Gadd 



PART 2 - EXPERIMENTS ON INUIDS 



One possible technique for handling the wavemaking side of design is that 

 adopted by Inui (1) and Pien (5). A distribution of wavemaking sources, giving 

 suitably low resistance, is chosen and the corresponding hull surface calculated 

 by, in effect, assuming the water surface to remain flat. Thus the source dis- 

 tribution is treated as one of ordinary potential sources reflected in the undis- 

 turbed water plane. This procedure cannot be rigourously justified, even for 

 very low Froude numbers, but it has been plausibly argued by Pien that for 

 forms of low wavemaking resistance, in which we are mainly interested, the 

 theory should be adequate, since the water surface then does remain fairly flat. 



Inui has made experimental comparisons between models (which have been 

 termed "Inuids") derived from source distributions in this way and other models 

 derived from the same source distribution by the Michell thin- ship approxima- 

 tion. The latter is equivalent to relating the waterline slope to the source den- 

 sity on the central plane. In general the Inuids showed better agreement with 

 experiment, but discrepancies remained and were attributed by Inui largely to 

 viscous effects. 



A third possible way in which the source distribution may be related to hull 

 shape is that proposed by Guilloton (2). Although this method looks very prom- 

 ising it has not been investigated here. Rather, it was attempted to establish 

 whether in experiments with Inuids the discrepancies between theory and exper- 

 iment could reasonably be ascribed to viscous effects or whether instead the 

 discrepancies resulted mainly from the Draconian treatment of the free surface 

 condition. 



With normal ship models viscous interference with wavemaking arises 

 mainly at the stern. If, then, a head form is used, followed by a long parallel 

 afterbody from which it is separated by a small gap, the viscous effects on the 

 wavemaking of the head should be small. Such a head form in deeply submerged 

 potential flow would have zero drag if the pressure in the gap were equal to that 

 in the undisturbed stream. In a free surface, therefore, the net pressure force 

 on the head, acting over all its surfaces, plus the force represented by the ex- 

 cess of mean pressure in the gap over free-stream pressure times the wetted 

 area of the rear face of the head abutting the gap should equal the wave resist- 

 ance. A net total force measurement on the head will, of course, include a fric- 

 tion component, but this ought to be reliably calculable, since the boundary layer 

 should remain thin and well-behaved. 



Experiments on models of this kind were independently suggested by Bres- 

 lin (6). At the NPL, two such head-form models were designed as shown in 

 Figs. 4 and 5. Each model was generated, by the Inui method, from source dis- 

 tributions defined over a rectangular area of length l and depth D = L/^-n on the 

 central plane. The source distribution was uniform for the first model, which 

 did not have a particularly low theoretical wave resistance. For the second, 

 quasi bulbous-bow, model the source distribution varied both in the longitudinal 

 or X direction and the depth or z direction, as indicated in Fig. 5. However, 

 the source density was everywhere positive to ensure that the waterlines were 

 everywhere inclined at a positive angle to the oncoming flow, in the interests of 



716 



