A Hardness-Normalised Model of Creep Rupture for P91 Steel

The use of creep data in life assessment is complicated by performance differences between different material casts. Direct tests on suspect components are not always feasible. This paper describes a novel analysis of P91 creep rupture data by reference to both generic materials flow stress data in the creep range, and hardness data for each individual cast at ambient temperature. The basic principles are to normalise the creep test stress by reference to the flow stress at the test temperature, and to use cast-specific hardness data to estimate the cast-specific flow stress. Data on a range of new P91 materials of varying hardness >200HV fell naturally into two regimes at higher and lower normalised stress values. The analysis showed substantially reduced data scatter compared with PD6525. An empirical refinement, allowing the hardness exponent to vary from unity, further reduced scatter and eliminated bias. The model is also compared with Further data on softer P91 materials which had been abnormally heat treated and/or softened in service. A reasonably good data fit is obtained. The application of the model to estimate the remanent life of soft materials in service is discussed.