Internship assignment: Physics-Based Modelling of Failure Modes in Hydrogen Fuel Systems to Support Smart Maintenance

Hydrogen fuel systems are pivotal to the future of aviation, offering a pathway to net-zero carbon emissions. Their adoption, however, introduces new maintenance challenges: hydrogen involves harsher operating conditions, stricter safety requirements, and more intricate system behaviour than conventional fuels. Prognostics and Health Management (PHM) can address these challenges by enabling condition-based, predictive and prescriptive maintenance strategies that reduce unplanned downtime, extend component life, and improve safety.

A key building block of PHM is a good understanding of how components degrade and fail. Physics-based models of failure mechanisms capture this understanding explicitly: they describe, from first principles, how a component's condition evolves under given operating conditions. Just as important as building such a model is knowing when to trust it. Every model rests on assumptions and simplifications, and a model is only truly useful when its domain of validity, its limitations, and the confidence in its predictions are clearly understood.

This internship focuses on exactly that: developing a physics-based failure model for a hydrogen fuel system component, and systematically establishing where it works, where it does not, and how much its predictions can be trusted. The resulting model is also intended as a building block for Physics-Informed Machine Learning (PIML), in which physical models are combined with data-driven methods. Embedding the model in a PIML framework is foreseen as a follow-up project; the model developed in this internship must therefore already be shaped to be compatible with such frameworks.

What will you be doing?

The goal of the internship is to develop, calibrate, and validate a physics-based model of a relevant failure mode in a hydrogen fuel system, and to characterise its applicability and limitations. The specific sub-system and failure mode are deliberately not fixed in advance: identifying a suitable and feasible target is part of the assignment. Two requirements apply to this selection: the failure mode and its associated sub-system or component must be relevant to the context of hydrogen fuel systems, and data for validating the model must be available. The latter is critical, as suitable data is quite often lacking. The work includes the following tasks:

•       A literature study on hydrogen fuel systems and their degradation and failure mechanisms, complemented by consultation with experts at NLR, to map candidate sub-systems and failure modes.

•       A structured feasibility assessment of which physics can realistically be modelled within the internship, considering the maturity of the available models, the complexity of the underlying physics, and the availability of data for calibration and validation.

•       A justified selection of one failure mode (and its associated sub-system or component) as the modelling target. In short: list candidates from the literature and expert input, screen them on relevance to hydrogen fuel systems, criticality, maturity of the physical understanding, and availability of validation data, and choose the strongest candidate that satisfies all criteria. If no suitable experimental data exists for an otherwise strong candidate, validation data may instead be generated through high-fidelity simulations or lab tests at NLR facilities.

•       Design and implementation of a physics-based model of the selected failure mode, including a clear motivation of the modelling assumptions, simplifications, and design choices. It is a requirement of the internship that the model is structured for compatibility with PIML frameworks, with clearly defined and documented inputs, outputs, parameters, and governing equations, so that it can be embedded in a PIML approach in a follow-up project.

•       Calibration of the model parameters and validation of the model against experimental data from literature, public datasets, or high-fidelity simulations.

•       A systematic characterisation of the model's domain of validity: the operating conditions and scenarios in which it performs well, and those in which it does not.

•       Where possible, quantification of the uncertainty in the model outputs, so that predictions come with an indication of confidence: low uncertainty signals high confidence, while high uncertainty signals that the prediction should not be trusted.

The emphasis of the assignment is not on obtaining a perfect model, but on obtaining a usable, well-documented model whose strengths and limitations are clearly understood.

Methodology

The internship will follow a systematic model development workflow that is well established in engineering practice:

•       Failure mode identification. Candidate failure modes are identified and prioritised using literature review and established techniques such as Failure Mode, Effects and Criticality Analysis (FMEA/FMECA), supported by expert consultation. Candidates are screened on relevance to hydrogen fuel systems, criticality, and availability of validation data.

•       Model conceptualisation. The governing physics of the selected failure mode are described, and the modelling assumptions, inputs, outputs, and required fidelity are defined.

•       Implementation and verification. The model is implemented (for example in Python) and verified, i.e. checked that the equations are solved correctly.

•       Calibration. Model parameters are estimated from available data using standard parameter estimation techniques.

•       Validation. Model predictions are compared against independent data, following recognised verification and validation (V&V) practice for computational models, to establish how well the model represents reality.

•       Uncertainty and sensitivity analysis. The sensitivity of the outputs to inputs and parameters is analysed, and the uncertainty in the predictions is quantified where feasible.

Result

The final outcomes of this assignment will be:

•       A structured assessment, as part of a literature study, of candidate hydrogen fuel system sub-systems and failure modes, together with a justified selection of the modelling target.

•       A working, documented physics-based model of the selected failure mode, implemented in code in a form compatible with PIML frameworks, calibrated and validated against available data.

•       A clear characterisation of the model's domain of validity and its limitations, including, where feasible, uncertainty estimates on the model outputs.

•       An internship report, following the guidelines of your faculty, describing the approach, results, and conclusions of the work.

Duration of the internship?

In accordance with the internship requirements of your faculty.

What do we expect from you?

•       You are an MSc student in Aerospace Engineering, preferably in the Structures and Materials track, or a closely related programme.

•       You have a solid physics/engineering background that allows you to understand and model physical degradation and failure mechanisms.

•       You have experience with programming, preferably in Python.

•       You are able to work independently, structure an open problem, and clearly document and communicate your choices and results.

What do we offer you?

•       A challenging internship in a high-tech, result-oriented work environment

•       Weekly supervision and availability of the technical staff for support

•       An internship allowance

•       Working in an actual R&D project as part of the team

•       Internship results to be used in current and future projects

About NLR

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.

This assignment will be managed by the Vertical Flight group within the Aerospace Vehicles Vertical Flight and Aeroacoustics (AVVA) department.

Want to apply?

Nice! We look forward to get to know more about you. Apply with your application, together with your (i) Motivation letter and (ii) CV to the NLR vacancy portal and we will contact you as soon as possible. For more information about the assignment contact Lisandro Jimenez (lisandro.jimenez@nlr.nl).

A positive VOG screening is necessary for this position.

Recommended literature

On hydrogen fuel systems:

•       Tiwari, Saurav, Michael J. Pekris, and John J. Doherty. "A review of liquid hydrogen aircraft and propulsion technologies." International Journal of Hydrogen Energy 57 (2024): 1174-1196.

•       Massaro, Maria Chiara, et al. "Potential and technical challenges of on-board hydrogen storage technologies coupled with fuel cell systems for aircraft electrification." Journal of Power Sources 555 (2023): 232397.

On physics-based modelling, validation, and uncertainty:

•       Sun, Danning, Jiangfeng Cheng, and Yupeng Wei. "Physics-informed machine learning for prognostics and health management: foundations, advances and prospects." Digital Engineering (2026): 100126.

•       Sargent, Robert G. "Verification and validation of simulation models." Journal of Simulation 7 (2013): 12-24.

•       Oberkampf, William L., and Christopher J. Roy. Verification and Validation in Scientific Computing. Cambridge University Press, 2010.

•       Vachtsevanos, George, et al. Intelligent Fault Diagnosis and Prognosis for Engineering Systems. Wiley, 2006.