Issue 48

L. Reis et alii, Frattura ed Integrità Strutturale, 48 (2019) 318-331; DOI: 10.3221/IGF-ESIS.48.31 319 I NTRODUCTION n general, fatigue occurs in materials that are subjected to cyclic loadings. The consecutive loading and unloading cycles applied to the component, with amplitudes below the material’s yield stress, will damage the component. Cycle by cycle the damage done to the material will accumulate, promoting fatigue crack’s nucleation and growth, degradation of material properties and, ultimately, fracture [1]. Unlike static fracture, that occurs with visible plastic deformation, fatigue fracture will occur without warning, except for the cases where a crack is detected and its growth monitored before fracture. Depending on the role of a component in a mechanical structure, fatigue failure might have huge consequences, which can include life losses and great financial costs. One example, where mechanical fatigue failure is a major cause of incidents, is the aviation industry where, between 1927 and 1984, more than half of the documented incidents were due to fatigue failure [2]. The damage mechanism in uniaxial fatigue is well understood, and fatigue life is reasonably estimated. However, for multiaxial fatigue, which consists in the combination of stresses acting in two or more axes, is yet to be fully understood. Damage mechanisms depending on stress combinations, loading interaction effects, sequence effects, and cyclic plasticity are some of the many variables needed to be considered to assess fatigue damage in multiaxial loading conditions. Several studies were already conducted for the referenced parameters and many others related to multiaxial fatigue behavior [3-7]. Along the last decades, a great effort was made to correctly model the damage mechanism of multiaxial fatigue. However, most of the proposed criteria found in the literature has serious limitations to its usage. Most of the criteria proposed in the literature use parameters based in uniaxial testing data, or parameters that require extensive testing to determine them. Moreover, most of these criteria are evaluated using simple block loadings. However, when applied to complex non- proportional loadings, or random loadings, with variable stress amplitude, these models give poor or inconsistent results. When a new criterion is proposed, it should be validated with the most challenging task, which is to estimate fatigue life with non-proportional random loadings with variable stress amplitude. Many important authors in the multiaxial fatigue area of research have compared experimental results with many of the created models and with new and improved ones by themselves relating to the parameter of study [8-13]. This work aims to validate the Stress Scale Factor model (SSF) recently proposed and developed by Anes et al. [14] under those loading conditions. T HEORETICAL B ACKGROUND SSF Criterion nes et al. [10] proposed a multiaxial fatigue criterion in a form of an equivalent shear stress. This criterion uses a new approach to calculate the stress scale factor. The stress scale factor (SSF) is often used in multiaxial fatigue criterions to convert an axial damage into shear’s damage scale, or conversely. For example, in the von Mises criterion, the shear damage is updated into the axial damage scale, by multiplying the shear stress by a factor of √3. In previous works, the authors found that the loading path, the stress amplitude ratio and the stress level, all have influence in fatigue life. Therefore, using a constant SSF can lead to poor estimations. Using six different loading paths, where five were proportional and one was nonproportional, the authors obtained, through regression of fatigue data, a polynomial function, as shown in Eq. (1). ሺ , ௔ ሻ ൌ ൅ ∙ ௔ ൅ ∙ ௔ ଶ ൅ ∙ ௔ ଷ ൅ ∙ ଷ ൅ ∙ ଷ ൅ ℎ ∙ ସ ൅ ∙ ହ (1) where a, b, c, d, f, g and h are the polynomial fitting constants and are material dependent. Therefore, they need to be estimated with experimental fatigue data from the referenced six different loadings paths and stress levels. The fatigue results will give the S-N trendline for each applied loading path and thought them a corresponding SFF value can be calculated. It is using the different SFF values for each loading path that the polynomial constants of Eq. (1) can be calculated. λ represents the stress amplitude ratio, and it is given by Eq. (2). ൌ tan ିଵ ቀ ఛ ೌ ఙ ೌ ቁ (2) where τ a and σ a are the shear stress amplitude and the axial stress amplitude, respectively. This polynomial function can estimate the SSF for all stress amplitude ratios and stress levels for a given material. I A