• The Bulletin of The Korean Astronomical Society
• /
• v.44 no.1
• /
• pp.82.3-82.3
• /
• 2019

In this study, we generate future full disk magnetograms in 12, 24, 36 and 48 hours advance from SDO/HMI images using deep learning. To perform this generation, we apply the convolutional generative adversarial network (cGAN) algorithm to a series of SDO/HMI magnetograms. We use SDO/HMI data from 2011 to 2016 for training four models. The models make AI-generated images for 2017 HMI data and compare them with the actual HMI magnetograms for evaluation. The AI-generated images by each model are very similar to the actual images. The average correlation coefficient between the two images for about 600 data sets are about 0.85 for four models. We are examining hundreds of active regions for more detail comparison. In the future we will use pix2pix HD and video2video translation networks for image prediction.

• Nuclear Engineering and Technology
• /
• v.53 no.4
• /
• pp.1181-1188
• /
• 2021

When abnormal events occur in a nuclear power plant, operators must conduct appropriate abnormal operating procedures. It is burdensome though for operators to choose the appropriate procedure considering the numerous main plant parameters and hundreds of alarms that should be judged in a short time. Recently, various research has applied deep-learning algorithms to support this problem by classifying each abnormal condition with high accuracy. Most of these models are trained with simulator data because of a lack of plant data for abnormal states, and as such, developed models may not have tolerance for plant data in actual situations. In this study, two approaches are investigated for a deep-learning model trained with simulator data to overcome the performance degradation caused by noise in actual plant data. First, a preprocessing method using several filters was employed to smooth the test data noise, and second, a data augmentation method was applied to increase the acceptability of the untrained data. Results of this study confirm that the combination of these two approaches can enable high model performance even in the presence of noisy data as in real plants.

• Structural Monitoring and Maintenance
• /
• v.7 no.4
• /
• pp.345-365
• /
• 2020

At present, many machine leaning and data mining methods are used for analyzing and predicting structural response characteristics. However, the platform that combines big data analysis methods with online and offline analysis modules has not been used in actual projects. This work is dedicated to developing a multifunctional Hadoop-Spark big data platform for bridges to monitor and evaluate the serviceability based on structural health monitoring system. It realizes rapid processing, analysis and storage of collected health monitoring data. The platform contains offline computing and online analysis modules, using Hadoop-Spark environment. Hadoop provides the overall framework and storage subsystem for big data platform, while Spark is used for online computing. Finally, the big data Hadoop-Spark platform computational performance is verified through several actual analysis tasks. Experiments show the Hadoop-Spark big data platform has good fault tolerance, scalability and online analysis performance. It can meet the daily analysis requirements of 5s/time for one bridge and 40s/time for 100 bridges.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.62 no.2
• /
• pp.155-163
• /
• 2013

As the increasing HIF(High Impedance Fault) with the arc cannot be easily detected for the low fault current magnitude compared to actual load in distribution line. However, the arcing current shows that the magnitude varies with time and the signal is asymmetric. In addition, discontinuous changes occur at starting point of arc. Considering these characteristics, wavelet transformation of actual current data shows difference between before and after the fault. Althogh raw data(detail coefficient) of wavelet transform may not be directly applied to HIF detection logic in a device, there are several developing methods of HIF monitoring data using the original wavelet coefficients. In this paper, a simple and effective developing methods of HIF monitoring data were analized by using the signal data through an actual HIF experiment to apply them to economic devices. The methods using the sumation of the wavelet coefficient squares in one cycle of the fundamental frequency as the energies of the wavelet coefficeits and the sumation of the absolute values were compared. Besides, the improved method which less occupies H/W resouces and can be applied to field detection devices was proposed. and also Verification of this HIF detection method through field test on distribution system in KEPCO power testing center was performed.

• Journal of the Korea Institute of Building Construction
• /
• v.10 no.1
• /
• pp.81-89
• /
• 2010

Since 2004, the government has changedthe cost estimate system to one of an actual cost basis in order to calculate the optimum construction cost by reflecting changes in circumstance on the construction site in a timely manner. Currently, this is being applied to public construction work forover a billion won of actual cost data in estimation by contract unit cost. However, directly reflecting actual cost, which for large-sized construction work was originally an average unit cost, to a small building, entails the application of a low discount rate for the cost of materials, labor, etc. and therefore can frequently give rise to cases in which the actual cost of work performed exceeds the contract sum, which in turn causes problems such as decreased revenues, bad effectson business operation, productivity, etc. Therefore, to apply actual cost to small-sized construction work (less than a billion won), there should be a plan to modify unit cost in a manner that can reflect project scale, etc. in order to resolve the problem of unit cost application of actual cost to small-building construction projects. The unit cost modification model for proper actual construction cost in small-scale construction projects developed by this study will help to increase the relevant productivity and proper gain, preventing the aggravation of business operations. Organizations placing orders are also expected to be able to secure a more realistic construction cost in arranging the budget.

