Mechanistic Modeling of Wall-Fluid Thermal Interactions for Innovative Nuclear Systems
Wednesday 09 December 2015
to 17:15 at
Roman Thiele (Reactor Technology; KTH Royal Institute of Technology)
Next generation nuclear power plants (GEN-IV) will be capable of not only producing energy in a reliable, safe and sustainable way, but they will also be capable of reducing the amount of nuclear waste, which has been accumulated over the lifetime of current-generation nuclear power plants, through transmutation. Due to the use of new and different coolants, existing computational tools need to be tested, further developed and improved in order to thermal-hydraulically design these power plants.
This work covers two different non-unity Prandtl number fluids which are considered as coolants in GEN-IV reactors, liquid lead/lead-bismuth-eutectic and supercritical water. The study investigates different turbulence modeling strategies, such as Large Eddy Simulation (LES) and Reynolds-Averaged Navier-Stokes (RANS) modeling, and their applicability to these proposed coolants. It is shown that RANS turbulence models are partly capable of predicting wall heat transfer in annular flow configurations. However, improvements in these prediction should be possible through the use of advanced turbulence modeling strategies, such as the use of separate thermal turbulence models. A large blind benchmark study of heat transfer in supercritical water showed that the available turbulence modeling strategies are not capable of predicting deteriorated heat transfer in a 7-rod bundle at supercritical pressures. New models which take into account the strong buoyancy forces and the rapid change of the molecular Prandtl number near the wall occurring during the transition of the fluid through the pseudocritical point need to be developed.
One of these strategies to take into account near-wall buoyancy forces is the use of advanced wall functions, which cannot only help in modeling these kind of flows, but also decrease computational time by 1 to 2 orders of magnitude. Different advanced wall function models were implemented in the open-source CFD toolbox OpenFOAM and their performance for different flows in sub- and supercritical conditions were evaluated. Based on those results, the wall function model UMIST-A by Gerasimov is recommended for further investigation and specific modeling tactics are proposed.
Near-wall temperature and velocity behavior is important to and influenced by the wall itself. The thermal inertia of the wall influences the temperature in the fluid. However, a more important issue is how temperature fluctuations at the wall can induce thermal fatigue. With the help of LES thermal mixing in a simplified model of a control rod guide tube was investigated, including the temperature field inside the control rod and guide tube walls. The WALE sub-grid turbulence model made it possible to perform LES computations in this complex geometry, because it automatically adapts to near-wall behavior close to the wall, without the use of ad-hoc functions. The results for critical values, such as the amplitude and frequency of the temperature fluctuations at the wall, obtained from the LES computations are in good agreement with experimental results.
The knowledge gained from the aforementioned investigations is used to optimize the flow path in a small, passively liquid-metal-cooled pool-type GEN IV reactor, which was designed for training and education purposes, with the help of 3D CFD. The computations were carried out on 1/4 of the full geometry, where the small-detail regions of the heat exchangers and the core were modeled using a porous media approach. It was shown that in order to achieve optimal cooling of the core without changing the global geometry a ratio of close to unity of the pressure drop over the core and the heat exchanger needs to be achieved. This is done by designing a bottom plate which channels enough flow through the core without choking the flow in the core. Improved cooling is also achieved by reducing heat losses from the hot leg through the flow shroud to the cold leg by applying thermal barrier coating similar to methods used in gas turbine design.