• Journal of Convergence for Information Technology
• /
• v.9 no.10
• /
• pp.101-107
• /
• 2019

This paper proposes a method for detecting faults in conveyor systems used for transportation of raw materials needed in the thermal power plant and cement industries. A small drone was designed in consideration of the difficulty in accessing the industrial site and the need to use it in wide industrial site. In order to apply the system to the embedded microprocessor, hardware and algorithms considering limited memory and execution time have been proposed. At this time, the failure determination method measures the peak frequency through the measurement, detects the continuity of the high frequency, and performs the failure diagnosis with the high frequency components of noise. The proposed system consists of experimental environment based on the data obtained from the actual thermal power plant, and it is confirmed that the proposed system is useful by conducting virtual environment experiments with the drone designed system. In the future, further research is needed to improve the drone's flight stability and to improve discrimination performance by using more intelligent methods of fault frequency.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2015.08a
• /
• pp.132-132
• /
• 2015

Virtual metrology (VM) model based on plasma information (PI) parameter for C4F8 plasma-assisted oxide etching processes is developed to predict and monitor the process results such as an etching rate with improved performance. To apply fault detection and classification (FDC) or advanced process control (APC) models on to the real mass production lines efficiently, high performance VM model is certainly required and principal component regression (PCR) is preferred technique for VM modeling despite this method requires many number of data set to obtain statistically guaranteed accuracy. In this study, as an effective method to include the 'good information' representing parameter into the VM model, PI parameters are introduced and applied for the etch rate prediction. By the adoption of PI parameters of b-, q-factors and surface passivation parameters as PCs into the PCR based VM model, information about the reactions in the plasma volume, surface, and sheath regions can be efficiently included into the VM model; thus, the performance of VM is secured even for insufficient data set provided cases. For mass production data of 350 wafers, developed PI based VM (PI-VM) model was satisfied required prediction accuracy of industry in C4F8 plasma-assisted oxide etching process.

• Journal of Electrical Engineering and Technology
• /
• v.8 no.2
• /
• pp.288-294
• /
• 2013

The diagnosis of motor failures using an on-line method has been the aim of many researchers and studies. Several spectral analysis techniques have been developed and are used to facilitate on-line diagnosis methods in industry. This paper discusses the first application of a motor flux spectral analysis to the identification of broken rotor bar (BRB) faults in induction motors using a multiple signal classification (MUSIC) technique as an on-line diagnosis method. The proposed method measures the leakage flux in the radial direction using a radial flux sensor which is designed as a search coil and is installed between stator slots. The MUSIC technique, which requires fewer number of data samples and has a higher detection accuracy than the traditional fast Fourier transform (FFT) method, then calculates the motor load condition and extracts any abnormal signals related to motor failures in order to identify BRB faults. Experimental results clearly demonstrate that the proposed method is a promising candidate for an on-line diagnosis method to detect motor failures.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2012.02a
• /
• pp.523-523
• /
• 2012

As the wafer geometric requirements continuously complicated and minutes in tens of nanometers, the expectation of real-time add-on sensors for in-situ plasma process monitoring is rapidly increasing. Various industry applications, utilizing plasma impedance monitor (PIM) and optical emission spectroscopy (OES), on etch end point detection, etch chemistry investigation, health monitoring, fault detection and classification, and advanced process control are good examples. However, process monitoring in semiconductor manufacturing industry requires non-invasiveness. The hypothesis behind the optical monitoring of plasma induced ion current is for the monitoring of plasma induced charging damage in non-invasive optical way. In plasma dielectric via etching, the bombardment of reactive ions on exposed conductor patterns may induce electrical current. Induced electrical charge can further flow down to device level, and accumulated charges in the consecutive plasma processes during back-end metallization can create plasma induced charging damage to shift the threshold voltage of device. As a preliminary research for the hypothesis, we performed two phases experiment to measure the plasma induced current in etch environmental condition. We fabricated electrical test circuits to convert induced current to flickering frequency of LED output, and the flickering frequency was measured by high speed optical plasma monitoring system (OPMS) in 10 kHz. Current-frequency calibration was done in offline by applying stepwise current increase while LED flickering was measured. Once the performance of the test circuits was evaluated, a metal pad for collecting ion bombardment during plasma etch condition was placed inside etch chamber, and the LED output frequency was measured in real-time. It was successful to acquire high speed optical emission data acquisition in 10 kHz. Offline measurement with the test circuitry was satisfactory, and we are continuously investigating the potential of real-time in-situ plasma induce current measurement via OPMS.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2010.02a
• /
• 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.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.9 no.7
• /
• pp.1588-1593
• /
• 2005

