435 



0.2 0.4 0.6 08 1.0 1.2 1.4 1.6 



Surface Coordinate over Diameter, s/D 



1.8 



20 



FIGURE 3. Computed pressure co- 

 efficient as a function of surface 

 coordinate over diameter for hemi- 

 spherical nose. Data points ob- 

 tained from measurements by Rouse 

 and McNown (1948) at Re = 2.1 x 

 10^ are included. 



with the theoretical contour over a distance , x/D 

 = 0-1.6, and next changes smoothly into a circular 

 cylinder with a diameter of 9.88 mm. The cross 

 sections of the SST models are shown in Figure 2. 

 For the Teflon hemispherical nose the dimensions 

 are the same as for the SST hemispherical nose. 

 However, the Teflon model was not made of solid 

 Teflon but consisted of a Teflon nose slipped on 

 a SST core. Extreme care has to be exercised in 

 manufacturing models for cavitation studies . An 

 accurate similarity of the model contour is essential, 

 but a smooth surface is even more critical. The 

 drastic effects of surface roughness, in particular 

 isolated irregularities, on cavitation inception 

 have been demonstrated by Holl (1960) and Arndt 

 and Ippen (1968) . The present models were made by 

 Instrumentum TNG in Delft. The models were inspected 

 by an optical comparator (magnification 50x) . For 

 the SST hemispherical nose the maximum deviation 

 from the true contour was within 5 ym, for the 

 Teflon hemispherical nose within 10 \na. For the 

 blunt nose , the maximum deviation for x/D < 0.3 

 was within a few microns and for x/D > 0.3 within 

 10 vim. The mean surface roughness height for the 

 SST models was 0.05 ym; for the Teflon model this 

 value was considerably higher. 



Computations of the pressure coefficient for the 

 hemispherical nose and the blunt nose were made at 

 the National Aerospace Laboratory NLR, The Nether- 

 lands. The velocity potential for irrotatlonal 

 flow along the model contour was computed with the 

 variational finite element method according to 



Labrujere and Van der Vooren (1974) . This method 

 is suitable for axisymmetric flows. The relation 

 between the pressure coefficient, Cp, and the 



velocity potential, 



dS 



is given by 



(1+Cp) 



(1) 



where s is the strearawise distance along the model 

 contour and V^ the free stream velocity. The pres- 

 sure coefficient was computed in the absence of 

 tunnel walls and with tunnel walls. In the latter 

 case , it was necessary to substitute the square 

 cross section with rounded corners by a circular 

 one (diameter 55.44 mm), having the same cross- 

 sectional area. For the hemispherical nose, the 

 results are plotted in Figure 3. Also given are 

 data points obtained from measurements by Rouse 

 and McNown (1948) at a Reynolds number of 2.1 x 10 . 

 The computed Cp-values are claimed to be accurate 

 within 0.1 percent. The Cp-value for irrotatlonal 

 flow in the absence of tunnel walls is 0.7746 at 

 s/D = 0.6825 (Y = 78.2°). With tunnel walls the 

 Cp • -value at the same location is 0.8367. For 

 the blunt nose, the results are plotted in Figure 

 4. The computed Cp-values are accurate within 1 

 percent. The Cp-value for irrotatlonal flow in 

 the absence of tunnel walls is 0.750, which is 

 consistent with the accurate computations by Schiebe 

 (1972). With tunnel walls the Cpjj^j^-value is 0.802. 

 Tabulated values of Cp are presented by Van der 

 Meulen (1976b) and Labrujere (1976). 



