Evaluation of turbulence models for the numerical prediction of transient cavitation around a hydrofoil

The objective of this paper is to evaluate the predictive capability of three turbulence models for the simulation of unsteady cavitating flows around a 2D Clark-y hydrofoil. Three turbulence models were standard k-ε model, hybrid model of density correction model (DCM) and filter-based model (FBM) and an improved partially-averaged Navier-Stokes model (PANS) based on k-ε model. Using the above-mentioned turbulence models and a homogeneous cavitation model, the unsteady cloud cavitation flows around the hydrofoil were numerically simulated and the time evolutions of cavity shape and lift evolutions over time were obtained. The results with comparison to a tunnel experiment data show that the hybrid model and PANS model can accurately capture unsteady cavity shedding details, fluctuation frequency and amplitude of lift and drag. The k-ε model has a poor agreement with the real experimental visualizations and this is mainly attributed to an over prediction of the turbulent viscosity in the rear part of the cavity, which limits the reentrant jet fully reaching the leading edge. The adverse pressure gradient plays an important role in the progression of the reentrant jet. Both the shock wave generated by the collapse of the cloud cavity and the growth of attached sheet cavity contribute to the increase of adverse pressure gradient.