The Critical Heat Flux (CHF) is a physical phenomenon that may cause the deterioration of the heat transfer in the core of nuclear reactors, potentially leading to core damage. Its accurate prediction is therefore a crucial issue in nuclear reactor safety. The proposed internship concerns the prediction of the CHF through the development of machine learning techniques and their uncertainty quantification. It will be realized in the framework of an international OECD/NEA project on the development and application of Artificial Intelligence methodologies for nuclear engineering.

Within this framework, the intern will:

·        Perform a preliminary literature review on artificial intelligence methodologies.

·        Analyze the available experimental CHF database to understand the physical phenomenon and to verify the coherence and exploitability of the experimental data.

·        Test the existing machine-learning algorithms developed at CEA for CHF in a heated circular tube on other geometries, i.e. rod-bundle.

·        Develop a neural network model to predict as accurately as possible the CHF for different geometries, i.e. rod-bundle. Different architectures will be tested to determine the optimal solution.

·        Quantify the uncertainty of the developed neural networks applying different state-of-the-art methodologies such as conformal prediction and techniques developed at CEA, i.e. CIRCE.

·        Present and discuss their work during technical meetings and write a final report.

The candidate is a master student who has a good knowledge of statistics/applied mathematics with an interest in artificial intelligence and neural networks. Ideally, the student has basis of fluid-dynamics/thermal-hydraulics and nuclear reactors. Knowledge of Python is required. Knowledge of one of the major deep learning frameworks (PyTorch, JAX, Tensorflow) and Jupiter is welcomed. A good knowledge of English is a must. French would be a plus.