The objectives of this study were to investigate the fatigue behavior of three high hardness steels, where each of the three steels had a hardness of 60 HRC (653 HB). Solid specimens of three different materials were tested under normal environmental laboratory conditions. Axial monotonic and fatigue properties were obtained through the use of a uniaxial testing machine. A large portion of the axial fatigue testing was performed under fully-reversed loading conditions, but a portion of the testing was conducted to evaluate the effects of introducing a mean stress. Fully-reversed pure torsion fatigue data were also generated using an axial-torsion test machine.
The focus of the testing was directed toward longer life tests with several shorter life tests to obtain the stress-life curves in both the low and high cycle regimes. Significant scatter in the experimental data in some instances necessitated a statistical investigation. The fully-reversed axial testing produced both surface and subsurface specimen failures for all three steels, while the pure torsion testing resulted in surface failure on the maximum principal plane.
The axial stress-life curves were predicted by using material hardness, and the predictions were found to be reasonable when compared with the experimental curves for two of the three steels. With the maximum principal stress, maximum principal strain, von-Mises, and Tresca criteria, the hardness predicted stress-life curves were converted into shear stress-life predictions. As a result, the maximum principal strain criterion was found to provide the best prediction. The same criteria were then used to correlate the axial and torsion fully-reversed data for each material, and the maximum principal strain criterion was found to provide the best predictions.
The area0.5 parameter for the axial fatigue limit prediction provided estimated fatigue limits for subsurface inclusion failures. This parameter was found to predict fatigue limits that were lower than the experimental fatigue limit at 107 cycles. The modified Goodman parameter and Gerber parameter predictions were found to exceed the 900 MPa stress amplitude median life. The modified Goodman equation resulted in a better prediction than the Gerber parameter. For the particular R ratio of -0.5 tested, the SWT (Smith-Watson-Topper) parameter was found to produce similar results to the modified Goodman equation.