The following text may have been generated by Optical Character Recognition, with varying degrees of accuracy. Reader beware!

VOL. 39. CLEVELAND, JANUARY 14, 1909. NEW YORK No. 2 Naval Architects and Marine Engineers. The ninth paper read at the annual. meeting of the Society of Naval Archi- tects and Marine Engineers was on the "Service Test of the Steamship Har- vard' by Prof. ©.2 A, Peabody and W. S. Leland. lt. was. read-by.Mr. Leland as follows: The attention. of members is par- ticularly called to the fact that 'this test was run under the conditions ob- taining in actual service, no attempt having been made to approximate the ideal conditions existing on the usual trial-trip runs. We are much indebted to Mr. R. McGregor, general superintendent 'of the Metropolitan Steamship Co., for the personal interest he manifested in the test and for the assistance he freely gave. Observations were taken by Messrs. Young, Fisher, Hubbard and Davis, students of the mechanical engineer- ing department of the Massachusetts Institute of Technology, for thesis work; but the test was made under the direction of the department of naval architecture, and under the per- sonal supervision of the authors. The curves shown on Plates 1 and 2 were plotted by the authors and the results 'reported were computed by assistants in the naval architecture department. The engine test began at 9 P. M., when the engine had reached full power, and was stopred at 3 A. M., as we approached Nantucket Shoals. The last reading was taken at 2:55, when the first half-speed bell sounded, and the curves: exterpolated to 3 o'clock. The boiler test was begun at 7:10 P. M., and continued until 7 o'clock the next morning, the curves being exterpolated to make 12 hours. HOE (00,00; 1th on Pressures were read, as far as pos- sible, on the engine-room gages, which were: tested, for the occasion by the Crosby Steam Gauge & Valve Co. Where engine-room gages were not available, Institute gages tested at our own laboratories were used. The horsepower was determined by of the Denny and Johnson torsion meter, belonging to the de- partment of naval architecture at the Institute of Technology. The induc- tors were 50 set aS to. give a clear length' of shaft between inductors center shaft, 47.52 ft... starboard, and' 49:20) it:5 aport, which gave readings of approximately means 95, 75, and. 70 'vespectively at tull power. Inductors were set on all three shafts, but during the test the. port recorder failed to work at the end of two hours, due to a rupture of the connections. From the few readings thus obtained, and from readings tak- en on a preliminary run, when it did work satisfactorily, it appears that the torque readings on this shaft were 93 per cent of those on the star- board, and in computing the horse- this figure has been taken. This, of course, throws some little uncertainty into the calculation, but there is every reason to believe that the assumption is very nearly cor- rect. It would make practically 3 per cent error on this shaft to affect the total horsepower by so much as 1 per cent. The torsion meter can be read only to single units, so that this' of itself means an _ uncer- tainty in the last digit which corres- ponds to an accuracy of about 1% per cent. When, however, the read- ings are plotted at 10-minute inter- power vals, as is here done, and the results faired, the probable error is likely to be much less. In computing the horsepower 1.506 based on a torsional modulus of elasticity of 11,600,000 was used for the constant K. in' the formula 2. =P. Kd'rR = --------,, in which d=the diame- Ck ter of the shaft im inches (3 in), y=torsion meter reading, =revolu- tions per minute, inductor con- stant. (12.5), Dejdensth of sham -feet between inductors. The 'shaft was not twisted to de+ termine the exact constant, but since. 11,600,000 is the mean angle of a large number of tests it cannot be far from right. We find 11,000,000 and 12, 000,000 to be practically the minimum and maximum values, showing a vari- ation of about 3 per cent only in ex- treme cases. That the actual modulus of the Harvard's shaft should vary from the mean value by more than 11%4 per cent is extremely improbable. All in all, taking into account. the uncertainty of the port shaft, there *is every reason to believe that the horsepower is subject to an error less than that of the ordinary steam en- gine indicator. The water consumption was meas- ured by a 6-in. Hersey hot water meter loaned by the Hersey Manufac- turing Co., of South Boston. It was installed in the suction line between the hot well and the feed pump, and gave exceedingly satisfactory re- sults. This meter was previously calibrated by the Hersey Co. and in- dicated an under-run of 1 per cent under the conditions of the test. The curve -of meter readings was plot- ¢