Optimizing the estimation of the mass of the nuclear material by advanced statistical methods

In order to comply with safety and security standards for nuclear waste storage and non-proliferation treaties, producers of waste containing uranium or plutonium often need to measure the amount of nuclear materials in their radioactive waste. The radiological characterization of nuclear materials by passive and active neutron measurement is one of the historical research activities of the Nuclear Measurement Laboratory (LMN) of the CEA/IRESNE Institute.

Proportional counters filled with 3He or covered with boron are the reference detectors used for these techniques, which are reference tools for measuring plutonium or uranium. In passive measurement, neutron coincidence makes it possible to discriminate spontaneous fission events associated in particular with 240Pu from neutrons resulting from (a, n) reactions. In active measurement, the active neutron interrogation technique (DDT) provides information on the amount of fissile isotopes inside a waste package.

In order to reduce the sensitivity of neutron measurement techniques to matrix attenuation and contaminant localization effects, one of the objectives of the thesis is to study the coupling of different types of measurements, such as channel-by-channel measurement, emission tomography or high-energy X-ray radiography, within a framework of advanced statistical methods. The thesis also aims to evaluate the contribution of advanced statistical methods, such as regression algorithms, Bayesian approaches (among which the Gaussian process), and neural networks, to reduce the uncertainty associated with the plutonium mass.

Particular attention will be paid to the treatment of heterogeneities in the matrix and the distribution of the radioactive contaminant. The influence of these heterogeneities can be particularly difficult to quantify, requiring not only the use of advanced statistical methods, but also an in-depth experimental study using the SYMETRIC neutron measurement station of the CEA/IRESNE Institute.

The thesis work will be carried out at the CEA site of Cadarache Nuclear Measurement Laboratory, which is a professional laboratory, expert in non-destructive methods of radiological, elementary and physical characterization of objects whether radioactive or not. It is equipped with leading technological platforms, located in the TOTEM facility (neutron and gamma measurements) and the INB Chicade (SYMETRIC platforms for neutron measurement and CINPHONIE for high-energy RX imaging). Finally, the doctoral student will work in a collaborative environment where the different teams interact closely with each other.

Study of fuel assembly creep in fluid-structure interaction

In the context of the energy transition and the low-carbon mix, controlling the performance and safety of the nuclear reactors in the fleet is an imperative that still opens up avenues for research and development with high added value. This is particularly true for the optimization of fuel elements.
Indeed, during its stay in the core of a power reactor, the fuel assembly is subjected to mechanical, thermal and hydraulic constraints. It undergoes a change in its geometry, in particular an elongation and a lateral deformation, due to the creep phenomenon linked jointly to the irradiation and the flow of water in the core. With the increase in the residence times of fuel assemblies in reactors and due to increasingly demanding conditions, the need to understand the phenomenon is necessary to improve the robustness of the design. This is in particular a problem of fluid-structure interaction where the flow plays a role in the creep behavior of the structure and where the deformation of the structure modifies the flow.
A previous study made it possible to implement an experimental device to obtain rapid creep on reduced-scale fuel assembly models. These tests were able to highlight a significant effect of the fluid inlet conditions on the creep behavior of the assemblies. The objective of the proposed thesis work is then to analyze the experimental results using simulation tools in order to understand and quantify the phenomenology of the coupling in fluid-structure interaction under creep. This analysis could lead to the realization of additional tests. Another important aspect will be the transposability of the results to real conditions. The thesis will be carried out at the IRESNE institute of the Cadarache center, in collaboration with the industrialist Framatome, bringing its operational vision in the monitoring and orientation of the research work. The proposed work therefore opens up solid perspectives at the end of the thesis both in research centers and in industrial environments.

Mass transfers and hydrodynamic coupling: experimental investigation and models validation and calibration

During a core meltdown accident, a pool of molten material, known as corium, can form at the bottom of the reactor vessel. The composition of the pool can change over time. The corium bath is not homogeneous and can stratify into several immiscible phases. As the overall composition of the corium changes, so do the properties of the different phases. The vertical stratification order of the phases may change, leading to a vertical rearrangement of the phases. During this rearrangement, one phase passes through the other in the form of drops. The order of the phases and their movements are of prime importance, as they have a major influence on the heat flows transmitted to the tank. A better understanding of these phenomena will enable us to improve the safety and design of both current and future reactors.

