Développement d’un nouveau modèle d’explosion de vapeur pour le logiciel MC3D

​Linkei WEI a soutenu sa thèse le 17 mars 2023 à Nancy.

Résumé

Steam explosion is one of the most critical and complex phenomena that may occur during severe accident in a Nuclear Power Plant. This explosive phenomenon is due to the fine fragmentation of the molten fuel and the very fast release of the associated energy during the passage of the shock wave. However, the precise mechanisms of the pressurization process are still very uncertain and were the subject of works carried out under the program RSNRICE (2014-2023). These works validated the general principle of MC3D-EXPLO model, based on a principle of direct boiling at the surface of the corium fragments, through a film boiling mechanism. However, it appears that the fragmentation may be achieved without significant dispersion of the fragments in the water, which is contrary to the current MC3D model assumptions. The aim of the thesis is to finalize these works through an in-depth analysis of the current model and supplementary DNS simulations of combined heat transfer and fragmentation to propose an improved modeling of the explosion for MC3D.

Following an extensive bibliography, a detailed analysis of the behavior of the current MC3D model is performed, using a simple and ideal test-case. The detailed mechanisms of the pressure escalation are evidenced. The crucial role of initial and created void (vapor) is highlighted: the initial void may easily damp the trigger shock and avoid the escalation. Some numerical deficiencies are highlighted, and major needs of improvements are outlined.

The Basilisk software was then used to perform DNS simulations to get details of combined fragmentation and heat transfer. Comprehensive analysis about fragmentation regimes is performed and statistics including Sauter Mean Diameter and Probability Density Function of fragment mass and drag coefficient are obtained. The simulations highlight the heat transfer mechanisms, indicating that heat is transferred mostly at the front of the fragmenting drop and transported to the rear in the wake. Results for the same test-case using MC3D are then compared with that of Basilisk, indicating the main needs of modification of MC3D model.

Finally, the main lines of a new model are presented and discussed. Those that can be validated from the DNS results were implemented in the current MC3D version. This concerns the introduction of a new Non-Equilibrium Micro-Interaction (NEMI) mechanism, by improving the model of the entrainment processes and considering the effect of deformation on the heat transfers. The other proposals concerning the behavior of the void may be implemented in the future version (V4) of the code. Indeed, as the DNS simulations are limited to non-boiling conditions, the coalescence of the small bubbles into a large one, remains to be clarified. The analysis suggests improvements of modeling in this direction