Abstract

The new scheme of proficiency testing of laboratories for determination of main quality characteristics of oil cake was developed. In pilot round was determined such quality characteristics as moisture content (I), mass fraction of fat (II) and mass fraction of crude protein (III). The control sample was prepared from the sample of rapeseed oil cake. It was determined that control sample was sufficiently stable and homogeneous. In the round participated 22 laboratories, which used different test methods. The results were processed by the method described in ISO 13528. Based on these calculations as assigned value for characteristics I and III was chosen robust mean value and for characteristic II - mode of results. As standard deviation of proficiency testing for characteristic I was chosen robust standard deviation and for characteristics II and III was chosen value which is predicted by equation of Horwitz-Thompson. It was determined, that for characteristics I and III different test methods gives equivalent results, so some unsatisfactory results of laboratories can be attributed to laboratory bias. For characteristic II clustering of results according to measurement method in observed. It was established, that outdated standard GOST 13496.15-97 gives higher mass fraction of fat compared to other test methods.

Keywords: proficiency testing, oil cake, statistical methods

Article language: Ukr