Modeling detrimental effects of high surface roughness on the fatigue behavior of additively manufactured Ti-6Al-4V alloys

Abstract

This article presents a fatigue model that can robustly capture the effect of high surface roughness on the fatigue crack initiation of additively manufactured Ti6Al4V (AM Ti64) alloys in the moderate cycle fatigue regime based on nonlinear finite element (FE) analyses and the theory of critical distances. We propose two methods to use surface measurements of AM components to create FE models, viz., surface topography of the component is explicitly represented in the FE models or a simplified geometrical model, which has only a single sinusoidal-shaped notch, is proposed to replace the surface topography in the FE models. The geometrical parameters of the simplified model are derived based on the MOTIF method (ISO 12085) and its parameters are determined by using the leave-one-out cross-validation method. In the moderate cycle fatigue regime, the stress applied to components is below the yield stress, but the high surface roughness acts as a stress raiser and causes local plasticity at surface valleys. Thus, a nonlinear hardening elasto-plasticity model is employed in the FE models to capture the cyclic mechanical behavior of AM Ti64 alloys. Afterward, the fatigue analyses are performed by employing our in-house C++ software within which fatigue indicator parameters (FIPs) are computed by post-processing the finite element results. In the present work, the Smith-Watson-Topper (SWT) parameter is used as an FIP and is computed at every element in the FE mesh, i.e., the local FIP. Subsequently, to account for the stress gradient effect in the proposed fatigue model, the SWT parameter is averaged over a so-called critical area to compute a nonlocal FIP. The physical fidelity of the proposed fatigue model is shown by a good agreement between the predicted fatigue life cycles based on the nonlocal FIP and the corresponding experimental data. Moreover, the predicted fatigue cycles obtained from the simplified geometrical model also yield good accordance to the experimental data, thus we can model the effect of the surface roughness on the fatigue properties of AM Ti64 in an efficient way.