419 



leading edge roughness on the flow and on the 

 boundary layer. Holographic methods, as applied by 

 van der Meulen (1976) in studying the effects of 

 polymers can be attractive for these experiments. 



When the effect of roughness at the leading edge 

 is studied three regions on the model propeller can 

 again be distinguished. At radii larger than the 

 critical radius , where inception on the smooth 

 blades takes place due to laminar separation, the 

 cavitation behavior is unaffected by roughness. 

 Cavitation was always present on the roughened 

 blades. It is unknown if the sensitivity to nuclei 

 in the unsteady case increases, as is suspected on 

 the smooth blade. Experiences with several other 

 propellers behind a model indicate that this is 

 not the case and that nuclei have very little effect 

 when roughness is applied. 



In the laminar region, at radii smaller than the 

 critical radius , roughness at the leading edge has 

 its major effect, as described above. In some 

 cases, however, problems appeared in the form of 

 streaky cavitation as shown in Figure 27a. When 

 the pressure on the blade sections was constant , 

 as was the case for propeller A at r/R=0.7 and for 

 propeller C at r/R=0.8, both at 30^ slip (Figure 9) , 

 and when the Reynolds number was low, cavitating 

 streaks were formed behind the roughness elements. 

 In Figure 27b the same blade in the same condition 

 at a higher Reynolds number is shown. Here a smooth 

 cavity is seen. The roughness elements apparently 

 suffer from laminar separation at low Reynolds 

 number and cavitation occurs in the separated regions 

 behind the roughness. The length of the spots is 

 strongly dependent on the cavitation index, as is 

 shown in Figure 28b, where the same situation as 

 in Figure 27a is shown at a somewhat higher cavita- 

 tion index. The spots disappeared and the propeller 

 is near inception. Figure 28 also shows that in- 

 ception of the sheet at the leading edge is not far 

 from the vapor pressure, because the ideal inception 

 radius in this case was 0.78. When roughness was 

 applied, electrolysis had no further effect at 

 radii smaller than the critical radius. 



In the region with shockfree pressure distribution, 

 bubble cavitation was seen to be promoted in some 

 cases by roughness at the leading edge. The influ- 

 ence of roughness, however, was inconsistent again 



SMOOTH 



60 M-m CARBORUNDUM 



°NT 



; 2 72x10^ 

 : 1.0 



FIGURE 29. Effect of leading edge roughness on bubble 

 cavitation. Propeller A at 30% slip in the cavitation 

 tunnel. 



in this region, as it was with electrolysis. When 

 there was cavitation at the leading edge due to 

 the roughness , again bubble cavitation appeared at 

 midchord, as is illustrated in Figure 29, where 

 nuclei generated by cavitation at the leading edge 

 created bubble cavitation at midchord. The cavita- 

 tion index at 0.7R in Figure 29 is 0.18 and the 

 minimum pressure coefficient from Figure 9 is 0.20, 

 so the situation with roughness seems to be the 

 situation without scale effects on cavitation 

 inception. Nuclei in the flow, however, did not 

 create bubble cavitation. 



8. CONCLUSIONS 



The results of the present test program can be 

 summarized as follows: 



1. On the suction side of a model propeller a 

 critical radius can exist outside of which 

 the boundary layer is turbulent from the 

 leading edge. This critical radius is due 

 to laminar separation, as was seen from some 

 observations, from calculations (Figure 14) , 

 and from the Reynolds independency of the 

 critical radius. (Figure 15). 



2. To obtain natural transition near the leading 

 edge on a propeller model , high Reynolds 

 numbers {Rejj>2 . 5x10^) are required. 



3. The critical radius is a limit for the radial 

 extent of sheet cavitation from the leading 

 edge. An increase of nuclei by electrolysis 



is ineffective in the laminar region (Figure 22). 



4. Outside the critical radius, cavitation is not 

 inhibited (the inception pressure was not 

 accurately determined) , but a lack of nuclei 

 at low Reynolds number seems to decrease the 

 frequency of inception (Figure 23) . In the 

 unsteady case the nuclei content of the water 

 is probably important in this region. 



5. Roughness at the leading edge can effectively 

 remove the critical radius , thus simulating a 

 higher Reynolds number. Inception of cavitation 

 at the roughness elements occurs close to the 

 vapor pressure, which is assumed to be also 



the case on the prototype. 



6. When the pressure distribution is very flat 

 and the Reynolds number is low, the roughness 

 elements can induce spots of cavitation. The 

 length of these spots is strongly dependent on 

 the cavitation index and is different from the 

 cavity length at high Reynolds numbers. This 

 is probably due to laminar separation at the 

 roughness elements (Figure 27) . 



7. The inception of bubble cavitation near mid- 

 chord at inner radii is not consistent. There 

 seems to be an interaction between the pressure 

 distribution, the nuclei distribution, and 

 even the boundary layer. When cavitation at 

 the leading edge is present, bubble cavitation 

 occurs near midchord when the pressure is below 

 or near the vapor pressure in that region. 



8. Lifting line and lifting surface calculations 

 can adequately predict the open-water character- 

 istics of a propeller. For the calculation of 

 the pressure distribution, however, lifting 

 surface calculations are necessary. The corre- 

 lation between calculations and the results of 

 paint tests and cavitation observations is good. 



