考慮量測誤差下的製程能力指標

Abstract

製程能力指標廣泛的應用在製造業上，由於這些指標精確地量化製程能力與製程表現，因此這些指標對於成功的改善製程品質與製程流程上，扮演重要的角色。然而，在大多數的製程能力分析上卻未考慮量測誤差，在衡量製程能力時，描述品質特徵的資料卻是與量測工具息息相關的，因此，若不考慮量測誤差下，只依照單一的製程能力指標值來衡量製程能力的結論將是不可靠的。在我們的研究中，我們討論量測誤差對於製程能力指標Cp、Cpk、Cpu和Cpl的估計值有哪些影響，並且提出修正的信賴區間與檢定臨界值，以應用於量測誤差在不可避免下的製程能力分析。我們的研究可以幫助實行者做精確的製程能力評估，並做較可靠的決策。

Process capability indices have been widely used in the manufacturing industry. Those capability indices, quantifying process potential and performance, are important for any successful quality improvement activities and quality program implementation. Most research works related to process capability analysis have assumed no gauge measurement errors. However, the quality of data on the process characteristics relies very much on the gauge measurement. Conclusions about capability of the process just only based on the single numerical value of the index are not reliable. In our research study, we conduct the performance of the estimators of the indices, Cp, Cpk, Cpu and Cpl with gauge measurement errors, and present adjusted confidence interval bounds and critical values for capability estimation or testing purpose of those indices with unavoidable measurement errors. Our research would help practitioners to determine whether the factory processes meet the capability requirement, and make more reliable decisions.