Study On Repeatability And Reproducibility Of ISO Standards For Chemical Composition Analysis Of Iron And Steel

ISO TR 21074 was published in 2016 according to a series of ISO 5725 standards by ISO/TC17/SC1(Steel--Methods of determination of chemical composition of International Organization for Standardization).The formal rules for ISO standards were established.The technical report(ISO TR 21074:2016)was first time to apply the coefficient of variation of interlaboratory reproducibility standard deviation to confirm the lower and upper limit of analytical range for the test method.The trueness criterion of the results of collaborative test was also proposed.The requirement of the correlation coefficients of the regression equations of the precision parameters was described.Chinese Steel and Alloy Chemical Composition Measurement Sub-committee(SAC/TC 183/SC 5)which is the counterpart of ISO/TC 17/SC 1 also make standards according to the latest requirements of ISO.It was found that in the technical report the data amount was too limited in the model fitting of the interlaboratory reproducibility;the maximum variation coefficient of interlaboratory reproducibility which was used to decide the analytical scope of test method was not rigorous;the requirement of correlation coefficient of regression equation was ambiguous.Based on 63 ISO effective test standards,technical reports and technical specifications,the validity of standard data involved in statistical analysis was verified.The factors of within-laboratory repeatability limit(),interlaboratory reproducibility limit()and within-laboratory reproducibility limit(_w)were analyzed.Mathematical models were established of,and_w versus the analyte levels.The ranges of the coefficients of variation of within-laboratory repeatability,interlaboratory reproducibility and within-laboratory reproducibility were set up.The following application researches were carried out according to the models of within-laboratory repeatability,interlaboratory reproducibility and within-laboratory reproducibility:the ratio between the parameters(,,_w),the estimation of the quantitative limit and the semi-quantitative limit in the laboratory,the precision target and the maximum precision limit of the analytical instrument for the standard method and the non-standard method,the evaluation rules of the long-term stability of the instrument,the tolerance limit of the test method under different measurement conditions,the simple estimation of measurement uncertainty,decision of the analytical range in the standard method,the requirement of the correlation coefficients of the regression equations of the precision parameters.Finally,the revision draft of the ISO TR 21074:2016 was proposed through these researches.