Short-Term Creep Data Based Long-Term Creep Life Predictability for Grade 92 Steels and Its Microstructural Basis

Abstract

Long-term creep life (5000-100,000 h) predictabilities, based on the short-term creep life data (similar to 5000 h), for Grade 92 steel were investigated among major creep life prediction models, Larson-Miller parameter (LMP), normalized power law (NPL) and Wilshire models. The NPL and Wilshire models showed superior short-term creep data based long-term creep life predictabilities to the LMP model. In particular, the Wilshire model showed relatively accurate predictions (within an error range of 7%-10%), which seemed to be due to reasonable coupling of the normalized stress with the temperature-dependent rupture life term in a form of the cumulative distribution function. Both NPL and Wilshire models, calibrated by the short-term creep life data, showed a transition in creep mechanism at a similar normalized stress. Thermodynamics-based kinetic simulation (MatCalc (TM)) results for major precipitates (M23C6, MX and Z phases) of Grade 92 steel suggested that the creep transition is associated with coarsening of M23C6 precipitates at high temperatures (above 600 degrees C), which led to the degradation of the creep property.