SESAME Newsletter #2 (February 2017)
Wire-wrap Direct Numerical Simulation
Liquid metal fast reactors (LMFRs) are expected to play an important role in the future of nuclear energy. Thermal-hydraulics of a fuel assembly in a LMFR is recognised as a key scientific topic for the development of such reactors. This subject has been a challenge for researchers over several decades.
The fuel pins in each fuel assembly of most liquid metal cooled reactors are separated by wire wrap spacers, which are wires that are helically wrapped around the pin along the pin axis. The wrapped wire introduces additional turbulence and establishes a bulk rotation of the fluid within the assembly. The flow becomes fully turbulent, with Reynolds numbers typically in the range of 40000–65000. A good prediction of the flow and heat transport in such a complex flow configuration is a challenge for the available CFD turbulence models and computational power. Moreover, no extensive validation of these modelling approaches has been performed because a very limited reference database is currently available suitable for validation of pragmatic CFD approaches.
In Work Package 2 of the SESAME project, the Nuclear Research and Consultancy Group (NRG) has performed a high fidelity simulation for an infinite wire wrapped fuel assembly based on the MYRRHA design. As a first step, an extensive effort has been put forward to design this infinite wire wrapped fuel assembly. Accordingly, the high fidelity simulation is performed by using the well-known CFD code STAR-CCM+. The computational cost of such a numerical simulation is extremely high. For instance, the aforementioned simulation has ran for ~2.6 million core hours on NRG’s cluster. If you run this simulation on 400 processors of 2.3 GHz, it takes approximately 9 months to complete. Nonetheless, after an enormous amount of man and computational power, the simulation has yielded in creating one of a kind reference database and will serve the nuclear community to validate lower order pragmatic turbulence modelling approaches.