Search

The large scatter in high-cycle fatigue (HCF) life poses significant challenges to safe and reliable in-service assessment of additively manufactured metal components. Previous investigations have indicated that inherent manufacturing defects are a critical factor affecting the fatigue performance of the components, and the HCF life is significantly influenced by the geometric parameters of the critical defects inducing crack nucleation. Therefore, it is highly important to elucidate the correlation of the HCF life with the geometric parameters of critical defects. This study proposes a new fatigue life prediction model for laser additively manufactured AlSi10Mg alloys by including the combined effects of loading stress and defect geometries (size, location and morphology) in terms of domain knowledge-guided symbolic regression (SR). Domain knowledge is extracted from the semi-empirical Murakami, Z-parameter and X-parameter fatigue life models to establish the variable subtrees. The results show that compared with these semi-empirical models, the domain knowledge integration-based SR model has higher prediction accuracy and generalization ability. Moreover, compared with traditional 'black box' machine learning models, SR excels at balancing prediction accuracy and model interpretability, which provides useful insights into the relationship between fatigue life and defect geometries. This article is part of the theme issue 'Physics-informed machine learning and its structural integrity applications (Part 2)'. [ABSTRACT FROM AUTHOR]

• Fatigue strength and life of 2219 welded joints at 77 K are improved compared to RT. • The maximum pore defects near the weld surfaces are predicted by EVS. • Fatigue strengths at RT and 77 K are predicted using EVS and El-Haddad model. • A fatigue life prediction model at room and low temperatures is proposed. This study investigated the tensile and high-cycle fatigue (HCF) properties of a tungsten inert gas (TIG)-welded 2219 aluminum alloy at room temperature (RT) and 77 K for the first time. Defect-sensitivity-based fatigue strength and life prediction models were proposed to link welding defects with fatigue properties at cryogenic temperatures. The findings revealed that the yield strength, ultimate tensile strength, and fatigue strength of 2219 welded joints at 77 K were significantly improved by 29.7 %, 34.1 %, and 60.4 %, respectively, compared to those at RT. The largest possible pore defects near the weld surface were predicted using a combination of laboratory X-ray computed microtomography and the extreme value statistics (EVS) method. Furthermore, the fatigue strength was predicted by considering the maximum defect size in terms of the El-Haddad-based Kitagawa–Takahashi (K–T) diagram. The predicted fatigue strength was 95.5 MPa with an error of 5.4 % at RT, whereas at 77 K, the predicted value was 144.5 MPa with an error of 10.8 %. Finally, a fatigue life prediction model at both RT and 77 K was proposed based on a modified K–T diagram and Murakami's S - N prediction model. The combination of the fatigue life prediction model and the maximum possible defect size based on EVS can provide slightly conservative predictions of the fatigue life without conducting a large number of HCF tests. [ABSTRACT FROM AUTHOR]

• Construct a neural network with physical constraints in loss function. • Effects of temperature and strain rate on fatigue life are involved. • The effect of the strength of physical constraints on prediction accuracy is discussed. • Prediction accuracy of fatigue life is located within the twice error band. Degraded fatigue performance has been frequently reported for metallic materials at elevated temperatures, which is further deteriorated by reducing the loading frequency due to creep deformation. The evaluation of fatigue life for materials and structures typically relies on a large number of experiments and simulations, which is time-consuming and expensive. To address these challenges, a novel physical informed-based multi-layer nested neural network framework is proposed in this work to predict the low-cycle fatigue life of 316 stainless steels at different temperatures and strain rates. The proposed framework integrates physical constraints into a multi-layer nested neural network. Introducing physical constraints into the neural network can enhance the prediction accuracy of the machine learning approach. All predicted fatigue lives fall within the twice error band, with the majority of them falling within the 1.5 times error band. This approach has the potential to significantly improve computational efficiency and reduce resource consumption in the fatigue life prediction of materials. [ABSTRACT FROM AUTHOR]

• Post-heat treated LPBF 316L steel exhibits a cyclic hardening followed by non-saturated cyclic softening stage. • The cyclic deformation is primarily influenced by the evolution of dislocation density and dislocation patterns. • Anisotropy in fatigue life is attributed to the anisotropic microstructure, defects and yield strength. • Ratchetting strain and ratchetting strain rate are highly dependent on the applied stress levels. This study has investigated the cyclic deformation and ratchetting-fatigue interaction of 316L stainless steel fabricated by laser powder bed fusion (LPBF) and subjected to a 900 °C/2 h post-heat treatment, considering both the vertical and horizontal orientations. Strain- and stress-controlled low cycle fatigue (LCF) tests are conducted to explore the cyclic deformation and the effects of the mean stress and stress amplitude on the fatigue life and ratchetting deformation. A detailed analysis of the cyclic deformation mechanism is conducted through X-ray microtomography and transmission electron microscope (TEM) observations. The results show that post-heat treated LPBF 316L steel exhibits cyclic hardening followed by a non-saturated cyclic softening stage. This behavior is attributed to the evolution of dislocation density and dislocation patterns, and the formation of surface cracks. The ratchetting strain and its rate are sensitive to the mean stress and stress amplitude. The fatigue lives of vertically built specimens are slightly higher under strain-controlled loading conditions, but significantly lower under stress-controlled loading conditions than those of horizontally built ones. Furthermore, ratchetting deformation can promote fatigue damage, resulting in a reduction in fatigue life. [ABSTRACT FROM AUTHOR]

