Document Scope
A foundational metrological supplement provides guidance on the propagation of probability distributions for the evaluation of measurement uncertainty. This JCGM 101 document introduces numerical methods that serve as an alternative to the traditional law of propagation of uncertainty. It focuses on cases where the linear approximation of the measurement model is inadequate or the output distribution is highly asymmetric.
Numerical Implementation
The document establishes a framework for applying Monte Carlo techniques to assign probability distributions to output quantities. Instead of relying on analytical solutions, the method generates a large number of random samples from the input distributions.
Validation Process
Verification of the traditional GUM uncertainty framework is a primary application of this international supplement. If the output of the analytical method agrees with the numerical results within a defined numerical tolerance, the simpler GUM method can be safely used. This step prevents the misapplication of linear approximations in highly non-linear measurement scenarios.
Output Analysis
The results of the calculation provide the mean and the standard deviation for the output quantity. By avoiding the assumption of a symmetric Gaussian distribution, this approach yields more realistic coverage intervals for non-linear models. This makes the method indispensable for high-precision calibrations where asymmetric physical limits exist, and it prevents the underestimation of risk in industrial testing.