Thesis title: Computational Fluid Dynamics Simulations of Cavitating Flows and Numerical Modelling of Cavitation Damage
In this thesis, Computational Fluid Dynamics (CFD) simulations are used to numerically model cavitation on different geometries: a NACA-66 (mod) hydrofoil profile, a horizontal axis tidal turbine (HATT) and an impinging nozzle test case. The Singhal (FCM) and the Schnerr and Sauer (SSCM) cavitation models have been selected from the literature, they are both based on the non-compressible bubble dynamics equation of Rayleigh-Plesset. Turbulence assessments involve the use of the Transitional Shear Stress Transport (TSST), the Scale Adaptive Simulation (SAS) and the Shear Stress Transport (SST) k-omega turbulence models to achieve the Reynolds Average Navier Stokes (RANS) equations closure. Based on the theory of Plesset, Chapmann and Lush a new cavitation damage index, capable to map the damage of solid boundaries due to vapour collapse has been proposed. Initially, in Chapter 1 a review of the major works contributing to the state of the art of cavitation and cavitation damage modelling is presented. Chapter 2 describes the physics of cavitation and cavitation induced damage, treating in depth phenomena ranging from bubble dynamics to water jet formations due to single bubble implosions. In Chapter 3, fundamental laws of turbulent and cavitating flows have been presented. The theoretical treatment involves derivation of the Navier-Stokes equations, presentation of the RANS approach used to model the cavitating flows and description of the turbulence models selected to close the RANS system. Then, the main equations ruling the mass transfer mechanisms are mentioned for both the FCM and SSCM, detailing peculiarities of each cavitation model. At the end of the Chapter, the formulation of a numerical framework aimed at defining a new cavitation damage index is explained. Its theoretical assumptions and the implementation are reported as well. In Chapter 4, the standard TSST and SAS turbulence models are used in conjunction with the FCM to detect pressure coefficients on the NACA-66 (mod) profile. Predicted results are compared to the experimental findings of Shen and Dimotakis. In Chapter 5 the Reboud density function is implemented to adjust the eddy viscosity computation with respect to the standard turbulence modelling. New results obtained on the same hydrofoil profile used in Chapter 4 are substantially improved in the cavitation region when the FCM is selected in combination with the SST k-omega turbulence model. The combination of the Reboud corrected turbulence model and the FCM is then exported to simulate cavitation occurrence for four different rotating conditions on a selected HATT. Cavitating simulations are performed at the end of a Mesh Sensitivity Analysis (MSA) aimed at identifying the optimal computational grid in terms of both turbine power and thrust. Final cavitation simulations reveal that the severest risk of cavitation inception occurs outside of the machine operating regime. In the end (Chapter 6), the work focuses on a campaign of simulations dedicated to validate the proposed cavitation damage model on the nozzle test case of Franc. The optimal computational grids are selected performing a MSA in terms of both static pressure and velocity magnitude. Results have been obtained by testing the corrected SST k-omega turbulence model and the cavitation damage model in conjunction with the FCM and the SSCM, with the latter returning the best performance in terms of damage peak location when the finest time step condition is implemented. Conclusions and future developments, exposed in Chapter 7, complete the thesis argumentation reporting the main outcomes of the work and its possible industrial applications.
Keywords: bubble dynamics, cavitation, cavitation damage, CFD, damage index, FCM, SSCM.