• Journal of Korean Society for Quality Management
• /
• v.51 no.1
• /
• pp.55-65
• /
• 2023

Purpose: The purpose of this study is to propose an optimization process to improve product yield in the process using process data. Recently, research for low-cost and high-efficiency production in the manufacturing process using machine learning or deep learning has continued. Therefore, this study derives major variables that affect product defects in the manufacturing process using eXplainable Artificial Intelligence(XAI) method. After that, the optimal range of the variables is presented to propose a methodology for improving product yield. Methods: This study is conducted using the injection molding machine AI dataset released on the Korea AI Manufacturing Platform(KAMP) organized by KAIST. Using the XAI-based SHAP method, major variables affecting product defects are extracted from each process data. XGBoost and LightGBM were used as learning algorithms, 5-6 variables are extracted as the main process variables for the injection process. Subsequently, the optimal control range of each process variable is presented using the ICE method. Finally, the product yield improvement methodology of this study is proposed through a validation process using Test Data. Results: The results of this study are as follows. In the injection process data, it was confirmed that XGBoost had an improvement defect rate of 0.21% and LightGBM had an improvement defect rate of 0.29%, which were improved by 0.79%p and 0.71%p, respectively, compared to the existing defect rate of 1.00%. Conclusion: This study is a case study. A research methodology was proposed in the injection process, and it was confirmed that the product yield was improved through verification.

• Korean Journal of Remote Sensing
• /
• v.34 no.6_3
• /
• pp.1383-1398
• /
• 2018

Sea surface wind is one of the most fundamental variables for understanding diverse marine phenomena. Although scatterometers have produced global wind field data since the early 1990's, the data has been used limitedly in oceanic applications due to it slow spatial resolution, especially at coastal regions. Synthetic Aperture Radar (SAR) is capable to produce high resolution wind field data. KOMPSAT-5 is the first Korean satellite equipped with X-band SAR instrument and is able to retrieve the sea surface wind. This study presents the validation results of sea surface wind derived from the KOMPSAT-5 backscattering coefficient data for the first time. We collected 18 KOMPSAT-5 ES mode data to produce a matchup database collocated with buoy stations. In order to calculate the accurate wind speed, we preprocessed the SAR data, including land masking, speckle noise reduction, and ship detection, and converted the in-situ wind to 10-m neutral wind as reference wind data using Liu-Katsaros-Businger (LKB) model. The sea surface winds based on XMOD2 show root-mean-square errors of about $2.41-2.74m\;s^{-1}$ depending on backscattering coefficient conversion equations. In-depth analyses on the wind speed errors derived from KOMPSAT-5 backscattering coefficient data reveal the existence of diverse potential error factors such as image quality related to range ambiguity, discrete and discontinuous distribution of incidence angle, change in marine atmospheric environment, impacts on atmospheric gravity waves, ocean wave spectrum, and internal wave.

• Journal of the Korean association of regional geographers
• /
• v.22 no.1
• /
• pp.240-250
• /
• 2016

Monitoring for water quantity and quality was conducted in this study for 2 years (2012~2013) in Oenam Stream which is a tributary of Seomjin River and upstream of Juam Lake. Suspended solid and total phosphorous(TP) were monitored and analyzed, then water quantity and quality as well as their relation with landuses were identified based on the previous study. Flow showed the similar pattern with precipitation but some discrepancies existed due to the distance between weather station(Gwangju) and study area. Watershed was modeled based on observed data using SWAT(Soil and Water Assessment Tool). Model calibration was conducted using data obtained in 2012 and validation was conducted using data in 2013. The coefficient of determination ($R^2$) between observed and modeled showed 0.6644 and 0.5176 for flow and TP, respectively for model calibration period. For validation period, $R^2$ was 0.7529 for flow and 0.7057 for TP, which were higher than calibration period. Hot spots were determined for watershed management by analyzing the amount of sediment and TP outcome from each sub-watershed. TP loading by landuse determined that cropland, of which the area takes only 5% from entire watershed, generated 53.6% of TP and residential and cowshed was responsible for 23.5% of TP loading.

