The partial cavity on a 2D foil revisited

The partial cavity on a 2D NACA0015 foil at 6 degrees angle-of-attack is studied numerically. Assuming the fluid to be a continuum of variable density, we solve the RANS equations, complemented with turbulence and cavitation models. Some important details of the mathematical model are pointed out first. We study then carefully what occurs in the numerical simulations in and near the cavity from the inception phase to the stage well before serious unsteadiness (cavity shedding) starts. By making the computations on grids of different densities we get an impression of numerical uncertainties. This is important for the interpretation and the subsequent comparison with what experimental investigations have learned us about the physics of these almost steady partial cavities on foils. The results show that close to inception a cavity exists while the boundary layer is non-separating. The liquid-vapour interface turns out not to be a material surface, neither at the front end nor at the tail of the cavity. It also appears that the widely accepted re-entrant jet model as conceived from freestreamline theory is not a good description of the flow at the tail. The confrontation of the numerical results with information from experiments indicates that there is agreement and corroboration in several respects, but also intriguing discrepancies are found which require further elucidation.