Although the high cycle fatigue (HCF) loading conditions experienced in gas-turbine engines are characterized by low stress amplitudes and long fatigue lives in terms of number of cycles (i.e. > 10⁹), such failures in engine components involve extremely high cyclic loading frequencies (1 kHz or greater), such that fatigue lives are very short in terms of time to failure. The situation is further complicated by complex loading conditions, including varying degrees of mode-mixity, high mean stresses, and superimposed HCF and LCF (low cycle fatigue) spectra, as well as the presence of surface conditions which are believed to hasten crack initiation, including foreign object damage (FOD) and fretting. Accordingly, for life prediction and design procedures, it is necessary to characterize the critical states of microstructural and geometric damage that will lead to crack initiation, growth, and failure by HCF. One approach to such a characterization is the definition of lower-bound thresholds for fatigue crack propagation.

Damage tolerant approaches to fatigue design commonly employ crack propagation rates measured for long cracks (e.g., crack size, a > 5 mm). Given the widely reported disparity between long and small (a < 0.5 - 1.0 mm) fatigue cracks [1-4], this approach must be deemed non-conservative. Indeed, a pronounced small-crack effect has been observed in Ti-6Al-4V (wt. %) with a variety of microstructures, with small cracks exhibiting i) faster propagation rates than long cracks at equivalent DK and ii) significant crack-growth rates at applied DK below the long-crack DK_TH. In Figure 1, this point is illustrated using a compilation of literature data [5-9] for propagation rates of small, semi-elliptical surface cracks and long, through-thickness cracks in Ti-6Al-4V exhibiting both fully lamellar and bimodal (equiaxed primary a in a transformed b matrix) microstructures. In addition to clearly demonstrating the presence of a small-crack effect, Figure 1 also reveals that the relative fatigue-crack propagation resistance of these two common Ti-6Al-4V microstructures is dependent on defect size. In the case of long, through-thickness cracks, fully lamellar microstructures exhibit superior cyclic crack growth properties, with higher DK_TH and lower growth rates at equivalent DK than the bimodal microstructures. Conversely, for small crack sizes (full surface crack length, 2c ~ 5-500 mm), the bimodal microstructures are better, with lower growth rates at equivalent values of DK. Additionally, small crack propagation rates in the lamellar microstructure exhibit far greater scatter. Although such small crack behavior has been well documented for constant amplitude, mode I loading on naturally initiated flaws, a detailed examination of this behavior is needed under the aforementioned surface and loading conditions which characterize gas-turbine engine HCF.

Accordingly, in the present work, the initiation and propagation of small, semi-elliptical surface cracks in Ti-6Al-4V is characterized under HCF conditions, i.e., at high mean stress (R = 0.8) and high frequencies (n = 1-2 kHz). Various microstructures are examined, having been thermo-mechanically processed to achieve varied volume fractions of primary a while maintaining equivalent a lath dimensions in the transformed b matrix. Specifically, microstructures containing primary a contents of ~50% (STOA), 15%, and 0% (fully lamellar) are investigated. Thresholds for crack formation, Ds_TH, and growth, DK_TH, from sites of surface damage, including cylindrical notches (produced by both traditional machining and electrical discharge machining), hardness indents, and ion milled notches, are observed and compared with available fatigue-crack propagation data for long cracks and naturally initiated small cracks. The influence of superimposed low mean stress, low frequency loading (i.e. HCF/LCF interactions) on the threshold conditions for crack initiation and growth are also observed. In the case of HCF/LCF interactions, a comparison is made between a loading spectrum in which the minimum applied stresses remain tensile and a spectrum in which the LCF minimum stress is compressive. This compression loading is expected to alter crack initiation and growth conditions by generating residual tensile stresses in the vicinity of notch-like defects and by suppressing crack closure effects.

Acknowledgments

This work was funded by the U.S. Air Force Office of Scientific Research through the Multidisciplinary University Research Initiative on High Cycle Fatigue, Grant No. F49620-96-1-0478

References

1. S. Suresh and R.O. Ritchie, Int. Metals Rev. 29 (1984), 445.
2. K.J. Miller, Fatigue Engng. Mater. Struct. 5 (1982), 223.
3. R.O. Ritchie and J. Lankford, Mater. Sci. Eng. A84 (1986), 11.
4. J. Lankford, Fatigue Engng. Mater. Struct. 8 (1985), 161.
5. A.W. Thompson, J.C. Williams, J.D. Frandsen, and J.C. Chesnutt in Titanium and Titanium Alloys: Scientific and Technological Aspects, ed. J.C. Williams and A.F. Belov, (Plenum Press, 1982), 691.
6. M.D. Halliday and C.J. Beevers, Int. J. Fract. 15 (1979), R27.
7. G. Lütjering, J. Albrecht, and A. Gysler in Titanium '92: Science and Technology, ed. F.H. Froes and I. Caplan, (TMS; Warrendale, PA, 1993), 1635.
8. R.K. Bolingbroke and J.E. King in Small Fatigue Cracks, ed. R.O. Ritchie and J. Lankford, (TMS; Warrendale, PA, 1986), 129.
9. G. Lütjering, A. Gysler, and L. Wagner in 6th World Conference on Titanium, (Societé Francaise de Metallurgie, Les Illis, France, 1988), 79.