Abstract
Creep rupture life in the time range of 105 h and longer has been evaluated on the basis of creep rupture data available on modified 9Cr–1Mo (Gr.91) steel. The steel is strengthened by a tempered martensite lath structure, which degrades by means of strain induced recovery in short-term creep (Region H). The degradation leads to creep rupture. In long-term creep (Region L), static recovery of the martensite lath structure takes place in addition to the strain induced recovery, resulting in premature creep rupture and the consequent decrease in stress exponent n for rupture life. The premature creep rupture always occurs in Gr.91 steel, but the activation energy QL for rupture life in region L is usually close to the value QH in Region H in the steel. In such a case, long-term rupture life can be predicted correctly by simple extrapolation of short-term data. In some heats of Gr.91 steel, however, another region with low values of n and Q, namely Region L2, appears in the time range longer than 104 h. Simple extrapolation of short-term data results in overestimation of rupture life in Region L2. The static recovery and the consequent premature creep rupture are controlled by migration of sub-boundaries stabilized by precipitates. M23C6 as well as MX precipitates in Gr.91 steel have higher thermal stability than those in other high Cr ferritic steels, resulting in the later appearance or absence of Regions L and L2 in the steel.
