화학공학소재연구정보센터
International Journal of Multiphase Flow, Vol.86, 35-55, 2016
A mechanistic model of bubble entrainment in turbulent free surface flows
A framework for the development of models of bubble entrainment in free surface turbulent flows is presented. The framework uses mechanistic processes to model the stages involved in the entrainment of bubbles due to interaction of turbulence with a free surface. Entrainment is modeled as a chain of events for a bubble that is formed at the free surface, then pulled into the fluid against buoyancy, and interacts with vortices that break it up into smaller bubbles, or with other bubbles by coalescence. The main entrainment mechanism is modeled as a vortex/free surface interaction process that can entrain bubbles if the vortices located a given distance from the surface are strong enough. This approach overcomes limitations of approaches where the entrainment is determined only by turbulence parameters, which in the case of objects interacting with a free surface entrain bubbles on the boundary layers irrespective of the distance to the free surface. Depending on the computational fluid dynamics approach used to solve the flow, these processes may need different levels of modeling; more resolved approaches like large-eddy simulation with a volume of fluid method of the free surface will require less modeling complexity than a less resolved RANS method, since some of the involved processes of entrainment are directly accounted for. In this paper a standard RANS approach is used with the free surface modeled using a single-phase level set method, and models are presented for each of the relevant processes to produce a complete mechanistic model of turbulent bubble entrainment The model was calibrated and tested for two relevant problems: a 2D + T breaking wave in model scale, and the full scale bubbly flows around the US Navy Research Vessel Athena. For the second case a grid study is carried out to analyze grid convergence performance of the model. Comparisons with experimental data show that the model predicts well location and magnitude of entrainment. (C) 2016 Elsevier Ltd. All rights reserved.