Hydrodynamic Cavitation (HC) is a highly turbulent, unsteady, multi-phase flow that has been useful in many processing applications like wastewater treatment and process intensification and hence needs to be studied in detail. The aim of this study is to investigate the mechanisms driving HC inside a Venturi tube using numerical simulations. The numerical simulations are conducted in the form of both two-dimensional (2D) and three-dimensional (3D) simulations using the Detached Eddy Simulation (DES) model database to simulate the cavitation-turbulence interplay, and the results are validated against high-fidelity experimental data. Initial 2D calculation results show that though URANS models are able to show unsteady cavitation, they are unable to reproduce the correct cavity morphology while the DES models reproduce the cavity morphology accurately. After extending to 3D simulations and the resulting vorticity budget analysis highlight the cavitation-vortex interactions and show the domination of velocity gradients and the growth and shrinking of the fluid element terms over the baroclinic torque for vortex production. Finally, localized scale comparisons are conducted to evaluate the model's ability to simulate the cavitation-turbulence interaction. It is observed that the 3D DES simulations are able to predict accurately the cavitation-turbulence interaction on a localized scale for turbulence properties like Reynolds shear stress and Turbulent Kinetic Energy (TKE), emphasizing the 3D effects of turbulence and their influence on the cavitating flow. However, significant discrepancies continue to exist between the numerical simulations and experiments, near the throat where the numerical simulations predict a thinner cavity. Therefore, this study offers new insights on simulating HC and highlights the bottleneck between turbulence model development and accurate simulations of HC to provide a reference for improving modeling accuracy.

In this work, laser-induced florescent particle image velocimetry was performed to measure simultaneously the liquid and vapor velocity fields at the mid-span of a small-scale Venturi type section to determine the presence of a slip velocity between the phases. Various dynamic behavior and Kelvin–Helmholtz (K-H) instability involved in the cloud cavity shedding regime are discussed at four different cavitation numbers. The velocity, vorticity, and turbulence field information of the two phases are analyzed. The liquid–vapor mixture in a cavitating flow is usually considered a homogeneous medium in currently used computational models, but it is shown in this study that the two phases have very different dynamics. The measurements of the time-averaged velocities highlight the existence of a noteworthy slippage between the liquid and the vapor phases, especially in the upstream part of the cavitation region, where the slippage between the two phases can reach about 50% of the liquid velocity. Using phase-locked average, it is shown that the slip velocity in the upstream region is mainly located at the upper liquid–vapor interface, while the slip velocity in the closure area is near the bottom wall, due to the reentrant jet. These results contradict a primary assumption of the current models, where the medium is usually considered as a homogeneous mixture with a unique velocity field, thus providing a reference for future computational model improvement.

This paper investigates cavitation over a wedge using numerical simulations for cavitation numbers ranging from σ=1.47 to 2.06 with a special emphasis on the influence of cavitation and turbulence models. A cross-platform study is conducted to investigate the influence of the mechanisms driving the cavitating flow using an in-house, compressible code and OpenFOAM. This is followed by a detailed comparison of the cavitation and turbulence models in OpenFOAM. The study focuses initially on the model's ability to simulate the sub-stages of a cavitating cycle across various cavitation numbers followed by their ability to reproduce the cavity dynamics on a localized scale using void fraction values at different stations. The results show that both solvers simulate the cavity dynamics reasonably well with minor differences. Next, a discussion is conducted on the cavitation models and the turbulence models where it is observed that while both Merkle and Schnerr-Sauer can reproduce the cavity dynamics accurately, the standard k-ω Shear Stress Transport model is able to reproduce the unsteady cavity shedding without the need for an empirical correction that is usually needed in Unsteady Reynolds-Averaged Navier–Stokes (URANS) turbulence models to predict sheet to cloud cavitation transition. Finally, it is observed that the URANS models over-predict the void fraction at higher σ numbers while the Delayed Detached Eddy Simulation model can predict the void fraction accurately. Thus, this study reveals how current modeling strategies can predict the cavity dynamics observed in experiments, both on a global and a localized scale.