July, 1931 AIRCRAFT ENGINEERING 171 An Account of Experiments on the Precision of the U.S. Army Air Force Specification By Graham Edgar† HE N the U.S. Army Air Force adopted its Th e Information Sought specification for anti-knock petrol Th e determinations made were: (1) the octane W (Y-3557-B) involving th e use of high jacket numbe r of each fuel; and (2) the amount of temperatures, relativel y littl e information was avail Tetracthy l Lead necessary to make each fuel equal able concerning the precision of the test, or the 87 octane. The octane number determinations agreement to be expected among different labor were made against actual octane and heptane atories. A short co-operative programme was blends in all cases. arranged, therefore, among several laboratories interested in the subject, to determine for the types I t was recommended that the knock intensity of fuel likely to be employed in meeting the Army employed bo that given by an 87 octane number specification the agreement to bo expected from fuel at a compression pressure of about 150 lb. different laboratories using the specified technique Several laboratories also mad e determinations under and a brief study of some of the variables likely to somewha t different conditions of test in order to be met. get some idea of the influence of different variables The co-operating laboratories were the U.S. on the determinations. Army Air Force, the Ethyl Gasoline Corporation's Humidity , air temperature, and barometer were laboratories a t Detroit and Yonkers, the Standard reported , although no attemp t was made t o control Oil Company of California, the Standard Oil thes e variables except in the case of the U.S. Army Company of Indiana, and the Standard Oil Com Ai r Force data, in which the inlet air was dried with pany of New Jersey. calcium chloride. Th e Fuels Studied Th e main experimental data are given in Table I, The fuels studied were : (a) a sample of straight- togethe r with average figures and deviations from rim California aviation petrol; (b) sample (a) plus th e average . The U. S. Arm y dat a arc omitte d from 10 per cent of benzol ; (c) a sample of straight-run th e averages, since their engine was not standard Smackover petrol, containing about 10 per cent of an d their inlet air was dried. Mid-Continent petrol; (d) sample (c) plus 30 per I t will be noted that sample (c) shows wide cent of benzol; and (c) a sample of 100 per cent variation in th e lead requirements to equal 87 octane. cracked petrol from a Mid-Continent crude. A comparison of the gravity and volatility data The Ethyl " Series 30 " engine was employed in from different laboratories indicated that this all tests, as this is specified by th e Army Air Force. sampl e was not uniform, and later exchange of All engines were standard except tha t used by the retaine d samples confirmed the fact, so tha t knock Air Force, which wa s fitted with thermocouples and rating s on sample (c) must be disregarded as far special gasket arrangements. (It was later found a s agreement among different laboratories is con that the valve adjustment was also not standard.) cerned. The engine conditions were:— Speed—600 r.p.m. Th e following additional data were obtained:— (c) The U. S. Army Air Force carried out some Spark advance—14 deg. (a) The Standard Oil Company of California experiment s at other speeds than 600 r.p.m., and Fuel-air ratio for maximum knock. an d the Standard Oil Company of Indiana made also mad e some measurements with the temperature Jacke t temperature—300 deg. F . some determinations at 22 deg. spark advance, as plu g instead of the bouncing pin. In