The behaviour of fatigue cracks under variable amplitude spectra has been studied intensively over the past 20 years. [3,4] For many load sequences, delay effects are found, compared with the growth rates predicted by a linear summation. It is well known that fatigue crack growth rates caused by small cycles following larger load excursions are slower that the linear summation damage prediction. A linear summation of crack growth is that obtained by summing the crack growth increments for constant amplitude loading, assuming no interactions between succeeding load cycles. Fatigue crack acceleration effects have been reported previously, [5] and these were found when high R ratio cycles were followed by underloads and overloads. More recently Buller et al [2] have demonstrated crack acceleration effects, compared with a linear summation, when precracked CT samples of titanium alloy Ti-10-2-3 and aluminium alloy 7010 were subjected to helicopter rotorhead spectra consisting of large numbers of near threshold cycles of R between 0.7-0.83, interspersed with excursions to zero load. To further investigate these effects, in this study, fatigue crack growth rates were measured under two different Simple Variable Amplitude Loading (SVAL) spectra. These are simplified versions of the real helicopter rotorhead spectrum. The spectra consisted of small cycles with R ratio of 0.7-0.9, and excursions to zero load. The sequence, number and R ratio of the cycles were systematically varied. The study was conducted on Ti-10-2-3 alloy. Experimental crack growth rates were compared with those calculated using a linear summation technique, with input data from constant amplitude testing on the Ti-10-2-3 alloy. EXPERIMENTAL PROCEDURE Tests were performed on Ti-10-2-3 which belongs to the near beta (β) or solute-lean beta alloy class where the body centred cubic (bcc) β phase is predominant [6]. Mechanical properties of the test material are shown in Table 1. TABLE 1 NOMINAL MECHANICAL PROPERTIES OF THE Ti-10-2-3 TEST MATERIAL Parameter Description Ti-1023 σUTS Ultimate Tensile Strength 1146 MPa σYS Yield Strength 1048 MPa KIC Fracture Toughness 63.9 MPa m^1/2 Compact tension (CT) specimens (W = 70.0 mm, t = 17.5mm) were used for the research. Constant amplitude and decreasing ∆K crack growth rate data for this alloy were measured at R ratios of 0.1, 0.4, 0.7 and 0.9. In addition, fatigue tests were performed using the two load spectra shown in figure 1. In SVAL-1, constant amplitude cycles at R* ratios of 0.7 and 0.9 were applied for n cycles, followed by n΄ cycles of underloads at R=0 (figure 1a). The starting ∆K value of the R* cycles was equal to the threshold of 1.9 MPa m1/2 at R* of 0.7 and 1.6 for R* of 0.9. The ∆K of the R = 0 cycles was 6.3 MPa m1/2 for R* of 0.7, and 16 MPa m1/2 for R* of 0.9. For R* = 0.7, the number of underloads n΄ was varied from 1 to 50, keeping n constant at 1000 cycles. Additional tests were performed with n values between 2,000 and 50,000. A further sequence applied 1 underload every 20 small cycles with R* = 0.7. For R*=0.9, the number of test sequences was reduced to 3, with n of 20, 1000 and 2000; n' being 1, 50 and 50 respectively. All sequences were repeated for a total of 106 small cycles. In SVAL-2, CAL at R=0.9 was applied for k cycles followed by CAL at R=0.7 for k'' cycles (figure 1b). The ∆K of cycles at R=0.9 and R=0.7 was the same and was equal to the threshold value of the stress intensity factor ∆Kth (=1.6 MPa m1/2) at the beginning of the test. Tests were performed such that the ratio (k''+1)/k was 0.1. The range of small cycles k at R=0.9 was from 100 to 10,000. This sequence was repeated for 106 small cycles. Unlike SVAL-1, this sequence contained three cycle ranges. These were k
RkJQdWJsaXNoZXIy MjM0NDE=