382 SECTIONAL TRANSACTIONS.— G. 



for greater actual engine -weight, a weight/power ratio of 3 must be aimed at. Com- 

 pression-ignition engines built for other purposes give good economy, but only at 

 low speeds and low m.e.p.'s, and consequently great weight per H.P. : for example, 

 the lightest engine for submarine work weighs about 40 lb. per H.P. In order to keep 

 down weight, it is necessary to limit the maximum cyUnder pressure. 



Research work has taken place along parallel but independent lines with ' jerk- 

 pump ' injection and with a mechanically operated fuel-valve. In single-cylinder 

 work a B.M.E.P. of 115 lb. per square inch has been maintained, combined with 

 good economy and low maximum pressure, up to a mean piston speed of 2,200 ft. p.m. 

 This compares with a B.M.E.P. of about 70 lb. per square inch at 1,400 ft. p.m. in 

 some commercial engines and with 130 lb. per square inch at about 2,300 ft. p.m. in 

 aircraft petrol engines. 



The chief problem is that of producing a fuel-jet which penetrates the compressed 

 gas sufficiently to give adequate mixing of the fuel and air, and of obtaining at the 

 same time sufficient pulverisation of the fuel to promote its vaporisation and com- 

 bustion in the extremely short time available (about 3/l,00Oths of a second at 1,200 

 r.p.m.). When working at the maximum cyUnder output it has been found possible 

 to burn 70 per cent, of the oxygen in the cylinder. Owing to the extreme difiBculty 

 of getting perfect mixing and so of burning all the available oxygen, and also to the 

 nature of the combustion problem at high speeds, it is tolerably certain that the 

 compression-ignition engine can never compete on level terms with the petrol engine 

 as regards weight per H.P. Nevertheless, the potential benefit of eliminating aU the 

 complications and danger incidental to electric ignition and the possibility that 

 better economy may compensate for extra engine-weight on a long flight, make the 

 development of a compression-ignition aircraft engine an aim well worth striving for. 



9, Dr. G. D. Bengough and Mr. H. Sutton. — The Protection of 

 Aluminium and its Alloys against Corrosion hy Anodic Oxidation. 



A process has been developed for the protection of aluminium and its alloys against 

 corrosion b}^ means of a coating produced electrolytically. The article to be treated 

 is made the anode in a suitable electrolyte, and its potential is gradually raised in a 

 prescribed manner. 



The effect of the treatment is to produce on the aluminium or alloy a thin, hard, 

 and adherent coating of oxide or hydroxide of aluminium. The coating renders the 

 metal much more resistant to corrosion in the atmosphere, in contact with sea-water 

 and in other corrosive media ; it affords a good base for greases, paints, or enamels 

 which further increase resistance to corrosion ; it can be coloured by dyes. 



The process has proved to be capable of industrial application, and has already been 

 operated on a considerable scale. 



Tuesday, August 10. 



10. Prof. r. C. Lea. — The Effect of Superimposing a Torsional Stress on 



Repeated Bending Stresses. 



11. Mr. S. TiMOSHENKO. — Stress Concentration produced by Fillets and 



Holes. 



In many practical cases a very high stress concentration is produced by holes, 

 grooves, notches, and sharp variation in cross-sections. This stress concentration is 

 particularly undesu-able where materials undergo reversal of stress. The majority 

 of fractures in service can be attributed to progressive cracks starting from the regions 

 of high stress concentration. In this paper it is shown that the elementary theory of 

 tending of curved bars gives an approximate solution for the stress distribution in 

 the case of a plate with a circular hole with bead reinforcement. 



In the case of two-dimensional problems, the investigation of transparent models 

 in polarised light gives the complete picture of stress distribution. Applying this 

 method to the study of tlie stress at the root of gear-teeth, it is shown that with usual 

 proportions the maximum stress is about 1 .6 times higher than that as shown by the 

 usual beam formula. This root stress increases with decrease in radius of the fillet 

 at the tooth-root. 



