401 



behaviour of a propeller is independent of size, 

 provided that no additional parameters play a role 

 in the cavitation process. 



The choice of the cavitation index as a parameter 

 implies the assumption that inception occurs when 

 the local pressure is equal to the vapour pressure. 

 When the inception pressure deviates from the vapour 

 pressure these deviations are called "scale effects 

 on cavitation inception" . 



Two scaling problems do arise now. First it is 

 impossible to maintain the Froude number and the 

 Reynolds number at the same time. The Reynolds 

 number is abandoned and is lowered on model scale 

 by a factor of X^/2, where A is the scale ratio. 

 Even if the Froude number is not maintained it is 

 practically impossible to obtain the full scale 

 Reynolds number on model scale. The second scaling 

 problem is that nuclei play a role in cavitation 

 inception. Both problems manifest themselves as 

 scale effects. 



Pure water can withstand very high tensions and 

 nuclei are necessary to generate inception of 

 cavitation. Nuclei are mostly considered to be gas 

 pockets in the fluid, possibly trapped in small 

 crevices of hydrophobic particles. For a review 

 see Holl (1970) . In a cavitation tunnel, however, 

 the flow will also contain free air bubbles which 

 come out of solution at the pump, at sharp corners, 

 or at the cavitating propeller in the test section. 

 Resorbers are used to bring the free gas back into 

 solution, or the tunnel can be prepressurized. 

 When no large nuclei are present, however, scale 

 effects on cavitation become larger [Hill and 

 Wislicenus (1961)]. Inception of cavitation becomes 

 related to the pressure at which the largest gas 

 bubbles become unstable and start to expand, and 

 this pressure is lower than the vapour pressure 

 when the nuclei are small [Daily and Johnson (1956) ] . 

 In a towing tank there are very few nuclei since 

 they will rise to the surface or to go into solution. 

 Therefore Noordzij (1976) created additional nuclei 

 in the NSMB Depressurized Towing Tank by electrolysis 

 and showed the "stabilizing" influence of nuclei on 

 propeller cavitation behind a ship model. A similar 

 effect was reached by Albrecht and Bjorheden (1975) 

 who injected additional nuclei into the water of 

 their free surface cavitation tunnel after the low 

 pressure in the test section had deaerated the 

 water so much that nuclei were no longer formed in 

 the tunnel. 



It is very difficult to control the nuclei content 

 of the incoming flow [Schiebe (1959)]. When the 

 nuclei are large enough, the inception pressure 

 will be close to the vapour pressure. However, 

 when the nuclei are too large they can lead to 

 "gaseous cavitation" [Holl (1970) ] with inception 

 above the vapour pressure, or they can be removed 

 from the region of lowest pressures by the pressure 

 gradient in the flow, as was theoretically shown by 

 Johnson and Hsieh (1966) . 



Variation of the Reynolds number leads to viscous 

 effects on cavitation inception. Arakeri and Acosta 

 (1973) and Casey (1974) showed the effect of the 

 boundary layer on cavitation inception. Laminar 

 separation was shown to be especially important. 

 Arakeri and Acosta (1973) visualized the boundary 

 layer by a schlieren technique and they tentatively 

 related the cavitation index at inception and the 

 pressure coefficient at laminar separation or at 

 transition. Increased pressure fluctuations in 

 the reattachment region of a laminar separation 



bubble and in the transition region were measured 

 by Arakeri (1975) and by Huang and Hannan (1975) . 

 Van der Meulen (1976) also observed the inception 

 process on headforms by means of holography. He 

 showed that suppression of laminar separation by 

 polymers also could suppress cavitation inception. 

 The relation between the inception pressure and 

 the pressure at laminar separation or transition 

 was not always confirmed. In a recent case study 

 [Kuiper (1978)], it was shown that viscous effects 

 were responsible for a delay in cavitation inception 

 on a propeller model. Additional nuclei had no 

 effect in this case, but it was not yet clear if 

 nuclei did interact with the boundary layer to 

 create cavitation inception. 



In this study, scale effects on cavitation on 

 three propellers with different characteristics 

 were investigated. When a propeller operates in 

 a wake, scaling problems of the incoming flow and 

 of cavitation cannot be separated. Therefore the 

 propellers were tested in uniform axial flow. The 

 tests were carried out mainly in the Depressurized 

 Towing Tank. A description of this facility is 

 given by Kuiper (1974) . The advantages of this 

 tank for the research on scale effects on cavitation 

 inception are the, supposedly, very low and constant 

 turbulence level and nuclei content, the uniform 

 inflow of the propeller, and the absence of wall 

 effects. Both advance speed and propeller revolu- 

 tions can be controlled very accurately. The range 

 of Reynolds numbers which can be tested is lower 

 than in a cavitation tunnel (maximum carriage speed 

 is 4 m/sec.) but is not smaller. 



The aim of the present study is to gain insight 

 into the occurrence of scale effects on cavitating 

 propellers and to develop means to improve the 

 correlation with full scale observations. Paint 

 tests were carried out to visualize the boundary 

 layer flow on the propeller blades. Methods to 

 calculate the pressure distribution on the blades 

 are discussed and the calculated pressure distri- 

 butions are used for the interpretation of the 

 results of the paint tests and the cavitation 

 observations. The nuclei content is varied by 

 using electrolysis, and roughness at the leading 

 edge of the propeller blades is applied to make the 

 boundary layer on the blades turbulent, thus simu- 

 lating a higher Reynolds number. The relation 

 between the boundary layer on the blades and 

 cavitation inception is shown and the effect of 

 leading edge roughness and electrolysis is investi- 

 gated. 



2. TEST PROGRAM 



Propellers and Test Conditions 



Four propellers were investigated in uniform flow. 

 Propeller A is the propeller which was investigated 

 behind a model in a case study by Kuiper (1978) . 

 This propeller showed viscous scale effects on 

 cavitation inception but was insensitive for 

 electrolysis (Figure 1) . Behind the model, this 

 propeller operated in a nozzle. In this study it 

 was tested without a nozzle. 



Propeller B is the propeller which was tested by 

 Noordzij (1976) behind a model. This propeller 

 was very strongly influenced by electrolysis. 

 Without electrolysis the sheet cavitation varied 

 per revolution, (Figure 2) . With electrolysis the 



