• 한국기계가공학회지
• /
• 제19권7호
• /
• pp.7-15
• /
• 2020

Recently, the fourth industrial revolution has been presented as a new paradigm and additive manufacturing (AM) has become one of the most important topics. For this reason, process monitoring for each cross-sectional layer of additive metal manufacturing is important. Particularly, deep learning can train a machine to analyze, optimize, and repair defects. In this paper, image classification is proposed by learning images of defects in the metal cross sections using the convolution neural network (CNN) image labeling algorithm. Defects were classified into three categories: crack, porosity, and hole. To overcome a lack-of-data problem, the amount of learning data was augmented using a data augmentation algorithm. This augmentation algorithm can transform an image to 180 images, increasing the learning accuracy. The number of training and validation images was 25,920 (80 %) and 6,480 (20 %), respectively. An optimized case with a combination of fully connected layers, an optimizer, and a loss function, showed that the model accuracy was 99.7 % and had a success rate of 97.8 % for 180 test images. In conclusion, image labeling was successfully performed and it is expected to be applied to automated AM process inspection and repair systems in the future.

• 대한산업공학회지
• /
• 제38권1호
• /
• pp.46-56
• /
• 2012

The purpose of virtual metrology (VM) in semiconductor manufacturing is to predict every wafer's metrological values based on its process equipment data without an actual metrology. In this paper, we propose novelty detection-based reliability estimation models for VM in order to support flexible utilization of VM results. Because the proposed model can not only estimate the reliability of VM, but also identify suspicious process variables lowering the reliability, quality control actions can be taken selectively based on the reliance level and its causes. Based on the preliminary experimental results with actual semiconductor manufacturing process data, our models can successfully give a high reliance level to the wafers with small prediction errors and a low reliance level to the wafers with large prediction errors. In addition, our proposed model can give more detailed information by identifying the critical process variables and their relative impacts on the low reliability.

• 한국시뮬레이션학회논문지
• /
• 제19권4호
• /
• pp.21-29
• /
• 2010

본 연구는 시뮬레이션 방법을 이용한 자동차용 열 수축 튜브 생산 공정을 가상의 생산 공정 모델로 구현하고, 구현된 모델을 기반으로 열 수축 튜브 생산 공정을 분석하는 연구이다. 자동차용 열 수축 튜브 생산 공정을 분석하기 위해서 공정별로 생성되는 데이터를 수집하고, 수집된 데이터 분석을 통한 자동차용 열 수축 튜브 생산 공정을 분석할 수 있는 가상의 생산 공정 모델을 구현하였다. 구현된 모델을 통해 공정 내에서 발생되는 병목현상 파악 및 원인분석, 공정별 사이클 타임, 부품 생산량 등을 산정함으로써 현 공정 분석 및 개선방안을 모색할 수 있어 기업의 생산 공정관리 효율성을 높일 수 있다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권12호
• /
• pp.3330-3344
• /
• 2023

This paper proposes an antenna performance prediction model in the autonomous driving radar manufacturing process. Our research work is based upon a challenge dataset, Driving Radar Manufacturing Process Dataset, and a typical AutoML machine learning workflow engine, Pycaret open-source Python library. Note that the dataset contains the total 70 data-items, out of which 54 used as input features and 16 used as output features, and the dataset is properly built into resolving the multi-output regression problem. During the data regression analysis and preprocessing phase, we identified several input features having similar correlations and so detached some of those input features, which may become a serious cause of the multicollinearity problem that affect the overall model performance. In the training phase, we train each of output-feature regression models by using the AutoML approach. Next, we selected the top 5 models showing the higher performances in the AutoML result reports and applied the ensemble method so as for the selected models' performances to be improved. In performing the experimental performance evaluation of the regression prediction model, we particularly used two metrics, MAE and RMSE, and the results of which were 0.6928 and 1.2065, respectively. Additionally, we carried out a series of experiments to verify the proposed model's performance by comparing with other existing models' performances. In conclusion, we enhance accuracy for safer autonomous vehicles, reduces manufacturing costs through AutoML-Pycaret and machine learning ensembled model, and prevents the production of faulty radar systems, conserving resources. Ultimately, the proposed model holds significant promise not only for antenna performance but also for improving manufacturing quality and advancing radar systems in autonomous vehicles.

• 대한안전경영과학회지
• /
• 제12권3호
• /
• pp.333-339
• /
• 2010

Data mining is the process of finding and analyzing data from a big database and summarizing it into useful information for a decision-making. A variety of data mining techniques have been being used for wide range of industries. One application of those is especially so for gathering meaningful information from process data in manufacturing factories for quality improvement. The purpose of this paper is to provide a methodology to improve manufacturing quality of fuel tanks which are auto-parts. The methodology is to analyse influential attributes and establish a model for optimal manufacturing condition of fuel tanks to improve the quality using decision tree, association rule, and feature selection.

