417 



_T 



PROP D 



J I I I I I I I I I I I L 



0,01 



02 

 ilOKQo 



O 0,2 4 6 0,8 1.0 12 1-1 



ADVANCE COEFFICIENT J 



FIGURE 25. Effect of leading-edge roughness on torque 

 and thrust coefficients. 



surface, as studied, e.g., by Holl (1965) and 

 Benson (1966) . Application of their results is 

 also difficult, because the ratio between grainsize 

 and boundary layer thickness without roughness, 

 which is required for the calculations, varies 

 rapidly in this region. The boundary layer thickness 

 on the smooth blades near the end of the roughness 

 was about 30 ym in all conditions , when no separation 

 took place. At the position of laminar separation 

 the boundary layer thickness was only a few ym. 

 Thus , the ratio of grainsize to boundary layer 

 thickness easily varies by a factor of ten. Appli- 

 cation of inception calculations on distributed 

 roughness [e.g., Arndt and Ippen (1958)] seems 

 more appropriate, but this is difficult, because a 

 friction coefficient is required for the calculations, 

 as well as an "equivalent sandroughness" . Both 

 are strongly interrelated [Bohn (1972) ] and espe- 

 cially near the leading edge these quantities are 

 difficult to estimate. 



The roughness elements do form a massive distur- 

 bance of the boundary layer and an increase in the 



torque coefficients is given in Figure 25. These 

 measurements were carried out with a special dyna- 

 mometer inside the propeller hub to assure that the 

 differences were not insignificant due to inaccuracy 

 of the measurements. The accuracy in Figure 25 is 

 still only about ± 0.005. 



Using Lindgren (1972), the value of AKq between 

 fully turbulent and fully laminar boundary layer 

 flow on the propeller is 0.0035. The actual 

 influence of the roughness at the leading edge is 

 smaller, so that we can conclude that the resistance 

 due to the carborundum was very small. An analysis 

 of the effect of roughness at the leading edge on 

 the performance of the propeller is beyond the 

 scope of this paper. 



The effect of leading edge roughness on cavitation 

 is sketched in Figure 18. The radial extent of the 

 cavitation is increased in those cases where the 

 critical radius was a limit for cavitation. The 

 risk of scale effects on cavitation inception due 

 to laminar boundary layer flow is largest at low 

 propeller loadings, when the risk of laminar 

 separation is smallest. But it still can be 

 considerable at high loadings , as is shown in Figure 

 26, where propeller A is shown with and without 

 roughness at 60^ slip. 



Application of roughness at the leading edge is 

 expected to cause two problems. First the geometry 

 of the leading edge may be altered, having a pro- 

 found influence on the minimum pressure peak. 

 Secondly, the local inception index may be changed 

 due to roughness. The effect on the shape of the 

 leading edge can only be minimized by using small 

 grain sizes. However, to obtain a turbulent boundary 

 layer the current 30 \im grainsize was about the 

 minimum and no differences in cavitation behavior 

 were observed between blades roughened with 30 ym 

 and 60 ym roughness. The effect of surface irreg- 

 ularities on cavitation inception can be large, as 

 was shown by Holl (1965). Moreover, Holl points 

 out that "the most disastrous place to locate 

 surface roughness is at the point of minimum 

 pressure of a parent body". This is exactly what 

 cannot be avoided at the rather sharp leading edge 

 of thin propeller sections. The situation very 

 close to the leading edge, however, is different 

 from the situation of an isolated roughness at a 



60|J.m CARBORUNDUM 

 = N -' 



Re„ = 0.73 xlO° 



"NT 



= 1.5 



FIGURE 26. Effect of leading edge roughness on cavi- 

 tation. Propeller A at 50% slip. 



