Fatigue life evaluation model for various austenitic stainless steels at elevated temperatures via alloy features‐based machine learning approach

Abstract

An alloy features-based and chemical compositions-based machine learning method was used to examine the low cycle fatigue life of austenitic stainless steels at different elevated temperatures employing one model. Furthermore, eight algorithms were used to examine the impact of algorithms on the precision of constructed models. As input, physicochemical features of elements were transformed from chemical compositions. After being conducted by the feature screening process, electronegativity deviation (E2.sd), ionization energy deviation (E6.sd), testing conditions, and tensile strength were chosen as input.The results show that algorithms affect accuracy and the models with the highest accuracy are SVR and ANN for alloy features and chemical compositionsbased method, respectively. Chemical composites-based model demonstrates relatively lower precision than the alloy feature model. Almost all testing data distribute within two-factor band lines predicted by alloying feature-based model. The validation testing results indicate that 83% data plots distribute within two-factor band lines.

A unified creep and fatigue life prediction approach for 316 austenitic stainless steel using machine and deep learning

“…Duan et al 21 proposed a support vector ML model with 94% accuracy to forecast the fatigue life of 316 AusSS based on influencing parameters, including stress intensity factor, strain amplitude, and residual stress. He et al 22 employed several ML techniques to predict the LCF life of AusSS based on alloy features and chemical compositions. The choice of algorithms significantly influenced the accuracy of the models, with support vector regression and artificial neural network identified as the most accurate for the respective approaches.…”

Novel fatigue life prediction approach combined with rain‐flow cycle counting process for random multiaxial non‐proportional loading

“…Fatigue failure frequently occurs in metallic structures and components when subjected to cyclic loading beneath the tensile or yield strength of the materials. As one of the main rupture modes, many articles have been published concerning fatigue strength, fatigue crack growth behavior, fatigue life evaluation, etc 1–8 . In engineering applications, components typically undergo random loading during the service period; thus, studies on fatigue life prediction under axial random loading have attracted attention over the past several decades 9–15 .…”

“…As one of the main rupture modes, many articles have been published concerning fatigue strength, fatigue crack growth behavior, fatigue life evaluation, etc. [1][2][3][4][5][6][7][8] In engineering applications, components typically undergo random loading during the service period; thus, studies on fatigue life prediction under axial random loading have attracted attention over the past several decades. [9][10][11][12][13][14][15] Generally, fatigue life evaluation under random amplitude loading can be achieved using cycle counting methods such as the level crossing method (LCM), 16 three-point method (TPM), 16 and rain-flow method (RFM), 17 by transforming irregular loading waveforms into regular loading patterns and by calculating fatigue damage via a cumulative fatigue damage rule such as the Palmgren-Miner rule (also called liner cumulative fatigue damage rule, LDR) 18,19 based on S-N or ε-N curves.…”

Guest editorial of virtual special issue: Structural integrity in the context of carbon neutrality

“…By considering the adverse effects of oxidation on creep–fatigue endurance, a physically based oxidation damage equation was developed, and its accuracy was proved. He et al 5 conducted a machine learning‐based fatigue life prediction process to evaluate the low cycle fatigue life for a series of stainless steels at different elevated temperatures employing one model. Furthermore, alloy feature‐based and chemical composition‐based approaches were used to compare the accuracy of the constructed models.…”