Abstract: In Volume 2/3 of R5 ``An assessment procedure for the high temperature response of structures", the strain based ductility exhaustion method is suggested to calculate the creep damage of stress concentration region, which involves description of creep dwell initial stress, stress drop and creep ductility. Considering lots of uncertainty existed in these assessment, some sensitivity analysis is required in R5 procedure to ensure the conservatism of assessment results. In this paper, laboratory creep-fatigue test data of 316H at 550℃ with different loading conditions are selected as a special case to investigate whether the basic R5 approach is conservative, and the different material data combinations of cyclic stress-strain, creep deformation and creep ductility are used to identify those significant parameters affecting the assessment results. The analytical results indicate that the creep deformation model and creep ductility data are more significant to the results comparing with cyclic stress-strain data. If the upper bound of creep deformation law and lower bound of creep ductility data are used to predict their creep-fatigue lives, the degree of conservatism can be as large as a factor of -300, but if the modified creep deformation model and cast specific ductility data are used, very well prediction results can be gained within a factor of ±2.0 although there is slight non-conservatism.

Key words: Creep, Fatigue, Damage, Life prediction, Sensitivity study