Accounting for the distributions of input quantities in the procedure for the measurement uncertainty evaluation when calibrating the goniometer

Abstract

The discords concerning the measurement uncertainty evaluation in the Guide to the Expressing of Uncertainty in Measurement (GUM) and its Supplement 1 are considered. To overcome these discords, the authors of the paper propose to use the kurtosis method and the law of the propagation of the expanded uncertainty. Using the example of the goniometer calibration, the features of accounting for the distribution laws of input quantities in the procedure for the measurement uncertainty evaluation are shown. A model for direct measurements of the value of a reference measure of the angle using a goniometer is written, the procedures for the measurement uncertainty evaluation are described, and uncertainty budgets for each of the methods are given. An example of the measurement uncertainty evaluation when calibrating a digital goniometer using a 24-sided reference prism is described. An estimate of the expanded measurement uncertainty for this example was made based on the web-based software application NIST Uncertainty Machine, which showed a good agreement with the estimates obtained by the considered methods. The technology of applying this software application for the confidence level of 0,9545, which the software lacks, is shown. The estimates of the measurement uncertainty obtained by the proposed methods, Monte Carlo method and methodology of the Guide to the Expressing of Uncertainty in Measurement are compared.