碩士 / 國立成功大學 / 工業與資訊管理學系專班 / 101 / In a continuous manufacturing process, the reliability and stability of machines is very critical to the quality of products. Many studies therefore investigate the period of preventive maintenance to enhance the reliability of continuous manufacturing processes. The state of a machine is assumed to be either under-control or out-of-control. The binary logistic regression can be used not only to predict the state of a machine, but also to find the critical factors for machines to be out-of-control.
In this study, a continuous manufacturing process is assumed to be supervised by three shifts in one day, and every shift will collect inspection data from the products of the process. The statistics calculated from the inspection data obtained from the previous two shifts for a machine are employed to generate a binary linear regression model to predict whether the machine will be under-control in current shift.
The experimental results on eight data sets show that the prediction accuracy of the binary linear regression model built by our method is between 0.93 and 0.98. The variance and the error rate of the inspection data obtained from the previous two shifts are the most important factors in identifying whether a machine is under-control or out-of-control.
| Identifer | oai:union.ndltd.org:TW/101NCKU5041057 |
| Date | January 2013 |
| Creators | Ya-ChihChang, 張雅智 |
| Contributors | Tzu-Tsung Wong, 翁慈宗 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 44 |
Page generated in 0.0095 seconds