製程能力指標應用於多品質特性及工具磨耗之製程

Abstract

製程能力指標被廣泛地應用在製造業，做為衡量產品品質的標準。製程能力指標能簡單且直接地表示產品品質，是建立在某些前提假設之下：製程穩定且不受特定因素影響而改變（即管制內）、製程符合常態分配、抽樣觀察值皆獨立。現實中，這些假設並非都能滿足，製程常會因為外在因素而產生變動。 本文中分別針對多品質特性及工具磨耗提出製程能力指標之修正，並分別以兩個實際的產品做案例，以說明如何應用修正之製程能力指標。 第一部分，利用複式抽樣法估計 並比較四種方法的估計準確度。結果顯示BCPB方法的準確度較高。本文以BCPB方法計算，將符合信賴下界及估計準確度的抽樣數列表，提供使用者可容易查出需要的抽樣數，以進行抽樣計畫。 第二部分，考量受外在因素影響的製程，提出監控製程的程序，並依據製程能力需求提供指標臨界值。這個監控製程的程序類似其他管制圖的監控程序，可用來判斷製程是否要停止生產或更換零件。

Process capability indices have been widely used in the manufacturing industry and provided numerical measures on process performance. However, process capability indices seem to be easy and straightforward to apply. That is because of some assumption must be satisfied: the process under investigation is free from any special or assignable cause, the process characteristic is normal distributed, and the observations of quality characteristics are statistically independent. However, these conditions are not always fulfilled in many manufacturing situations. In real world, process is always influenced by some assignable cause. In addition, capability measure for processes with single characteristic has been investigated extensively, but capability measure for processes with multiple characteristics is comparatively neglected. In the thesis, the process capability indices for multi-characteristics of one-sided process and the modified index for tool wear process are proposed. Two real-world examples from multiple characteristics process and tool wear process respectively are taken to illustrate the applications of the propose approaches. For the first example, we used four bootstrap methods to estimating and compare the estimation precision of these methods. The results indicated that the BCPB method has better performance based on estimation precision than others. The table of the lower confidence bound values and sample sizes required for specified precision of the estimation is provided for the practitioners. For the second example, a procedure of capability measure with assignable causes is developed, and the critical value for various values capability requirements and sample size are provided. The proposed procedure is similar to those used in monitoring a process with control chart, and used to monitor the process and decide if the process should stop and replace the tool to avoid producing unacceptable products.