Required the water-rate of a simple Corliss engine with 80 pounds steam by gauge, and working against 2 pounds absolute back pressure, that is with a vacuum of 10 pounds. Rated horsepower of the engine supposed to be 100.

The values of a would be taken as follows, when the engine is loaded to 10 per cent. of its rated capacity x = 10, at 25 per cent. x = 4, at 125 per cent x = 1.25.

A computation made in this manner, gives the probable water consumption in pounds per 1. H. P. for various loads, and classes of engines, which are given in the following table.

From the figures given, it will be noted that the increase in water consumption is small, when the extreme variation in load is confined between the limits of one half, and one and one half the load which gives the best results, and that the water consumption increases much more rapidly for light than heavy loads.

TABLE XXI.-PROBABLE WATER CONSUMPTION FOR VARIOUS CLASSES OF ENGINES WITH DIFFERENT LOADS.

In conclusion the writer wishes to say that the paper which has been presented is to be considered simply as the result of a study of a large number of tests of actual engines working under various conditions. It differs from previous papers principally in the fact that the basis of all the results is a comparison. of actual, with the theoretical steam consumption of an engine, which converts all heat received in the steam and not discharged with the liquid into the exhaust, into work.

FIG. 3. Curves of Probable Water Consumption for Various Classes of Engines with Different Loads.

The author is well aware that this base of comparison may be open to criticism, and at present is willing to say that the only reason for its use, is the fact which has already been shown, that it gives nearly a constant ratio when compared with the result of the actual engine, and this was far from being true, when the engine was considered as working in a reversible cycle.

As to the final computations, it may be said that their value

depends upon agreement with actual tests. They are of course to be considered as only average results applying to well built engines, since they do not take into account differences due to poor or good construction, which in practice is of great importance, and may give results which differ 10 per cent., or even more, from those given.1

In addition to what has been already stated, it may be said that the general result of my study would indicate that the more economical an engine, even when loaded at its best ratio of expansion, the less will be the change in economy, due to increase or decrease of load, and I am fully satisfied that the results of actual tests provide a strong argument in favor of the most economical type of engine, even for conditions under which the load varies rapidly, and through wide limits.

The coefficients which I have been able to gather by compilation of tests are no doubt subject to modification and correction, still in the form presented I believe they may prove useful in pointing out the probable economy of an engine with its most economical load, and its modification due to change of load. It should be understood that the only claim for consideration that these values possess is due to agreement with the results of reliable and careful tests.

FRICTION OF ENGINES, ECONOMY ON BASIS OF DELIVERED

HORSE-POWER.

The results given in the paper are based on the work of the steam, as shown by the indicated horse-power developed, but in many cases it is essential to know the effect on the economy, caused by variation in delivered horse-power.

This result could easily be obtained from the preceding table, and the frictional work for various classes of engines known,

1. As illustrating the great difference due to excellent construction, the author would state that since writing the above he has conducted a test of 24 hours duration, of a Corliss triple expansion pumping engine, working between the limits of 136 pounds absolute steam pressure and 1.2 pounds back pressure, in which the actual water-rate was 11.79 pounds. The standard engine for this case (not given in table) would require 8 pounds, so that the equation is 8+ 3.76 V, and the ratio r is 3.79 ÷ 8 = .47; a result very much below that of any triple expansion engine tabulated, and also probably the best result ever attained with any engine of any kind.

The probable water-rate for this engine for different loads would be as follows:

because it is a thoroughly well established fact in steam engineering that the friction of an engine remains practically constant regardless of the work (see Thurston Manual Steam Engine, vol. i, art. 132-3), and is in every case essentially the same as that shown by the indicator diagram when the engine is working without load.

To determine the amount of friction that may reasonably be expected for various classes of engines, the author has gathered the following records of actual tests. These records are not as numerous as possibly might be desired, due, no doubt, to the great difficulty of obtaining an accurate record of the friction horse-power of large engines.

Average 12.3 per cent.

Deduct probable friction of air pump 3 per cent., gives net

friction 9.3 per cent.

Trials were made by Walther, Meniner and Ludwig to determine the friction of a compound engine run in various ways.

The results bring, Compound engine with condenser and air pump. High-pressure cylinder with condenser and air pump.

High-pressure cylinder without condenser.

་

248.97 I. H. P., 39.48 friction H. p.

Per cent. of friction 13.7.

153.1 1. H. P., 24.74 friction H. P.

Per cent. of friction 16.0. 128.4 I. H. P., 17.49 Friction H. P. Per cent. of friction 12.0.

It may perhaps be unsafe to draw conclusions from such a limited number of facts, but a study of the tests cited, show at once that there is no great difference as to the amount of friction in various classes of engines, the increased range of steam pressure and the greater return in work, being sufficient to make ample compensation for the extra work required to move the more complex mechanism of the compound or triple expansion engine.

The friction varies to a considerable extent in different engines; in general it is less in large than small engines.

It seems to be in every case between 5 and 9 per cent. of the rated load for non-condensing engines of 150 horse-power or over, whether simple, compound or triple, and to lie between 6 and 12 per cent. for condensing engines from 50 to 150 horsepower.

The condensing engines, to which an air pump is attached, usually have 2 to 3 per cent. greater friction than those without an air pump.

The average friction may be taken for engines of 50 to 150 horse-power as 10 per cent. non-condensing, and 13 per cent. condensing, those exceeding 150 H. P. as 7 per cent. noncondensing, and 10 per cent. condensing, with very little error. The following special test may be of interest as bearing on this subject

EXPERIMENT WITH SIBLEY COLLEGE EXPERIMENTAL ENGINE. Sibley College has an experimental engine of the Corliss type, with three cylinders respectively, 9, 16 and 24 inches in dia