使用資料探勘方法在偵測製程缺陷

Abstract

近年來，製造業製程變得越來越複雜。為了提高生產良率，如何快速地偵測出製程中造成產品缺陷的主因，已經成為一個重要的議題。但是複雜的製程所造成的產品缺陷會間歇性的出現且因素良多，加上這些因素間存在著非線性關係，因此傳統上以統計為基礎的一些方法難以精確的找出缺陷的主因。在這篇論文中我們正式地定義了製程缺陷偵測問題(Manufacturing Defect Detection Problem)來說明如何偵測造成缺陷產生的主因機器，同時我們提出一個稱為Root cause Machineset Identifier (RMI) 的方法來解決這個問題。最後，實驗的結果顯示RMI在處理真實的製造業案例時，確實可以正確且有效率地找出造成缺陷的主因機器。

In recent years, the procedure of manufacturing has become more and more complex. In order to meet high expectation on quality target, quick identification of root cause that makes defects is an essential issue. Traditional statistic-based methods are still difficult to identify the root cause due to the resulting multi-factor & nonlinear interactions or intermittent problem. In this thesis, Manufacturing Defect Detection Problem is formally defined and a corresponding methodology, called Root cause Machineset Identifier (RMI), is also proposed. RMI has three procedures to handle such Manufacturing Defect Detection Problem. Finally, the results of experiment show the accuracy and efficiency of RMI are both well with real manufacturing cases.