observations. The model included pads with a 3.175 mm blending radius at both ends, resulting in an undeformed contact length of 6.35 mm for the short pad and 19.05 mm for the long pad. In two of the three test conditions simulated numerically [1,3,4], the coefficient of friction, µ, was taken as 0.3. The third case was run with µ=1.0, to allow comparison of results where only µ was changed. Assuming cracks normal to the specimen surface under pure mode I loading, and using the stresses from the FEA, the stress intensity factor, K, was computed as a function of crack length. The K distributions were calculated using a program developed to model a single edge tension (SE(T)) specimen geometry under a non-constant σx stress field [12]. Some success has been reported in modeling fretting fatigue behavior as an edge notch, since both conditions produce similar stress fields [13]. Characterization results from a prior investigation [4] indicated that cracks nucleated at or near the edges of contact in specimens tested in the existing apparatus. The objective of these calculations was to compare the K values from two different cases where the peak stresses were significantly different, yet the stress fields both matched conditions corresponding to failure in 107 cycles. The K solution program required material and crack dimension input parameters in addition to the stress distribution data. 120 GPa was used for the material modulus. The depth position of the final stress value given in the stress distribution, which was equal to half of the specimen thickness, was used for the final crack length. Stress intensity values are only reported for the first 100µm into the specimen thickness, although calculations were performed as far into the model specimen as possible using the available stress data from the FEA simulation. TABLE 1 SUMMARY OF TEST CONDITIONS AND RESULTS FOR FINITE ELEMENT ANALYSIS. Specimen thickness (mm) Pad length (mm) µ Clamping Load (kN) Applied σaxial (MPa) Average σy (MPa) Average τxy (MPa) Peak σx (MPa) Max. Relative Displacement (µm) 1 12.7 0.3 49 275 770 21.6 1300 4.0 4 25.4 0.3 34 275 140 28.9 540 19.0 4 25.4 1.0 34 275 140 28.9 1100 3.4 RESULTS & DISCUSSION The first results presented are the Haigh stresses corresponding to a 107 cycle fatigue life for the 1 mm and 4 mm thick specimens over a range of contact conditions (Figs. 2 and 3). The data are presented as a function of average applied clamping stress, which is taken to be positive. Two contact radii (CR) and two stress ratios are represented, and the baseline uniaxial fatigue limits for this material: 660 MPa @ R=0.5, 558 MPa @ R=0.1, are included. The numerical simulation conditions area also noted. Much of the data for the 4 mm thick specimens was reported earlier [1] and seemed to indicate no appreciable trend as a function of applied clamping stress. However, the data for the 1 mm thick specimens do show a trend of increasing fatigue strength with decreasing applied clamping stresses, which was not noted in previous investigations [1,4]. The trend is not marked and conducting tests with clamping stresses much below 100 MPa to verify the trend is not possible, since clamping stress has to be applied to keep the specimen from pulling out of the grips in the current test apparatus. Also, the trend is obscured by experimental scatter in tests involving the smaller (0.4 mm) contact radius. Another feature to note in comparing Figures 2 and 3 is an apparent effect of thickness. The 4 mm thick specimens produced much lower fatigue limit stresses than 1 mm thick specimens under similar contact conditions, emphasizing the need for investigators to consider test specimen geometry issues closely in the development of life prediction models for service components.
RkJQdWJsaXNoZXIy MjM0NDE=