Subject Ph. D.

1. Context and industrial objectives Interest in fourth-generation reactors was shaken up with the end of the ASTRID program and is regaining strength, partly thanks to the launch of several research and funding programs, both in France and internationally, aimed at developing innovative concepts related to the 4th generation. Fourth-generation reactors are distinguished by safety objectives that are superior to those of pressurized water reactors, thanks to the integration of passive systems. Probabilistic risk assessment methods, usually reserved for the safety demonstration and licensing phases, could play a key role from the design phase and throughout operation. The industrial objectives of the PhD thesis are as follows: - To develop a methodology adapted to 4th generation reactors for: - optimizing the design (justify simplifications, reduce conservatism) - optimizing the operating phase (adjust operating parameters, maintenance policies, process control instructions, etc.). - Capitalize on existing safety studies (Superphénix) in order to identify common points and those to be reassessed in the safety demonstration of a 4th generation reactor. - Continue the industrialization of dynamic methods within Probabilistic Safety Assessment Studies.

2. State of the art, objectives of the PhD thesis and scientific obstacles Boolean static methods (Fault Tree, Event Tree) are used to support the safety demonstration of nuclear reactors. However, the static approach remains very conservative when accident scenarios are characterized by a strong impact of the physics and temporality of events or if the kinetics is slow and the repair of components could be valued. To reduce these conservatisms, dynamic methods have been developed (Wiltbank, 2021). EDF R&D, within the PERICLES Department, has extensive experience in the use and development of these methods (Rychkov et al., 2025), (Massoulier et al., 2023). In recent years, in particular, the choice has shifted towards the use of statecharts and several studies and publications have been carried out. The MPSI (Modeling and Control of Industrial Systems) department of CRAN also has solid experience in the development of dynamic methods and their implementation in systems control and their dependability, e.g. (Babykina et al. 2016) which uses stochastic hybrid automata or (Gouyon et al. 2020) which uses timed automata. Statecharts extend the formalism of automata with mechanisms of hierarchy, orthogonality, and communication between states, allowing for concise modeling of complex systems. These dynamic methods do not model unavailability, but they do model the true behavior of a system when events occur. By modeling the behavior of all systems, dynamic methods would make it possible to: quantify the probability of scenarios occurring, identify the causes of their occurrence, and optimize operations in order to reduce these probabilities. The specificities of 4th generation reactors make them compatible with dynamic modeling, due to the greater thermal inertia and longer grace periods of each component. The PhD thesis aims to explore the use of statecharts to optimize the operation of 4th generation reactors. From a scientific point of view, this aims to address three development points: - Operation: dynamic methods are currently reserved for safety; the thesis aims to apply them to the field of operation - Technological target: no study has yet focused on 4th generation reactors - a first attempt was carried out on ASTRID - Model scale: existing dynamic models only cover the systems involved in the accident studied; the thesis aims to develop a dynamic model that integrates the entire installation. The innovative features of the thesis are: - Extended scope model: unlike existing models which are limited to a scope reduced to the components or systems affected by the accident analyzed, the proposed model will cover all relevant subsystems. - Application to operations: in the nuclear industry, dynamic models have so far been used exclusively for safety; the project transposes them to operational process optimization. The main obstacles are: - Scientific obstacle - Applicability of dynamic methods: Dynamic methods, developed for safety analysis, could prove unsuitable for other types of applications. - Scientific obstacle - Exploration of critical scenarios: It is necessary to generate scenarios that can impact the lifetime of the vessel and the availability of the reactor. - Technical obstacle - Modeling scope: Dynamic modeling must cover a broad scope (number, size, complexity of the system) while remaining calculable. - Technological obstacle - Tool capacity: the itemis CREATE (Statechart) tool was not originally designed for models of this magnitude; it must be verified whether it can handle the required complexity. - Technical obstacle - Data availability and quality: The data needed for a current design is not always accessible. As soon as it is available, it will be possible to use the model already in the design phase to guide design choices.

3. Technical program and detailed tasks of the doctoral thesis To meet the objectives, the steps of the thesis will consist of: - To study the SPX documentation: identify scenarios likely to impact the lifespan of the plant and determine levers for optimizing design and operation - To build a dynamic model for representing the behavior of SPX and the interaction of the different systems - To simulate the model for assessing the occurrence probabilities of each critical scenario and identify the operating parameters to reduce these probabilities - To evaluate the possibility of extending the scope of the model to include safety analysis - To analyze the results - To reproduce the process for a reactor other than SPX with the possibility of extending the application of the model to the design.