AN ANALYSIS OF TESTS OF WATER-TIGHT BULKHEADS. 93 



the bulkhead must be broader or double, and we shall thus have a great 

 accumulation of material at the neutral axis, which is not rational. 



Thus we may say that double stiffeners are less efficient, heavier and 

 more expensive than single stiffeners, and to this must be added that double 

 stiffeners interfere with the caulking of the seams of the bulkliead plating. 



Single stiffeners should therefore be used in all cases except where 

 double stiffeners must be used for some special reason. 



In test No. 12 the load was excessive as evidenced by the great per- 

 manent set and by the high theoretical stress, which exceeded the elastic 

 limit. The curve for elastic deflection followed, however, also in this case 

 closely the theoretical curve. 



Test No. 13 was on a bulkhead constructed according to the Rules 

 of Germanischen Lloyds. The head was also here very great, and the 

 brackets were not of an effective type; hence the great deflection, which, 

 as shown on the diagram, exceeds that of a freely supported stiffener if 

 the length is reckoned from bracket to bracket. By reckoning the length 

 slightly greater, 14.75 feet instead of 13.0 feet, the calculated curve of 

 deflection may be made to correspond very closely with the observed curve. 

 In calculating the stress this case has therefore been dealt with as that of 

 a stiffener freely supported at the ends and of a length equal to 14.75 feet. 



In test No. 14, which was on a bulkhead of the Russian Cruiser Bojarin, 

 the bracketing was feeble and not connected to a rigid structure. Hence 

 the length used in the calculation is here the total length of the main bar 

 of the stiffener, but even so the observed deflection exceeds the calculated 

 deflection somewhat. This shows the importance of strengthening and 

 supporting the inner bottom where brackets are fitted to it. 



Allowing for the reinforcement given by the brackets, the maximum 

 stress was, as stated above, found to exist at or near the middle of the 

 stiffeners. The stress .was calculated first for pure bending, supposing 

 the stiffeners to carry the full load on one span. In accordance with the 

 above the formula for a freely supported stiffener was used in case of most 

 of the double stiffeners, while the formulas for stiffeners fixed at one or 

 both ends, in accordance with the system of bracketing, were used in case 

 of most of the single stiffeners. Thereafter the stress due to pure tension 

 was calculated corresponding to the observed deflection, as explained in 

 the Appendix, and added to the stress due to bending. 



It was not found necessary or useful, in determining the stress, to 

 apply the complex formulas for combined tension and bending to the tests 

 here analyzed. 



The following table gives the results of the calculations: — 



