HMM for Monitoring Tool Wear and Yield Evaluation

• Main Authors: Chun-Hung Chien, 簡俊宏
• Other Authors: Chen-Ju Lin
• Format: Others
• Language: zh-TW
• Online Access: http://ndltd.ncl.edu.tw/handle/94857498931341677460

碩士 === 元智大學 === 工業工程與管理學系 === 101 === With the development of high-tech industry, the manufacturing industry requires a greater emphasis on the quality of products. Since tool wear can gradually lead to defects, it is necessary to monitor the state of tool wear. The existing technologies for monitoring tool wear are diversified, where some methods rely on high-precision equipment. However, not all of the tooling processes can support expensive testing instruments. Instead, this paper diagnoses the state of tool wear by monitoring the quality characteristics of products under a hidden Markov model (HMM). The decision of tool replacement is based on the viewpoint of yield. The probability that the next product will in control of the specification is used to determine whether to change a tool or not. Under the proposed decision scheme, decision maker can fully make use of a tool and control yield.