• Journal of Intelligence and Information Systems
• /
• v.2 no.1
• /
• pp.109-120
• /
• 1996

TV 부라운관 유리를 생산하기 위한 용해 공정은 원료의 투입에서부터 유리물이 생성될 때 까지 고온의 화확적 작용을 거친다. 유리몰을 생성하기 위한 여러 가지 조건중에서 용해로 내부에서의 유리몰의 온도 변화 및 이에 따른 제반 공정변수의 조정 설정치(Set-up)는 불량률에 밀접한 영향을 미친다. 그러나 고온의 밀폐된 환경에서 반응이 진행되므로 공정의 운전 요원들은 그들의 경험을 바탕으로 용해로의 운전상태를 파악하고 운전해 나간다. 본 연구에서는 이러한 경험적 판단에 따른 위험성을 가능한 한 줄이고 용해로의 안정적인 운영을 통하여 불량률을 감소시키기 위하여 용해로 최적 Set-up 예측 시스템을 개발하였다. 이 시스템은 일정 기간 동안의 용해로서 운전 상태와 불량률간의 관계를 신경망 기법을 이용하여 학습한 후에, 이를 이용하여 불량률을 유지하기 위한 Set-up 값을 예측하여 준다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.10a
• /
• pp.290-291
• /
• 2021

Defects in the coating process of plastic automotive components are caused by various causes and phenomena. The correlation between defect occurrence rate and meteorological and environmental conditions such as temperature, humidity, and fine dust was analyzed. The defect rate data categorized by type and cause was collected for a year from a automotive parts coating company. This data and its correlation with environmental condition was acquired and experimented by machine learning techniques to predict the defect rate at a certain environmental condition. Correspondingly, the model predicted 98% from fine dust and 90% from curtaining (runs, sags) and hence proved its reliability.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1995.09a
• /
• pp.387-395
• /
• 1995

롤(Roll)의 주물 가공 시스템의 성능(Performance)분석의 문제는 일반적으로 관련된 많은 요인들 때문에 매우 복잡하다. 롤 제조 시스템의 성능과 관련된 주요 요인으로서 주형(Moulding) 제작, 원재료의 용해, 후처리 및 가공 공정 의 요인들을 들 수 있다. 본 연구에서는 이러한 복잡한 롤의 주물 및 가공공 정상의 요인들로부터 롤 제조 시스템의 불량률을 평가하기 위하여 발견적인 방법인 GMDH(Group Method Data Handling)-Type 모델링 방법을 이용하 였다. 롤 주물 가공 시스템의 성능을 불량률로 두고 이에 주요 영향 요인들 의 입력 Data를 위하여 현장 자료로부터 상하한 값을 구하여, Hyper-Cube 프로그램을 이용하여 필요한 수의 Data를 보완하여 사용하였다. 시스템 성능 과 관련된 인자들을 2개식 가능한 조합을 하고 이들 각각의 조합들에 대하 여 6개항으로 된 예측식으로 회귀분석하고 일정 수준 이상의 결과들만을 다 음 단계의 자료로 사용하였다. GMDH 방법은 매 단계마다 영향이 적은 변 수조합을 제외시키므로 최종 해는 그 정확성이 매우 높다. 본 연구를 위하여 GMDH 알고리즘에 따라 계산할수 있는 전산 프로그램을 개발하여 사용하였 으며, 적용예를 롤 주물제조공정에 응용하여 보였다.. 분석된 자료에 의하면 예측 오차가 매우 적음을 보였다.

• Journal of the Korean Society of Radiology
• /
• v.12 no.6
• /
• pp.785-791
• /
• 2018

Recently, Disk array is widely used as a long term storage device in PACS, but reliability is not enough in relation to annual failure rate of disk. Between October 2016 and February 2017, we scanned the serial port of the hard disk while reading or storing medical images on a PACS reader. The data rate was calculated from the data stored in HDD 99ea that were used in the PCAS image storage device and in HDD 101ea that were used in the Personal Computer. When a CT image was read from a PACS reader, Reading was 87.8% and Writing was 12.2% in units of several tens of megabytes or less. When the CT image was stored in the PACS reader, Reading was 11.4% and Writing was 88.6% in units of several tens of megabytes or less. While reading the excel file on the personal computer, Reading was 75% and Writing was 25% in less than 3 MB, and In the process of storing the excel file on the personal computer, Reading was carried out by 38% and Writing was carreid out 62% in the units of 3 MB or less. The transfer rate of the hard disk used in the PACS image storage device was 10 GB/h, and the transfer rate per hour of the hard disk of the personal computer was 5 GB / h. Annual failure rate of hard disk of image storage system is 0.97 ~ 1.13%, Annual failure rate of Hard Disk of personal computer is 0.97 ~ 1.13%. the higher transfer rate is, the higher annual failure rate is. These results will be used as a basis for predicting the life expectancy of the hard disk and the annual failure rate.

