Uncertainty evaluation in measurement processes
The main objective is the dissemination of the experimental statistics principles, developing their application in the industrial area. In the early ISO 9000 standards, the references to the statistical tools use in products and process development were unspecific and limited; the latest standards are more evident in the requirements of the statistical methodologies use, particularly when referring to the management of laboratory tests and measurements.
A number of factors may affect the measurements; therefore the ultimate report can not be a single digit, but rather an interval containing the measurement result in which we are confident to find the characteristics of interest. Such interval is just estimated using methodologies, sometimes very complex, requiring a rigorous statistical analysis.
• Definition of the mathematical model, i.e. definition of the relationships between the magnitudes determined by the measurement procedure chosen.
• Identification of sources of uncertainty, quantification of the uncertainty components, and their estimation; computation of the combined standard deviation as measure of uncertainty.
• Uncertainty evaluation due to the sampling errors.
• Definition of efficient sequential design strategies to reduce sample size.