• Smart Media Journal
• /
• v.9 no.3
• /
• pp.80-89
• /
• 2020

Out - of - distribution (OOD) samples are frequently encountered when deploying a classification model in plenty of real-world machine learning-based applications. Those samples are normally sampling far away from the training distribution, but many classifiers still assign them high reliability to belong to one of the training categories. In this study, we address the problem of removing OOD examples by estimating marginal density estimation using variational autoencoder (VAE). We also investigate other proper methods, such as temperature scaling, Gaussian discrimination analysis, and label smoothing. We use Chonnam National University (CNU) weeds dataset as the in - distribution dataset and CIFAR-10, CalTeach as the OOD datasets. Quantitative results show that the proposed framework can reject the OOD test samples with a suitable threshold.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.29 no.11
• /
• pp.66-73
• /
• 2015

AC-based power systems, having the advantages that voltage transformation and long distance transmission are easy, have been constructed since the last 19th century. However, DC-based power system is paid attention these days because of the development of power electronic devices as well as the increase of digital loads and distributed generation. For instance, the transmission systems using High Voltage DC (HVDC) are commercially operated in the world and the researches on distribution system using Low Voltage DC (LVDC) are gradually increased. This paper analyzes voltage sag, resulted from faults, in LVDC distribution system according to the number of poles. Modeling and simulation with various conditions are conducted by using ElectroMagnetic Transients Program (EMTP). Moreover, some countermeasures to reduce voltage sag in LVDC distribution system are suggested briefly.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.18 no.12
• /
• pp.98-103
• /
• 2019

With the increasing need for environmental protection, it is particularly important to improve the energy saving and reliability of refrigerators. Generally, the cold air flowing into the freezer compartment transits to the bottom of the refrigerating compartment, which can lead to uneven temperature distribution. This paper proposes two design solutions for improving the temperature distribution problem. Of these, the optimal refrigeration design was selected and tested using Computational Fluid Dynamics (CFD) modeling and simulation. The results showed improved uniformity of the temperature distribution inside the refrigerator, thus benefitting food storage while reducing energy consumption.

• Journal of Distribution Science
• /
• v.20 no.3
• /
• pp.59-69
• /
• 2022

Purpose: The purpose of this study is to provide information for other journals as well as the continuous development of distribution science research by confirming the globalization performance of the Journal of Distribution Science (JDS), the main journal of KODISA. Research Design, Data, and Methodology: A total of 863 papers published in JDS from 2011 to 2021 searched by scienceON were divided into 4 periods and analyzed under the headings of submission system, standardity, collaboration, and degree of achievement of publication goals. SPSS 24.0 and R 4.1.1 package were used to perform the publication frequency analysis, crosstab-analysis, keyword frequency analysis, and LDA topic modeling were performed. In addition, trend analysis with weight applied to each word was performed. Results: It was found that the ratio of English-written papers, which is the indicator of a journal's starndardity, is continuously increasing, and the ratio of overseas authors, which is the indicator of collaboration, is also continuously increasing. It was confirmed through keyword trend analysis by period and LDA topic modeling results - which were weighted to confirm the degree of achievement of the journal's publication goal - that the articles published by the journal has been in agreement with monthly research topic proposed by JDS. Conclusion: By examining the five criteria for globalization, it can be concluded that JDS's efforts for globalization are achieving significant results and providing effective directions for other academic journals. However, in order for JDS to become a top academic journal, it was suggested that efforts should be made to establish a system for collaborative research by domestic and foreign authors, as well as to provide a clear definition for the monthly research topics and classification of sub-topics.

• Journal of Electrical Engineering and Technology
• /
• v.12 no.2
• /
• pp.586-593
• /
• 2017

Modeling busbar distribution system as a transmission line is an important subject of power line communication in the smart grid concept. This requires extraction of busbar RLGC parameters, accurately. In this study, a comparison is made between conventional and modified method for the aspect of optimum RLGC parameters extraction in the 1 MHz to 50 MHz frequency band. The usefulness of these methods is shown both in time and frequency-domain analysis. The frequency-domain analyzes show that the inherent power of modified method can eliminate the errors especially due to the discontinuities arise in conventional method. This makes the modeling approach of modified method more advantageous for the busbars due to its robustness against disturbances in the S-parameters measurements which cannot be eliminated with the calibration procedure. On the other hand, time-domain simulations show that the transmission line representation of the modified method is closer to physical reality by handling causality issues.

