Measurement uncertainty evaluation by kurtosis method at calibration of a household water meter

The investigation of measurement uncertainty evaluation at water meter calibration was considered. The peculiarities of this problem is non-linear measurement model and the significant uncertainty, related with the poor repeatability at the small number of the measurements, which makes it difficult to expanded measurement uncertainty evaluate based on the Bayesian approach. The paper checks the influence of the nonlinearity of the model equation on the accuracy of evaluation the numerical value and the standard uncertainty of the measurand. The application of the kurtosis method in the evaluation of the expanded uncertainty of type B and the application of the law of propagation of the expanded uncertainty in the evaluation of the expanded uncertainty of the measurand are described. The obtained values of the expanded uncertainty are very close to the one obtained from Monte Carlo method.