Electron beam melting (EBM) is a promising three‐dimensional printing technology for the fabrication of components with high complexity and freedom of structural design. However, the major limit for its wider industrial applications is to preclude manufacturing the large‐scale part in one piece. It is because the part size is largely limited by the enclosed vacuum chamber. An alternative option is the use of welding to join EBM‐processed subparts together. In this study, fatigue crack growth (FCG) rate and high‐cycle fatigue (HCF) performance were measured for both EBM‐processed Ti–6Al–4V alloy and its welded joints. An X‐parameter model was then proposed to correlate stress amplitude and geometric parameters (location, size, and shape) of the critical defects at the crack origins with fatigue life. Research framework and result from this work are expected to serve a reference for the defect‐based fatigue life evaluation of welded parts individually manufactured using additive manufacturing. [ABSTRACT FROM AUTHOR]

• The corrosive environment accelerates failure at stresses two thirds of in-air fatigue limit. • A conditional probability density function is used for the statistical assessment of fatigue life. • The environmentally induced cracks tend to initiate from localized corrosion pitting. • The theoretical fatigue limit is obtained based on a low probability event. The effects of environment (in air and 3.5 wt% NaCl solution) and artificially-induced surface scratches are investigated on the fatigue properties of railway EA4T (also called 25CrMo4) alloy steel. This steel is found to be markedly sensitive to the chloride-containing environment, with cracking at stresses two thirds of the in-air fatigue limit (+/−326 MPa) under rotating bending (Wöhler) tests. The scratched specimens show even shorter lifetime at stresses below 300 MPa and no fatigue limit is found down to 100 MPa. The fracture surfaces are characterized by the initiation of cracks from multiple sources and intergranular cracking. The fatigue failure of smooth specimens tends to initiate from localized corrosion pitting, while the corrosive environment facilitates crack initiation without pitting for scratched specimens. The fatigue crack growth rate in NaCl solution is slightly higher than that in air. Within the framework of defect-tolerant design, El-Haddad modified Kitagawa-Takahashi (KT) diagrams are constructed to explore the variation in fatigue strength and the defect assessment for both conditions. The modified KT model, which takes the short crack behavior into account, is in better agreement with the experimental results than the original KT diagrams. [ABSTRACT FROM AUTHOR]

• Gradient microstructure and mechanical properties were observed in S38C axles. • Compressive residual stress was rebuilt by a proportional integral iterative method. • Crack arrest happens due to compressive residual stresses to enhance the fatigue life. • The critical non-propagating crack depth at the axle body is 4 mm. The remaining life assessment of damaged railway axles is an important topic primarily because of extending the maintenance interval. However, the fatigue resistance of surface-strengthened materials is extremely challenging due to the gradient-distributed microstructure, tensile properties, and residual stresses across the axle. To elucidate the fatigue cracking behaviors of induction-hardened S38C railway axles, the residual stress field ranging from −400 MPa to zero near the axle surface were rebuilt using a new unit pressure and proportional-integral approach (UPA) instead of complicated heat simulation. The fatigue life of full-scale S38C railway axles was then predicted by employing a real vertical load spectrum. The results show that in the presence of a 4.0-mm-deep crack, the remaining life of a damaged axle with compressive residual stresses is more than three times that of an axle without compressive residual stresses. In addition, the crack front is surrounded by a compressive stress field when the crack depth does not exceed 4.0 mm, which is the probable detection limit for the maximum crack depth extension. The present work is expected to serve a basic reference for the optimization and extension of the inspection intervals for induction-hardened S38C railway axles. [ABSTRACT FROM AUTHOR]

• Based on the critical defects data measured by SEM, a machine learning model is established to predict the fatigue life, and the model is in good agreement with the experimental data. • Use the trained machine learning model to quantitatively analyze the importance of critical defects characteristics on fatigue life and analyze the effect of aspect ratio on fatigue life. • The ML model suggests that circular defects of the same area to be less potent than elongated defects of the same area in terms of the HCF lifetime. Given their preponderance and propensity to initiate fatigue cracks, understanding the effect of processing defects on fatigue life is a significant step towards the wider application of additively manufactured (AM) parts. Here a novel machine learning (ML) based approach has been developed to predict the fatigue life of laser powder bed fused AlSi10Mg alloy. The four most important parameters, treferred to here as the Wu-Withers parameters, were found to be the applied stress and the projected area, location and morphology of the critical defects. It was found that an Extreme Gradient Boosting model was able to predict the fatigue lives with high accuracy with the importance of these characteristics in limiting fatigue life ranked in the order given above. The model was able to predict the very different lives of samples tested parallel and perpendicular to the build direction in terms of these four W-W parameters indicating that microstructure was of minor importance. In particular the large projected area of the defects on the crack plane when testing parallel to the build direction was found to be primarily responsible for the shorter lives observed for this testing orientation. The fatigue lives were adequately predicted by the more general two variable (stress and projected area) Murakami model, and even more closely predicted by an empirical model (using essentially the same four W-W parameters) for which the ML model corroborated the empirical dependences. [Display omitted] [ABSTRACT FROM AUTHOR]