Systematic Evaluation of Creep-Fatigue Life Prediction Methods for Various Alloys

Failure under creep-fatigue interaction is receiving an increasing interest due to an increased number of start-up and shut-downs in fossil power generation plants as well as development of newer nuclear power plants employing low-pressure coolant. Such situations have prompted the studies on creep-fatigue interaction and the developments of various approaches for evaluating its significance in design as well as remaining life evaluation, but most of them are fragmental and rather limited in terms of materials and test conditions covered. Therefore, applicability of the proposed approaches to different materials or even different temperatures is uncertain in many cases. The present work was conducted in order to comparably evaluate the representative approaches used in the prediction of failure life under creep-fatigue conditions as well as their modifications, by systematically applying them to available test data on a wide range of materials which have been used or are planned to be used in various types of power generation plants. The following observations have been made from this exercise: (i) The time fraction model has a tendency to be nonconservative in general, especially at low temperature and small strain ranges. Because of the large scatter of the total damage, this shortcoming would be difficult to cover by the consideration of creep-fatigue interaction in a simple manner. (ii) The classical ductility exhaustion model showed a general tendency to be overly conservative in many situations, especially at small strain ranges. (iii) The modified ductility exhaustion model based on the redefinition of creep damage showed improved predictability with a slightly nonconservative tendency. (iv) Energy-based ductility exhaustion model developed in this study seems to show the best predictability among the four procedures in an overall sense although some dependency on strain range and materials was observed.