Conventional time-temperature-parameter (TTP) methods often overestimate long-term creep rupture life of creep strength enhanced high Cr ferritic steels. The cause of the overestimation is studied on the basis of creep rupture data analysis on Gr.91, 92 and 122 steels. There are four regions with different values of stress exponent n for creep rupture life commonly in stress-rupture data of the three ferritic steels. Activation energies Q for rupture life in the regions take at least three different values. The values of n and Q decrease in a longer-term region. The decrease in Q value is the cause of the overestimation of long-term rupture life predicted by the conventional TTP methods neglecting the change in Q value. Therefore, before applying a TTP method creep rupture data should be divided into several data sets so that Q value is unique in each divided data set. When this multi-region analysis is adopted, all the data points of the steels can be described accurately, and their long-term creep life can be evaluated correctly. Substantial heat-to-heat and grade-to-grade variation in their creep strength is suggested under recent service conditions of USC power boilers.

A study of Grade 91 steel's creep rupture behavior at 600°C (up to 90,000 hours) and 650°C (up to 23,000 hours) reveals that static recovery of tempered martensite lath structures leads to decreased stress exponent and breakdown of creep strength. While M 23 C 6 and MX particles initially stabilize lath structures by hindering sub-boundary migration, the progressive aggregation of M 23 C 6 particles reduces their pinning force, triggering static recovery. Although Grade 91 steel shows better M 23 C 6 thermal stability compared to Grade 122 type steels (9-12%Cr-2W-0.4Mo-1Cu-VNb), coarsening of M 23 C 6 particles and subgrain width is expected to occur slightly beyond 100,000 hours at 600°C, potentially leading to creep strength breakdown.