Year
1997
Abstract
An essential part of any measurement control program is adequate knowledge of the uncertainties of the measurement system standards. Only with an estimate of the standards’ uncertainties can one determine if the standard is adequate for its intended use. The estimate is also necessary to calculate the total uncertainty of the measurement process. Purchased standards usually have estimates of uncertainty on their certificates. However, when standards are prepared and characterized by a laboratory, variance propagation is required to estimate the uncertainty of the standard. Traditional variance propagation typically involves tedious use of partial derivatives, unfriendly software and the availability of statistical expertise. As a result, the uncertainties of prepared standards are often not determined or determined incorrectly. For situations meeting stated assumptions, easier shortcut methods of estimation are now available that eliminate the need for partial derivatives and require only a spreadsheet or calculator. A system of simplifying the calculations by dividing into subgroups of absolute and relative uncertainties is utilized. 1 These methods also incorporate the International Standards Organization (ISO) concepts for combining systematic and random uncertainties as published in their Guide to the Expression of Measurement Uncertainty.2 Details of the simplified methods and examples of their use are included in the paper.