• Korean Journal of Remote Sensing
• /
• v.36 no.6_2
• /
• pp.1493-1507
• /
• 2020

The LEOP Cal/Val (Launch and Early Operation Phase Calibration/Validation) was carried out during 6 months after KOMPSAT-3A (KOMPSAT-3A Korea Multi-Purpose Satellite-3A) was launched in March 2015. After LEOP Cal/Val was successfully completed, high resolution KOMPSAT-3A has been successfully distributing to users over the past 8 years. The sub-meter high-resolution satellite image data obtained from KOMPSAT-3A is used as basic data for qualitative and quantitative information extraction in various fields such as mapping, GIS (Geographic Information System), and national land management, etc. The KARI (Korea Aerospace Research Institute) periodically checks and manages the quality of KOMPSAT-3A's product and the characteristics of satellite hardware to ensure the accuracy and reliability of information extracted from satellite data of KOMPSAT-3A. To minimize the deterioration of image quality due to aging of satellite hardware, payload and attitude sensors of KOMPSAT-3A, continuous improvement of image quality has been carried out. In this paper, the Cal/Val work-flow defined in the KOMPSAT-3A development phase was illustrated for the period of before and after the launch. The MTF, SNR, and location accuracy are the key parameters to estimate image quality and the methods of the measurements of each parameter are also described in this work. On the basis of defined quality parameters, the performance was evaluated and measured during the period of after LEOP Cal/Val. The current status and characteristics of MTF, SNR, and location accuracy of KOMPSAT-3A from 2016 to May 2020 were described as well.

• Journal of Bio-Environment Control
• /
• v.3 no.1
• /
• pp.10-19
• /
• 1994

Wind load is known to be one of major forces to influence the stability of agricultural structures. General flow fields were calculated to determine flow characteristics over the envelop of the following three types of greenhouses with arched roof : single span, twin span greenhouses, and two single span greenhouses apart 3m inbetween. Pressure coefficients along the envelop of greenhouse were numerically calculated by the k-$\varepsilon$ turbulence model, which lead to determine wind forces on it. Curvilinear coordinate for an arched roof and the upwind scheme were adopted for the study. The calculated pressure coefficients were validated with the avaliable data of Japanese Standard and NGAM Standard. The Magnitude of calculated forces over the envelop was not in good accordance with data except the windward wall. Even tile data of Japanese and NGAM Standard for validation deviated a lot from each other in quantity and quality. Such discrepancy may be attributed to different geometric and/or flow configuration conditions for experiments, or the insenstivity of the k-$\varepsilon$ turbulence model to recirculation flow.

• Structural Engineering and Mechanics
• /
• v.68 no.6
• /
• pp.691-700
• /
• 2018

Ground granulated blast furnace slag (GGBS) is a by product obtained from iron and steel industries, useful in the design and development of high quality cement paste/mortar and concrete. This paper investigates the applicability of relevance vector machine (RVM) based regression model to predict the compressive strength of various GGBS based concrete mixes. Compressive strength data for various GGBS based concrete mixes has been obtained by considering the effect of water binder ratio and steel fibres. RVM is a machine learning technique which employs Bayesian inference to obtain parsimonious solutions for regression and classification. The RVM is an extension of support vector machine which couples probabilistic classification and regression. RVM is established based on a Bayesian formulation of a linear model with an appropriate prior that results in a sparse representation. Compressive strength model has been developed by using MATLAB software for training and prediction. About 70% of the data has been used for development of RVM model and 30% of the data is used for validation. The predicted compressive strength for GGBS based concrete mixes is found to be in very good agreement with those of the corresponding experimental observations.

