PSI - Issue 75

Available online at www.sciencedirect.com Available online at www.sciencedirect.com ScienceDirect Structural Integrity Procedia (2025) 000 – 000

www.elsevier.com/locate/procedia

ScienceDirect

Procedia Structural Integrity 75 (2025) 102–110

Fatigue Design 2025 (FatDes 2025) Machine learning model for estimating the stress concentration factor based on 3D scans of notched specimens Kris Hectors a,b, *, Seppe Vanheulenberghe a , Jelle Plets a , Wim De Waele a,b

a Laboratory Soete, Department of Electromechanical, Systems and Metal engineering, Faculty of Engineering and Architecture, Ghent University, Technologiepark 46, 9052 Zwijnaarde, Belgium b FlandersMake@UGent – corelab MIRO

© 2025 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under the responsibility of Dr Fabien Lefebvre with at least 2 reviewers per paper Abstract This study demonstrates the development of machine learning models capable of accurately predicting elastic stress concentration factors in V-notched specimens using geometry data derived from 3D scans. A comprehensive training database was established based on 3D scan data of 175 specimens. The -value of each individual specimen was determined from finite element models based on the as-machined profiles. Rigorous preprocessing and recursive feature elimination proved crucial for optimizing model performance. Among the various models tested, voting regression and gradient boosting regression exhibited the highest predictive accuracy. In conclusion, this research validates the combination of 3D scanning and machine learning as a powerful approach for rapid and precise estimation of stress concentration factors in as-machined components, offering a significant improvement over methods based on idealized geometry. © 2025 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under responsibility of the scientific committee of the Fatigue Design 2025 organizers Keywords: Stress concentration factor, Machine learning, 3D scanning, Finite element analysis, As-machined geometry 1. Introduction Notches act as stress raisers in metal structures and components. The localized increase in stress strongly depends on the geometry of the notch. The most influential geometric factors are the notch opening angle, the notch radius,

* Corresponding author. E-mail address: kris.hectors@ugent.be

2452-3216 © 2025 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under responsibility of the scientific committee of the Fatigue Design 2025 organizers

2452-3216 © 2025 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under the responsibility of Dr Fabien Lefebvre with at least 2 reviewers per paper 10.1016/j.prostr.2025.11.012