Numerical simulation of cavitating bubble-laden turbulent flows

Many engineering devices and propulsion systems suffer from undesirable effects of cavitation; such as degradation in the efficiency of pumps and turbines, generation of noise and vibration on ship propeller, increased drag and erosion of propeller blade, etc. In spite of decades of research on this problem, detailed study of cavitation physics is still a technical challenge for the current experimental and numerical approaches. Further development of robust and accurate numerical methods for cavitating flow simulation, is the basis of this research.;The main objective of this research is to investigate traveling bubble cavitation in turbulent flows. Variable density Navier-Stokes equations are solved in an unstructured grid finite volume solver. Large Eddy Simulation technique with dynamic Smagorinsky subgrid model is used for turbulence modeling. Bubble cavitation is modeled in a Lagrangian framework with subgrid models for forces acting on bubbles. Bubble size variation is modeled using local flow field hydrodynamic pressure for solving Rayleigh-Plesset equation. An adaptive time stepping method is devised for the efficient solution of this highly non-linear equation. Turbulent flow over an open cavity is reproduced numerically from the experimental work by [Liu and Katz, PoF, 08]. Flow statistics agreed very well to those of experiment. Cavitation inception was studied using two different models: (i) discrete bubble model, and (ii) scalar transport model. Severe cavitation was predicted near on top of the trailing edge for cavitation index of sigma < 0.4, in agreement to experiments. Dynamics of inception from both models are in good agreement with experiment. A parametric study is performed to study effect of initial gas nuclei size and cavitation number. Cavitation inception is found to occur on top of the trailing edge, which is in agreement with experiment. Inception happens in different pressure coefficient (Cp) values expected from the classical inception criterion. A PDF analysis on the bubble distribution shows that the larger bubbles are more susceptible to cavitation. Most of the cavitating bubbles are in the regions of negative Cp.