619 



PLASTIC FAILURE OF CYLINDRICAL SHELLS. 5 



when working in that position, and the loads were assessed by the diametral 

 strains of a calibrated ring placed between the jack and the specimen, the 

 strains being measured by an Ames dial. The ram which applied the 

 load to the specimen was spherically ended, its diameter being 

 1| inch. 



Displacements of the ram were observed by a telescope carrying a 

 cross-hair focused on a scale attached to the bottom of the calibrated 

 ring. Rods attached to the bottom of each end of the cylinder also 

 carried scales and the readings of the scales by separate telescopes gave 

 the bodily sinking of the cylinder in the sand and so enabled the true 

 amount of indenting of the specimen to be determined. 



Two Ames dials supported from the bottom of the sand box at opposite 

 ends of a horizontal diameter of the cylinder enabled the diametral spread 

 under load to be measured, but those measurements were not found to bo 

 of particular importance. 



In some tests the longitudinal alterations in length of the cylinders 

 were measured ; steel balls were attached near the top of the cylinder at 

 each end and provided bearing points for the ends of a micrometer 

 measuring bar ; a sufficiently accurate overall strain measurement was 

 thus obtained. 



Standard Tbst Cylinders. 



The standard test cylinder is shown in Figs 4 and 5. It consisted of 

 a 14-gauge cylindrical shell 2 feet IJ inch in length and 16J inches external 

 diameter, internally stiffened by eleven rings made of ^-inch by ^-inch 

 by 14-gauge T-section steel. The rings were evenly spaced along the 

 length of the cylinder and the legs of the T's were tack-welded to the shell, 

 the welds having a pitch of approximately 2 inches. That was found by 

 trial to be more satisfactory than continuous welding which introduced 

 a serious danger of damaging the shell by over-heating it. 



The end plates were also of 14-gauge steel and were stiffened by ff-inch 

 by ^-inch T-sections and brackets arranged as shown in Figs 4. 



The cylinders were made by Messrs. Benham and Company of Wands- 

 worth and in the first place six were completed. An experiment in manu- 

 facture, however, provided an additional unfinished specimen which gave 

 useful information on test. It had a central stiffening ring continuously 

 welded to the shell and a ring spot-welded to the shell at 3 inches from 

 either side of the central ring. No end-plates were provided in that 

 cylinder. 



At a later date further cylinders were made, three to the original design 

 and three as shown in Figs 6. The feature of those was the replacement 

 of the normal end-plates by 1-inch disks to give resistance to bending out 

 of their planes and the discs were connected by four 1 J-inch-diameter steel 

 rods. 



