Experimental evaluation of biaxial out-of-phase fatigue life failure criteria

Abstract

Complex multiaxial fatigue loading conditions often arise in mechanical components like crankshafts, propeller blades and gas turbines. The loading conditions are often out-of-phase, and the out-of-phase fatigue limit of the material is also unknown. Designing a component for fatigue resistance under these loading conditions using conventional in-phase fatigue failure criteria is considered less appropriate and requires cautious and conservative assumptions. In the existing literature, several multiaxial fatigue criteria are proposed and reviewed. The criteria are evaluated using fatigue data of metals under complex multiaxial loading. However, the available data sets are argued to be old and insufficient, which makes the data sets questionable. The criteria, with reported differences in fatigue limit estimation and material applicability, are consequently also questioned based on the reliability of the data sets with which the criteria are validated. This research study aimed to generate new fatigue data for a single ferrous metal under uniaxial and out-of-phase biaxial loading conditions. The required experimental verification measurements and considerations were made for these data sets. Supplemented and compared with the available literature data, these material and fatigue strength data were used to evaluate three critical plane fatigue criteria. It was shown from experimental testing for 90 [°] out-of-phase bending-torsion loading with ! = 0.5 [-] on EN19 (T-condition) that the dominant Von Mises stress component (bending) seemingly determined the material's fatigue failure instead of the combination of the components. It was also observed for the 90 [°] phase difference that the biaxial stress components can be considered independent of each other. The experimental uncertainties were shown to be sufficiently low. A comparison between the literature and experimental investigation data showed that the validity of the multiaxial fatigue literature data set is questionable. It was shown that all three fatigue criteria made conservative predictions of the experimental data, and one of them is considered more appropriate for the tested material and loading. However, for the literature data set, it was shown that only one of the criteria made a comparative prediction, whilst the others overestimated the fatigue limit. The criteria validity was, therefore, also questionable. Consequently, it is suggested that more experimental test data is required on the same or different loading and material to investigate further the criteria evaluation differences observed.