XFEM-Calibrated Crack-Growth Surrogates for Variable-Amplitude Fatigue Loading in Steel

Authors

  • Samuel Thompson Walker Department of Mechanical Engineering, Cockrell School of Engineering, The University of Texas at Austin, Austin, Texas, USA Author
  • Dylan A. Torres Department of Mechanical Engineering, College of Engineering, University of Maryland, College Park, College Park, Maryland, USA Author
  • Mia Shah Walker Department of Mechanical Engineering, Cockrell School of Engineering, The University of Texas at Austin, Austin, Texas, USA Author

DOI:

https://doi.org/10.66525/ijictt152

Keywords:

Fatigue Crack Growth, Variable-Amplitude Loading, Surrogate Modeling, Machine Learning

Abstract

The accurate prediction of fatigue crack growth under variable-amplitude loading remains a critical challenge in the structural integrity assessment of steel components. Traditional numerical methods, such as the eXtended Finite Element Method, offer high-fidelity approximations of crack propagation trajectories and stress intensity factors by enriching the standard finite element approximation space. However, the computational cost associated with resolving cycle-by-cycle crack increments under complex loading spectra renders the eXtended Finite Element Method impractical for real-time monitoring, digital twin applications, or extensive probabilistic assessments. To address this computational bottleneck, this paper presents a comprehensive framework for developing computationally efficient surrogate models calibrated using high-fidelity eXtended Finite Element Method datasets. By generating extensive synthetic databases of crack growth under randomized and structured variable-amplitude loading scenarios, machine learning techniques, specifically deep feedforward neural networks, are trained to map loading histories and current crack states to future crack growth increments. The surrogate model explicitly accounts for load sequence effects, such as crack growth retardation following tensile overloads, which are critical in variable-amplitude fatigue. Extensive validation against both independent numerical simulations and experimental data on standard steel specimens demonstrates that the proposed surrogate model achieves an accuracy comparable to the eXtended Finite Element Method while reducing computational time by several orders of magnitude. The findings suggest that eXtended Finite Element Method-calibrated surrogate models can seamlessly integrate into structural health monitoring frameworks, enabling highly accurate, real-time remaining useful life estimations for aging steel infrastructure.

Downloads

Download data is not yet available.

References

1. Wang, X., Ma, X., Zhu, P., Ng, W. S., Zhang, H., Xia, X., ... & Au, K. W. S. (2024). Design, optimization, and experimental validation of a handheld nonconstant-curvature hybrid-structure robotic instrument for maxillary sinus surgery. IEEE/ASME Transactions on Mechatronics, 29(4), 3074-3082.

2. Jin, X., Qiao, K., Bu, M., Wang, J., Wang, M., & Fang, C. (2025). Intelligent safety evaluation of tunnel lining cracks based on machine learning. Engineering Failure Analysis, 167, 109082.

3. Wu, B., Sun, F., Wang, L., Chen, M., Lu, Y., & Jiang, D. (2023). Characterization of mechanical equivalent properties for node enhanced graded lattice structure. Modelling and Simulation in Materials Science and Engineering, 31(6), 065016.

4. Silva, R. V., de Brito, J., & Dhir, R. K. (2015). Tensile strength behaviour of recycled aggregate concrete. Construction and Building Materials, 83, 108–118. https://doi.org/10.1016/j.conbuildmat.2015.03.034

5. Wang, Y. Y., Lou, M., Wang, Y., Wu, W. G., & Yang, F. (2022). Stochastic failure analysis of reinforced thermoplastic pipes under axial loading and internal pressure. China Ocean Engineering, 36(4), 614-628.

6. Guo, F., Zhang, R., Zhou, H., Zhang, J., & Mo, H. (2026). Dynamic temporal scale decoupling and variable interaction modeling for photovoltaic power forecasting. Solar Energy, 318, 115021.

7. Jin, X., Bu, M., Guo, P., & He, M. (2026). Review on energy tunnels thermal performance and thermo-mechanical behavior. Applied Thermal Engineering, 131520.

8. Ouyang, Z., Mayne, P. W., & Agaiby, S. (2024). In situ test-based evaluation of soil effective stress strength properties and stress history. In Databases for Data-Centric Geotechnics (pp. 344-368). CRC Press.

9. Ouyang, Z., & Mayne, P. W. (2021). Variable rate piezocone data evaluated using NTH limit plasticity solution. Geotechnical Testing Journal, 44(1), 174-190.

10. Ouyang, Z., & Mayne, P. W. (2016). New DMT method for evaluating soil unit weight in soft to firm clays. Geotechnical and Geophysical Site Characterisation; Lehane, B., Acosta-Martínez, HE, Kelly, R., Eds, 785-790.

11. Yu A, Pang F, Yuan Y, et al. Simultaneous current and vibration measurement based on interferometric fiber optic sensor[J]. Optics & Laser Technology, 2023, 161: 109223.

12. Yang, Y., & Zhang, S. (2024). Pixelwise calibration method for a telecentric structured light system.Applied Optics,63(10), 2562-2569.

13. Wang, L., Liu, F., Chen, M., Mao, Z., Chen, G., Zhang, Z., & Xiang, Z. (2025). Synergizing machine learning and multiscale shakedown method for shakedown loading capacity evaluation of parameterized lattice structures. Extreme Mechanics Letters, 75, 102297.

14. Malik, B. A., Jie, S., Zeyu, L., Ouyang, Z., & Wang, D. (2025). Experimental investigation of suction installation behavior of caissons in loose saturated sands. Canadian Geotechnical Journal, 62, 1-21.

15. Wang, Y., Lou, M., Liang, W., & Zhang, C. (2023). Numerical and experimental investigation on tensile fatigue performance of reinforced thermoplastic pipes. Ocean Engineering, 287, 115814.

Downloads

Published

2026-09-24