chimney gases per pound of "combustible" were figured from samples taken in the combustion chamber. The general direction of the curve was determined by curve No. 1.

Low difference of draft readings occurs in two ways—by carrying a low-stack draft accompanied by a thin fire or by carrying a highstack draft and a thick fire. On most of these tests the stack draft is comparatively high as regards the amount of draft obtainable, hence the average points on the low difference of draft may be assumed to result from thick fires.

The use of a thick fire causes resistance to the passage of air through the fuel bed. When high-stack draft is used the difference between the pressure of the gases on the inside of the boiler setting and the atmospheric pressure outside is large, and thereby the absolute amount of leakage of air into the setting is increased. However, since, in general, high-stack draft has been carried, it follows that on the chart the per cent of leakage is a maximum at the points of low-draft differences. That is to say, as the draft difference increases the pounds of dry chimney gases per pound of combustible increase, the per cent of air leakage decreases, and more of the air comes through the fuel bed. This condition holds up to a point where the fire is thin, as indicated by a high-draft difference. The air now enters more freely through the fuel bed, increasing the rate of combustion, thereby decreasing the pounds of dry chimney gases per pound of "combustible." Another cause for the decrease of the pounds of dry chimney gases at the high-draft difference points is the very low per cent of air leakage. Thus it is shown that with two widely varying conditions we may obtain the same weight of dry chimney gases per pound of "combustible."

The curves have no bearing on efficiency. The variation in the slope of curves Nos. 1 and 2 is probably due to air leakage.

The following table gives figures on pounds of dry gases per pound of carbon, calculated from gas analyses taken simultaneously at different parts of the boiler:

Average per cent increase from surface of fire to combustion chamber, 52.33; from combustion chamber to front leg, 8.43; from front leg to stack, 21.25; from combustion chamber to stack, 28.48.

CLASSIFICATION OF COALS.

CARBON-HYDROGEN RATIO IN AIR-DRIED COAL.

In the belief that coals of approximately the same proportions of constituents, as shown by chemical analysis, should behave more nearly alike when subjected to similar treatment than coals in which the proportions of the constituents vary widely, the following attempts have been made to classify the coals burned by the boiler division, according to their performance under a boiler.

In accordance with the method proposed by Marius R. Campbell," the ratios of carbon to hydrogen in the ultimate analyses of airdried car samples were computed. One hundred and seventy-four tests on about 75 different coals were employed for this purpose, washed, dried, and briquetted coals being rejected. It was not expected that this classification (fig. 37) would be of as much service as one based on the same ratio of either coal as received or dry coal, since the amount of drying to which a sample is subjected is dependent on the local conditions, which vary from day to day. The statement just made seems to be borne out by the fact that the average

a Prof. Paper U. S. Geol. Survey No. 48, 1906, pp. 156-173.

points from which the general efficiency curve (No. 1) is derived fall farther from such a curve than is the case with general curves derived from the ratio of carbon to hydrogen, based on either coal as received or dry coal.

The ratio of carbon to hydrogen should have been obtained from the ultimate analysis of the boiler-test sample after air drying, but these data were not available. The ratio used is calculated on the

10

NO. I

8

No. 2

65

FIG. 37.--Coal classification based on carbon-hydrogen ratio from ultimate analysis of air-dried car samples as related to efficiency 72*, tests 120-401 (curve No. 1); per cent of rated capacity developed feurve No. 2); and efficiency 72* at about rated capacity, tests 1-78 (curve No. 3). General curves, omitting tests of washed, dried, and briquetted coals.

air-dried coal of the car sample-a fact which might account for some of the variations of some of the points from the curve.

The broken line, curve No. 2, is drawn through the average per cent ratings for all tests coming in each group. The per cent rating attained does not seem to influence the efficiency, as it does in other classifications which follow.

Curve No. 3 shows the relations between carbon-hydrogen ratios on air-dried coal and efficiency from boiler tests conducted during 1904. The tests in which washed and briquetted coals were burned are omitted. It was decided to abandon this classification and to use the ratio of carbon to hydrogen of coal, as received, and of dry coal. The curves and conclusions on classifications based on carbon-hydrogen ratios of coal, as received, and of dry coal, were derived from approximately 250 boiler tests run in one series. The tests included raw, washed, briquetted, and a few mixed coals. The relative values of the classifications were based on the uniformity of the general efficiency curves, efficiency here meaning the efficiency of the boiler on the combustible basis (the combustible consumed, as shown by proximate analysis, less the combustible lost through the grate, as determined by an analysis of the refuse). The item used for the efficiency is 72* of the A. S. M. E. code.

CARBON-HYDROGEN RATIO IN COAL AS RECEIVED.

The carbon-hydrogen ratios of fig. 38 were computed from the ultimate analyses made by the chemical division on boiler-test samples. These ratios were divided into six groups, as follows: Group K, including all values up to 11.2; group L, from 11.2 to 12.5; group M, from 12.5 to 14.5; group N, from 14.5 to 15.5; group O, from 15.5 to 17.0, and group P, all values over 17.0. The several items in each group were then averaged and curves plotted on the basis of (1) efficiency 72*; (2) per cent of CO loss; (3) combustion-chamber temperature (by Wanner optical pyrometer); (4) per cent of rated capacity of boiler; (5) pounds of dry chimney gases per pound of "combustible" (coal free from ash and moisture), and (6) B. t. u. per pound of dry coal.

A comparison of the averages just mentioned shows that no one item was maintained, even approximately, constant, so that whatever conclusions may be drawn therefrom must necessarily be considered anything but absolute. Particular mention should be made of the per cent of rated capacity attained in the trials, which, if maintained constant, might have aided materially in defining more closely the relative causes and effects as indicated by the other items, although it is believed that these items are as much dependent on the kind of coal as on the per cent of rated capacity attained.

The averages of all of the like items of each group are graphically represented on the accompanying charts (figs. 39-45), on which the ratio of carbon to hydrogen appears as an abscissa.

The curves of efficiency 72* in figs. 38 and 39 do not agree; but it is hardly to be expected that items should vary alike when tests are classified on such an illogical basis as the carbon-hydrogen ratio in coal as fired, inasmuch as the hydrogen of the free moisture is also included.

FIG. 38-Coal classification based on carbon-hydrogen ratio as related to theoretical relative efficiency of boiler as a heat absorber at various temperatures (curve No. 1); per cent of CO loss (curve No. 2); efficiency 72* (curve No. 3); per cent of rated capacity developed (curve No. 4); combustion-chamber temperature (°F.) (curve No. 5); B. t. u. per pound of dry coal (curve No. 6); and pounds of dry chimney gases per pound of "combustible" (curve No. 7). General curves, samples as received, tests 120-380.

The following list gives the coals appearing in each group and their respective carbon-hydrogen ratios:

Average carbon-hydrogen ratios determined from analysis of boiler-test samples of coal as fired, including moisture.