Evaluation of turbulence modeling approaches for the prediction of cross-flow in a helical tube bundle

Abstract

Helical coil steam generators are being adopted in a number of advanced nuclear reactor designs because of their increased heat transfer efficiency and compactness when compared to traditional configurations. Due to the limited operational experience of helical coil steam generators, and the expected high cost of dedicated experiments, it is imperative to demonstrate the capability of modeling the flow phenomena present in such systems, in order to support their development. Computational fluid dynamics (CFD) is ideally suited for this problem, due to its high resolution and accuracy. However, traditional low-cost URANS turbulence models have shown limited applicability for cross-flow in helical tube bundles. On the other hand, LES methods provide high-accuracy and high-fidelity data with prohibitively high computational cost for design use. Hybrid, and second-generation URANS turbulence models aim to bridge the gap between URANS and LES, by resolving additional turbulence, without the extremely high cost of LES. In this work, the recently proposed second-generation model STRUCT is compared to classic URANS turbulence formulations in terms of pressure drop, turbulent kinetic energy and flow velocity fields generated from a high-fidelity LES simulation in Nek5000. The STRUCT model produces accurate predictions for all relevant quantities, demonstrating high potential for reproducing helical tube bundle flow phenomena accurately, at an affordable computational cost. Further work should extend the validation scope to a wider range of flow velocities and geometrical arrangements, to confirm the demonstrated applicability of the method.