Risk of false conformity assessment applied to automotive fuel analysis: A multiparameter approach

Abstract

The low-quality of automotive fuels may lead to the generation of pollutants harmful to both environmental and human health. The quality evaluation of automotive fuels requires a multiparameter conformity assessment, which may lead to an increased total risk of false conformity decisions even if all parameters comply with the acceptance limits. Thus, the aim of this work was to propose the establishment of multivariate acceptance limits in order to ensure a reduced total risk of false conformity decisions applied to automotive fuels analysis. Particular and total (consumers' and/or producers') risks were estimated using frequentist (specific) and Bayesian (global) approaches. Multivariate acceptance limits were estimated using Monte Carlo method, adopting an appropriate multivariate coverage factor (k') defined using MS Excel Solver function. The definition of multivariate acceptance limits ensures a total risk below the maximum admissible risk (typically 5%) and was successfully employed in the conformity assessment of automotive fuels (diesel and gasoline). The employment of the multivariate acceptance limits may be useful in the conformity assessment of several multiparameter products.

Keywords: Automotive fuel; Conformity assessment; Measurement uncertainty; Multiparameter approach; Multivariate analysis.