Next Article in Journal
Experimental and Numerical Investigation of Electromagnetic Wave Propagation Through Conductive Multilayer Coatings
Previous Article in Journal
Sky’s-Eye Perspective: A Multidimensional Review of UAV Applications in Highway Systems
Previous Article in Special Issue
Applications of Computational Mechanics Methods Combined with Machine Learning and Neural Networks: A Systematic Review (2015–2025)
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
This is an early access version, the complete PDF, HTML, and XML versions will be available soon.
Article

Application of FEM Analyses and Neural Networks Approach in Multi-Stage Optimisation of Notched Steel Structures Subjected to Fatigue Loadings

Department of Machine Design and Composite Structures, Faculty of Mechanical Engineering, Cracow University of Technology, ul. Warszawska 24, 31-155 Cracow, Poland
*
Author to whom correspondence should be addressed.
Appl. Sci. 2025, 15(20), 11194; https://doi.org/10.3390/app152011194
Submission received: 15 September 2025 / Revised: 6 October 2025 / Accepted: 16 October 2025 / Published: 19 October 2025

Abstract

The stress concentration, which appears in loaded structural elements with voids, holes or undercuts, is the main source of premature fatigue failure. So, an increase in fatigue life can be achieved by reducing stress concentrations around the notches. Different techniques can be used to reduce the stress concentration. One of them is the application of additional stress relief undercuts or holes, while a second one relies on the application of overlays glued in the vicinity of notches. The proposed study is focused on the optimisation of notched specimens using a multi-stage optimisation process, including the use of artificial neural networks (ANNs). On this basis, the comparison of the effectiveness of various modern finite element optimisation tools is made. Here, special attention is paid to samples with elliptical holes and the application of the ANN technique in determining the optimal solution for the configuration of stress relief holes. The proposed study is illustrated by the example of a steel specimen with an elliptical opening. Specimens without stress relief holes and with an optimal configuration of stress relief holes are subjected to fatigue tests to confirm the effectiveness of the proposed approach. The performed study revealed that the cutting of additional circular stress relief holes reduces the stress concentration around the elliptical opening by about 12% and leads to an increase in fatigue life by about 79% for the applied material. Moreover, the comparison of the possibilities of the reduction in SCF by the application of stress relief holes, composite overlays and the simultaneous application of composite overlays and stress relief holes for the investigated notched samples is performed. Following the numerical results, it is observed that the use of composite overlays additionally decreases the stress concentration factor in relation to specimens with stress relief holes by an additional 6%.
Keywords: optimisation; artificial neural networks; notched structural elements; finite element method; fatigue life; stress concentrations; composite overlays; stress relief holes optimisation; artificial neural networks; notched structural elements; finite element method; fatigue life; stress concentrations; composite overlays; stress relief holes

Share and Cite

MDPI and ACS Style

Romanowicz, P.J.; Szybiński, B.; Barski, M.; Stawiarski, A.; Pałac, M. Application of FEM Analyses and Neural Networks Approach in Multi-Stage Optimisation of Notched Steel Structures Subjected to Fatigue Loadings. Appl. Sci. 2025, 15, 11194. https://doi.org/10.3390/app152011194

AMA Style

Romanowicz PJ, Szybiński B, Barski M, Stawiarski A, Pałac M. Application of FEM Analyses and Neural Networks Approach in Multi-Stage Optimisation of Notched Steel Structures Subjected to Fatigue Loadings. Applied Sciences. 2025; 15(20):11194. https://doi.org/10.3390/app152011194

Chicago/Turabian Style

Romanowicz, Paweł J., Bogdan Szybiński, Marek Barski, Adam Stawiarski, and Mateusz Pałac. 2025. "Application of FEM Analyses and Neural Networks Approach in Multi-Stage Optimisation of Notched Steel Structures Subjected to Fatigue Loadings" Applied Sciences 15, no. 20: 11194. https://doi.org/10.3390/app152011194

APA Style

Romanowicz, P. J., Szybiński, B., Barski, M., Stawiarski, A., & Pałac, M. (2025). Application of FEM Analyses and Neural Networks Approach in Multi-Stage Optimisation of Notched Steel Structures Subjected to Fatigue Loadings. Applied Sciences, 15(20), 11194. https://doi.org/10.3390/app152011194

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop