PSI - Issue 79

Available online at www.sciencedirect.com

ScienceDirect

Procedia Structural Integrity 79 (2026) 298–305

© 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 IGF28 - MedFract3 organizers Keywords: machine learning; neural networks; physics-informed neural networks; fatigue life; error estimate; safety factors; model uncertainty A physics-informed neural network (PINN) model was trained on experimental data with physical principles modelled mathematically, such that physically feasible predictions are made possible outside of the training data. By utilizing these functions, the PINN model mitigates the influence of limiting experimental data. Predictions of the elastic and plastic deformation from the material testing gives an underlying understanding of the material behaviour for fatigue evaluation. Based on the probability intervals from the fatigue model, a variable safety factor can be estimated, which allowed for a smaller safety factor compared to conservative conventional methods. 28th International Conference on Fracture and Structural Integrity - 3rd Mediterranean Conference on Fracture and Structural Integrity Physics Informed Modelling of Fatigue Safety Factors Henrik Petersson a , Daniel Leidermark a , Mattias Tiger b , Robert Eriksson a, ∗ a Division of Solid Mechanics, Linko¨ping University, SE-58183 Linko¨ping, Sweden b Artificial Intelligence and Integrated Computer Systems, Linko¨ping University, SE-58183 Linko¨ping, Sweden Abstract In this work, AI algorithms are utilized to predict fatigue life based on known stress and strain states, representative of a small-data paradigm with limited experimental data. The work also explores variable safety factors of fatigue models, derived from the model uncertainty induced by the available training data. The approach enables adjustment of fatigue safety margins based on model confidence, reducing unnecessary conservatism against failure while maintaining structural reliability.

1. Introduction

The structural integrity of components subjected to cyclic load is critical to ensure operational performance, relia bility and safety. This is mainly accomplished through design processes involving finite element (FE) simulations and traditional fatigue life estimation models. These models rely on material parameters quantified from costly and time consuming experimental fatigue test campaigns. Insu ffi cient knowledge of fatigue behaviour often leads to the use of large safety factors, with the consequence of strongly conservative designs. This results in increased costs and loss in performance and e ffi ciency.

∗ Robert Eriksson. Tel.: + 46 13-281139. E-mail address: robert.eriksson@liu.se

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 IGF28 - MedFract3 organizers 10.1016/j.prostr.2025.12.337