摘要

品質的問題正日益受到重視，許多廠商為了增進產品及服務品質，依照 ISO 9001、QS-9000 等標準來建立廠商自身的品管系統，其中QS-9000 為美國三大汽車廠開發出一套屬於汽車業的驗證制度，而在QS-9000 之參考手冊 (Reference Manual) 中，量測系統分析(Measurement System Analysis) 則為分析量測系統能力與精準必備方法。
由於量測之精準與否在品質改善上扮演一個相當重要的角色，唯有精確之量測技術，才能保証對產品及製程品質的判定無誤。在量測系統分析中，量具之再生性及再現性（Gage Repeatability and Reproducibility，GR&R）主要以實驗設計分析，並以所得數據分析變異來自於檢測人員、檢測設備、或產品本身，依此改善系統之量測能力。然而在此分析中，通常有一重要假設即為所量測物件之特性必須為連續值。但在工業上，許多受檢測特性為離散值，例如在半導體晶圓片上缺點數、LCD面板上之亮點數，此類量測系統能力與精確度如何評估，是值得研究的主題。
本研究即為建立一評估離散型量測系統其再現性與再生性變異的方法，以量測此類量測系統中之能力與精確度，並提供一評估此類量測系統鑑別力方法，而本研究結果顯示在本研究模式中鑑別力會隨著Reference Value增加而增加，且當GR&R變異升高時，量測系統的鑑別力會逐漸下降。研究中以模擬方式比較兩模式在離散型量測系統下，估計母體參數之能力，而以本論文所提出模式之估計能力較佳。在個案中，以兩種估計方式評估個案公司量測系統精準度，對於個案公司量測系統精準度皆判定為極差。且在個案公司採用本研究模式估計再現性與再生性之變異，有相當高的鑑別力，證明本研究模式對於一離散型之量測系統為一適當的衡量方式。

Abstract

To increase the quality level of their products and service, businesses many organizations have established their own quality systems based on ISO 9000, QS-9000, and many international standards. QS-9000, or the ISO/TS 16949 is a certification system initiatively introduced in the automotive industry. As described in the reference manual of QS-9000, Measurement System Analysis (MSA) is a method which can evaluate the accuracy and capacity of a specific measurement system.

The accuracy of measurement plays an important role on quality improvement. The so called gage repeatability and reproducibility (GR&R) method is developed to analyze when the variance is due to operators, equipments or parts. In the ANOVA method of GR&R, it is assumed that the measurement data is continuous and follows normal distribution. However, many measurement data belongs to discrete type and the ANOVA method cannot be applied.

The main purpose of this research is to develop a method to evaluate the accuracy and capacity of a measurement system which has discrete data. A discrimination power method and it gage performance curve is also developed. The empirical analysis indicates that the discrimination power will increase when reference value increases and decrease otherwise. Compared to the ANOVA method, the proposed method is a better estimator of the parameters.

In the case study, the case measurement system has poor accuracy and precision but has a high discrimination power when using proposed method. It is concluded that this research develops a more suitable measurement method to evaluate a measurement system of discrete data.

參考文獻
中文部份
1.中華民國實驗室認證體系, http://www.cnla.org.tw/
2.台灣光電校正實驗室 (2001)， “量具之重覆性與再現性討論,” 台儀資訊, 第21期, 33-39 頁。
3.江巧玉 (2002)， “量測系統重覆性與再現性的分析研究,” 成功大學, 統計研究所碩士論文。
4.周國雄 (1991)， 「計量計測管理」， 中華民國品質管制學會。
5.俞佳伶 (1999)， “印刷電路板之完稿底片製程變異評估,” 元智大學, 工業工程研究所碩士論文。
6.韋孟育 (1990)， 「材料實驗方法：金相分析技術」, 全華書局。
7.陳源樹 (2002)， “三次元量床量測能力之研究,” 中華民國品質學會第三十八屆年會暨第八屆全國品質管理研討會論文集， 493-501 頁。
8.褚龍印 (2002)，“儀器校正對於量測結果的影響,” 量測資訊, 第77期, 116-118 頁。
9.鄭希龍 (1997)， “檢測數據分析與管制作業,” 量測資訊, 第58期, 47-54 頁。
10.蕭名男 (2002)， “平面顯示器研發階段量測系統不確定度分析,” 中華民國品質學會第三十八屆年會暨第八屆全國品質管理研討會論文集， 551-559 頁。
11.蕭堯仁 (1999)， “機器視覺之重覆性與重現性研究,” 中華民國品質學會第三十五屆年會暨第五屆全國品質管理研討會論文集， 201-210 頁。

