http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
A/S 및 필드데이터를 활용한 공작기계 시스템 신뢰도 산출
김근수(Keunsu Kim),황태완(Tae Wan Hwang),김수지(Su Jii Kim),전병주(Byungjoo Jeon),윤병동(Byeng D. Youn) 대한기계학회 2016 대한기계학회 춘추학술대회 Vol.2016 No.12
Machining tools of manufacturing equipment is often called “mother machine” as they make machines that support industries. Currently, there are in challenges of rapid, high accurate and durable processing. To satisfy these challenges, the reliability of the machine is needed. So we develop the reliability growth process and reliability evaluation techniques using warranty data and field failure data for the horizontal machining center. System reliability is calculated based on data with 8 years collected of manufacturing company and improved reliability result is estimated when fault diagnosis techniques is applied. test method and to use it with care by noticing the unreasonable changes in it.
전병주(Byungjoo Jeon),하종문(Jong Moon Ha),이준민(Junmin Lee),박찬희(Chanhee Park),김수지(Su Jii Kim),윤병동(Byeng Dong Youn) 대한기계학회 2016 대한기계학회 춘추학술대회 Vol.2016 No.12
Chemical-Mechanical planarization (CMP) is a process to flatten a surface of a substrate with combination of chemical and mechanical effect. For a successful CMP process, various operating parameters (e.g. rotating speed and pressure) of CMP should be optimized. However, relationship between the operating parameters and the performance of the CMP process has not been fully identified. To address this challenge, this paper introduces a model that estimates the removal rate of Chemical-Mechanical planarization (CMP) process which is known to represent performance of the CMP machine using various time-series operating parameters. First, 135 features were extracted from the time-series variables for physical representation of the polishing process. Second, Gaussian process (i.e. Kriging) and regression tree were employed to estimate the removal rate of the CMP process. To avoid the overfitting problems, various feature selection schemes were used to define candidates for the best feature subsets. Finally, ensemble regression model was developed to integrate the regression models with the feature subsets.