Application of large-eddy simulation to pressurized thermal shock problem: A grid resolution study

Life time extension assessment is a very important issue for the nuclear community. A serious threat to the life time extension of a Reactor Pressure Vessel (RPV) is an occurrence of the Pressurized Thermal Shock (PTS) during an Emergency Core Coolant (ECC) injection in a loss-of-coolant accident. Traditional system codes fail to predict the complex three-dimensional flow phenomena resulting from such ECC injection. Improved results have been obtained using Computational Fluid Dynamics (CFD) analysis based on the Reynolds-Averaged Navier-Stokes (RANS) equations. However, it has been also shown that transient RANS approaches are less capable to predict the complex flow features in the downcomer of the RPV. More advanced CFD methods like Large-Eddy Simulation (LES) are required for modeling of these flow phenomena. This paper addresses the feasibility of the application of LES for single-phase PTS. Furthermore, the required grid resolution for such LES analyses is identified by evaluation of the solution on different mesh sizes. A buoyancy-driven PTS experiment has been considered here. This experiment has been performed in the 1:5 linear scale Rossendorf Coolant Mixing Model (ROCOM) facility. In the applied numerical model, the incompressible Navier-Stokes equations are solved in the LES formulation, with an additional transport equation for a scalar, which is responsible for driving the embedded buoyancy term in the momentum equations. Instantaneous mixing characteristics are investigated based on evaluation of the scalar concentration. The analysis presented in this paper indicates that the application of LES is feasible nowadays for single-phase PTS. It is demonstrated that the mixing in the downcomer is quite sensitive to small turbulent disturbances at the ECC inlet, i.e., two simulations performed with slightly different fluctuations at the inlet result in substantially different flow in the downcomer. This complicates the analysis of the data from simulations and suggests that evaluation of the results should be performed in the frequency-amplitude domain instead of the classically employed temporal data analysis.