• 한국정보시스템학회지:정보시스템연구
• /
• 제26권2호
• /
• pp.105-121
• /
• 2017

Purpose It is a major goal to improve the product yields during production operations in the manufacturing industry. Therefore, factory is trying to keep the good quality materials and proper production resources, also find the proper condition of facilities and manufacturing environment for yields improvement. Design/methodology/approach We propose the hybrid framework to analyze to dataset extracted from MES. Those data is about the alarm information generated from equipment, both measurement and equipment process value from production and cycle/pitch time measured from production data these covered products during production. We adapt a data warehousing techniques for organizing dataset, a logistic regression for finding out the significant factors, and a association analysis for drawing the rules which affect the product yields. And then we validate the framework by applying the real data generated from the discrete process in secondary cell battery manufacturing. Findings This paper deals with challenges to apply the full potential of modeling and simulation within CPPS(Cyber-Physical Production System) and Smart Factory implementation. The framework is being applied in one of the most advanced and complex industrial sectors like semiconductor, display, and automotive industry.

• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회/대한산업공학회 2005년도 춘계공동학술대회 발표논문
• /
• pp.960-968
• /
• 2005

The high cost in maintaining complex manufacturing process makes it necessary to enhance an efficient maintenance system. For the effective maintenance of manufacturing process, precise fault diagnosis should be performed and an appropriate maintenance action should be executed. This paper suggests an intelligent fault diagnosis system using hybrid data mining. In this system, the rules for the fault diagnosis are generated by hybrid decision tree/genetic algorithm and the most effective maintenance action is selected by decision network and AHP. To verify the proposed intelligent fault diagnosis system, we compared the accuracy of the hybrid decision tree/genetic algorithm with one of the general decision tree learning algorithm(C4.5) by data collected from a coil-spring manufacturing process.

• 산업경영시스템학회지
• /
• 제34권4호
• /
• pp.57-65
• /
• 2011

Monitoring auto correlated processes is prevalent in recent manufacturing environments. As a proactive control for manufacturing processes is emphasized especially in the semiconductor industry, it is natural to monitor real-time status of equipment through sensor rather than resultant output status of the processes. Equipment's sensor data show various forms of correlation features. Among them, considerable amount of sensor data, statistically autocorrelated, is well represented by Box-Jenkins autoregressive moving average (ARMA) model. In this paper, we present a design method of statistical process control (SPC) used for monitoring processes represented by the ARMA model. The proposed method shows benefits in the power of detecting process changes, and considers robustness to ARMA modeling errors simultaneously. We prove benefits through Monte carlo simulation-based investigations.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제4권9호
• /
• pp.357-370
• /
• 2015

제조현장에서 발생하는 대량의 데이터는 제조공정의 개선 등을 위해서 매우 중요한 역할을 수행한다. 국내 제조업은 제조환경의 변화에 대응하기 위하여 다각적인 노력을 하고 있으나 구조적, 기술적 취약성으로 인해 많은 어려움을 겪고 있다. 코팅제는 도료의 일종으로 제품에 피막층을 형성하여 제품을 보호하고 다양한 특성을 부여하는 고분자 산업에서 활발하게 연구되는 분야 중의 하나이다. 코팅제는 다양한 산업 분야에서 중요성이 더욱 커지고 있으나 실제 제조업체에서는 여전히 작업자의 경험에 의존하여 배합공정을 수행하는 실정이다. 본 논문에서는 HW-SW 통합 프레임워크를 활용한 제조공정 개선을 위한 실시간 모니터링 시스템과 데이터 분석검증 TA설계를 제안한다. 제안된 프레임워크를 통한 분석 결과는 보다 정량적인 작업 기준 데이터를 확보하고 작업 현장에 제공함으로써 코팅제 배합 공정을 개선시킬 수 있다. 특히 정확한 배합 기준이 되는 표준 데이터의 부재로 인한 품질 저하와 원가 손실을 감소시키고, 배합 공정에서 발생한 오차 데이터에 대하여 R과 실험 계획법을 이용한 분석을 통하여 표준 보정 관계식을 도출함으로써 차후 발생 가능한 오차에 대한 대응 방안을 제시한다.

• International Journal of Precision Engineering and Manufacturing
• /
• 제7권3호
• /
• pp.56-59
• /
• 2006

UV laser micromachining is generally used to create microstructures for micro-products through a sequence of lithography-based photo-patterning steps. However, the micromachining process is not suitable for rapid realization of complex 3D micro-products because it depends on worker experience. In addition, the cost and time required to make many masks are excessive. In this paper, a more effective and rapid micro-manufacturing process, which was developed based on laser micromachining, is proposed for fabricating micro-products directly using UV laser ablation and phase-change filling. The filling process is useful for holding the micro-products during the ablation step. The proposed rapid micro-manufacturing process was demonstrated experimentally by fabricating 3D micro-products from functional UV-sensitive polymers using 3D CAD data.