Internship: Physics-informed machine learning for variable amplitude loading in fatigue crack growth

Marknesse Master Thesis

Fatigue in metallic structures is still considered a major threat to the continuing airworthiness of aircraft. Existing models to predict the fatigue life of aircraft structural components under variable loading conditions have limitations due to a limited understanding of the interaction between load cycles. Current maintenance programmes are therefore conservative to account for the limitations in predictions. The aim of this project is to enhance current physical models for predicting fatigue life under variable operational conditions by exploring the application of Physics-Informed Machine Learning (PIML) within the context of Prognostics and Health Management (PHM).

The assignment will include the following tasks:

• Preliminary assessment of available PIML for variable amplitude (VA) fatigue crack modelling.

• A literature review to analyse the application of the PIML framework within the context of PHM to support aircraft health management.

• Designing and implementing a PIML framework to model VA fatigue crack growth or the use of machine learning to understand load interaction effects by finding relationships between different terms in a physical model for VA fatigue crack growth.

• Validating the framework using multiple real-world datasets and benchmarking its performance against state-of-the-art methods.

Available model and VA fatigue crack growth datasets

NLR has created physical model for VA fatigue crack growth, but it is unclear how VA loading changes the interaction between different terms in the equation that is validated for constant amplitude data.

Relevant fatigue crack growth datasets to this research problem are:

• Internal VA fatigue crack growth rate data for different spectra

• PHM Data Challenge 2019

• ASSIST long crack growth challenge

Additional data can also be obtained through laboratory tests conducted at the NLR facilities, and utilising these resources is encouraged to support the project's objectives.

The final outcome of this assignment will be:

• An PIML-based method and/or workflow that can:

o Predict VA fatigue crack growth in metallic structures

o Support aircraft structural analysis

o Account for interpretability

o Include uncertainty quantification.

• A technical thesis report, following the guidelines of your faculty, describing the approach, results and conclusions of the work.

• Optional: a conference and/or journal publication

Standard duration of a thesis at your faculty.

• You are an MSc in Aerospace or Mechanical Engineering, Physics, Computer Science or a master on a related topic.

• You have experience with programming in Python and packages for machine learning within Python, for instance PyTorch.

• You have completed a Machine Learning course (e.g., DSAIT4005, DSAIT4115, AE2224-II).

• You have completed a course on Fatigue of Structures and Materials or similar.

Royal NLR has been the ambitious research organisation with the will to keep innovating for over 100 years. With that drive, we make the world of transportation safer, more sustainable, more efficient and more effective. We are on the threshold of breakthrough innovations. Plans and ideas start to move when these are fed with the right energy. Over 800 driven professionals work on research and innovation. From aircraft engineers to psychologists and from mathematicians to application experts.

Our colleagues are happy to tell you what it’s like to work at NLR.

The assignment will be managed by the Aerospace Vehicles Integrity & Life Cycle Support (AVIL) department. This department provides operational, inspection and maintenance advice to manufacturers and users of aerospace structures and components of (alternative) propulsion sources. It focuses on many facets of MRO, including predictive maintenance, and carries out materials research and conducts engineering failure analysis.

Send your application, together with your motivation letter and CV to Paul.Stuiver@nlr.nl and we will contact you as soon as possible.

Keith, Brendan, Thomas O’Leary-Roseberry, Benjamin Sanderse, Robert Scheichl, and Bart van Bloemen Waanders. "Scientific machine learning: a symbiosis." Foundations of Data Science 7, no. 1 (2025): i-x.

Hao, Zhongkai, Songming Liu, Yichi Zhang, Chengyang Ying, Yao Feng, Hang Su, and Jun Zhu. "Physics-informed machine learning: A survey on problems, methods and applications." arXiv preprint arXiv:2211.08064 (2022).

Fink, Olga, Vinay Sharma, Ismail Nejjar, Leandro Von Krannichfeldt, Sergei Garmaev, Zepeng Zhang, Amaury Wei et al. "From Physics to Machine Learning and Back: Part I-Learning with Inductive Biases in Prognostics and Health Management." Reliability Engineering & System Safety (2026): 112213.

Thiyagalingam, Jeyan, Mallikarjun Shankar, Geoffrey Fox, and Tony Hey. "Scientific machine learning benchmarks." Nature Reviews Physics 4, no. 6 (2022): 413-420.

Psaros, Apostolos F., Xuhui Meng, Zongren Zou, Ling Guo, and George Em Karniadakis. "Uncertainty quantification in scientific machine learning: Methods, metrics, and comparisons." Journal of Computational Physics 477 (2023): 111902.

Schijve, J. (Ed.). (2009). Fatigue of Structures and Materials. Springer Netherlands. https://doi.org/10.1007/978-1-4020-6808-9

Amsterdam, E., Wiegman, J. W. E., Nawijn, M., & De Hosson, J. Th. M. (2022). The effect of crack length and maximum stress on the fatigue crack growth rates of engineering alloys. International Journal of Fatigue, 161, 106919. https://doi.org/10.1016/j.ijfatigue.2022.106919

Amsterdam, E., Willem E. Wiegman, J., Nawijn, M., & De Hosson, J. Th. M. (2023). On the strain energy release rate and fatigue crack growth rate in metallic alloys. Engineering Fracture Mechanics, 286, 109292. https://doi.org/10.1016/j.engfracmech.2023.109292