Cavitation inception on ship propeller models

In model propeller testing it is assumed that cavitation starts when the minimum pressure on the blades is at the vapor pressure. Scaling of cavitation by maintaining Thoma's cavitation index is based on this assumption. However, to start cavitation (cavitation inception), nuclei are required to break the strong bond between the water molecules. The nuclei contents of the water can, therefore, be a factor in cavitation inception. Also the character of the boundary layer on the propeller blades may influence cavitation inception. The purpose of this investigation is to determine the inception conditions on model propellers and to devise test techniques to improve the prediction of cavitation on the prototype.
Three propeller models were designed and made to serve this purpose. Each propeller model exhibited a specific type of cavitation, viz. bubble, sheet and tip-vortex cavitation. The experiments were carried out in the Depressurized Towing Tank and in the large Cavitation Tunnel of the Netherlands Ship Model Basin in order to compare cavitation inception in both facilities. The experiments were carried out in uniform axial inflow.
A strongly simplified model of a nucleus is a spherical gas bubble in static equilibrium. This approach is described in section 1 and arguments are given why dynamic effects and gas diffusion can be ignored in cavitation inception on propeller models.
In section 2 the calculation of the pressure distribution on propellers in an undisturbed flow is described. The sensitivity of the calculation results to various assumptions made in the calculation method, is investigated.
The boundary layer on the blades of propeller models was investigated by paint tests. In section 2 regions with various types of boundary layer flow are given. Laminar flow and laminar separation are shown to be very important on model scale and very high Reynolds numbers are required to avoid these phenomena. Therefore the leading edge of the propeller blades was roughened with 60 micrometre carborundum to make the boundary layer turbulent. This technique is described in section 4. The influence of the roughness on the blade geometry is within the manufacturing accuracy of the propeller blades. Roughness may, however, cause early inception of cavitation, as is shown by tests with a cylinder in cross-flow.
The nuclei contents of the water was varied by electrolysis. The size of bubbles generated by wires is investigated in section 5.
The cavitation patterns on the propellers with bubble and sheet cavitation are shown in sections 6 and 7 respectively. These patterns are related with the calculated pressure distribution, with the character of the boundary layer as observed with paint tests and with the nuclei content of the water both with and without electrolysis. From the calculated pressure distribution cavitation inception can be predicted, provided that inception takes place at the vapor pressure. For the occurrence of bubble cavitation on model propellers additional nuclei must be generated in the tank as well as in the cavitation tunnel. Sheet cavitation is inhibited if the boundary layer is laminar. An increase of the Reynolds number does not improve this, although surface irregularities become effective at high Reynolds numbers. Application of roughness has a similar effect and may also produce nuclei. This technique may reduce scale effects on cavitation inception.
Tip-vortex cavitation is described in section 8. For the analysis of a cavitating tip vortex an inviscid description is used and a method to determine cavitation inception by measuring the radius of the cavitating core has been developed.
The results of this investigation may explain some of the scale effects which regularly occur in model testing, as discussed in section 9. This investigation also gives guidance to the application of electrolysis and to the use of roughened leading edges of model propeller blades.
The most important conclusions are summarized in section 10.