| Cavitation plays an important role in the design and operation of fluid machinery and devices because it causes performance degradation, noise, vibration, and erosion. Cavitation involves complex phase-change dynamics, large density ratio between phases, and multiple time scales. Noticeable achievements have been made in employing homogeneous two-phase Navier-Stokes equations for cavitating computations in computational and modeling strategies. However, these issues pose challenges with respect to accuracy, stability, efficiency and robustness because of the complex unsteady interaction associated with cavitation dynamics and turbulence.; The present study focuses on developing and assessing computational modeling techniques to provide better insight into unsteady turbulent cavitation dynamics. The ensemble-averaged Navier-Stokes equations, along with a volume fraction transport equation for cavitation and turbulence closure, are employed. To ensure stability and convergence with good efficiency and accuracy, the pressure-based Pressure Implicit Splitting of Operators (PISO) algorithm is adopted for time-dependent computations. The merits of recent transport equation-based cavitation models are first re-examined. To account for the liquid-vapor mixture compressibility, different numerical approximations of speed-of-sound are further investigated and generalized. To enhance the generality and capability of the recent interfacial dynamics-based cavitation model (IDM), we present an improved approximation for the interfacial velocity via phase transformation. In turbulence modeling, a filter-based model (FBM) derived from the k-epsilon two-equation model, an easily deployable conditional averaging method aimed at improving unsteady simulations, is introduced.; The cavitation and turbulence models are assessed by unsteady simulations in various geometries including square cylinder, convergent-divergent nozzle, Clark-Y hydrofoil, and hollow-jet valve. The FBM reduces eddy viscosity and captures better unsteady features in single-phase flow, and yields stronger time-dependency in cavitating flows, than the original k-epsilon model. Various cavitation models show comparable steadystate pressure distributions but exhibit substantial variations in unsteady computations. The influence of speed-of-sound treatments on the outcome of unsteady cavitating flows is documented. By assessing lift and drag coefficients, pressure and velocity distributions from inception to cloud cavitation regimes, the present approach can predict the major flow features with reasonable agreement to experimental data. |
