Measurement Uncertainty
- PDF / 91,425 Bytes
- 5 Pages / 504.567 x 720 pts Page_size
- 47 Downloads / 203 Views
Measurement Uncertainty Han Haitjema Mitutoyo RCE, Best, The Netherlands
Synonyms Measurement measurement
uncertainty;
Uncertainty
of
Definition A nonnegative parameter characterizing the dispersion of the quantity values being attributed to a measurand, based on the information used (Source: JCGM 200:2012 (2012) (VIM)).
Theory and Application Introduction Few subjects have generated so much debate in the recent decennia as the concept “uncertainty,” better specified as “measurement uncertainty.” The need for a harmonized approach came from industry, standardization organizations, accreditation organizations, national metrology institutes, etc. Historically, measurement uncertainty is closely related to the progress of science and adoption/rejection of theories. From the industrial
side, there is the fitting of products or, if that cannot be done directly, specifying products and testing whether these specifications are met. On the other hand, there is the field of statistics, taking samples out of a – limited or unlimited – range of possible outcomes and estimating what the distribution will be behind the outcomes when a limited set – down to one – of outcomes is available. Basic Documents Where in the past – and still – physicists are taught “correct for systematic errors and add all random errors (i.e., standard deviations) quadratically,” it was not uncommon that mechanical engineers had as a rule “add all maximum errors to minimize the risk.” Nowadays, it is agreed that “systematic” and “statistical” errors are not so essentially different and that a harmonized approach is possible. The approach as outlined in the “GUM” (GUM 1995; JCGM 100:2008 2008) is most generally agreed and available. This document has a supplement 1 (JCGM 101:2008 2008) that looks less basic from the title; but, in fact, this second part is considered the most fundamental basis. The difference is that in supplement 1, uncertainty is basically treated as the propagation of uncertainty distributions, rather than uncertainties, and that these can be simulated by Monte Carlo methods. The approach of these documents may look complicated, and for this reason, several handson documents have appeared that claim to have a more practical approach. Typical examples are the
# CIRP 2016 The International Academy for Production Engineering et al. (eds.), CIRP Encyclopedia of Production Engineering, DOI 10.1007/978-3-642-35950-7_6599-3
2
Measurement Uncertainty
EA-document EA-4/02 M (2013) and the “PUMA” approach explained in ISO 14253-2 (2011). Basic Method In practice, the precise uncertainty distribution, correlation, and degrees of freedom considerations may not be too relevant, and we will omit these aspects from now on. With the risk of simplifying things too much and without the pretention of giving a definite simplified approach, the basic method can be summarized as follows: In general, a measurement result y is a function of n input quantities xi (i = 1,2,. . .n). These input quantities can be measured values, known constants, etc. This
Data Loading...
