General quantification of fatigue damage with provision for microstructure: A review

Surface integrity influences the material fatigue performance significantly, but most micromechanical models based on crystal plasticity only consider the microstructure of the matrix. In this paper, a crystal plastic finite element modeling method was proposed, which simultaneously considered surface roughness, residual stress, and gradual microstructure. It was verified through multiaxial fatigue experiments of materials processed by different multistage machining processes. The results show that larger axial residual compressive stress and smaller grain size can lead to higher multiaxial fatigue life. The existence of residual compressive stress causes the location of material crack initiation to decrease from the surface layer to the subsurface layer. The proposed model can predict the multiaxial fatigue life of materials within a 50% error band.

Section: Crystal Plasticity Modeling and Parameter Calibrationmentioning

confidence: 99%

A crystal plastic finite element model for the effect of surface integrity on multiaxial fatigue life after multistage machining processes

Liang,

Jiao,

Yan

et al. 2024

“…In fact, the reliable assessment of the fatigue life of parts produced by new manufacturing technologies, such as AM, requires extensive experimental testing campaigns, and, since microstructural defects appear in large quantities, the role of microstructure has become increasingly important. 16 Additionally, the fatigue properties of SLMed parts highly depend on the surface roughness and mainly on the process-induced defects (pores and lack of fusion) commonly located mainly within 250 μm of the surface and tend to decrease to the interior of the component. 17 Hence, high fatigue resistance improvements can be obtained by machining external layers of the parts at a depth greater than the indicated and by removing subsurface defects 17 since this has proven to be an easy and reliable post-treatment process of stainless steels.…”

Section: Introductionmentioning

confidence: 99%

The influence of printing strategies on fatigue crack growth behavior of an additively manufactured AISI 316L stainless steel

Camacho

Martins

Branco

et al. 2023

Selective laser melting (SLM) is an additive manufacturing powder‐bed fusion process that allows producing complex metallic parts with a relative density of up to 99.9%. Nevertheless, the mechanical properties of these components depend on numerous process variables, and there is a lack of systematic studies focused on SLM stainless steels. In the work herein presented, three specific printing strategies were X‐Y patterned to manufacture Compact Tension specimens transversally, longitudinally, or chess‐oriented, without any post‐processing heat treatment. The specimens were made of AISI 316L austenitic stainless steel. Thereafter, fatigue crack growth rate (FCGR) tests were carried out at room temperature, at constant amplitude loading (R = 0.2), to determine the fatigue crack propagation constants of the Paris Law associated with each print strategy. It was possible to conclude that transversal additively manufactured specimens showed the lowest FCGRs, and all results were well below the fatigue design curve given in ASME BPVC, Section XI.

“…For example, some critical plane models can predict the fatigue crack directions, 6,7 and energy-based models include thermodynamical effects. [8][9][10][11] Castelluccio and McDowell, 12 Stopka and McDowell, 13 and Przybyla et al 14 showed that the Fatemi-Socie (FS) parameter could be used as the local driving force for fatigue crack formation and early growth in mesoscale simulations.…”

Section: Introductionmentioning

confidence: 99%

Application of machine learning methods in multiaxial fatigue life prediction

Pałczyński

Skibicki

Pejkowski

et al. 2022

This paper compares the results of multiaxial fatigue life estimation using machine learning methods and classical fatigue models. The fatigue life of PA38-T6 aluminum alloy under uniaxial, proportional, and non-proportional loading, including asynchronous loading, is studied. Machine learning methods are trained only on basic loadings, namely, axial, torsional, and 90 out-of-phase. The results obtained with the machine learning algorithms, dense neural networks, support vector regression (with linear and radial basis functions), decision tree, random forest, and XGBoost algorithms, are comparable to the results of classical models like Ellyin-Goło s, Fatemi-Socie, and Ince-Glinka. The best results are achieved for dense neural networks. In that case, they are often slightly more accurate than those obtained using classical methods.