Assessment of erosion sensitive areas via compressible simulation of unsteady cavitating flows

The objective of this paper is the assessment of the numerical predictability of erosive events arising in cavitating flows. First, a numerical method and an efficient thermodynamic model for the simulation of cavitating flows are briefly described. The prediction of typical flow details is evaluated by simulating the 3-D flow around a quasi 2-D NACA hydrofoil. We find that the maximum length of the attached cavity, the Strouhal number, and the average diameter of detached clouds are essentially grid independent. Scale enrichment and enhanced 3-D flow details are observed on refined grids. Even delicate flow features, such as cavitating vortices and irregular 3-D break-up patterns, are reproduced, provided that the spatial resolution is sufficiently high. The simulation of cloud collapses and resulting instantaneous peak pressures is assessed in a second investigation. Here, we analyze the effect of the computational grid resolution with respect to typical collapse characteristics, such as the collapse duration, and the instantaneous maximum pressure within the flow field and at walls. The proposed methodology is confirmed by a third investigation, where an experimental setup to investigate cavitation erosion is simulated, and regions of experimentally observed cavitation damage are compared with numerical predictions of strong collapses. The excellent agreement of numerically predicted collapse positions and experimentally observed damage justifies the proposed methodology.