Stimulated Boundary Layers 



From the previous discussion, we would expect that any effect on cavi- 

 tation of a laminar separation would be eliminated if the laminar boundary 

 is sufficiently "stimulated" to transition. This stimulation may be via a 

 variety of freestream factors, but in the laboratory it is easiest to use 

 a mechanical "isolated roughness" or boundary layer "trip." Studies were 

 carried out, first on the hemisphere body (Arakeri and Acosta 1976), with 

 trips of tape and then with machined steps. These trips were rather large 

 (several times larger than the boundary layer displacement thickness) and 

 were located near the nose to forestall cavitation at the trip. At low 

 speeds a laminar separation was still observed (Figure 22), but at higher 

 ones, about 30 ft/sec, the separation gave way as transition occurred. The 

 results of this tripping were indeed spectacular, as the normal course of a 

 cavitation experiment was completely reversed. We see in Figure 23 a 

 sequence of photographs of cavitation on a hemisphere body taken as speed 

 is increased. As this happens the established laminar separation disappears 

 and with it, in the Caltech HSWT, the cavitation, too 1 . In fact, cavitation 

 inception indices show a decrease with tunnel speed (Figure 24) instead of 

 the usual trend. We should emphasize that the form of cavitation in this 

 tunnel was always that of an attached cavity; freestream travelling cavi- 

 tation events were extremely rare. However, this kind of cavitation 

 observation is not unique to the Caltech facility, as the interesting 

 photograph of Figure 25 shows. More recently, boundary-layer stimulation 

 has been proposed as a standard laboratory technique in cavitation testing 

 to avoid unwanted scale effects from laminar separation (Kuiper 1978). As 

 the findings of Figure 23 shows there still may be other important factors. 



These findings on tripped boundary layers raise many issues concerning 

 inception scaling laws. These issues are of two types; in the first, we 

 have (as already sketched) all of those questions affecting the fully- 

 wetted viscous flow itself. The second type may be said to be those having 

 a direct bearing on the cavitation process itself. Among these is the 

 presence (or absence) of nuclei, and the state of the liquid (i.e., thermo- 

 dynamic factors). There is also the large effect of polymer solutions on 



40 



