Davis and English 



modulus. Several materials were considered for the Bras d'Or application, in- 

 cluding titanium, beryllium nickel, and various steels. The eventual choice, 

 however, was for forged billets of a high nickel chromium alloy known as Inconel 

 718. This material possesses excellent physical properties; its main disadvan- 

 tage is that it is difficult to fabricate, since it is an extremely tough alloy. Its 

 main advantages over titanium are its higher elastic modulus which is nearly 

 double that of titanium and its superior fatigue life characteristics, particularly 

 in sea water. A screw made from Inconel 718 would therefore deflect about half 

 as much as a similar screw in titanium and would presumably last longer. 



In line with the choice of an exotic material for the propeller construction, 

 the blades have been made separate from the hub and are retained by bolts, thus 

 enabling a blade to be changed without a complete screw replacement in the 

 event of damage. 



It was essential that the propeller blades were forged separately, as it 

 would be impossible to forge a single propeller of this size in this alloy, i.e., 

 4-ft. diameter and weighing 1000 Ibf finished. Thus the blades are bolted to the 

 hub using special Inco 718 bolts. This method of construction therefore has 

 some merit from both the manufacturing and blade replacement points of view. 



Structural Testing on Static Models 



Structural testing has been conducted at De Havilland by Mr. S. Morita 

 (Refs. 14 and 15), to explore the effects of blade geometry and external blade 

 loading on the stresses and deflections induced in the blades. The experiments 

 were performed in two stages, the first of which employed simple plane models 

 without pitch, while the second used a more representative model in which pitch 

 was included. In both cases static models of single blades were used and hence 

 centrifugal effects were not included. Only a brief survey of the experiments 

 and the results is given here. 



The simple models were used to aid in the rapid production of results and 

 to permit a large coverage of parameters to be made relatively easily. These 

 models consisted of cantilevered, trapezium-shaped aluminum specimens rep- 

 resenting the blade under test. Pitch was not incorporated in these specimens. 

 The specimens had a common radial distribution of chord widths, but three val- 

 ues of leading edge sweepback were covered, while two variations in section 

 shape were used. The sections were either or triangular shape with blunt trail- 

 ing edges or triangular apart from a trailing edge chamfer on the back. Camber 

 was not included in the sections. The radial blade loading was the same in all 

 the tests, but two chordwise loading distributions were used in which the centre 

 of pressure position was at 39 percent and 25 percent of the chord from the 

 leading edges. 



In all the experiments the external load was applied using rubber pads 

 bonded to the specimen and loaded in tension through a system of cables, pul- 

 leys, and "whiffle trees" by a single hydraulic jack. Such a system is commonly 

 used in the aircraft industry, and Fig. 22 shows the test setup em.ployed. Strain 

 measurements were made using strain gauges attached to the "wetted" side of 

 the blades, while blade deflections were made with dial gauges. 



988 • 



