103 



'0,7 



a: 



UJ 

 CD 



0,6 



§0,5 



0.4 



p 03 ■ 

 g 



» 



O 50 WPPM 

 • 500 WPPM 



7 8x10"* 



FIGURE 39. The data of figure 38 replotted against the 

 dimensionless injection parameter G. 



of cavitation index. There is a limit, however, 

 beyond which no further increase in cavitation 

 suppression occurs. In the present experiments on 

 the hemisphere nose body this limiting value of G 

 is approximately 7 >: 10~^ which also coincides with 

 the removal of the laminar separation. These 

 results and others are summarized in Figure 40 

 where the maximum percent reduction in cavitation 

 index has been plotted versus the Reynolds number. 

 These include the "polymer ocean" results of Baker 

 et al. (1973) , Holl et al. (1974) , and Ellis et al. 

 (1970) . However, the information from their re- 

 ports is limited and all that can be said is that 

 they give values approximately the same as those 

 noted in the present case. The agreement is be- 

 lieved to be reasonably good for experiments of this 

 type insofar as the maximum effect goes. We presume 

 that similar effects in "ocean" experiments could 

 be achieved at much smaller concentrations if the 

 G parameter has significance. 



During their cavitation tests Baker and Holl 

 noted a change in the appearance of the developed 

 cavitation. From photographic observations of these 

 changes they speculated that the cavitation attenu- 

 ation was due to a "flow reorientation in the region 

 of the laminar separation bubble." They further 

 suppose [Arndt et al. (1975) ] that the amount of 

 attenuation might depend on a Deborah number, 

 TV_^/6g, where T is the molecular relaxation time 

 of the molecule, V the freestream velocity, and 



^1, 



UJ > 



a: 



mn d« MEULENII976)-500WPPM 



• PRESENT STUDY - 500WPPM 

 A 20 WPPM 

 A 50 WPPM 

 D 20 WPPM 

 ■ 80WPPM 



• 20WPPM HOLL el 01 (53) 



ELLIS 610111970) 

 BAKER elol (1973) 





O 

 O 

 O 



IxlO' 234 56789 IX 



BODY REYNOLDS NUMBER - UD/U 



FIGURE 40. Maximum cavitation inception index suppres- 

 sion by polyox WSR 201 on the hemisphere nose. The 

 Ellis and Baker results are for polymer "oceans." 



6g is the boundary layer displacement thickness at 

 separation. It now seems clearly established in 

 our opinion, that the overall gross effect caused 

 by the polymer in the flow about these bodies is a 

 removal of the laminar separation by stimulation of 

 transition and that this is indeed the origin of the 

 flow "reorientation" noted by Baker and Holl. Pre- 

 sumably, the molecular relaxation time has an im- 

 portant role in boundary layer stability, but as 

 yet this appears to be unknown; it may be that the 

 parameter proposed by Arndt is important for some 

 laminar flows with separation (as it is indeed for 

 the flow about a circular cylinder) , but we think 

 not in the context of the present experiments. 



Since the suppression of cavitation upon these 

 bodies is a result of the elimination of the laminar 

 separation by the polymer it is worthwhile to com- 

 pare the present results with those in which the 

 separation is eliminated by another method. Arakeri 

 and Acosta (1976) carried out a series of tests with 

 a hemisphere nose body and an ITTC body using bound- 

 ary layer trips to reduce the critical Reynolds num- 

 ber in the HSWT. It was, briefly, found that with 

 the trip present and at velocities above the new 

 critical velocity, the occurrence of cavitation was 

 significantly suppressed, and that at higher ve- 

 locities the tunnel would choke from the model 

 support before the body could be made to cavitate! 

 The present polymer tests show a very similar large 

 effect on inhibiting cavitation but not quite as 

 dramatic as the tripped tests. 



8. FREESTREAM NUCLEI AND CAVITATION INCEPTION 



Some Observations in the LTWT 



As will be recalled from the description of the 

 LTWT, this facility has no resorber which neces- 

 sitated cavitation data acquisition before pump- 

 generated bubbles entered the test section. On a 

 number of occasions the cavitation on the NSRDC and 

 hemisphere bodies was deliberately )maintained and 

 the pump-generated gas bubbles allowed to pass 

 through the test section. As the number of free 

 gas bubbles increased, the initially-occurring band 

 type cavitation was gradually destroyed and replaced 

 by travelling bubble type cavitation. An alterna- 

 tive procedure was to lower tunnel static pressure 

 so that the cavitation number had a value below 

 ~^Pmin ^^^ above the inception value and again 

 allow the pump-generated bubbles to accumulate. 

 The body would then eventually cavitate with in- 

 ception then always being of the travelling bubble 

 type. Schlieren observations of the basic viscous 

 flow on the hemisphere nose were made at these 

 gradually increasing freestream bubble populations 

 and nuclei populations were measured when band type 

 inception occurred and when this above deliberately- 

 promoted bubble type inception occurred. The 

 schlieren observations show (see Figure 41) that 

 as the number of freestream nuclei increased, the 

 laminar separation on the hemisphere nose became 

 unsteady and was finally greatly diminished if not 

 eliminated. Thus, in effect, the free-stream bub- 

 bles serve to trip the boundary layer. 



Nuclei populations obtained when band type incep- 

 tion occurred (Oj^ = 0.44) are shown with distribu- 

 tions obtained when deliberately promoted travelling 

 bubble inception occurred (o-j^ = 0.58, 0.73) in 

 Figure 42. As can be seen in this figure, for 



