Interphase Momentum Closure Analysis for Multiphase Computational Fluid Dynamics of Horizontal Bubbly Flow

Abstract

Multiphase Computational Fluid Dynamics (M-CFD) offers a high-fidelity method for analysis of safety and performance in helical coil steam generators (HCSGs). Of the M-CFD frameworks, Eulerian multiphase (EMP) is the industrial standard, its efficacy determined by the ability of implemented closures to represent physical phenomena. Research has overwhelmingly focused on the development and validation of EMP closures for vertical bubbly flow, rather than the near-horizontal orientations of HCSGs. Advancement of horizontal bubbly flow modeling would provide a foundation for modeling HCSGs, as well as bubbly flows more broadly. Evaluation of conventional closures for vertical flow indicates that a similar approach is generally inappropriate for horizontal configurations. Lift and drag-based turbulent dispersion closures control the radial void distribution in vertical flows, both of which are based on slip velocity and bubble shape due to deformation. Analyses demonstrate these are both negligible in horizontal flows, thus lift and drag-based turbulent dispersion play at most a minor role. Analysis also shows that the standard EMP momentum equations are erroneous, and an alternative stress term for the dispersed phase is recommended for inclusion. In the absence of established approaches, a rarely discussed closure, virtual mass stress (VMS), is explored. VMS is affected by a range of factors: virtual mass coefficient, turbulence response coefficient and turbulence anisotropy. The performance of VMS and the influence of these factors, along with the recommended particle-averaged stress term, were investigated by modeling flows in microgravity, horizontal and vertical configurations. Modeling results suggest implementation of VMS has potential to capture void fraction distributions, however it remains strongly influenced by the aforementioned factors. Of these, the turbulence response coefficient yields the largest effects, and cannot be reliably predicted by available models. Further work into the turbulence modeling of both phases is required to realize the potential of VMS and accurately model horizontal bubbly flow.