Page images

per pound coal and the weight charged, both figured to a moisture and ash free basis, is shown graphically in Plate II. The only point, which is far off the curves in any of the tests, is the one for test 35, made on the Illinois coal with a retort heavily coated with carbon and at an extremely low temperature.

If it be granted that the yield of gas from each retort is a constant, which is the product of the weight of the charge and the rate of gas evolved per pound of coal, the equation becomes of the form xy=c, which gives a hyperbola of the general form shown by the small curve in the upper corner of Plate II.



[ocr errors]



PLATE I. PLATE II. Relation between Weight of Coal in Charge and Gas per Hour

from One Pound of Coal.

The range of our experiments covers only a small portion of this hyperbolic curve, but the data as plotted in Plate II show five roughly parallel curves, whose general direction indicates a harmony with the mathematical expression. The data are not

[merged small][merged small][merged small][ocr errors][merged small][graphic][subsumed][ocr errors][ocr errors][ocr errors][subsumed][ocr errors][ocr errors][ocr errors][subsumed][ocr errors][ocr errors][ocr errors][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][ocr errors][ocr errors][subsumed][subsumed][subsumed][subsumed][ocr errors][subsumed][subsumed][subsumed]

PLATE W. PLATE III. Curves showing Lack of Relation between Weight of Coal

Charged and Quality and Quantity of Products.

accurate enough to make it worth while to try to determine the extent of the deviation.

There is not sufficient evidence to show whether a rapid or a slow rate of gas evolution per pound of coal is better for coinmercial purposes.

INFLUENCE OF Weight OF CHARGE. A special effort was made to determine the influence of size of charge. Each of the five coals mentioned in this report was tested on a charge of 300, 400 and 500 pounds, the three tests being made consecutively on a single long working day, in order that variables due to the condition of the retort and to the firing mighi be eliminated so far as possible. It is unfortunate that pyroinetric troubles, mentioned elsewhere, prevented temperature measurements being taken on some of these tests, but it is not unreasonable to assume that ordinarily the temperature of the retort remained fairly constant during the day. The results are given in figures in Table I-V, and are summarized graphically in Plate IlI.

The percentage yield of coke is largest for the heavy charge, and decreases for the light charges, except in the case of the Pittsburgh coal, where the 300 pound charge is reported with a higher percentage yield of coke than the 400. The difficulty of raking the retorts perfectly clean, the danger of spilling in drawing the coke, and the necessity of using the less sensitive scales on the retort house floor for weighing the hot coke, introduce a considerable probable error into these figures for coke, but the evidence is sufficiently in one direction to make it probable that the heavy charges do give a higher percentage of coke than light ones. This may be due to reactions within the retort, but is perhaps more probably to be attributed to the fact that the loss in drawing, etc., is a relatively constant weight, and hence appears as a higher percentage on the light charges.

No constant effect of size of charge can be traced in the quantity of gas produced, nor in its candie power or heating value. The curves of Plate III show simply a haphazard variation in yield of gas, candle power, heating value, candle feet, and B. T. U. in gas from one pound of coal. The yield of tar, ammoniacal liquor, NH3 and unaccounted for loss varies also without any uniformity. It is apparent that there is some other variable much more important than the size of the charge.


In our tests we have desired to keep the retort under a back pressure which would not be over one-tenth of an inch of water. The gauges were closely watched, and every effort made to keep the pressure constant, but we are free to confess that the results were not altogether satisfactory, although we believe that our control of the pressure was much better than is usual in retort house practice. We have no means of estimating quantitatively the nican variation in pressure for one test as compared with another, but there is evidence that it has been an important unaccounted for variable. Our reasons which are purely theoretical, are as follows:

A retort has a porous body readily permeable to gases. Under the usual manufacturing conditions, where a slight pressure is kept on the inside of the retort, some of the gas is forced out into the fire space and lost. This gas becomes badly cracked as it strikes the hot wall of the retort, and builds up the dense wall of carbon which accumulates on the inside of the retort. Conversely, the presence of this dense coat of carbon is proof that gases are being forced out of the retort. If gases were being sucked in, the inside of the retort would be clean. If there were no interchange in either direction, at the most only a thin scale of carbon could form.

Let us consider how much variation from the mean amount of gas there is to be accounted for, and, to eliminate the influence of ash and moisture, calculate the figures on a basis of coal free from moisture and ash.

Number Tests Cu. ft. gas per 1b. Max'm Deviar'n Name of Coal Averaged Coal, ni. & a. free from Average Illinois


5.0% Kentucky

5.30 Pennsylvania Tennessee

5.79 West Virginia





This shows that in any of our tests a loss of 7.5 per cent. of gas through the retort wall, or a gain, through sucking 7.5 per cent. of smokc gases into the retort, would have been sufficient to account for all the variation from the average yield of gas. It is not probable that 7.5 per cent. of sinoke gases were sucked into the retort in any test, but it is entirely possible that more than 7.5 per cent. may have been forced out. We have no data on which to base an estimate of the amount of gas which might be forced through a retort by a pressure of one-tenth of an inch of water, and our conclusion as to the probable importance of this variable is reached mainly because it seems to be the only cause whose unknown variation is sufficient to account for the results. Indirect evidence of the correctness of our view is afforded by the close agreement of the figures for heat value in the gas from a pound of coal—a value where the retort pressure is of less importance. This will be discussed in a following section.


The question of influence of temperature nlust be treated with many reservations. It is extremely difficult to measure retort temperature. The readings of a thermocouple placed inside of the retort are almost valueless, since they record only the immediate local surroundings of the couple, which may be couled by contact with a projecting piece of coal or be unduly heated through proximity to the roof of the retort. We have found the temperature of the outside of the retort to give the most reliable indications, but have had vexatious and expensive experiences in attempts to measure it. The retort which we use is in the end bench of the battery, and there is an inch hole in the masonry opposite the middle of the retort, designed for the insertion of a thermocouple. The temperature to be measured runs as high as 2200 degrees F. A platinum thermocouple must be protected absolutely from the furnace gases, and it is difficult to find a satisfactory protective covering for continuous service. The Hoskins thermocouple has been equally unsatisfactory on account of the tendency to break when the wires contract on cooling. It was difficult to use optical pyrometers on account of the presence of smoke in the peep hole, but this was overcome by luting into

« ՆախորդըՇարունակել »