• 한국진공학회:학술대회논문집
• /
• 한국진공학회 2009년도 제38회 동계학술대회 초록집
• /
• pp.413-413
• /
• 2010

In semiconductor manufacturing field, all equipments have various sensors to diagnosis the situations of processes. For increasing the accuracy of diagnosis, hundreds of sensors are emplyed. As sensors provide millions of data, the process diagnosis from them are unrealistic. Besides, in some cases, the results from some data which have same conditions are different. We want to find some information, such as data and knowledge, from the data. Nowadays, fault detection and classification (FDC) has been concerned to increasing the yield. Certain faults and no-faults can be classified by various FDC tools. The uncertainty in semiconductor manufacturing, no-faulty in faulty and faulty in no-faulty, has been caused the productivity to decreased. From the uncertainty, the rough set theory is a viable approach for extraction of meaningful knowledge and making predictions. Reduction of data sets, finding hidden data patterns, and generation of decision rules contrasts other approaches such as regression analysis and neural networks. In this research, a RGB sensor was used for diagnosis plasma instead of optical emission spectroscopy (OES). RGB data has just three variables (red, green and blue), while OES data has thousands of variables. RGB data, however, is difficult to analyze by human's eyes. Same outputs in a variable show different outcomes. In other words, RGB data includes the uncertainty. In this research, by rough set theory, decision rules were generated. In decision rules, we could find the hidden data patterns from the uncertainty. RGB sensor can diagnosis the change of plasma condition as over 90% accuracy by the rough set theory. Although we only present a preliminary research result, in this paper, we will continuously develop uncertainty problem solving data mining algorithm for the application of semiconductor process diagnosis.

• 제어로봇시스템학회논문지
• /
• 제4권6호
• /
• pp.765-771
• /
• 1998

An artificial neural network(NN) technique is developed for hardware redundant sensor validation. Since the measurement space is a continuous space with many operating regions, it is difficult to train a NN to correctly detect failure in an accurate measurement system. A conventional backpropagation NN is modified to include an additional preprocessing layer that extracts classification features from scalar measurements. This feature extraction means transform the measurement space to parity space. The NN is independent of the state variable being measured, the instrument range, and the signal tolerance. This NN resembles the parity space approach to signal validation, except that analytical parity equations are unneeded and the NN pattern recognition capability is utilized for decision making.

• 한국전산구조공학회논문집
• /
• 제25권4호
• /
• pp.287-292
• /
• 2012

본 연구에서는 다채널 자속누설 센서를 이용하여 강케이블의 국부손상을 검색하였다. 먼저 자속누설 기법을 고정된 케이블 구조물에 적용하기 위해 프로토타입의 8채널 자속누설 센서헤드를 제작하였고, 국부손상이 발생한 케이블을 구현하기 위하여 PVC 파이프에 강케이블을 채워 강케이블 다발 시편을 제작하였고, 케이블 시편 외부 및 내부에 다양한 크기 및 방향을 가지는 국부손상을 단계적으로 발생시켰다. 이와 같이 제작된 강케이블 시편을 대상으로 각 손상단계에서 자속누설 센서헤드를 이용하여 자속신호를 스캔하고 출력전압으로 표현하였다. 이어서 일반극치분포를 이용해 손상유무를 판단할 수 있는 기준이 되어줄 임계값을 설정하였고, 이를 각 채널에서 계측된 자속신호와 비교하여 객관적인 손상판단을 수행하였다. 또한 케이블 모니터링에 있어 가장 중요한 정보인 손상의 길이방향 위치를 효과적으로 검색하기 위해 모든 채널의 자속값을 합하여 총합값의 형태로 임계값과 함께 나타내었다. 최종적으로 임계값을 초과한 부분의 길이방향 및 원주방향 위치를 실제 손상과 비교함으로써 본 기법의 국부손상 검색 가능성을 살펴보았다.

• 한국정밀공학회지
• /
• 제9권4호
• /
• pp.22-35
• /
• 1992

In highly automated machining centers, intelligent sensor fddeback systems are indispensable on order to monitor their operations, to ensure efficient metal removal, and to initate remedial action in the event of accident. In this study, an on-line tool wear detection system for thrning operations is developed, and experimentally evaluated. The system employs multiple sensors and the signals from these sensors are processed using a multichannel autoegressive (AR) series model. The resulting output from the signal processing block is then fed to a previously tranied artificial neural network (multiayered perceptron) to make a final decision on the state of the cutting tool. To learn the necessary input/output mapping for tool wear detection, the weithts and thresholds of the network are adjusted according to the back propagation (BP) method during off-line training. The results of experimental evaluation show that the system works well over a wide range of cutting conditions, and the ability of the system to detect tool wear is improved due to the generalization, fault-tolearant and self-ofganizing properties of the neural network.