Comparison of Creep-Fatigue Evaluation Methods With Notched Specimens Made of Mod.9Cr-1Mo Steel

In a component design at elevated temperature, creep-fatigue is one of the most important failure modes, and assessment of creep-fatigue life in structural discontinuity is important issue to evaluate structural integrity of the components. Therefore a lot of creep-fatigue life evaluation methods were proposed until now. To compare and assess these evaluation methods, a series of creep-fatigue tests was carried out with notched specimens. All the specimens were made of Mod.9Cr-1Mo steel, which it is a candidate material for a primary and secondary heat transport system components of JSFR (Japan Sodium-cooled Fast Reactor). Mechanical creep-fatigue tests and thermal creep-fatigue tests were performed by using conventional uni-axial push-pull fatigue test machine and thermal gradient generating system with an induction heating coil. Stress concentration levels were adjusted by varying the diameters of notch roots in the both tests. In the test, creep-fatigue lives, crack initiation and propagation processes were observed by digital micro-scope and replica method. Besides those, a series of elastic Finite Element Analysis (FEA) were carried out to predict the number of cycles to failure by several creep-fatigue life evaluation methods. Then these predictions were compared with test results. Several types of evaluation methods which are stress redistribution locus (SRL) method, simple elastic follow-up method and the methods described in JSME FR (Fast Reactor) code were applied. The applicability and conservativeness of these methods were discussed. It was appeared that SRL method gave rational prediction of creep-fatigue life with conservativeness when the factor of κ = 1.6 was applied for all the conditions tested in this study. Comparison of SRL method and simple elastic follow-up method indicated that SRL method applied factor of κ = 1.6 gave the smallest creep-fatigue life in practicable stress level. JSME FR code gave an evaluation 70∼100 times conservative lives comparing with the test results.