they compare very favorably with the respective experimental collapse 

 pressures , 



It is recalled that examination of Models OV-3 and OV-5 after col- 

 lapse (see Figures 16 and 18) revealed extensive corrugation of the inner 

 and outer shells between adjacent webs, indicative of gross yielding of 

 the shell elements. Also, the observed collapse pressures were in excel- 

 lent agreement with the computed values p^^ ; see Table 2. Since the in- 

 elastic general-instability pressures are also in good agreement with the 

 experimental collapse pressures, it appears logical to conclude that pos- 

 sibly the short one -diameter model (OV-3) of the original design and the 

 four-diameter model (OV-5) of the suggested redesign may be of such a 

 balanced nature that in each case the two modes of axisymmetric plastic 

 collapse and inelastic general instability occurred simultaneously. 



The aforementioned procedure for determining the inelastic general- 

 instability pressure p for a web-stiffened sandwich cylinder may appear 

 to be somewhat empirical. However, it does define the important feature 

 of the dependence of the inelastic general-instability strength on the stress - 

 strain characteristics of the material, and also affords a means for ob- 

 taining a good estimate of strength for this mode of collapse. Moreover, 

 Models OV-3, OV-4, and OV-5 were machined from nascent bars of tita- 

 nium alloy so that no appreciable residual stresses existed in these small- 

 scale models, as would occur in a fabricated structure due to rolling and 

 welding of flat plating. Such built-in stresses would effectively alter the 

 shape of the stress -strain curve so that collapse strength of a fabricated 

 sandwich hull may be considerably lower than that observed from tests of 

 initially stress -free models. The question of how much strength reduction 

 would occur can only be answered by testing some large-scale fabricated 

 models in which the influences of residual stresses and out -of -roundness 

 are present. 



If the observed collapse pressure of 11, 100 psi for Model OV-4 is 

 extrapolated to a yield strength of 120, 000 psi, a pressure of 9650 psi is 

 obtained. This value is 3.5 percent below the minimum design collapse 

 pressure of 10, 000 psi. Since the mode of failure observed for Model 

 OV-4 appears to be greatly influenced by the shape of the stress -strain 

 curve of the material, the collapse pressure of the fabricated prototype 



32 



