membrane Z inner outer 



bending Z inner outer 



It can be seen from Figures 24 to 26 that the theoretical values for (a^) agree well with 



bending 



the measured values and that the discrepancy between theoretical and measured stresses a^ 

 is due to the theoretical values for (a^) . Their poor agreement is attributed to the 



membrane i ■ j 



use of two different local shell thicknesses (see Figure 1) in a solution for an oval cylinder 

 of uniform thickness subjected to hydrostatic pressure. The results of such a calculation 

 lead to two different axial contractions, with the thinner shell contracting more than the 

 thicker shell. If it is assumed that both portions of the shell contract the same (as may be 

 the case for the models tested) and that the net end load does not change, then relative to 

 the calculated contractions, the thinner shell would have to be stretched and the thicker shell 

 compressed. This would increase the axial membrane stress in the thin shell and decrease 

 (algebraically) the corresponding stress in the thick shell. Such a correction would shift the 

 theoretical curves for a toward the experimental results. 



COLLAPSE PRESSURES 



It was found from previous hydrostatic tests ^° of ring-stiffened circular cylinders which 

 failed by axisymmetric yielding that experimental collapse pressures agreed best with theory 

 based on the Hencky-Von Mises criterion of failure and allowing for the plastic reserve 

 strength after initiation of yielding. Collapse pressures computed from the Model Basin 

 plastic hinge theory^ ^ applied very well with circular cylinders fabricated from steels exhibit- 

 ing a plateau-type stress-strain curve. Based on the local radius of curvature for the oval 

 cylinder shown in Figure 1, plastic-hinge collapse pressures of 2356 and 3049 psi were com- 

 puted for the large and small radial section of the model, respectively. Comparison of these 

 pressures with the experimental collapse pressure of 1670 psi for Model EC-IB shows that 

 the strength of the oval cylinder tested is lower than a comparable circular cylinder of the 

 same dimensions and radius of curvature. 



Another interesting point worthy of mentioning is the fact that an oval cylinder has less 

 enclosed volume than a circular cylinder of equal peripheral length. By virtue of this fact, the 

 ratio of weight of pressure hull to weight of displaced water of the hull is very high for 

 Model EC-1. A weight-displacement ratio of 0.598 was computed for this model. Based on 

 least-weight calculations, a circular cylinder with the same weight-displacement ratio and 

 fabricated from HY-100 steel would have a collapse pressure on the order of 4000 psi. This 

 can be compared with the collapse pressure of Model EC-IB which was only 1670 psi. Thus, 

 on a strength-weight/displacement basis, the test of Model EC-IB lends further evidence that 

 a ring-stiffened oval cylinder with a major to minor axis ratio of 1.5 is a very inefficient 

 structure when compared to a circular cylinder. 



