Multi-Fluid Computational Fluid Dynamic Predictions of Turbulent Bubbly Flows Using an Elliptic-Blending Reynolds Stress Turbulence Closure

• Main Authors: Marco Colombo, Michael Fairweather
• Format: Article
• Language: English
• Published: Frontiers Media S.A. 2020-03-01
• Series: Frontiers in Energy Research
• Subjects: bubbly flows; nuclear thermal hydraulics; computational fluid dynamics; multi-fluid model; Reynolds stress model
• Online Access: https://www.frontiersin.org/article/10.3389/fenrg.2020.00044/full

The accurate prediction of bubbly flows is critical to many areas of nuclear reactor thermal hydraulics, mainly, but not only, in relation to the key role bubble behavior plays in boiling flows. Large scale computations of flows with hundreds of thousands of bubbles are possible at a reasonable computational cost using computational fluid dynamic, multi-fluid Eulerian-Eulerian models. The main limitation of these models is the need to entirely model interfacial transfer processes with proper closure relations. Here, the capabilities and advantages provided by a model that includes an elliptic-blending Reynolds stress turbulence closure (EB-RSM), allowing fine resolution of the velocity field in the near-wall region, are tested over a large database. This database includes mostly monodispersed bubbly flows over a wide range of operating conditions and geometrical parameters, including upward and downward pipe flows, large diameter pipes and a square duct. The model shows encouraging accuracy and robustness, with good agreement over most void fraction distributions and accurate prediction of the magnitude and position of the near-wall void fraction peak. The model does not include any wall force, avoiding all the related uncertainties, and the prediction of the void fraction peak relies on the fine resolution of the near-wall pressure gradient induced by the turbulence field. Overall, the EB-RSM allows accurate resolution of the velocity and turbulence field near the wall, and the transition to this and similar turbulence closures is of value in assisting the ongoing quest for thermal hydraulic models that are accurate and of general applicability. Additional modifications to the near-wall modeling approach, which is still based on its single-phase counterpart, may be required to deal with high void fraction conditions and, in the overall model, additional improvements to momentum and, most importantly, bubble-induced turbulence closures are desirable.