stiffener in the middle, as well as the end stiffeners (Figure 3), while 

 another had a series of stiffeners in addition to those at the ends 

 (Figure 4) . All the shells were tested with flat 7075-T6 aluminum alloy 

 end closures that permitted the ends of the shells to slide upon them. 

 Besides being subjected to slowly increasing hydrostatic pressure, some 

 of the shells were pressure cycled, some were subjected to dynamic pressure 

 loadings, while some were notched prior to hydrostatic pressure tests. 



The results of the tests conducted with the exploratory cylindrical 

 shell series in 1963 at the Southwest Research Institute, San Antonio, 

 Texas, were very encouraging. It was found that the elastic stability 

 and stress distribution formulas available for rib- stiffened metallic 

 cylinders apply in identical manner to ceramic cylinders, except that in 

 the ceramic cylinders all of the deformations take place in the elastic 

 stress region of the material. When stress raisers, such as deep scratches 

 or sharp fillets, are located in an area of the shell where only compres- 

 sive stresses occur, no degradation of the ceramic rib-stiffened cylinder's 

 collapse strength occurred. No creep has been observed at biaxial stresses 

 below 200,000 psi generated in the ceramic cylinders by hydrostatic pres- 

 sure for one day. The implosion resistance of ceramic rib-stiffened 

 cylinders to underwater shock waves has been found to increase with depth, 

 while that of metallic cylinders of identical dimensions decreased. When 

 the #9606 Pyroceram rib-stiffened cylinder (Figure 4) was subjected to 

 3000 pressure cycles which generated compressive stresses of 200,000 psi 

 magnitude in the cylinder, no failure of the cylinder occurred although 

 some flaking of the material was observed in areas where the highest com- 

 pressive stresses occurred (Figures 5-7). No failure or flaking of any 

 kind occurred at the joints where the ceramic cylinders rested on the 

 smooth 7075-T6 aluminum closures. 



The ceramic cylinders tested were not of large diameter; however, their 

 buoyancy was of sufficient magnitude that, if several of these cylindrical 

 shell sections were joined together into a long hull, it would support a 

 useful oceanographic instrumentation payload. Using an L/D ratio of 6-8 

 which insures minimum hydrodynamic drag, a 60- inch- long cylindrical hull 

 made up of the already proven 8-inch-OD ceramic shell sections could carry 

 a 30-pound payload to a depth of 20,000 feet and yet retain 5-10 pounds of 

 buoyancy for the return trip to the surface. But this could be feasible 

 only if a mechanical joint were developed that not only would withstand 

 the stresses generated by hydrostatic loading, but also would insure the 

 hull's integrity during launch and retrieval operations. 



Experimental Joints for Ceramic Shells 



Several approaches were tried to the design of. a reliable cylin- 

 drical shell section joint (Figure 8) that, in addition to withstanding 



266 



