Derivation of minimum 0,2%-proof strength values of steels at elevated temperatures. Background information and practical usage of the method 'A' presented in prEN 10314

The method A recommended in prEN 10314 for the derivation of minimum 0,2%-proof strength values of steels has been developed by combining the linear regression model as already set up in ISO 2605-1 and 2605-3 with the cumulative frequency model and the requirement that the standard deviation of the residual scatter shall not be in excess of 5% of Rp0,2/20C/min. The ratio values appear to be strongly related to the microstructure of the steels. The method has been shown to be applicable to small data groups with more than 80% coverage as required in prEN 10314, but also to large data groups with insufficient coverage when the mean ratio values at specified temperatures were determined by means of the correlation analysis. The some procedure proved successful in the evaluation of inhomogeneous data groups where the regression line of the controlling subgroup was assumed to be placed above the lower boundary of the scatter band by not more than 10% of Rp0,2/20C/min. Data groups related to nominally the same grade of steel but obtained from different sources should be kept separate as a measure to save as much as possible of information about the process route, and as a measure of precaution against the dominance of an individual subgroup when these are different in the number of data. (orig.)