Pien 



area so obtained is too big, we can add one more restraint on the design condi- 

 tion so that a desirable prismatic coefficient can be obtained. 



From the experience obtained from Model 4946 and many models tested by 

 Inui at the University of Tokyo, we can anticipate a phase shift between theoreti- 

 cal and measured wavemaking resistance curves. The experimental curve is 

 always shifted to the right of the theoretical curve. Therefore, in this design 

 the optimization is done at F = 0.28 rather than 0.32. 



At this point the computer programs are used to carry out the lengthy, tedi- 

 ous numerical computations. After a few trials and errors, we obtain the fol- 

 lowing singularity distribution for our final design. 



7] = ±[.013^^ + .04813^* - .16113^2 + . 1] . 

 On the side of t?- surface we have 



Ej = [6.5504<f - 29.0186^2 + 67.4020^^ - 74.7203^'' + 30.2692^^] 



E3 = [10.9934<f - 61.3783^2 + 158.8456^^ - 183.8859^'' + 75.0307^^] C^ . 

 or 



M(^,0 = Ej + E3. 

 On the end of 17 -surface we have 



Eg = .008 + .04^2 ^ 04^^ . 

 On the bottom of T7-surface we have surface source 



MgC^,^) = [-.1882^ - 2.3841^2 + 15.9590cf^ - 28.8343^'' + 13.6760^^] 



+ [2.2616^ - 5.3769^^ - 625.5772^^ + 163.9480<f'* - 82.5272^^]^ 



+ [-13.3743^ + 107.5863^2 _ 208.1482^^ + 151.7266^'* - 88. 7849^^] 77^ 



+ [12.2606^ - 121.1156^2 ^ 338.7764^^ - 419.1692^'' + 232. 2801^^] ^^ 

 surface doublet 



M^(<f,^) = [.1131^- .4594^2 ^ 1.5838^^ - 2.7567^'* + 1.6938^^] 



+ [-2.5399^ + .9383^2 .4. 1500^^ + 7.5077^"* - 5.1727^^]^ 

 + [-.3740^ + 2.4491^2 _ 8.8036^^ + 12.8106^'' - 5.2930^5]-2 

 + [.4157^ - 3.4692^2 + 1.3092^^ - 20.3145^'* + 1.0316<f5]^^ 



1132 



