DOI QR코드

DOI QR Code

Comparison of ISO-GUM and Monte Carlo Method for Evaluation of Measurement Uncertainty

몬테카를로 방법과 ISO-GUM 방법의 불확도 평가 결과 비교

  • Ha, Young-Cheol (Gas Quality and Flow Measurement Lab, R&D Division, Korea Gas Corporation) ;
  • Her, Jae-Young (Gas Quality and Flow Measurement Lab, R&D Division, Korea Gas Corporation) ;
  • Lee, Seung-Jun (Gas Quality and Flow Measurement Lab, R&D Division, Korea Gas Corporation) ;
  • Lee, Kang-Jin (Gas Quality and Flow Measurement Lab, R&D Division, Korea Gas Corporation)
  • 하영철 (한국가스공사 연구개발원) ;
  • 허재영 (한국가스공사 연구개발원) ;
  • 이승준 (한국가스공사 연구개발원) ;
  • 이강진 (한국가스공사 연구개발원)
  • Received : 2014.02.19
  • Accepted : 2014.05.29
  • Published : 2014.07.01

Abstract

To supplement the ISO-GUM method for the evaluation of measurement uncertainty, a simulation program using the Monte Carlo method (MCM) was developed, and the MCM and GUM methods were compared. The results are as follows: (1) Even under a non-normal probability distribution of the measurand, MCM provides an accurate coverage interval; (2) Even if a probability distribution that emerged from combining a few non-normal distributions looks as normal, there are cases in which the actual distribution is not normal and the non-normality can be determined by the probability distribution of the combined variance; and (3) If type-A standard uncertainties are involved in the evaluation of measurement uncertainty, GUM generally offers an under-valued coverage interval. However, this problem can be solved by the Bayesian evaluation of type-A standard uncertainty. In this case, the effective degree of freedom for the combined variance is not required in the evaluation of expanded uncertainty, and the appropriate coverage factor for 95% level of confidence was determined to be 1.96.

본 연구에서는 ISO GUM(불확도 표현 지침서)의 불확도 평가 방법을 보완하기 위해, 몬테카를로 방법(Monte Carlo Method, MCM)을 적용한 불확도 해석 프로그램을 개발하고, MCM과 GUM의 평가 결과를 비교하였다. 그 결과 다음과 같은 결과를 도출하였다. 첫째, 측정량의 확률 분포가 정규 분포가 아닌 때에도 MCM 방법은 정확한 포함 구간을 제공한다. 둘째, 정규 분포가 아닌 다른 분포들 몇몇 개가 합성되는 경우 그 확률 분포가 정규로 보이더라도 실제로는 정규가 아닌 경우가 있으며, 이의 판단은 합성 분산의 확률 분포로 할 수 있다. 셋째, 자유도가 낮은 A형 불확도가 불확도 평가에 포함된 경우 GUM은 포함 구간을 저평가하는 것을 알 수 있었고, 이러한 저평가 문제는 A형 표준 불확도에 t-분포의 표준 편차를 곱해주면 사라지는 것을 알 수 있었다. 이 경우 합성 분산의 유효 자유도는 확장 불확도 계산에 불필요하고, 신뢰의 수준 95 %의 포함 인자는 1.96이 적정한 것을 알 수 있었다.

Keywords

References

  1. ISO GUM, 1993, "Guide to the Expression of Uncertainty in Measurement," International Organization for Standardization, Geneva.
  2. JCGM 100, 2008, "Evaluation of Measurement Data-Guide to the Expression of Uncertainty in Measurement," BIPM.
  3. Satterthwaite, F. E., 1946, "An Approximate Distribution of Estimates of Variance Components," Biometrics Bulletin 2, pp. 110-114. https://doi.org/10.2307/3002019
  4. Welch, B. L. 1947, "The Generalization of "Student's" Problem When Several Different Population Variances are Involved," Biometrika 34, pp. 28-35.
  5. Park, C. H., 2004, "Random Processes," ISBN 89-7050-390-0.
  6. Robert, C. P., Casella, G., 2005, "Monte Carlo Statistical Methods," Springer New York, ISBN 0387212396.
  7. Hellekalek, P., 1998, "Good Random Number Generators are (not so) Easy to Find," Mathematics and Computers in Simulation, Vol. 46, pp. 485-505. https://doi.org/10.1016/S0378-4754(98)00078-0
  8. James F., 1990, "A Review of Pseudorandom Number Generators," Computer Physics Communications, Vol. 60, pp. 329-344. https://doi.org/10.1016/0010-4655(90)90032-V
  9. http://en.wikipedia.org/wiki/Box-Muller_transform.
  10. JCGM 101, 2008, "Evaluation of Measurement Data-Supplement 1 to the "Guide to the Expression of Uncertainty in Measurement"-Propagation of Distributions using a Monte Carlo Method," BIPM.
  11. Gelfand, A. E. and Smith, A. F. M., 1990, "Sampling-Based Approaches to Calculating Marginal Densities," The Journal of the American Statistical Association, Vol. 85, pp. 398-409. https://doi.org/10.1080/01621459.1990.10476213