Initial models have already been produced, but they lack validation and calibration. Prototype experiments are difficult to set up and none are planned in the short term. This thesis proposes to fill this gap by carrying out an experimental study of the phenomenon using a water-based simulating system that allows local instrumentation and large-scale test campaigns. The aim is to validate and calibrate the existing models, and even develop new ones, with a view to capitalising on these results in the PROCOR software platform, which is used to estimate the probability of a reactor vessel breach. The experimental set-up would be built and operated at the LEMTA laboratory at the University of Lorraine, where the PhD student would be seconded. In terms of experiments, two cases will be studied, the single drop case, and the stratified case with drop formation via Rayleigh-Taylor instabilities.

The work will be mainly experimental, with a component involving the use of code for calibration and validation, and may include a modelling component. It will be carried out entirely at the LEMTA laboratory in Nancy. The PhD student will benefit from LEMTA's expertise in the development of simulating experimental devices, fluid transfers and metrology. They will be part of a dynamic environment made up of researchers and other PhD students. The candidate should have knowledge of transfer phenomena (mass transfer in particular), as well as a definite interest in experimental science.

Local understanding of the corium-concrete interface through experimentation

In the context of the energy transition and the crucial role of nuclear power in a low-carbon energy mix, understanding and then mitigating the consequences of any accident leading to even partial core meltdown is an imperative research direction.
In the event of a severe core meltdown accident, the amalgam of materials produced by core meltdown, or corium, may interact with the concrete of the plant's floor. The lack of understanding of local and interfacial physical phenomena during corium-concrete interaction (ICB) has led to the development of various international simulation tools. None of them has been able to explain the recent observations at the Fukushima Daiichi accident site. It is therefore crucial to improve the ICB simulation tools.
The aim of this thesis is to carry out a detailed, local experimental study of the corium/concrete interface with prototypical corium (depleted uranium). To this end, the candidate will design a test device to be introduced into the VITI inductive furnace of the PLINIUS platform dedicated to the study of severe accidents at the Cadarache center. After qualification of the experimental set-up, local corium/concrete interaction tests in VITI will be carried out on different types of concrete (including a sample from Fukushima) and with different coriums, enabling an incremental approach using separate effects. Ablation will be characterized via mass loss and hydrogen release. The interface will also be characterized after rapid corium removal. Samples will also be X-rayed (e.g. tomography). As the work progresses and the phenomenology of the Molten Corium Concrete Interaction is understood, a model may be developed and integrated into a simulation tool.
The thesis work will be carried out jointly in the experimental and severe accident modeling laboratories of the IRESNE institute at Cadarache, in a research environment of the highest international standard for the study of multiphysical phenomena at very high temperatures. This work will also be enriched by research carried out within the framework of the ANR IMMOC, in partnership with academics (CNRS Laboratoire Navier, AMU-CNRS Madirel...).

fragmentation of a liquid jet against an obstacle : application to sodium spray fires

The sodium spray fires, generated by a leak in a sodium pipe, are of major concern for the safety of the Sodium Fast Reactors. The main consequences of these fires are the temperature and pressure increase within the involved building or containment as well as the production of a chemical source term that could be released into the environment.
The key phenomenon to consider for spray fire studies is the fragmentation of the liquid sodium jet against an obstacle (room wall or ceiling). This mode of fragmentation has to be carefully assessed with a special focus on the droplet population generated at the impact.
To do so, the LESC lab in Cadarache is currently preparing an experimental loop involving a water jet impacting a horizontal or vertical surface.
The Ph-D work will then consist in:
- Performing the impact experiments with varying jet initial diameter, velocity and orientation;
- Analyzing the images from the high-speed camera with an in-house software and extract the droplet size and velocity distributions;
- Suggesting a model that could represent this mode of fragmentation and the resulting droplet population;
- Implementing this model into the CEA canoP CFD software;
- Validating the developed model against the results from the above image analysis.