 플랫폼 간의 리소스 접근이 가능하도록 한다.

• Journal of Advanced Marine Engineering and Technology
• /
• v.37 no.4
• /
• pp.358-367
• /
• 2013

In the present paper, we apply a porous modelling technique to accurately predict the pressure drop through the strainer by replacing all or some of the filter composed of perforated plates with porous media and there imposing the streamwise and transverse loss coefficients required according to the Forchheimer law and then confirm its effectiveness. At first, the streamwise coefficient is obtained by performing a simple simulation on the pipe flow mimicking the hole flow. Subsequently, the transverse coefficient is obtained by setting a unit pattern to have common flow loss characteristics with the repeated shape patterns in the filter, then performing numerical simulations on the prototype and porous model of the unit shape pattern, and finally comparing their results of pressure drop. To validate the applied modeling technique, we perform the numerical simulation with the two specified loss coefficients on a whole shape of strainer and compare the modeling results with those of the corresponding prototype numerical simulation. Comparison indicates that the modeling technique can predict the pressure drop and flow characteristics comparatively accurately and save the number of nodes closely related to the computational cost (CPU and memory) by about 3~4 times compared with the prototype simulation.

• The Journal of the Acoustical Society of Korea
• /
• v.41 no.3
• /
• pp.268-277
• /
• 2022

In this study, each component of flow noise source of underwater propeller installed to the scale model of the KVLCC2 is investigated and the effect of each noise source on underwater-radiated noise is quantitatively analyzed. The computation domain is set to be the same as the test section of the large cavitation tunnel in the Korea Research Institute of Ship and Ocean Engineering. First, for the high-resolution computation of flow field which is noise source region, the incompressible multiphase Delayed Detached Eddy Simulation is performed. Based on flow simulation results, the Ffowcs Williams and Hawkings integral equation is used to predict underwater-radiated noise and its validity is confirmed through the comparison with the tunnel experiment result. For the quantitative comparison on the contribution of each noise source, the spectral levels of sound pressure and power levels predicted using propeller tip-vortex cavitation, blade surface and rudder surface as the integral region of noise sources are investigated. It is confirmed that the cavitation which is monopole noise source significantly contributed to the underwater-radiated noise than propeller blades and rudder which is dipole noise source, and the rudder have more contribution than propeller blades due to the influence of the propeller wake.

• Journal of Korea Water Resources Association
• /
• v.35 no.5
• /
• pp.553-561
• /
• 2002

The objectives of this study are to introduce flash flood forecasting system in Korea and to develop a system for computing threshold runoff on very fine catchment scale. The developed GUI system composed of 9 steps starting from input data preparation to Input file creation for flash flood forecasting compute basin subdivision, hydrologic subbasin characteristics, bankfull flows, unit peak flows and threshold runoffs on about 5 $\textrm{km}^2$ scale. When the developed system was applied on Pyungchang IHP basin, the computed 1-hour threshold runoffs ranged 18.72~81.96mm with average value of 46.39mm. Judging from the comparison of the computed threshold runoffs between this study area and three other basins in United States, the computed results in this study were reasonable. It can be concluded that the developed system on ArcView/Avenue are useful for computing threshold runoff on small catchment and can be used as a component of flash flood forecasting system.

• Journal of the Society of Naval Architects of Korea
• /
• v.33 no.4
• /
• pp.60-65
• /
• 1996

Mathematical model of maneuvering motion for a single-screw single-rudder ship established and versatile applications to the special situations of maneuvering are attempted. While, the mathematical model for twin-screw twin-rudder ship is not presented so much, because that type of ship is not popular. Lee et al. have examined the characteristics of such ship by captive model tests in 1988. This paper treats new mathematical models for propeller effective wake ($1-w_p$) and effective neutral rudder angle ${\delta}_R$ in the case of twin-screw twin-rudder ship. And some maneuvering motions are calculated with proposed models and compared with exact simulations.

• Journal of Korea Port Economic Association
• /
• v.34 no.2
• /
• pp.51-68
• /
• 2018

This paper compares the differences in management performance in the logistics market and analyzes the differences in business characteristics depending on the industry types. In addition, the effects of industry types and business characteristics on management performance are examined. The analysis method used is ANOVA and K-means clustering. The implication of the study are as follows. First, in the logistics market in Japan, there was a difference in management performance among the types of industry. The warehousing service type had the highest profitability and stability among all the industry types. Second, differences in business characteristics by industry types were tested. It was found that offshore cargo transportation type was more capital intensive than the other types. In addition, warehousing service type had higher business leadership and credit transaction than others. Third, industry types and clusters based on business characteristics had a significant impact on management performance through interaction effects. For the profitability variables in detail, other clusters had a significant effect between transportation types(onshore and offshore cargo) and warehousing service type. On the other hand, in stability variables, one cluster was effective in all types, which is a characteristic that lowers both capital intensity and business leadership.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.3
• /
• pp.512-522
• /
• 1996

To provide broadband ISDN service for the users in scattered locations, the application of satellite communications network is seriously considered. To trans mit ATM cells efficiently in satellite communications, it is effective to use TDM A method. However, it is necessary to have a method to compensate the cell delayvari-ation caused by the difference between TDMA and ATM. This paper optimized the cell control time(Tc) when traffic inputs have poisson or markov modulated poisson process by applying cell delay variation characteristics of time stamp method, which has the most advantages among compensation methods or cell delay variation. This paper also intorduces a method of reducing the cell clumping phenomena by adapting discrete time stamp method, including the analysis and evalutation of the range of required quality of CDV distribution by ATM transmission.The result of the experiment shows that CDV distribution-range can be controlled to 1.2$\times$Tc which reduces overall cell delay variation by discrrete time stamp method.

• Journal of the Korea Society for Simulation
• /
• v.25 no.4
• /
• pp.109-115
• /
• 2016

This study intended to testify the causality among three variables such as accident, depression and cognitive failure of the train drivers. For this purpose, two research models were suggested. Model 1 hypothesized the causality among three variables as 'depression ${\rightarrow}$ cognitive failure ${\rightarrow}$ accident'. On the other hand, model 2 hypothesized the causality among three variables as 'accident ${\rightarrow}$ depression ${\rightarrow}$ cognitive failure'. Results based on AMOS using 416 train drivers' questionnaire showed that model 2 is more valid than model 1. The statistical result of model 1 showed that depression has a positive effect on cognitive failure, however no significant relationship between depression and accident as well as between cognitive failure and accident. In model 2, the result showed that the accident has a positive effect on cognitive failure mediated by depression. This result suggests the necessity for establishment of countermeasures to mitigate mistake and cognitive failure caused by train drivers in a wider context, considering the causality between accident and depression.