general, the Compression Pressure at full throttle—190- well as at 14 deg., and obtained results essentially sam e order of results was obtained at 600 r.p.m. 200 lb . per sq. in. th e same. wit h both bouncing pin and temperature plug, Knock evaluated b y the Bouncing Pin. (b) The Yonkers Laboratory of the Ethyl bu t change in speed altered some of the ratings * A paper read during the Nineteenth National Aeronautic Gasoline Corporation determined th e octane number- appreciably, and a t 900 r.p.m. the temperature plug Meeting of the S.A.E., held in Detroit, April 15-16, 1931. lead requirement curves for the fuels, up to an d bouncing pin did not always agree. † Dr. Edgar is Director of Research of the Ethyl Gasoline 87 octane number. These dat a are shown in Fig . 1. Corporation. Goo d Agreement Obtained TABLE I. Fo r fuels a, b and d, representing straight-run fuels and benzol blends, the results of the different Octan e Octan e Octan e Octan e Engine Octan e laboratories agree to an average deviation of about Sampl e Sampl e Sampl e Sampl e Laborator y Sampl e Numbe r Numbe r Numbe r Numbe r Numbe r Numbe r one-half an octane number, and a maximum deviation of abou t one octane number. The agree a 74 — b 7 5 c 71 d 80 — e 73 S.O. Cal. ... ... ... 27 men t of different engines in the same laboratory is b 7 5 c 71 d 80 + e 77 S.O. Cal. ... ... ... 51 a 74 — distinctly better than this. On sample (c), 100 per S.O. N.J. ... ... ... 1 a 74 — b 75 — c 72 d 7 8 e 73 — 2 a 74 b 75 + c 72 d 78 + e 73 + cen t cracked stock, the agreement is not so good, S.O. N.I. ... ... ... a b 76 c 72 d 78 e 72 S.O. Ind. ... ... ... 74•5 an d in one instance a discrepancy of four octane b 75•5 c 71 d 79•5 e 72•5 E.G. Corp. (D) ... ... a 74 — number s appears, although it is always possible E.G. Corp. (Y) ... ... 67 a 74 b 76 c •72 d 79•5 e 71 tha t some undetected gross error affected this U.S.A.F ... ... ... 26 a 7 5 b 76•5 c 73•5 d 80•5 e 71 particula r determination. a 75•4 c d 79 e 73•1 Average ... ... 74•1 — — — Th e effect of moderate variations in compression pressure does not seem large, and it is not possible t o show any definite effect of air temperature or humidity . c.c. Lead c.c. Lead c.c. Lead c.c. Lead c.c. Lead t o t o t o t o I t was noted that the time lag in attaining t o Equa l Equa l Equa l Equa l Engine Equa l equilibrium was sometimes several minutes when Sampl e Sampl e Sampl e Sampl e Sampl e 87 87 87 87 87 Laborator y Numbe r shifting from a lead-treated fuel to heptane-octane, Octan e Octan e Octan e Octan e Octan e Numbe r Numbe r Numbe r Numbe r Numbe r bu t there was little lag when shifting from an untreate d fuel to heptane-octane. 2•5 0•85 27 a 1•5 b 1•1 c d e 5•0 S.O. Cal. ... ... ... 51 a 1•5 b 1•1 c 2•6 d 0•75 e 4•0 S.O. Cal. ... ... ... Th e fuels show great sensitivity to changes in 1 a 1•5 b 1•2 c 2•8 d 1•1 e 4•2 S.O. N.J. ... ... ... fuel level in the carburettor, and the fuel-air ratio a 1•6 b 1•2 c 2•9 d 0•8 e 3•8 S.O. N. J. ... ... ... mus t be adjusted with extreme care. a 1•6 b 1•1 c 2•5 d 1•25 e 4•5 S.O. Ind. ... ... ... E.G. Corp. (D) ... ... a 1•8 b 1•4 4•9 d 1•1 e 5•2 — Altogether, considering that most of the labor 32 a 1•6 b 1•4 4•2 d 0•95 e 4•7 E.G. Corp. (Y) ... ... c a 1•8 b 1•3 c 4•2 d 1•15 e 4•5 atorie s have had relatively little experience with E.G. Corp. (Y) ... ... 6 7 a 1•7 b 1•15 c 2•3 d 0•9 e 6•5 U.S.A.F. ... ... ... 26 high temperature ratings of aircraft fuels, the precision of the determinations seems reasonably a 1•61 b 1•23 c d 1•0 e 4•8 Averag e ... ... — — satisfactory.
Aircraft Engineering and Aerospace Technology – Emerald Publishing
Published: Jul 1, 1931