Validating RANS to predict the flow behavior in wire-wrapped fuel assemblies

Abstract

Liquid metal fast reactors (LMFRs) represent one of the most promising technologies in terms of sustainability, safety and reduction of nuclear waste and are expected to play a relevant role in the coming future. In liquid metal fast reactors, wire wraps convoluted helically along the fuel pins are often used as a spacer design, in order to avoid pin-to-pin contact and maintain a constant spacing among the fuel rods, with the result of altering noticeably the flow and heat transport, by introducing additional turbulence and establishing a bulk rotation of the fluid.Three different reference cases have been used for validation of the flow behaviour in engineering simulations, an infinite bundle, a 7-pin bundle, and a 61-pin bundle. In addition, the infinite bundle data also provides the opportunity to validate the thermal behaviour.A high fidelity Direct Numerical Simulation (DNS) database has been generated for a wire-wrapped fuel assembly configuration. This database is used for validation of engineering simulation approaches based on Reynolds Averaged Navier Stokes (RANS) modelling. Periodicity is conveniently applied thanks to the opportune definition of the computational domain. The dimensions are selected such that they are representative for the MYRRHA reactor fuel assembly design. Extensive analyses of flow and thermal fields, which in a DNS database are available simultaneously, while are hard to achieve in liquid metal experiments, have been performed and used in the validation process for RANS calculations. Mesh sensitivity as well as a sensitivity analysis on different turbulence models are performed. Both qualitative and quantitative analyses are presented in terms of velocity, temperature and turbulent kinetic energy profiles. Discussion on the different turbulence models adopted is presented.Simulations are compared with experimental measurements as well. A series of experiments have been performed for a 7-pin wire-wrapped rod bundle at varying flow rates. A matched-index-of-refraction technique was used employing water as a working fluid in the rod bundle. RANS calculations reproduced the same experiments and results are compared with the available measurements.Finally, a 61-pin wire-wrapped rod bundle case is simulated using steadystate RANS. The simulations reproduced isothermal flow in exactly the same geometry as the high-fidelity Large Eddy Simulation (LES) performed at Argonne National Laboratory, USA. Results of the LES are used as a reference data for validation of the RANS calculations. Parallel to this work, the LES results are being compared to matched-index-of-refraction experiments performed at the Texas A&M university. Even though this is outside the scope of the work presented in this paper, it will provide added value to the comparison in the future.