It has been generally accepted that the reduction of the fatigue life for the specimen tested under the creep-fatigue loading condition is thought to be attributed to the interaction between the fatigue damage resulting from the initiation of the fatigue crack at the material surface and its growth and the creep damage resulting from the cavity nucleation and growth at the grain boundaries.
On the other hand, in some recent works for 340L stainless steel and 1Cr-Mo-V steel, it has been reported that the fatigue lives were significantly reduced under the tensile hold low cycle fatigue conditions though the cavitational damage were not observed. In our experiments for the same materials, this phenomenon is also observed under the test conditions of 823K and 865K and several different duration of hold times. That is, the fatigue lives are abruptly reduced even though the main cracks are transgranularly propagated independent of the grain boundaries. Also, SEM micrographs clearly show the fatigue striations which are typical feature of the transgranular propagation. The degradation of the fatigue lives in tensile hold condition is reflected on the enlarged striation spacings and the increased population of the nucleated cracks. Especially, the ratios of the striation spacing in LCF with tensile hold to that in continuous LCF are similar to the ratios of the fatigue lives under the same conditions. The enlarged striation spacings are attributed to a certain damage within the crack-tip region.
In this paper, according to the investigations for the feature of the crack propagation mode and microstructure of the crack-tip region, this crack-tip damage is considered to be related with a stress relaxation within crack-tip region during hold time. The behaviour of this relaxation plays two roles in fatigue crack advance. One is the additional crack growth during hold time caused by stress relaxation. Another is the accelated crack advance during subsequent cycling caused by crack tip stress relaxation during hold time. That is, the crack-tip stress and strain distribution, readjusted during hold time caused by relaxation, enlarges the amount of crack tip sliding-off process during subsequent cycling. These two processes are correlated with each other through the stress relaxation and cyclic work hardening characteristics of the materials.
A model for the creep-fatigue crack advance is suggested from the new idea for the continuous fatigue crack growth. There is a fatigue process zone in front of the crack in which the actual degradation of the material is occurred with crack advance. Under the concept that the crack advance per cycle is occurred by the average amount of a irreversible shearing deformation ($\epsilon ^{avg}_p$) within the crack-tip shearing process zone ($D_p$), the quantitative expressions for the crack growth rate and life prediction are suggested. According to the comparison with the experimental results, the proposed life prediction is successful for the several materials.
Applying the concept for the continuous crack growth, the expression for the creep-fatigue crack growth rate with tensile hold time is suggested as following equation.
$da/dN=(\epsilon ^{avg}_p+\epsilon ^{avg}_{pr})Dp$
The accumulated average strain during hold time ($\epsilon ^{avg}_{pr}$) is obtained from the time dependent equation at high temperature condition suggested by Riedel and Rice. The relative importance of the respective terms which depend on each other is different from one material to another. According to the analysis for the model, in 304L stainless steel which has high cyclic stress-strain exponent, the first term which represents the accelated crack growth during cycling is more important than the second term which represents the additional crack growth during hold time. On the other hand, in 1Cr-Mo-V steel which has very low cyclic stress-strain exponent, the reverse is the case.
In this paper, this model for the crack growth is also applied to explain the degradation of the life under the condition of a compressive hold cycling for 1Cr-Mo-V steel and 12Cr-Mo-V steel. The consideration for the compressive hold cycle is based on the idea that the compressive hold results in the decrease of the maximum crack-tip flow stress for the crack-tip shearing deformation and induces tensile mean stress.
As a conclusion, while the previous reports suggested that the life degradation in tensile and compressive hold cycles was only attributed to the oxidation effect when excepting the cavitation, this model suggests that it is also attributed to the stress relaxation effect in the crack-tip region. Also, on these basis, the life predictions for 304L stainless steel and 1Cr-Mo-V steel are successful.
