NACA Report No. 108

NACA Report 108 - Factors of Airplane Engine Performance was issued by the United States National Advisory Committee for Aeronautics in 1921.

Summary
NACA Report 108 is an analysis of airplane engine tests made at the Bureau of Standards, attempting to identify fundamental relations between the many variables of engine operation. The engines were the Liberty L-12 and three models of the Hispano-Suiza 8F. The tests were made in the altitude chamber, where conditions simulated altitudes up to about 30,000 feet, with engine speeds from 1,200 to 2,200 rpm. The compression ratios of the different engines ranged from under 5:1 to over 8:1. The data taken on the tests were exceptionally complete, including many pressures and temperatures, besides the brake and friction torques, rates of fuel and air consumption, the jacket and exhaust heat losses.

With the Liberty engine, operating at 500 to 2,000 rpm and with the Hispano-Suiza 300 hp operating at 1,400 to 2,200, it was found that the friction torque increases approximately as a linear function of engine speed at a given air density, and approximately as a linear function of density at a constant speed. This means that the friction horsepower increases approximately as the square of the speed. Actually the reIation of torque and speed is such that the friction horsepower increases with speed raised to a power between the first and second, this power increasing with speed, approaching the square. The relation depends upon the engine design and speed and the air density. Any statements as to the distribution of the friction losses are based upon incomplete evidence; the indications are, however, that the pumping losses are about half of the total friction. There is no doubt that for a given process of combustion and at a constant speed the engine power is directly proportional to the weight of charge supplied; in other words, proportional to the charge density at the beginning of compression. As a consequence, if operating conditions are sensibly constant except for altitude, the engine power -will be closely proportional to the air density. The volumetric efficiency increases with increase of air temperature at constant pressure, so that power does not decrease as fast as the air density when the temperature is raised, due to changes in vaporization and heat transfer.

In order to compare the action of the gasoline engine with the theoretically perfect heat engine operating on the same cycle, it is necessary to base the heat balance of the actual engine on the heat actually made available by combustion, not upon the heat supplied in the fuel. By summing up the exhaust and jacket heat losses with the brake power, an approximation is made of the true heat available, accurate to within perhaps 5%. Basing the heat balance upon the heat thus accounted for, it is fairly well-established that the energy distribution is not appreciably altered by change of either altitude or speed. It is, of course, altered by compression ratio changes. The exhaust heat, as a percent of the heat accounted for, is practically the same as the theoretical rejection of heat computed for the same compression ratio.

An empirical and approximate statement, applicable to the speed and density ranges herein covered, is that the friction torque (m. e. p.) increases as a linear function of speed at Q constant, air density, and as a linear function of air density at a constant speed. This applies to both the Liberty 12-cylinder and the Hispano-Suiza 8-cylinder engines. The true relations of friction torque, engine speed, and air density are much more complex and cannot be deduced from the available information. It is known, however, that the friction torque is such that the friction horsepower wdl increase with engine speed raised to between the first and second power, approaching the square of the speed at the higher speeds. The exact relation is dependent on the engine, as well as on the speed and air density. Also, at a given speed, the friction horsepower increases with some varying power of the air density because of the change of the pumping work. The friction horsepower at constant speed is found to increase slightly more rapidly than does the density.

No systematic change of friction could be connected with change of compression ratio, from the data used in this report, although a slight change is to be expected.

Conclusions
The indicated mean effective pressure of the Hispano-Suiza engine is slightly greater than that of the Liberty 12, at the same air density and compression ration. This is explained by the less amount of heating with the H-S manifold jackets. When the air supply is at 32 °F the manifold temperature of the H-S is about 14 °F, while that of the Liberty is about 41 °F. For a given temperature and a given pressure of air supply, the actual density of the air entering the cylinders is reduced by heating, although it is also increased by the cooling resulting from the evaporation of the gasoIine. Using the manifold density instead of the supply density, the two engines give the same mean effective pressure for the same density. Apparently the relation between indicated mean effective pressure and air supply density is based upon the fact that the change of indicated mean effective pressure is proportional to the change of density of the charge at the beginning of compression, irrespective of whether the density is changed by altitude or by throttling.