ocean. The zones of failure for Spheres 

 no. 11 and 12 (Figures 15 and 16) 

 appeared to be from crushing of concrete 

 and were more localized than for the 

 non-preloaded spheres. Sphere no. 13 

 showed a failure zone (Figures 17 and 18) 

 that was more typical of the non- 

 preloaded spheres. A shear-compression- 

 type failure mode for a section of wall 

 was evident for about 50 inches (130 mm) 

 in length. 



The spheres that imploded in the 

 ocean were found totally fragmented. In 

 a pressure vessel, once failure of a 

 sphere began, the pressure load dropped 

 off rapidly. This did not occur in the 

 ocean, and, hence, a sphere was 

 subjected to extremely violent shock 

 forces when the water collapsed the 

 walls. Figure 19 shows the debris on the 

 seafloor for Sphere no. 1 that imploded 

 at a depth of "^,^15 feet (1,5^7 m). 



Strain Behavior. Strain behavior for 

 the spheres tested in the pressure vessel 

 was monitored by measuring the quantity 

 of water displaced from the interior of 

 the sphere while under load.* A change 

 of interior volume is a direct function of 

 the change in radius, and the change in 

 radius is a direct function of hoop strain. 

 The displaced water was measured to an 

 accuracy of +10 ml, which converted to a 

 strain accuracy of +2 in. /in. This is 

 applicable only when membrane displace- 

 ments account for most of the change in 

 volume. Near failure, displacements due 

 to flat-spot development can contribute 

 significantly to the displaced water. 



Figure 20 shows the raw data of 

 pressure versus displaced water for the 

 spheres. Other scales show the wall 



stress and strains. A similar presentation 

 of results is given in Figure 21 for the 

 non-preloaded spheres (after Ref 5). 

 The dry concrete spheres were able to 

 withstand about 20% more ultimate 

 strain than the saturated spheres before 

 imploding. 



It was quite apparent from the data 

 that saturated concrete behaved 

 differently than dry concrete. However, 

 at this time, data do not exist on the 

 stress-strain behavior of concrete 

 control cylinders that are saturated and 

 uniaxially loaded while in a hydrostatic 

 pressure environment. This sphere test 

 program (Figures 20 and 21) provides the 

 first data that show saturated concrete 

 under hydrostatic pressure does not 

 perform as well as similar dry concrete. 



Table 'f lists the stiffness, or elastic 

 modulus, for the walls of the spheres. 

 The average modulus value foe the three 

 ocean spheres was 5.1 x 10 psi (35.2 

 GPa); for the non-preloaded spheres, it 

 was kM X 10^ psi (30.3 GPa). This 

 difference was due to the effect of the 

 preload and is a well-known effect of 

 concrete creep (Ref 8). For 5.3 years 

 the concrete in the ocean spheres experi- 

 enced creep under a wall stress of about 

 5,000 psi. 



Concrete creep and a triaxial 

 loading condition explains why the 

 modulus of 5.1 x 106 psi (35.2 GPa) for 

 the wall of the preloaded sphere was 

 greater than that given in Table 3 for the 

 ocean-cured concrete block, where the 

 modulus was 3.77 x 10^ psi (26.0 GPa). 

 The concrete block in the ocean was not 

 stressed like the sphere wall; it was 

 saturated with seawater and in a state of 

 equilibrium with the environment. 



*This technique was used in Reference 5 with excellent results; the same 

 procedures were applied in these tests. 



20 



