Development of a numerical methodology for the prediction of cavitation-induced damage in hydrodynamic flow systems

A. Evangelisti*, G. Agati, D. Borello, P. Venturini, P. Capobianchi

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

A novel numerical model for the prediction of cavitation-induced damage is proposed and tested against cavitation erosion formation observed in a previous nozzle test case experiment. Turbulent flow is solved using a modified two-equations Shear Stress Transport k - ω model, while cavitation is modelled using the Singhal and the Schnerr and Sauer models. Firstly, a mesh sensitivity analysis is carried out considering grids having different azimuthal extensions and different values of y + to test the accuracy of the numerical approach. Subsequently, selected meshes are used to predict cavitation damage at the bottom plate of the nozzle. Results highlight that the damage model precision is highly sensitive to both time step size and type of cavitation model. Adopting a time step of the order of magnitude of the characteristic bubble implosion time, used in conjunction with the Schnerr and Sauer model, provides prediction of the damage pattern that aligns well with the extent of the experimental eroded area, confined between the annular limits of 19 mm and 32 mm. The model exhibits the highest accuracy under these conditions, returning a damage peak radial position of about 25 mm, which is in good agreement with the experimental peak radius, being located at around 22 mm.
Original languageEnglish
Article number105302
JournalInternational Journal of Multiphase Flow
Volume191
Early online date7 Jun 2025
DOIs
Publication statusE-pub ahead of print - 7 Jun 2025

Keywords

  • bubble dynamics
  • CFD
  • cavitation damage
  • cavitation

Fingerprint

Dive into the research topics of 'Development of a numerical methodology for the prediction of cavitation-induced damage in hydrodynamic flow systems'. Together they form a unique fingerprint.

Cite this