A. Hutson, W. J. Porter, University of Dayton Research Institute, Dayton, OH; J. M. Larsen, Air Force Research Laboratory, Wright-Patterson AFB, OH
Fatigue behavior of structural metals is known to vary widely, even with focused loading and environmental conditions. The mechanisms that drive this variability are unknown for most alloys; however, it has been suggested that the worst case fatigue behavior may be dominated by crack growth, with little contribution from crack initiation to the total fatigue life. Often, this worst case is not seen in laboratory experiments, because too few experiments are performed to capture the “1 in 1,000” behavior. In the present study, fatigue and crack growth samples were taken from a section of through-transus processed Ti-6Al-2Sn-4Zr-6Mo for the purpose of evaluating variability in behavior. Twenty-five constant load amplitude fatigue experiments were conducted for each of three test conditions. Two sets of tests were conducted on low-stress ground specimens, one at 827 MPa and one at 860 MPa. The third set of tests was conducted on electro-polished specimens at 860 MPa. All tests were performed in lab air at 260°C using a stress ratio of 0.05 at a frequency of 30 Hz. Statistical analysis of the resulting data provided a means of estimating a value for the “1 in 1,000” fatigue life. Fatigue crack growth experiments were also conducted, under similar laboratory conditions, to provide a basis for fatigue life prediction. Characterization of the microstructure and of the resulting fracture surfaces from S-N tests allowed estimates of the initial crack sizes for the predictions. A few fatigue experiments were conducted in which surface flaws with a ~ 0.010 mm and c ~ 0.020 mm were introduced, to validate the predictions. Life prediction analysis was conducted using a modified small plus large crack growth model. The results from the analysis were compared with the “1 in 1000” minimum fatigue lives from the statistical analysis. They were also compared with the minimum S-N fatigue lives and with the experimental fatigue lives from the micro-notched specimens. While the fatigue life variability was quite wide, exceeding a factor of ten in some cases, the minimum fatigue lives compared well with the crack growth predictions and with the fatigue lives obtained from the micro-notched samples.
Summary: This research describes experimental results and related crack growth life predictions on a through-transus processed Titanium alloy, and sources of variability in the experimental fatigue behavior.