In coagulant control of water treatment plants, rule extraction, one of datamining categories, was performed for coagulant control of a water treatment plant. Clustering methods were applied to extract control rules from data. These control rules can be used for fully automation of water treatment plants instead of operator's knowledge for plant control. To perform fuzzy clustering, there are some coefficients to be determined and these kinds of studies have been performed over decades such as clustering indices. In this study, statistical indices were taken to calculate the number of clusters. Simultaneously, seed points were found out based on hierarchical clustering. These statistical approaches give information about features of clusters, so it can reduce computing cost and increase accuracy of clustering. The proposed algorithm can play an important role in datamining and knowledge discovery.

• Journal of the Semiconductor & Display Technology
• /
• v.18 no.4
• /
• pp.30-34
• /
• 2019

We analyzed how the features in plasma information based virtual metrology (PI-VM) for SiO₂ etching depth with variation of 5% contribute to the prediction accuracy, which is previously developed by Jang. As a single feature, the explanatory power to the process results is in the order of plasma information about electron energy distribution function (PI_EEDF), equipment, and optical emission spectroscopy (OES) features. In the procedure of stepwise variable selection (SVS), OES features are selected after PI_EEDF. Informative vector for developed PI-VM also shows relatively high correlation between OES features and etching depth. This is because the reaction rate of each chemical species that governs the etching depth can be sensitively monitored when OES features are used with PI_EEDF. Securing PI_EEDF is important for the development of virtual metrology (VM) for prediction of process results. The role of PI_EEDF as an independent feature and the ability to monitor variation of plasma thermal state can make other features in the procedure of SVS more sensitive to the process results. It is expected that fault detection and classification (FDC) can be effectively developed by using the PI-VM.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.05a
• /
• pp.505-508
• /
• 2021

반도체 회로의 미세화로 단위 공정이 증가하면 TAT(turn-around time) 증가에 따른 제조 비용이 늘어난다. 반도체 공정 중 포토 공정은 마스크의 회로를 웨이퍼에 전사하는 공정으로 전사를 담당하는 노광장비의 성능에 의해 회로의 정확성이 결정된다. 이런 정확성을 검증하는 계측공정은 회로의 미세화가 진행될수록 필요성은 증가하나 TAT 증가의 주된 요인으로 최근 기계학습을 사용한 다양한 예측 모형들의 개발로 계측 결과를 예측하는 실험들이 진행되고 있다. 본 논문은 노광장비 센서들의 이상값을 감지하여 분류 후 계측공정을 진행하는 LFDC(Lithography Fault Detection and Classification) 시스템의 문제인 분류 성능이 떨어지는 것을 해결하기 위해 XGBoost를 사용하여 계측공정을 진행하지 않고 노광장비 센서의 이상값을 학습된 학습기를 통해 분류하여 포토 공정을 재진행하거나 다음 공정을 진행하는 방법을 실험하였다. 실험에서 사용된 계측 결과 예측 모형은 89%의 정확도를 확보하였고 반도체 데이터 특성인 심각한 불균형의 데이터에 대해서도 같은 정확도를 얻었다. 이런 결과는 노광장비 센서들의 이상값에 대해 89%는 정상으로 판단하였고 정상으로 판단한 웨이퍼를 실제 계측 시 예측과 같은 결과를 얻었다. 계측 결과 예측 모형을 사용하면 실제 계측을 진행하지 않고 노광장비 센서들의 이상값에 대한 판정을 할 수 있어 TAT 단축으로 제조 비용감소, 계측 장비 부하 감소 및 효율 향상을 할 수 있다. 하지만 본 논문에서는 90%의 성능을 보이는 계측 결과 예측 모형으로 여전히 10%에 대해서는 실제 계측이 필요한 문제에 대해 추후 더 연구가 필요하다.