ABSTRACT

In this study, cavitating flow modeling around the Delft hydrofoil by using Computational Fluid Dynamics (CFD) is presented. In this context, 2 different cavitating flow conditions around the 3-D Delft hydrofoil are simulated. Drag and lift forces, cavitation volume on the hydrofoil and cavitation pattern are processed via CFD analysis. The results obtained from the CFD analysis are validated by the cavitation tunnel test results besides the results of various numerical analysis studies obtained from the literature.

In order to model the cavitation accurately with CFD; all properties of cavitating flows such as turbulence, unsteady pressure and velocity fluctuations, two-phase flow, mass transfer from liquid phase to vapor phase, three-dimensionality, viscosity, dynamics of cavitation bubbles and interactions between bubbles should also be included in the solution. In this study, cavitating flow is simulated by using various models for the aforementioned properties by means of rapidly developing computational technology. Three-dimensional, unsteady cavitating flow around the hydrofoil is solved by the Detached Eddy Simulation (DES) technique with the SST Menter k-⍵ turbulence model. Two phase flow is modelled by the Volume of Fluid (VOF) method. Cavitation is modeled by the Schnerr-Sauer cavitation model, which solves the simplified Rayleigh-Plesset bubble equation. In the analysis, simulations are carried out initially using normal meshes. Thus the regions, where high pressure, velocity fluctuations and cavitation occur are determined. Then the mesh is refined in those regions. Eventually, the regions where high pressure and velocity fluctuations and cavitation occur, have better mesh resolution. Also, the mesh density in the all computational domain is increased and the mesh is enhanced to match the DES model. In this way, the computational errors related to the mesh has been minimized. In addition, the analyses are repeated with three systematically refined meshes and three different time steps. Numerical uncertainties of the analysis under simulated flow conditions are calculated by using lift force results obtained from these analysis and it is demonstrated that the study is independent of grid and time.