Application of large-eddy simulation to pressurized thermal shock: Assessment of the accuracy

Pressurized Thermal Shock (PTS) is identified as one of the safety issues where Computational Fluid Dynamics (CFD) can bring real benefits. The turbulence modeling may impact overall accuracy of the calculated thermal loads on the vessel walls, therefore advanced methods for turbulent flows are required. The feasibility and mesh resolution of LES for single-phase PTS are assessed earlier in a companion paper. The current investigation deals with the accuracy of LES approach with respect to the experiment. Experimental data from the Rossendorf Coolant Mixing (ROCOM) facility is used as a basis for validation. Three test cases with different flow rates are considered. They correspond to a buoyancy-driven, a momentum-driven, and a transitional coolant mixing pattern in the downcomer. Time- and frequency-domain analysis are employed for comparison of the numerical and experimental data.The investigation shows a good qualitative prediction of the bulk flow patterns. The fluctuations are modeled correctly. A conservative estimate of the temperature drop near the wall can be obtained from the numerical results with safety factor of 1.1–1.3. In general, the current LES gives a realistic and reliable description of the considered coolant mixing experiments. The accuracy of the prediction is definitely improved with respect to earlier CFD simulations.

► We compare large-eddy simulation with experiment on the single-phase pressurized thermal shock problem.
► Three test cases are considered, they cover entire range of mixing patterns.
► The accuracy of the flow mixing in the reactor pressure vessel is assessed qualitatively and quantitatively.