Analysis of the difference in creep strength between heats in Gr. 91 steel by data science

Abstract

Data analysis was conducted to investigate the causes of the difference in creep strength between the heats of Gr. 91 steel. Creep rupture data were divided into multiple regions and fitted with an exponential degradation model. The Larson-Miller constant of each heat obtained by the fitting was used as the objective variable, and chemical concentrations of important elements, hardness and grain size were used as explanatory variables. The stepwise method was used for the data analysis. As a result of the data analysis, it is found that hardness in region H, grain size in region G, and Cr and C concentrations in regions M and L are the causes of the difference in creep strength between the heats.