英文部份
1. Boyles, R. A. (2001), “Gauge Capability for Pass-Fail Inspection,” Technometrics, Vol. 43, pp. 223-229.
2. Burdick, R. K. and Larsen, G. A. (1997), “Confidence Intervals on Measures of Variability in R&R studies,” Journal of Quality Technology, Vol. 29, pp. 261-273.
3. Burdick, R. K. and Allen, A. E. (2002), “Comparing Variability on Two Measurement Processes Using R&R Studies,” Journal of Quality Technology, Vol. 34, pp. 97-105.
4. Chrysler, Ford, GM, (1998), Measurement System Analysis, Automotive Industry Action Group.
5. Dobson, A. J. (2002), An Introduction to Generalized Linear Models, Chapman & Hall, pp. 120-124.
6. Dolezal, K. K. and Burdick, R. K. (1998), “Analysis of a Two-Factor R&R Study with Fixed Operators,” Journal of Quality Technology, Vol. 30, pp. 163-170.
7. Floyd, D. A. and Laurent, C. J. (1995), “Gauging: An Underestimated Consideration In The Application Of Statistical Process Control,” Quality Engineering, Vol. 8, pp. 13-29.
8. Grubbs, F. E. (1973), “Errors of Measurement, Precision, Accuracy and the Statistical Comparison of Measuring Instruments,” Technometrics, Vol. 15, No. 1, pp. 53-66.
9. Hamada, M. and Weerahandi, S. (2000), “Measurement System Assessment Via Generalized Inference,” Journal of Quality Technology, Vol. 32, pp. 241-253.
10. Heuval, E. R. (2000), “Gage R&R Studies for Nonstandard Situation,” ASQ`s 54th Annual Quality Congress Proceedings, pp. 317-328.
11. International Organization for Standardization (1999), ISO/IEC Guide 17025 Quality Management System for Clinical and Related Laboratories - General requirement for the competence of testing and calibration.
12. International Organization For Standardization (2002), ISO/TS 16949-Quality System-Automotive Suppliers, Particular Requirements For The Application of ISO 9001:2000 For Automotive Production And Relevant Service Part Organization.
13. Jaynes, K. R. (1997), “Reviewing Gage R&R,” Manufacturing Engineering, Vol. 118, No. 8, pp. 12.
14. Levinson, W. A. (1996), “Do You Need a New Gage,” Semiconductor International, Vol. 2, pp. 23-27.
15. Lupo, C. (2002), “ISO/TS 16949 the Clear Choice For Automotive Suppliers,” Quality Progress, Vol. 35, No. 10, pp. 44-49.
16. Montgomery, D. C. and Runger, G. C. (1993), “Gauge Capability and Designed Experiment, Part I: Basic Methods,” Quality Engineering, Vol. 6, No. 1, pp. 115-135.
17. Morris, R. M. and Watson, E. F. (1998), “A Comparison of The Techniques Used to Evaluate The Measurement Process,” Quality Engineering, Vol. 11, pp. 213-219.
18. Ross, S. (1998), A First Course in Probability 5th, Prentice Hall, pp. 144-170.
19. Roussas, G. (1997), A Course in Mathematical Statistics 2th, Academic Press, pp. 72-76.
20. Runger, G. (2002), “Assignable Causes and Autocorrelation: Control Charts for Observations or Residuals,” Journal of Quality Technology, Vol. 34, pp. 211-218.
21. Stout, G. (1994), “Measurement Put to the Test,” Quality, Vol. 17. No. 9, pp. 41-48.
22. Tsai, P. F. (1988), “Variable Gauge Repeatability and Reproducibility Study Using the Analysis of Variance Method,” Quality Engineering, Vol. 6, No. 2, pp. 289-305.
23. Vardeman, S. B. and Vanvalkenburg, E. S. (1999), “Two-way Random-effects Analyses and Gauge R&R Studies,” Technometrics, Vol. 41, pp. 202-211.