Application of computational fluid dynamics codes for nuclear reactor design

Abstract

Development of computing power allows more sophisticated numerical simulation of single phase heat and mass transfer and makes two-phase simulation more likely. Direct numerical simulation (DNS) of single phase heat and mass transfer is accepted as a reliable replacement of high fidelity experiments with limited scale of the problem. In this paper, various turbulent models are assessed against the DNS to expand the role of computational fluid dynamics (CFD) in the design and analysis of sodium cooled fast reactor core. Also, the benchmarking of Oak Ridge National Laboratory’s (ORNL) liquid metal fast breed reactor (LMFBR) experiment is performed for further generation of virtual data to validate system code. This series of validating procedures are introduced as an example of CFD applications to design the high burnup sodium fast reactor.