• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.05a
• /
• pp.459-462
• /
• 2021

본 연구는 반도체 제조 공정중 발생하는 센서 데이터를 활용하여 딥러닝기반으로 불량을 예측하는 모델을 제안한다. 반도체 공장에서는 FDC((Fault Detection and Classification)라는 불량을 예측하는 시스템이 있지만, 공정의 복잡도가 높고 센서의 종류가 많아 공정 관리자가 모든 센서의 기준을 설정 및 관리하는데 한계가 있다. 이를 해결하기 위해 공정 설비의 센서 데이터를 딥러닝을 활용하여 학습시켜 센서 기준정보로 임계치를 제공하고, 가공중 발생하는 센서 데이터가 입력되면 정상 여부를 판정하는 모델을 제안한다.

• The Journal of Bigdata
• /
• v.3 no.1
• /
• pp.95-103
• /
• 2018

In recent years, many people in the manufacturing field have been making efforts to increase efficiency while analyzing manufacturing data generated in the process according to the development of ICT technology. In this study, we propose a data mining based manufacturing process using decision tree algorithm (CHAID) as part of a smart factory. We used 432 sensor data from actual manufacturing plant collected for about 5 months to find out the variables that show a significant difference between the stable process period with low defect rate and the unstable process period with high defect rate. We set the range of the stable value of the variable to determine whether the selected final variable actually has an effect on the defect rate improvement. In addition, we measured the effect of the defect rate improvement by adjusting the process set-point so that the sensor did not deviate from the stable value range in the 14 day process. Through this, we expect to be able to provide empirical guidelines to improve the defect rate by utilizing and analyzing the process sensor data generated in the manufacturing industry.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.6 no.1
• /
• pp.43-48
• /
• 2011

In this paper, We have investigated the weak waterproofing defection characteristic of polymer humidity sensor and identified the cause of failure. In high temperature and high humidity conditions(($60^{\circ}C/95%$), the defecting process is simulated about defective and improved samples which are modified for the crosslinking polymer structure's aspect. It is aimed at the defecting reason and suggestion of defection process mechanism.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.19 no.1
• /
• pp.95-100
• /
• 2011

The test was done on cars travelling at speeds of 20km/h, 60km/h and 100km/h, the performance testing mode for chassis dynamometer. In this test, the secondary waveform were measured, including those using faulty MAP sensors, oxygen sensors and spark plugs. The results from these measurements and their analysis of secondary waveform can be summarized as follows: 1) The secondary waveform measured from the faulty oxygen sensor showed a lot of noise around peak voltage and in the rising and falling sections during spark line which means that the air fuel mixture was non-homogeneous. 2) The secondary waveform from the faulty MAP sensor showed the worst shape compared to other sensors, including variation of spark line, state of air-fuel mixture and velocity of flame front. 3) The spark line time of secondary waveform using a faulty spark plug displayed the shortest and smallest energy spark line, which means that a misfire occurred.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2015.07a
• /
• pp.1-4
• /
• 2015

현재 고해상도 디스플레이 제품 생산은 대량 생산 공정 시스템으로 가동하고 있으며, 대량 생산 과정에서 WAFER의 제작 불량률을 낮추는 것이 생산업체에서 무엇보다도 주요한 목표이며 이와 함께 불량 제품을 정확하고 빠르게 검출하는 것이 매우 중요하다. 본 논문에서는 불량 WAFER을 정확하게 검출하기 위한 검출시스템으로 멀티 포인트 온도 검출 방법으로 구현된 면적형 온도 센서 기능과 검출된 데이터를 유/무선 통신방식으로 상위의 관리/모니터링 시스템으로 전송 할 수 있는 기능을 가진 마스터 콘트롤러 시스템을 제안한다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2004.04b
• /
• pp.697-699
• /
• 2004

LCD의 생산이 많아짐에 따라 LCD의 불량 검출이 중요해 지고 있다. 불랑 검사는 눈으로 확인할 수 있는 범위에서 검사가 이루어지고 있으며, 만약 눈으로 식별이 불가능한 경우 적외선 카메라나 초음파 센서를 사용하여 검사가 이루어진다. 본 논문에서는 카메라를 이용하여 LCD 패널의 표면에 있는 불량 검출을 위하여 각 Dot에 대한 R, G, B 값을 추출한 후, 추출된 픽셀을 제안된 알고리즘에 적용하여 불량을 검출하는 것을 목적으로 하고 있다.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.12 no.5
• /
• pp.529-536
• /
• 2019

In this paper, the signals of the sensor for extracting characteristic parameters of the rotor are collected and the performance of the extraction technique is analyzed. To this end, a vibration test league was developed for conducting model tests to analyze the signal characteristics under normal operation. As a result, it is judged that no change in the measured the raw data amplitude will occur in the acceleration sensor with the unbalanced mass measured from the acceleration sensor. Performing FFT showed a significant increase in amplitude of the rotational frequency of 20 Hz as the unbalanced mass increased. The analysis results according to the change in the unequal mass of the speed sensor also showed a significant increase in the 1X Harmonics component, such as the acceleration sensor. There was no change in the amplitude of the acceleration sensor data when the misalignment occurred, and for the Envelope data, the amplitude of 2X (40 Hz) was increased depending on the degree of misalignment. The velocity sensor at change of misalignment also showed similar results to the acceleration sensor, and the peak was reduced at 600 Hz as the load increased in the frequency spectrum.

• 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.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2014.04a
• /
• pp.43-46
• /
• 2014

CMOS 이미지 센서는 모바일 디바이스, 특히 스마트 폰에 내장된 카메라에 가장 광범위하게 사용된다. 이러한 이미지 센서의 정상 동작을 검사하기 위해서는 불량화소 검출과 같은 테스트가 수행되어야 하며, 테스트를 위해서는 센서에 의해서 캡처된 이미지를 대상으로 이미지 처리를 할 수 있는 함수제공이 필수적이다. 이 논문에서는 CMOS 이미지 센서의 동작을 효율적이고 엄격하게 판단할 수 있는 자동 검사 시스템을 구축하고 이미지 센서로부터 캡처되는 이미지 데이터에 대해서 목적에 맞는 테스트를 수행 할 수 있도록 이미지 처리 함수를 구현하고 실험하였다.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.44 no.4
• /
• pp.55-62
• /
• 2007

This paper proposes a defective pixel detection algorithm and its hardware structure for CCD/CMOS image sensor. In previous algorithms, the characteristics of image have not been considered. Also, some algorithms need quite a time to detect defective pixels. In order to make up for those disadvantages, the proposed defective pixel detection method detects defective pixels efficiently by considering the edges in the image and verifies them using several frames while checking scene-changes. Whenever scene-change is occurred, potentially defective pixels are checked and confirmed whether it is defective or not. Test results showed that the correct detection rate in a frame was increased 6% and the defective pixel verification time was decreased 60%. The proposed algorithm was implemented with verilog HDL. The edge indicator in color interpolation block was reused. Total logic gate count was 5.4k using 0.25um CMOS standard cell library.