Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
778222 | International Journal of Fatigue | 2014 | 10 Pages |
•Fatigue life data were analyzed for bimodal (FCG or initiation-dominated) behavior.•Short life populations were observed preferentially under more aggressive conditions.•Short lives were consistent with FCP lives, supported by micro-notched tests.•Relative sampling of two mechanisms affected accuracy of uni-modal predictions.•Crack growth-dominated lives were most readily assessed via electropolished samples.
The experimental conditions that produce bimodally-distributed fatigue populations were investigated in Ti–6Al–2Sn–4Zr–6Mo, with the aim of identifying conditions favorable for characterizing life-limiting fatigue behavior. In general, the more aggressive test conditions in this effort were identified as more favorable for producing bimodally-distributed lifetime populations in which the shorter life populations were dominated by the life-limiting behavior. Data sets obtained using electropolished specimens generally exhibited bimodal distributions, and this surface condition was determined to be most suitable for investigating life-limiting fatigue behavior for a wider range of test conditions, due to the elimination of machining-induced residual stresses. It was also determined that the presence of bimodally distributed data populations, and the relative sampling of the corresponding damage mechanisms, can have a significant effect on life predictions, if the bimodal behavior is not explicitly included.