Review and Assessment of Multiaxial Fatigue Limit Models

The purpose of this paper is to provide a comparison of multiaxial fatigue limit models and their correlation to experimental data. This paper investigates equivalent stress, critical plane and invariant-based multiaxial fatigue models. Several methods are investigated and compared based on ability to predict multiaxial fatigue limits from data published in literature. The equivalent stress based model developed by Lee, Tjhung and Jordan (LTJ), provides very accurate predictions of the fatigue limit under multiaxial loading due to its ability to account for non-proportional loading. This accuracy comes from the model constant which is calculated based on multiaxial fatigue data. This is the only model investigated that requires multiaxial fatigue testing to generate the model parameters. All other models rely on uniaxial test results. Of the five stress-based critical plane approaches investigated, the model proposed by Susmel and Lazzarin shows the greatest correlation with the multiaxial fatigue limit data. Overall, the invariant-based prismatic hull method has the best agreement with the data found in literature. Additionally, a new critical plane model has been proposed. This new model provides predictions nearly identical to the prismatic hull method. However, the prismatic hull method has the advantage of being less computationally intense compared to critical plane methods.