Cavitation Mechanisms Related to Erosion Studies on Kaplan Turbines, Foils, and Propellers

Doktorsavhandling, 2010

The overall aim of this thesis work is to study cavitation behaviours related to erosion. The focus is on processes of observable scale, i.e. the process prior to the micro scale level at which the actual damage is generated. For these studies, experiments has been carried out with model ship propellers in homogeneous flow, in inhomogeneous flow behind a ship model, hydrofoils, Kaplan model turbines and a prototype Kaplan turbine. By having cavitation behaviours of observable scale related to subsequent erosive processes, one can at an early stage, identify the importance of different global factors for the development towards an erosive cavitation process. This information can be useful knowledge at design, analysis of CFD simulations and at experimental assessment of cavitation. Experimental cavitation assessment normally involves a judging of the risk for erosion as regards to the corresponding full-scale device as well as a statement of possible underlying physical causes, which can be a difficult task. It is proposed that by using a suitable selection of several different tools for the cavitation assessment, the reliability of the analysis can be increased. Suggested tools are, visual observations (standard video, highspeed video), some kind of erosion test (paint test or similar) combined with the approach for cavitation analysis that is presented in the EROCAV Handbook, which has evolved during the evaluation of our experimental data. Principles for plotting of cavity collapses recorded by high-speed video for cavitation assessment are discussed and guidelines for plotting and analysis of cavity collapses are formulated. Important supporting facts for cavitation assessments are also the several different mechanisms that have been identified in the various studied applications. It is concluded that cavitation noise measurements alone is not good enough for judging the cavitation aggressiveness without further calibration. However, if noise recordings are complemented by for example visual observations, this limitation can partly be overcome by the information gained from the visual observations.