InternationaleJournals

Design and Development of a Methodology for Measurement-Uncertainty-Estimation through DoE & ANOVA

Koilakuntla Maddulety

National Institute of Industrial Engineering; Vihar Lake post; Mumbai – 400087 EMail:  koila@rediffmail.com

 

Abstract

In this paper the author developed a methodology with the applications of  DoE & ANOVA for estimating ‘standard components uncertainties’ and ‘combined uncertainty’ with minimum number of trails and minimum experimental efforts. The methodology had been validated with practical experimentation in which all the Type B uncertainty components are replaced with Type A uncertainty methods, which is more accurate and scientific.

In classical method of uncertainty estimation requires separate efforts for estimating type A uncertainty and Type B uncertainty. The accuracy level of Type B uncertainty is very low as it is estimated indirectly, unlike Type-A uncertainty estimation which is direct experimentation and statistical analysis. 

The suggested methodology in this paper is highly scientific, accurate, needs minimum   experimentation effort. Based on classical methodology separate measurement-uncertainty analyses are carried-out for each product range, in most of the laboratories and NABL approved laboratories. But as per this methodology, for all the ranges of the products or a group of ranges of products a single design, uncertainty experimentation & analysis produce highly reliable uncertainty-measurement results.

Hence it is highly recommended to the laboratories where the ranges of product are measured with same laboratory setup.

Key Words:  Measurement-Uncertainty; Combined-Uncertainty; Expanded Uncertainty; Experimental Design (DoE); Analysis of Variance (ANOVA); Factors; Factor Levels

 

 

 


International eJournal of Mathematics and Engineering

Volume 2, Issue 2, Pages:  937 - 946