~i r 1 1 



O tnception 1 



• desinencej 



O o o 



12 16 



Reynolds Number 



FIGURE 19. Cavitation inception and desinence number 

 as a function of Reynolds number for Teflon hem- 

 ispherical nose. 



19. The type of cavitation observed at inception 

 was spot cavitation. The spots were usually located 

 between the pressure minimum (y = 78°) and the 

 transition of hemisphere and cylinder (y = 90°). 

 The most striking differences between the inception 

 data for both models are: (a) the inception data 

 for the Teflon model are much higher than for the 

 SST model and (b) the Teflon model exhibits a strong 

 cavitation hysteresis [hoII and Treaster (1966) ] 

 whereas the SST model exhibits no hysteresis . Such 

 observations have been reported before by Reed 

 (1969), Gupta (1969), and Van der Meulen (1971). 

 Since the viscous flow behavior of the Teflon model 

 is the same as for the SST model (see Section 3) , 

 the above differences can only be explained by 

 surface effects. Teflon is a porous material and 

 has a high contact angle. Both properties are 

 essential features of the Harvey nucleus [Harvey 

 et al. (1944)]. Hence, the Teflon surface acts as 

 a host for surface nuclei, from which (gaseous) 

 cavitation is initiated. The mechanism most probably 



06 



Cp =0.76 



inception (or desinence) 



-Cp { I rrotationaL flow) 



10 1 4 18 22 



Reynolds Number x10~^ 



FIGURE 20. Comparison of cavitation inception (or 

 desinence) number with pressure coefficient at sepa- 

 ration, Ct3 , and at transition, C-n , for SST hem- 



S T 



ispherical nose. 



443 



involved with inception on the SST hemispherical 

 nose has been described by Arakeri (1973) . Inception 

 takes place in the transition and reattachment 

 region of the separation bubbles, where high pressure 

 fluctuations occur [Arakeri (1975a)]. The nuclei 

 may either originate from the surface (Arakeri) 

 or from the stream where they become trapped in . 

 the strong vortices occurring in the reattachment 

 region. 



When Oj^ (or a^) for the SST hemispherical nose 

 is to be compared with the pressure coefficient, 

 several problems arise. The most obvious pressure 

 to compare Oj^ with would be the pressure coefficient 

 at transition, Cp , since the onset of cavitation 

 takes place at the location of transition. Accord- 

 ing to Arakeri (1973), however, the important 

 pressure coefficient to compare a^ with would be 

 the pressure coefficient at separation, Cp^,. This 

 opinion is probably based on the assumption that 

 the pressure within the separation biibble is con- 

 stant (and thus Cp3 = Cp„) but, according to Van 

 Ingen (1975) , this is a good approximation only 

 at low values of Re. A mean curve of the present 

 inception (and desinence) data is plotted in Figure 

 20. Also plotted are Cp =0.76 and Cp^^, for 

 irrotational flow, derived from Figures 3 and 10 

 (with a^ = 7.5). The real (or viscous) values of 

 Cp are unknown and should be obtained from pressure 

 measurements. It can be estimated that the real 



values of Cp are considerably larger than those 

 for irrotational flow, but still smaller than Cp . 

 Thus it would seem that a^ (or a^) can be correlated 

 with the real value of Cp . In that case, it can 

 be argued that the peak pressure fluctuations , 

 measured by Arakeri (1975a) , are creating the 

 negative pressures necessary to overcome the sta- 

 bilizing pressure in stream nuclei, caused by the 

 surface tension. 



Cavitation inception data for the blunt nose are 

 plotted in Figure 21. Also plotted are inception 

 data with polymer injection. At inception, a 

 region of travelling bubbles was observed. The 

 approximate location of this region was x/D = 0.2 

 - 1.0. In Section 4, a further analysis will be 

 given of the type of cavitation occurring. The 

 inception data show that the a^^-and a^j-values are 

 almost identical and nearly constant {a ^ ^ = 0.46, 

 in the absence of polymers) . When aj_ is to be 

 compared with a suitable pressure coefficient, the 



Reynolds Number x 10"^ 



FIGURE 21. Cavitation inception and desinence number 

 as a function of Reynolds number for blunt nose with 

 and without polymer injection. 



