Simplified two-group two-fluid model for three-dimensional two-phase flow Computational Fluid Dynamics for vertical upward flow

Abstract

Listen

Translate article

Recent progress in nuclear thermal-hydraulics simulations has been largely focused on coupling with other computational packages, improved closure models for subcooled boiling and for bubbly flows, and the development of higher-fidelity simulation capabilities (Kulesza et al., 2016). While high-fidelity 3D simulation is important for model validation, scientific understanding, and some design calculations, it can be prohibitively expensive for system design applications or applications involving large geometries. Thus, there is also a need for practical, simplified approaches for those applications. The two-fluid model strikes a balance between detail and computational resources, but requires the accurate specification of several key constitutive models. These include (1) interfacial forces, (2) interfacial area concentration, (3) two-phase turbulence, and (4) wall and bulk boiling and condensation. In many modern CFD packages, uncertainties in the local interfacial area concentration can have strong effects on the ability to predict the other key parameters. This paper demonstrates that the drag force in 3D CFD can be formulated in much the same way as in 1D system analysis codes and that this approach can be used to formulate a model for interfacial area concentration. The method is also applied to two-group approaches to consider the difference in transport properties for different bubble size classes. This approach may open a method to calculate the interfacial forces without the need for interfacial area transport equations. This reduces the number of differential equations and avoids the modeling challenges associated with bubble breakup and coalescence kernels and the need to specify the inlet interfacial area concentration a priori. The new method is expected to decouple the effects of interfacial area uncertainty and calibrated coefficients, and should provide reasonable local bubble diameters for both group-1 and group-2 bubbles. The approaches proposed in this study are applicable to two-phase flow simulations in rather simple geometries such as upward two-phase flow in vertical channels. In view of many applications for upward two-phase flow in vertical channels, including nuclear reactor systems, the proposed methods are considered useful.