104 



FIGURE 41. Flow past the hemi- 

 sphere nose with many freestream 

 bubbles showing boundary layer 

 stimulation. 



nuclei with radii less than 100 microns all the dis- 

 tributions are essentially the same whereas for 

 nuclei greater than 100 microns radius the biibble- 

 type inception distributions have many more nuclei 

 than the band type inception distributions. Thus 

 it seems possible that in facilities with many 

 macroscopic freestream gas bubbles, the normally 

 occurring laminar separation on some bodies can be 

 eliminated. The subsequent cavitation index and 

 form of cavitation should then be controlled by the 

 nuclei population. 



If so, the experiments on the NSRDC body at that 

 facility and those tests on the same body in the 



10-" 10'= 10"'" 



NUCLEI RADIUS (m) 



FIGURE 42. Nuclei distributions measured by holography 

 in the LTWT (all microbubbles) and in the HSWT (essen- 

 tially only solid particles) . 



LTWT, when bubble type inception was deliberately 

 promoted, should be very similar. This is, in fact, 

 the case as the inception numbers are more-or-less 

 the same. Beyond that, nuclei distributions are 

 known for the two tests [Peterson (1972) and Fig- 

 ure 42] so that, following the philosophy of Silber- 

 man et al. (1974) , it is possible to estimate the 

 number of "cavitable" nuclei per unit volume for 

 each experimental point. A rough estimate of the 

 number of travelling cavitation events can be easily 

 made if we take Johnson and Hsieh's (1966) "capture" 

 radius of 0.01 body radius to determine the flux 

 of fluid through the cavitating region. These data, 

 calculated and measured events are tabulated in 

 Table 3. Peterson measured the event rate acous- 

 tically and chose one event/sec as the threshold 

 level because of the agreement with a "visual" in- 

 ception estimate. (Only the visual estimate was 

 made in the LTWT . ) 



Observation in the HSWT 



On the whole, the agreement of observations and 

 event rates is satisfactory and it seems clear in 

 this circumstance that viscous effects are not of 

 primary importance and that travelling bubble cavi- 

 tation, the type studied by the St. Anthony Falls 

 group, is the prevalent form. But, on all of the 

 bodies studied we have seen different forms of 

 cavitation occur, when separation was not present, 

 if the number of freestream nuclei is very small, 

 as it is presumably in the California Institute of 

 Technology HSWT and other resorber facilities. Then, 

 even on the Schiebe body we see attached forms of 

 cavitation at inception (see Figure 36) at very low 

 inception indices with only rare occurrences of 

 travelling bubble cavitation [see also Arakeri et al. 

 (1976)]. In these circumstances the fluid and the 

 nuclei that it contains pass through regions of 

 some tension (up to about 1/2 at m in the HSWT) . It 

 is conceivable then that the substantial pressure 

 fluctuations in transition regions [Huang and Hannon 

 (1975) ] can initiate cavitation. This is the ra- 

 tionale for Arakeri 's (1975) inception-transition 

 pressure coefficient correlation. Values of -Cp^j. 



