a salt water environment. The brine was then pressurized to 4450 psi and 

 pressure applied in cycles ranging from to 4450 psi to the oil. After 12' 

 cycles, the light jacket on Model DSRV-IF failed. Subsequently, the model 

 was removed from the tank and shipped to the Model Basin for reassembly 

 with a heavier jacket. When this was completed, the model was static tested 

 to 4900 psi at the Naval Research Laboratory (NRL) before being returned to 

 Southwest Research Institute. Then both models were again placed in the 

 SRI cyclic tank. After 500 cycles, the models were visually inspected and 

 static tests were made on both models at each subsequent 500-cycle inter- 

 val. The fatigue test continued until both models failed at a point be- 

 tween 11,500 and 12,000 cycles. The failure of these models was due to 

 circumferential crack propagation at the weld line joining the jacket to 

 the end ring. A longitudinal crack in the jacket of Model DSRV-IA is 

 apparent also. Examination of the 7079-T6 aluminum components of both 

 models failed to reveal any damage to these parts. See Figure 4 for photo- 

 graphs of Models DSRV-IA and DSRV-IF after cyclic testing. 



DISCUSSION AND CONCLUSIONS 



Several significant points were brought to light by these rather 

 exploratory tests. The results of the cyclic tests were fairly encouraging 

 since both models withstood almost 12,000 pressure cycles to 4450 psi. 

 Failure to discover any structural damage to the 7079-T6 rings and frames 

 after cycling is evidence that it is possible to design a composite hull, 

 weighing 53 to 58 percent of its displacement, that will withstand 10,000 

 excursions to a depth of 10,000 ft. 



There is a definite need, however, for further cyclic tests on 

 models of composite construction. More information is needed on the effect 

 of scale on fatigue life; larger models are required to establish a corre- 

 lation between size and structural characteristics. Certainly the small 

 scale of these two models contributed to fabrication and tolerance problems 

 and led to residual stresses which would be relieved or eliminated on 

 larger models. The unsatisfactory results obtained with the thin jacket on 

 Model DSRV-IF can be attributed to fabrication difficulties arising from 

 the thiimess and relative instability of this jacket. Conversely, it is 

 true that the fatigue life of plating decreases as the plate thickness 



