An Overview of Creep-Fatigue Damage Evaluation Methods and an Alternative Approach

Abstract

Renewed interest in elevated temperature nuclear reactors has occasioned a reassessment of creep-fatigue damage evaluation methods. Points to be improved in the current methods employed in Subsection NH of the ASME B&PV Code and other design codes are discussed as well as an alternate approach which avoids some of these problems. Most current creep-fatigue damage evaluation methods separately evaluate cyclic fatigue damage and creep damage and assess the combined damage through interaction diagrams. Typically test data are evaluated through a Miner’s Rule summation of fatigue damage and either a time fraction summation of creep damage or a ductility exhaustion approach in order to establish the appropriate interaction curve. In these approaches, cycles to failure can be counted directly but creep damage is a calculated parameter, subject the limitations of the evaluation technique. There can be considerable scatter in the results. The process is reversed for design and the methodology chosen to assess creep damage will have a major impact on the viability of the design process. This was found to be particularly true for advanced alloys such as Mod9Cr-1Mo-V, aka Grade 91. An alternate approach to determination of cyclic life has been proposed which avoids parsing the damage into creep and fatigue components. This approach, called the Simplified Model Test (SMT), employs a test specimen with elastic follow-up sized to represent the stress and strain redistribution encountered in more complex structures. The correlation parameter between test and design is the elastically calculated strain and the dependent test variable is the observed cycles to failure. The SMT approach has two major advantages. First, because the correlation parameter is elastically calculated strain, it is not necessary to calculate the inelastic stress-strain history for a design evaluation; either directly through inelastic analysis or indirectly through manipulation of elastic analyses. Second, because the test specimen itself incorporates the hardening, softening and aging effects of the structure it represents, it is not necessary do rely on theoretical modeling of these effects in an artificial separate accounting of creep and fatigue damage.