• Journal of Soil and Groundwater Environment
• /
• v.13 no.2
• /
• pp.54-61
• /
• 2008

The drawdown distribution due to pumping by horizontal or slanted wells is analyzed by numerical modelling. In the numerical modelling uses 1-D discrete element feature included in commercial groundwater modeling program FEFLOW (version 5.1) and the results are compared with the semi analytic solution which uses superposition of successive point sources proposed by Zhan and Zlotnik (2002). Results of the numerical modeling agree well with the semi analytic solution except for very near field region of sink sources. The drawdown distribution due to pumping in riverbank filtration(RBF) plan site can be evaluated quantitatively by the numerical modeling in this study.

• Journal of Korea Society of Industrial Information Systems
• /
• v.12 no.4
• /
• pp.107-118
• /
• 2007

A single-server queueing model with infinite buffer and batch arrival of customers is considered. In contrast to the standard batch arrival when a whole batch arrives into the system at one epoch, we assume that the customers of an accepted batch arrive one-by one in exponentially distributed times. Service time is exponentially distributed. Flow of batches is the stationary Poisson arrival process. Batch size distribution is geometric. The number of batches, which can be admitted into the system simultaneously, is subject of control. Analysis of the joint distribution of the number batches and customers in the system and sojourn time distribution is implemented by means of the matrix technique and method of catastrophes. Effect of control on the main performance measures of the system is demonstrated numerically.

• IE interfaces
• /
• v.22 no.3
• /
• pp.234-242
• /
• 2009

It is most important for distribution center in retail business to delivery commodities in a timely manner. Accordingly, many companies try to make distribution center effective using the Warehouse Management System(WMS) integrated legacy system. Also, the Customer Relationship Management(CRM) is the most typical paradigm in management lately. Even though the WMS and CRM are independent system of each other, WMS, coupled with CRM makes customer satisfied more effectively. In this paper, we proposed the methodology for inventory location after analyzing and applying customer buying pattern data in the CRM through the MBA(Market Basket Analysis), which is part of data mining. We used an example modeling a real distribution center in retail through a 3D simulation tool and examined correlation between commodities using customer buying pattern. After that, we applied it to the inventory location system through the MBA in an example. Finally, we identified decrease in the time for picking, which is the majority of distribution center. Besides, we proposed a simulation methodology before applying new methodology. Consequently, it removes potential errors in advance and makes a optimized inventory location system.

• The Transactions of the Korean Institute of Electrical Engineers C
• /
• v.54 no.9
• /
• pp.391-396
• /
• 2005

In this paper the modeling of interrupting characteristics of a high voltage current limiting fuse to be used in a power distribution system is introduced. In order to reduce the level of energy which can be absorbed by equipment during fault current flow, a high voltage current limiting fuse can generate a high voltage at the fuse terminals. Consequently it is necessary to model and analyze precisely the voltage and current variation during a CL fuse action. The characteristics of CL fuse operation modeled by electrical components have been performed with less than 6 [$\%$] errors. So the fuse designer or manufacturer can estimate the characteristics of CL fuse operation by using this modeling. The Electro Magnetic Transient Program (EMTP) is used to develop the modeling.

• Journal of Electrical Engineering and Technology
• /
• v.3 no.3
• /
• pp.331-336
• /
• 2008

Electric load components have different characteristics according to the variation of voltage and frequency. This paper presents the load modeling of an electric locomotive by the parameter identification method. The proposed method for load modeling is very simple and easy for application. The proposed load model of the electric locomotive is represented by the combination of the loads that have static and dynamic characteristics. This load modeling is applied to the KTX in Korea to verify the effectiveness of the proposed method. The results of proposed load modeling by the parameter identification follow the field measurements very exactly.