Machine learning-based predictions of fatigue life and fatigue limit for steels

作     者：steelsLei He ZhiLei Wang Hiroyuki Akebono Atsushi Sugeta Lei He;ZhiLei Wang;Hiroyuki Akebono;Atsushi Sugeta

作者机构：Department of Materials Science and EngineeringNagoya UniversityFuro-choChikusa-kuNagoya464-8601Japan Department of Mechanical Science and EngineeringHiroshima University1-4-1 KagamiyamaHigashi-HiroshimaHiroshima739-8527Japan 

出 版 物：《Journal of Materials Science & Technology》 (材料科学技术（英文版）)

年 卷 期：2021年第90卷第31期

页      面：9-19页

核心收录：

学科分类：08[工学] 080502[工学-材料学] 0805[工学-材料科学与工程（可授工学、理学学位）] 

基　　金：The author is grateful to Prof. Yoshitaka Adachi for supporting Shiny-MIPHA access 

主　　题：Machine learning Fatigue life prediction Inverse analysis Steels 

摘      要：To predict the fatigue life for oblique hyperbola-and bilinear-mode S-N curves of metallic materials with various strengths,a machine-learning approach for direct analysis was ***,to determine the fatigue limit of the utilized materials(AISI 316,AISI 4140 and CA6 NM series)with different S-N curve modes using finite-fatigue life data,a Bayesian optimization-based inverse analysis was *** results indicated that predictions of the fatigue life for the utilized datasets via the random forest(RF)algo rithm for AISI 4140 and CA6 NM,and artificial neural network(ANN)for AISI 316,distribute within 2 factor error lines for most *** the Bayesian optimization-based inverse analysis,the specific explanatory variables corresponding to the optimized maximum fatigue life were treated as the fatigue *** predicted fatigue limits either approximated to or slightly underestimated the experimental results,except for several cases with large *** the inverse analysis to predict the fatigue limit for both S-N curve modes is applicable for current employed ***,the explored maximum fatigue lives via BO corresponding to the predicted fatigue limit were underestimated for AISI 4140 and CA6 NM,and was overestimated for AISI 316 because of effect of shape of S-N *** combining the ANN or RF direct and BO inverse algorithms,whole S-N curves(including the fatigue limit)were evaluated for the S-N curve shapes of the oblique hyperbola and bilinear modes.