The axial compressive stresses of -63.3 psi/psi measured at midlength of the missile 

 tubes were slightly above the yield strength of the material when the tubes failed. Such high 

 stresses have not been previously observed in the missile tubes of circular pressure hulls 

 and can be attributed to the resistance of the tubes to a tendency of the oval cylinder to be- 

 come more elliptical under loading. It appears that for a tube arrangement such as that used 

 in Model EC-IA, thicker walled tubes are required for an oval cylinder than for an equivalent 

 circular cylinder of the same radius of curvature. 



COMPARISON OF MEASURED STRESSES WITH THEORY 



The solid-line curves for the stress sensitivities shown in Figures 17 to 26 were taken 

 from Reference 8 and represent the results of using the local thickness of the ring-stiffened 

 oval cylinder of Figure 1 in the theoretical solution obtained for an oval shell of uniform 

 thickness. The work done by Polytechnic Institute of Brooklyn, which led to the analytical 

 results shown in their figures, was discussed in a previous section of this report. The experi- 

 mental points appearing in the figures were obtained from strain-gage measurements during the 

 test of Model EC-1. The dashed curves shown in Figures 17 to 23 represent the stress distri- 

 bution based on using the local radius of curvature and shell thickness of the oval cross section 

 in the analysis of Von Sanden and Gunther^ for a ring-stiffened circular cylinder. This type 

 of approximate solution has been shown in References 3 and 5 to yield good results for simply 

 and clamped supported short oval cylinders. However, as has been indicated in Reference 5 

 and from an inspection of these figures, an equivalent circular cylinder solution based on the 

 local radius of curvature concept will not yield good results for the deformations and stresses 

 of an oval cylinder stiffened by elastic rings. 



Figures 17 and 18 are plots of the circumferential and longitudinal stress distributions 

 along the outside and inside surfaces of the shell at the major axis. Figures 19 and 20 are 

 corresponding plots at the minor axis. Figures 21 and 22 show the midbay stresses for a 

 quadrant of the oval shell; Figure 23 shows the circumferential flange stresses in the ring. 

 The abscissa 6 in these latter figures is the angle which the local normal to the median sur- 

 face of the shell makes with the major axis. 



The solid curves in Figures 17 to 23, which are based on the solution developed by 

 Polytechnic Institute of Brooklyn in which the local shell thickness was used, show excellent 

 agreement with the measured circumferential stresses oa. The theoretical longitudinal stresses 

 a (shown in Figures 17 to 22) differ from the test results by a translation in which the theo- 

 retical stresses are too low at the major axis and too high at the minor axis. Figures 24 to 

 26 isolate the discrepancy between the theoretical and measured stresses a . In these latter 

 figures, the stresses a^ have been separated into membrane and bending components by 

 the following relationships: 