• 한국신재생에너지학회:학술대회논문집
• /
• 2010.11a
• /
• pp.188.1-188.1
• /
• 2010

This paper describes the effort for the development of an actual wind simulation method on the wind turbine performance evaluation. It should be emphasized that the deep knowledge on real wind field is a key factor for both the design of a wind turbine and the performance evaluation. With this reason, there had been several simulation attempts to accurately match with the actual wind data. With an existing wind generation algorithm is under consideration, this study introduces several more new concepts including Van der Hoven spectrum being implemented in different methodology. Also this paper will compare the result from the wind simulations by using the basic formula with that by using MATLAB and SIMULINK previously developed. In addition, like the existing wind generation algorithm, random process for actual wind field simulation and white noise are incorporated to closely produce the actual wind field models.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.22 no.1
• /
• pp.61-71
• /
• 1998

The purpose of this study was to identify the factors which explain consumer's information search behavior. Data were analyzed by utilizing factor analysis and multiple-regression to investigate the relationship among information sources, benefit, and actual and ideal self- image and demographics. Based on the results, information sources for benefit, actual and ideal self-image were developed. Predictors of information sources, benefits, and self-image were identified. Marketing implication about information sources were discussed. The results were as follows. 1. Actual and ideal self-images and information sources were devised into three factors. And benefits were devised into five factors. 2. Actual self-image has more predicting power than ideal self-image to the benefits which consumer sought. Among five benefits, character pursuit was the best predicted factor according to the self-image. 3. Among three information sources, mass communication was the most effective source which can be explained by the benefit and self-image. Fashion pursuit factor was the most significant factor to the mass communication oriented source.

• Journal of the Korean Institute of Rural Architecture
• /
• v.11 no.1
• /
• pp.17-24
• /
• 2009

The purpose of this study is to clarify the actual condition of real certification schools of environment-friendly architecture per ecological environment elements through extracting ecological environment elements which is possible to analyze in quantity certification standards of environment friendly architecture. As the method of this study, firstly, certification as an examination on certification system of environment-friendly architecture, summary of certification system of environment-friendly architecture and eco-school pilot model project, ecological environment elements which is possible for quantitative analysis of ecological environment certification standards were extracted. Secondly, actual condition of ecological environment elements per school grade by collecting actual data of certification schools environment-friendly architecture were analyzed. As a result, total 35 environment-friendly school facilities are clarified that the average composition ratio of ecological area ratio is 23.0% in case of this certification regardless of school grade and region.

• Journal of Korean Society of Transportation
• /
• v.13 no.3
• /
• pp.5-17
• /
• 1995

One of interesting issues in urban land-use and transportation policy is about an unproven belief that imbalances between employment and residential sites strongly influence urban commuting patterns. The purpose of this study si to find how spatial distribution of jobs and housing affects commuting distances in Seoul Metropolitan Area using 1990 Census O-D Data. This study estimates cost-minimizing required commuting distance, given the actual distributions of jobs and housing , and compares required commuting distances with actual commuting distances. One of findings is that the standard urban economic theory explains only 1/3 of actual commuting in the case study area. The policy implication of this study is that jobs-housing balance policy will have only a minor effect on the actual commuting.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.2271-2274
• /
• 2005

The desired ZMP is different from the actual ZMP of a humanoid robot during actual walking and stand upright. A humanoid robot must maintain its stable posture although external force is given to the robot. A humanoid robot can know its stability with ZMP. Actual ZMP may be moved out of the foot-print polygons by external disturbance or uneven ground surfaces. If the position of ZMP moves out of stable region, the stability can not be guaranteed. Therefore, The control of the ZMP is necessary. In this paper, ZMP control algorithm is proposed. Herein, the ZMP control uses difference between desired ZMP and actual ZMP. The proposed algorithm gives reaction moment with ankle joint when external force is supplied. 3D simulator shows motion of a humanoid robot and calculated data.