PLASTIC FAILURE OF CYLINDRICAL SHELLS. 



621 



7 



correction for rebound. The amount of restitution was, however, mis- 

 judgeif and the actual energy absorbed in phistic deformation was only 

 132 inch-lb. Both the damaged surfaces of the cylinder were contoured 

 as cx|ilaincd later and Fig. S shows a comjiarison of the results of static 

 and dynamic loading. The general shapes of the contours agree well, but 

 the maximum indentation was less in the dynamic than in the static test 

 being under-estimated. 



Figs 6. 



-Tee framci at equal spacing 



i.J..li. |l.|l.AAAA 



lil l!l lil III III l!l lil 



lii lii i;i 1:1 I'l lii i!i iimi 



!£_£ 



LONGITUDINAL SECTION 



Staggered tack weld 



14 g3ijg« jhe 



'U'"'*-? '" ""8' 



4 dij MS rodi 

 I reduced for 1 ' nuts 



Scale 1 "ich = 1 foot 

 n 1 6 9 12 Inches 



f ■ [ — ! — 1—1 — t— 1 — I I » t I 



SECTION XX 



Cylinder with Heavily Reinforced Ends. 



Similar experiments were then made on a drum of soft copper, 24 inches 

 long, 15 inches in diameter and of thickness 0048 inch. The ends were 

 heavy wooden disks to which the shell was screwed. The static load- 

 displacement curve is shown in Fig. and a comparison of the contours 

 obtained in the static and dynamic tests in Fig. 10. The energy absorbed 

 in the dynamic test was about 10 per cent, greater than in the static test 

 and that is reflected in the contours, which show rather greater indentation. 



The close agreement between the results of static and dynamic tests 

 on the copper drum suggests that the rates of loading within those limits 



