Fatigue crack growth rates under variable amplitude load spectra containing tensile underloads

An extensive research program was performed to investigate the load interaction effect of the combined action of small amplitude high R ratio cycles and large amplitude low R ratio underloads on the crack growth of large cracks. The study was driven by the needs of the damage tolerance approach in the helicopter structures, which requires robust knowledge on the crack growth behaviour of the advance high strength alloys under the characteristic helicopter spectra loading.

The study was conducted on three metallic alloys, Ti-10V-2Fe-3Al, Al8090 T852 and Al7010 T76351 using compact tension specimens (w=70mm, t=17mm). The potential drop technique was used for the measurements of the crack length. The crack opening loads were determined from the applied load versus crack opening mouth displacement curve using a curve fitting technique and crack opening displacement gauge.

The experimental results show that cracks can grow faster than the life predictions with no load interaction effects under spectra containing tensile underloads. The acceleration effects are different depending on the number of the small cycles, the Kmax, the R ratio of the small cycles, the underload cycle and the material. Significant closure observations on the underloads and on the small cycles of variable amplitude loading spectra were made. Based on the test finding and on the studies of other researchers, it is suggested that the acceleration effects are mainly due to the reduction of crack opening point of the tensile underloads comparing with the Constant Amplitude Loading (CAL) data.

An extensive evaluation of the ability of FASTRAN model to predict the fatigue lives under the tested loading spectra was carried out. The evaluation focuses on the influence of the constraint factor a and the ∆Keff curve inputs on the predictions. The model produces very good and consistent predictions for the three alloys, when the inputs represent adequately the actual fatigue mechanism. The model predicts the measured acceleration effects by reducing the closure level of the underloads.