• Journal of Technology Innovation
• /
• v.26 no.3
• /
• pp.37-68
• /
• 2018

Since the 20th century, automobiles, which are the most common means of transportation, have been evolving as the use of electronic control devices and automotive semiconductors increases dramatically. Automotive semiconductors are a key component in automotive electronic control devices and are used to provide stability, efficiency of fuel use, and stability of operation to consumers. For example, automotive semiconductors include engines control, technologies for managing electric motors, transmission control units, hybrid vehicle control, start/stop systems, electronic motor control, automotive radar and LIDAR, smart head lamps, head-up displays, lane keeping systems. As such, semiconductors are being applied to almost all electronic control devices that make up an automobile, and they are creating more effects than simply combining mechanical devices. Since automotive semiconductors have a high data rate basically, a microprocessor unit is being used instead of a micro control unit. For example, semiconductors based on ARM processors are being used in telematics, audio/video multi-medias and navigation. Automotive semiconductors require characteristics such as high reliability, durability and long-term supply, considering the period of use of the automobile for more than 10 years. The reliability of automotive semiconductors is directly linked to the safety of automobiles. The semiconductor industry uses JEDEC and AEC standards to evaluate the reliability of automotive semiconductors. In addition, the life expectancy of the product is estimated at the early stage of development and at the early stage of mass production by using the reliability test method and results that are presented as standard in the automobile industry. However, there are limitations in predicting the failure rate caused by various parameters such as customer's various conditions of use and usage time. To overcome these limitations, much research has been done in academia and industry. Among them, researches using data mining techniques have been carried out in many semiconductor fields, but application and research on automotive semiconductors have not yet been studied. In this regard, this study investigates the relationship between data generated during semiconductor assembly and package test process by using data mining technique, and uses data mining technique suitable for predicting potential failure rate using customer bad data.

• Korean Journal of Optics and Photonics
• /
• v.17 no.5
• /
• pp.420-429
• /
• 2006

In the zoom lens of digital still cameras with the variation of the image plane generated by various symmetric error factors such as curvature, thickness and refractive index error of each lens surface about the optic axis, we induce a theoretical condition to fix constantly the image plane by translating the compensator group of the zoom lens by using the Gaussian bracket. We confirm the validity of this condition by using three examples of general zoom lens types with 3, 4, and 5 groups, respectively. When these error factors are randomly changed within the range of tolerance according to the Monte Carlo method, we verify that the distributions of the degree of moving of the compensator are normal distributions at three zoom lens types. From capability analysis using these results, we theoretically propose the method estimating the standard deviation, that is, sigma-level, as a function of the maximum movement of the compensator.

• Journal of the Korea Convergence Society
• /
• v.13 no.2
• /
• pp.249-255
• /
• 2022

This study proposed an analysis framework for real-time prediction of CNC processing defects using machine learning-based models that are recently attracting attention as processing defect prediction methods, and applied it to CNC machines. Analysis shows that the XGBoost, CatBoost, and LightGBM models have the same best accuracy, precision, recall, F1 score, and AUC, of which the LightGBM model took the shortest execution time. This short run time has practical advantages such as reducing actual system deployment costs, reducing the probability of CNC machine damage due to rapid prediction of defects, and increasing overall CNC machine utilization, confirming that the LightGBM model is the most effective machine learning model for CNC machines with only basic sensors installed. In addition, it was confirmed that classification performance was maximized when an ensemble model consisting of LightGBM, ExtraTrees, k-Nearest Neighbors, and logistic regression models was applied in situations where there are no restrictions on execution time and computing power.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.37 no.8
• /
• pp.975-981
• /
• 2013

A robust optimization is only one of the ways to minimize the effects of variances in design variables on the objective functions at the preliminary design stage. To predict the variances and to formulate the probabilistic constraints are the most important procedures for the robust optimization formulation. Though several methods such as the process capability index and the six sigma technique were proposed for the prediction and formulation of the variances and probabilistic constraints, respectively, there are few attempts using a percent defective which has been widely applied in the quality control of the manufacturing process for probabilistic constraints. In this study, the robust optimization for a lower control arm of automobile vehicle was carried out, in which the design space showing the mean and variance sensitivity of weight and stress was explored before robust optimization for a lower control arm. The 2nd order Taylor expansion for calculating the standard deviation was used to improve the numerical accuracy for predicting the variances. Simplex algorithm which does not use the gradient information in optimization was used to convert constrained optimization into unconstrained one in robust optimization.

• Journal of Internet Computing and Services
• /
• v.25 no.2
• /
• pp.21-28
• /
• 2024

Radar is an essential sensor component in autonomous vehicles, and the market for radar applications in this context is steadily expanding with a growing variety of products. In this study, we aimed to enhance the stability and performance of radar systems by developing and evaluating a radar performance prediction model that can predict radar defects. We selected seven machine learning and deep learning algorithms and trained the model with a total of 49 input data types. Ultimately, when we employed an ensemble of 17 models, it exhibited the highest performance. We anticipate that these research findings will assist in predicting product defects at the production stage, thereby maximizing production yield and minimizing the costs associated with defective products.

• Proceedings of the Optical Society of Korea Conference
• /
• 2007.02a
• /
• pp.62-63
• /
• 2007