Estimation of uncertainty in olfactometry.

Estimation of uncertainty in odour measurement is essential to the interpretation of the measurement results. The fundamental procedure for the estimation of measurement uncertainty comprises the specification of the measurement process, expression of the measurement model and all influences, evaluation of the standard uncertainty of each component, calculation of the combined standard uncertainty, determination of a coverage factor, calculation of the expanded uncertainty and reporting. Collaborative study such as interlaboratory comparison of olfactometry yields performance indicators of the measurement method including repeatability and reproducibility. Therefore, the use of collaborative test results for measurement uncertainty estimation according to ISO/TS 21748 and ISO 20988 is effective and reasonable. Measurement uncertainty of the triangular odour bag method was estimated using interlaboratory comparison data from 2003 to 2007 on the basis of the simplest model of statistical analysis, and the expanded uncertainty of odour index ranged between 3.1 and 6.7. On the basis of the establishment of the estimation procedure for uncertainty, a coherent interpretation method for the measurement results will be proposed and more effective and practical quality control of olfactometry will be available.