Numerical Modelling and Prediction of Cavitation Erosion Using Euler-Euler and Multi-Scale Euler-Lagrange Methods
Andreas Peters
Abstract
This thesis presents numerical predictions of cavitation-induced erosion, for hydrodynamic flows, using Euler-Euler and multi-scale Euler-Lagrange methods. Improved approaches to numerically predict hydrodynamic cavitation-induced erosion are developed and applied to internal and external flows that involve stationary and rotating components. Comparisons of numerically predicted erosion with experimentally obtained erosion patterns and qualitative experimental erosion predictions demonstrate the accuracy of the numerical methods. First, an efficient Euler-Euler approach predicts areas exposed to erosion based on information of pressure and vapour content in the flow. Second, a multi-scale Euler-Lagrange method uses calculated collapses of spherical Lagrangian bubbles to identify locations and time instances of cavitation bubble collapses and, thereby, predict erosion impacts. The multi-scale approach enables to correlate calculated bubble collapses with erosion pitting rates from measurements. The developed methods give more accurate predictions of cavitation-induced erosion beyond the state of the art.