Assessment of turbulence models and numerical schemes in CFD for nuclear applications

Computational Fluid Dynamics (CFD) plays a crucial role in nuclear engineering, providing an essential tool for simulating and analyzing fluid flows within nuclear power plants, including the reactor core. Specifically, CFD is used to optimize reactor designs, evaluate fuel cooling performance, and ensure safety by predicting thermal-hydraulic behavior of the coolant. In the context of nuclear applications, the use of Reynolds-Averaged Navier-Stokes (RANS) turbulence models is largely adopted, notwithstanding the complex nature of flow patterns. Many models and numerical discretization schemes can be selected for the analysis, and their choice is not an easy task. The scope of this thesis is to investigate the performances of different RANS turbulence models and numerical schemes in CFD, including Finite Volume Method (FVM) discretization techniques and pressure-velocity coupling algorithms. In order to perform a quantitative assessment of the turbulence models and numerical schemes available in different CFD software and with different wall-treatments, the study is done comparing the results obtained from simple RANS simulations, representing classical flow conditions, with Direct Numerical Simulation (DNS) data form validated databases. The results highlight the best options available, in terms of accuracy, for each case, software and approach.