• The Bulletin of The Korean Astronomical Society
• /
• v.45 no.1
• /
• pp.63.2-63.2
• /
• 2020

FLASH is a blind neutral hydrogen (HI) absorption line survey, eventually targeting about 100,000 background radio continuum sources in the entire southern sky using the full 36-antenna of the Australian Square Kilometre Array Pathfinder (ASKAP). Our primary goal is to search for associated and intervening HI absorption lines in the intermediate redshift range 0.4 < z < 1.0. The survey aims to understand the evolution of HI gas in galaxies as well as various physical mechanisms in active galactic nuclei, such as accretion and feedback processes. In this poster, we give an overview of the FLASH survey and present the preliminary results from our first 100-hrs of pilot observations. The latest survey data covers 1,000 square degrees and is ideal for validating observation and data processing in the continuous 300MHz-width low frequency ASKAP band (700-1000MHz). One of the crucial objectives of the pilot survey is to establish the analysis methodology that will be applied to upcoming large absorption surveys in the future. We discuss our data quality validation and present some detections of associated/intervening HI absorption lines. These absorption lines allow us to trace the cold gas properties of active and normal galaxies at higher redshifts where the HI emission line is too weak to be detectable.

• Journal of Korean Society of Environmental Engineers
• /
• v.34 no.5
• /
• pp.312-315
• /
• 2012

In this study, the model which can predict the quality of effluent has been implemented through multiple regression analysis to use operation data of a sewage treatment plant, to which a media process is applied. Multiple regression analysis were carried out by cases according to variable selection method, removal of outliers and log transformation of variables, with using data of one year of 2011. By reviewing the results of predictable models, the accuracy of prediction for $COD_{Mn}$ of treated water of secondary clarifiers was over 0.87 and for T-N was over 0.81. Using this model, it is expected to set the range of operating conditions that do not exceed the standards of effluent quality. In conclusion, the proper guidance on the effluent quality and energy costs within the operating range is expected to be provided to operators.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.39 no.2
• /
• pp.183-194
• /
• 2011

Synthetic aperture radar (SAR) is widely used because of high resolution imaging capability in all weather and day/night condition. In this paper development of Spotlight SAR simulator is proposed for image quality analysis. Proposed SAR simulator is based on the SAR system design parameters so that SAR image performance can be expected which is essential throughout the full system development procedure from the initial concept design stage to the final in-flight calibration and validation stage. The raw data of ideal point target is first generated by taking account of the flight and imaging geometry and the various SAR system design parameters, and the Spotlight image formation algorithm is implemented in order to obtain the point target response. Finally the image quality of the generated raw data is analyzed in terms of spatial resolution, peak to sidelobe ratio and integrated sidelobe ratio.

• Journal of Korean Society on Water Environment
• /
• v.35 no.6
• /
• pp.589-597
• /
• 2019

The purpose of this study was improve accuracy the IOPs inversion model(IOPs-IM) developed in 2016 for phycocyanin(PC) concentration estimation in the Nakdong River. Additionally, two optimum models were developed and evaluated with 2017 measurement field spectral data for the Geum River and the Yeongsan River. The used measurement data for IOPs-IM analyzation was randomly classified as training and verification materials at the ratio of 2:1 in all data sets. Using the training data set from 2015-2017, accuracy results of the IOPs-IM generally improved for the Nakdong River. The RMSE(Root Mean Square Error) decreased by 14 % compared to 2016. For the GeumRiver, the results of the IOPs-IM were suitable, except for some point results in 2016. Results of the IOPs-IM in the Yeongsan River followed the overall 1:1 line and MAE(Mean Absolute Error) was lower than other rivers. But the RMSE and MAE values were higher. As a result of applying the validation data to the IOPs-IM, the accuracy of the Nakdong River was reduced to RMSE 17.7 % and MRE 16.4 %, respectively compared with 2016. However, the MRE(Mean Relative Error) was estimated to be higher by 400 % in the Geum River, and the RMSE was more than 100 mg/㎥ of the Yeongsan River. Therefore, it is necessary to get the continuously data with various sections of each river for obtain objective and reliable results and the models should be improved.