자기상관이 있는 장치산업에서 공정 진단 및 부적합품률 제어모형에 관한 연구

• 영문명: A Study on the Diognostic & Nonconformity Rate Control Model in a Instrumental Process with Autocorrelation
• 발행기관: 한국산업경영시스템학회
• 저자명: 구자활 김대현 조진형
• 간행물 정보: 『한국산업경영시스템학회 학술대회』2009년 춘계학술대회 논문집, 91~97쪽, 전체 7쪽
• 주제분류: 공학 > 산업공학
• 파일형태: PDF
• 발행일자: 2009.05.30

영문 초록

Control chart is one of the most well-known tools used to monitor and control the process in real time. However, application of the chart is predicated on the assumption that each observed data has no correlation with one another. In recent years, manufacturing process has been automated and on-line monitor has become possible. They are widely used in chemical industry or instrumental process, where sensors monitor tanks, furnace, and flow of raw materials. In this kind of process, sampling interval for data collection tends to be short compared with the overall processing time, resulting in significant autocorrelation among data. The existence of the autocorrelation violates the premise of iid for the traditional control chart, whose application to such a case could lead to the wrong conclusion. The effect of autocorrelation can be understood employing a different statistical model, and it will provide the way to determine if the process is control-in or control-out. The best available model for analyzing time-series data with autocorrelation is ARMIA(autoregressive integrated moving average.) However, if manufacturing process itself is rather large in scale, and measuring physical properties, difficult, and there are large number of variables to be controlled, it is quite complicated to not only make diagnosis and control the process but also determine the relationship between control variables and occurrences of non-conforming products. These control variables tend to be interrelated. It is necessary to find out critical factors and realize effective process control by controlling those factors. In this study, we present a control model based on time series data with autocorrelation in instrumental process, and show how to effectively deal with non-conforming products, by applying it to data from real process.

목차

1. 서론

2. 연구모형 및 방법

3. 사례연구를 통한 검증

4. 결론 및 향후 연구과제

키워드

Control Chart Instrumental Process Autocorrelation ARIMA