convenient for this kind of work, are the large (16:1) contraction ratio, 

 the vaned elbows on each leg, the ready insertion of honeycomb straight- 

 eners and damping screens and, finally, the insertion of turbulence- 

 generating grids at the entrance of the working section (see Gates 1973 

 for a full description). It was possible, thereby, to vary the freestream 

 level by nearly a factor of 100 from a low value of about 0.05 percent. 

 (Unfortunately, it was not possible to carry out cavitation experiments 

 over the same range.) 



It was anticipated that these turbulence levels would have a profound 

 effect on bodies near transition but normally having a separation, such as 

 the DTNSRDC body. We see, in Figure 29, that a turbulence level of about 

 one percent is sufficient for transition to occur upstream of the 

 separation on this body. A similar upstream progression of the site of 

 transition at fixed body Reynolds number with increasing turbulence level 

 was also observed on the Schiebe body (not shown). Indeed, the observa- 

 tions of transition location together with those of van der Meulen agree 

 fairly well with the amplification computations of Wazzan and Gazley, Fig- 

 ure 30. Similar observations on the hemisphere body were surprising how- 

 ever, as the laminar separation appeared to be totally immune to the turbu- 

 lence level! This unexpected result is traced (Gates 1977) to significant 

 differences in the sensitive critical frequency of the respective boundary 

 layers on these bodies (deduced from linear stability theory), the highest 

 frequency being for the hemisphere body, DTNSRDC next, and then Schiebe 

 body. Because of this difference in frequency and reasoning from typical 

 grid turbulence spectra, it is argued that there is nearly two orders of 

 magnitude more disturbance energy available for boundary stimulation on 

 the DTNSRDC body than on the hemisphere. This seems a plausible explana- 

 tion and shows the desirability of providing freestream spectral data in 

 addition to intensity levels in future hydrodynamic work. 



51 



