Numerical simulation of a nuclear spent fuel pool including bubbles dynamics

(2025)

Files

Melot_16412000_2025.pdf
  • Open access
  • Adobe PDF
  • 16.29 MB

Details

Supervisors
Faculty
Degree label
Abstract
The thesis tests a low-cost Reynolds-Averaged Navier-Stokes (RANS) model coupled with Lagrangian bubble tracking in Ansys Fluent to predict post-boiling convection in nuclear spent-fuel pools under loss-of-coolant accident (LOCA). Validation and application cases show that the model reproduces the main circulation modes and the monotonic rise of the global Nusselt number caused by bubble injection at moderate and high Jakob numbers, though it misses some flow structures and slightly underpredicts heat transfer at very low Jakob numbers. Bubble-borne latent heat and momentum increase heat removal, but the RANS framework overestimates both thermal and momentum source terms while still keeping global heat-transfer errors within 5-20 % for the strongest forcing cases. The study concludes that this approach can serve as a practical, cheaper alternative to direct numerical simulation for engineering safety assessments, provided finer meshes, longer averaging windows and broader validation are added in future work.