with x = H <}>, which is zero at the trailing edge, and 



J J J 



f(p><l>) = y a p + y y b.. p 1 x 1 in the cavity plane (32) 

 i=l i=l j=2 



The two-dimensional solution is supplied from preliminary design of each 

 section. The cavity plane has to be finite numerically. A separate source distri- 

 bution along the truncation line of the cavity is considered with unknown coeffi- 

 cients to compensate for truncation of the cavity: 



f 2 (p) = HA^^V p 1 X (33) 



i=l 



In addition, the cavity thickness near the leading edge is specified to be equal to 

 that computed from cascade theory during preliminary design. This thickness is 

 chosen because many sets of cavity thicknesses give the same load distribution and 

 the same cavitation number for the linearized, two-dimensional cavity problem. The 

 leading edge conditions can be interpreted as follows: the source distribution at 

 the leading edge is exactly the same as that of a two-dimensional supercavitating 

 cascade, that is 



at the leading edge 



With the previously described representations of the source distribution, the 

 integration of each term in the integral equation is performed numerically. If the 

 propeller blades have rake or skew, E, is a function of p. This is treated by a 

 linear approximation in each Ap interval 



£ = K Q + P K ± (34-D 



where 



13 



