Some remarks on a bivariate analysis in the propagation of measurement uncertainty as an alternative approach to the Monte Carlo method

Recent literature shown that in cases where there is strong interdependence between direct measures involved in an indirect measurement, or when the uncertainties of direct measures are not as small as one would wish, or again, when a strong nonlinearity ties the measures involved to each other, the Monte Carlo method, proposed by Supplement 1 to the "Guide to the expression of uncertainty in measurement" (GUM), manages in most cases to obtain a more accurate measurement uncertainty evaluation than the one actually proposed by the GUM and bound to Taylor series developed. This paper aims to offer a new approach related to bivariate analysis of standard models such as sum and product of measures when it is either possible, or unsuitable, to apply the two methods quoted in Supplement 1.