Ensuring the validity of the results of calibrations and tests by statistical methods

Abstract

According to the requirements of ISO/IEC 17025:2017, the validity of test and calibration results is ensured, inter alia, by intralaboratory check of the results obtained. In this case, it is preferable to use statistical methods.
 The ISO 5725 standards define a number of such methods, but the choice of specific methods is left to the laboratory, taking into account the requirements for the adequacy of the effort, resources and time for the purposes of the work performed and the risks of obtaining inappropriate results. In this case, the laboratory itself must in a certain way determine which objects of calibrations (tests) should be predominantly used in checks and what frequency of checks should be foreseen.
 In connection with the increase in the accuracy and complexity of measuring systems, the need to apply the methods of the theory of random processes becomes more and more obvious. It is shown that the use of the Poincaré plot makes it possible to comprehensively, effectively and visually evaluate changes in the measuring process from the point of view of the dynamics of the obtained measurement results.
 The results of the check, in particular, the intermediate precision, make it possible to obtain a more realistic evaluation of measurement uncertainty in accordance with ISO 21748.
 The paper analyses some practical approaches (of varying degrees of complexity) to intralaboratory checks of the validity of calibration (test) results.