Creep life fraction assessment of 9Cr–1Mo–V steel using direct current potential drop technique

Abstract This study investigates the use of direct current potential drop (DCPD) measurements as a predictive tool for assessing creep life fractions in 9Cr–1Mo–V (P91) steel exposed to high-temperature and high-stress conditions. To optimize predictive reliability, the data analysis strategically focused on the mechanistically stable 20%–80% creep life fraction window. The continuous DCPD signals within this range were analyzed using machine learning-based regression models. These refined models achieved high predictive accuracy, yielding an average mean absolute percentage error of 8.14% for quadratic regression and 8.29% for linear regression. Furthermore, microstructural evolution was characterized through carbide precipitation using the on-site extraction replica technique, with an emphasis on the Cr/Fe ratio as a primary indicator of creep damage. Distinct DCPD signal patterns were identified across the creep stages: initial transient fluctuations during the primary stage, a steady decline in the secondary stage due to matrix purification, and a sharp increase in the tertiary stage driven by internal damage. These responses were strongly correlated with carbide precipitation and coarsening induced by prolonged stress exposure. Integrating DCPD with microstructural analysis provides a robust, non-destructive framework for assessing creep damage. This combined approach offers practical insights for the design and health monitoring of high-temperature components, supporting structural integrity and lifetime management in power plants and other critical applications.