CAVITATION IN SEPARATED FLOWS 



The foregoing remarks were concerned with the conditions near a hydraulically smooth 

 surface having unseparated boundary layers. It was pointed out 'in Reference 3 that care 

 must be exercised in the design of hydrodynamic systems to insure that separated regions do 

 not occur. In such regions, the very low pressures which can be developed in the essentially 

 vortex flows can lead to cavitation in spite of relatively high ambient pressures. This was 

 illustrated for wake flows from experiments at the Taylor Model Basin and for locally separ- 

 ated boundary layer flows on bodies of revolution from experiments at the Iowa Institute of 

 Hydraulic Research. An example of cavitation in a separated boundary layer in the vicinity 

 of the stagnation point is shown in Figure 1. In this case, the (probably laminar) boundary 



Figure 1 - Cavitation in the Separated 



Boundary Layer on the Upstream 



Side of a Toroidal Ring 



A half-toroid is fastened to a lucite plate through 

 wnich this view was taken. The flow is from left to 

 right. The two horseshoe-like vortices are swept down- 

 stream, one leg of each being clearly seen outside the 

 ring and the other leg passing through the hole but ob- 

 scured by the cavitation on the ring itself. 



This photograph was obtained by Mrs. E. A. Sykes 

 of the David Taylor Model Basin. 



layer separated in the region of the adverse pressure gradients approaching the toroidal ring. 

 The cavitating region was swept downstream in the form of two horseshoe vortices with legs 

 outside and inside the toroid. 



Conditions similar to these are evidently associated with cavitation about surface 

 roughnesses. It is clear that the flow about roughness elements with sharply varying contours 

 will separate and that cavitation may occur in the core of the separated regions. The strength 

 of the vortices will evidently depend upon the velocity near the top of the roughness element 

 so that the cavitation characteristics will depend upon the boundary layer configuration in the 

 vicinity of the elements considered. Although much work has been done on the noncavitating 

 flows about rough surfaces, results are not available in sufficient quantity to enable design 

 criteria to be- established for roughness limits acceptable from the standpoint of cavitation 

 prevention. Experiments have been made by Shalnev^ for roughness elements in a restricted 

 flow but lack of complete correlation with boundary layer characteristics make these results 

 of limited usefulness. Nevertheless, they are the only published results known to the writer, 

 and some empirical formulas are given which can be used when the roughness height is large 

 compared with the boundary layer thickness. 



