A new paradigm of fatigue variability behavior and implications for life prediction

The treatment of the fatigue variability behavior has traditionally been based on the understanding of the mean-lifetime behavior. With reference to two turbine engine materials, an α + β titanium alloy and a nickel-based superalloy, it is shown that the traditional approach may not accurately describe the fatigue variability behavior of these materials. Decreases in stress level, or microstructural change directed at increasing the mean lifetime, were found to affect mean and worst-case (life-limiting) fatigue behavior differently, and these differences could not be accounted for in the traditional understanding. In particular, the life-limiting mechanism was controlled by crack growth although the mean-lifetime response was increasingly dominated by crack initiation with decreasing stress level. A new paradigm of fatigue variability was therefore suggested, in which the total uncertainty in lifetime breaks down into the variability in (1) the worst-case mechanism and that in (2) the classical, mean-lifetime governing response. The effects of microstructure and temperature on the fatigue variability behavior were studied with respect to the new paradigm and found to have a very systematic effect on the worst-case and the mean behavior, depending on the degree of influence of these variables on the crack initiation and the growth regime.