About the size of measuring instruments sample in the conformity assessment of them with technical regulations on the module F

Abstract

The principles of measuring instruments sampling in the conformity assessment with technical regulations on the module F and the reliability of using the sample are considered in this paper. According to the technical regulation of the legislatively regulated measuring instruments and the technical regulation of measuring instruments, in the compliance assessment according to the module F, sampling control is possible with a confidence level of 0.95. In order to verify the fulfillment of the requirement to ensure a confidence probability of 0.95, the simplest but very laborious way is to take a sufficiently large batch size and verify that the amount of measuring instruments with an error not exceeding its normalized value is not less than 95 % of the batch size. Another way that requires a much smaller batch size is to find the root mean square deviation of errors (with their signs) for each measuring instrument, then based on the Student’s distribution, to find the deviation corresponding to the probability of 0.95. The test result can be considered positive if the deviation obtained does not exceed the normalized error. Confidence intervals for the root mean square deviation are about 22 % for 20 instruments and about 9 % for 100 instruments. There may be some additional reduction in sample size, if each instrument from a batch is tested at several points of the range and there are grounds for considering the results at these points to be independent. Then, not the maximum error in the range is taken, but the errors (with signs) at all points of the range are used for each instrument. In this case, the equivalent number of instruments is reduced in the number of measurement points of the range.