« ՆախորդըՇարունակել »
INTRODUCTION. The present paper is the tenth to be presented to the Michigan Gas Association, as a result of your support of a fellowship in Gas Engineering at the University of Michigan. The paper presented last year gave a description of the Experimental Station for gas manufacture and of the methods of conducting tests. The general scheme of the experimental plant has not been altered since the description a year ago, but details of arrangement have been greatly changed. The building has been enlarged to double its former size, and the apparatus rearranged so as to be more accessible and more readily controlled. The same retort has been used throughout the two years of the tests. The most important addition to the equipment has been a Lloyd Control for constant retort pressure—a gift from the Lloyd Construction Company, which we gratefully acknowledge. It was installed too late to be used in the work reported here. The Technologic Branch of the U. S. Geological Survey has co-operated in the work by making chemical analyses and determining heat value of the coals and cokes used in the tests, as it did last year.
The paper presented last year* reported eighteen tests on eleven different coals of widely different composition from all sections of the United States. It was recognized that it was unsafe to draw conclusions from so few tests, and the paper was almost entirely descriptive.
The work of the present year has been an extension of that of last year, but it has been confine:1 to a smaller number of coals, and an attempt has been made to determine the effect of variables in operating conditions. The samples of coal were portions of the identical lots used last year. Most of them had been stored in sacks under cover, but two of them had been stored in open bins exposed to the weather. This effect of weathering furnished one variable to be studied. Another variable was introduced by substituting for the 400 pound charge uniformly used last year
* Studies in the Manufacture of Coal Gas (Second Paper) by Alfred H. White, John H. Wyman, William A. Dunkley, and Perry Barker. Proceedings, Michigan Gas Associa
charges of 300), 400, 500 and even 600 pounds. These were the only two variables purposely introduced, but the retort temperature was unavoidably a variable, and, as will be shown later, so was the pressure on the retort.
The results of the tests failed to show any definite effect due to weathering. This is an important question, and it is planned to inaugurate during the coming year a weathering test which shall extend over a period of five years.
PRESENTATION OF DATA,
The five coals on which report is made come from the states of Illinois, Kentucky, Pennsylvania, Tennessee and West Virginia. At least one test on each of the coals except those from West Virginia was reported last year. There were then presented full data and curves showing the composition of the coals and the changes taking place during distillation.
The curves obtained this year were similar to those of last year, and are omitted from this report. There are here presented summarized data showing the average composition of each coal and the data of the tests which are used in the discussions. Tests 18 to 35 belong in last year's series and those with higher numbers in this year's. There is also presented a more detailed discussion and representation of the tests on an individual coal to serve as an introduction. The coal chosen is that from Hellier, Kentucky, whose tests afford a wide range of conditions.
COAL FROM HELLIER, KENTUCKY. A carload of this coal (U. S. G. No. A. A. 3) was received in January, 1909, and a portion of it was stored in a bin exposed to the weather. Five tests were made on it, two on charges of 400 pounds, and one each on a charge of 500, 360 and 300 pounds, Tests 21 and 30 were made in the spring of 1909. Tests 54, 55, and 56 were made consecutively in one day on the same coal a year later, it having been exposed all that time to the weather. Tests 21 and 30 were made in a retort heated to over 2000 degrees F, while the other three were distilled at about 1850 degrees There were thus many variables which might affect the results.
Condensed data are given in Table I, and a graphic representation of the changes in quality and quantity of gas evolved is given in Plate I. The average result of the five tests is as follows: Coke, per cent.
..63.9 Gas, cu. ft. per lb. coal
14.7 Candle feet
.. 74.4 B. T. U. per cu. ft. of gas
.642.0 B. T. U. in gas from one pound of coal .3261.0 NH3, lbs. per ton coal
.4.17 Tar, gallons per ton coal
1.11.2 The figures of Table I show a great lack of uniformity, The yield of gas per pound of coal varies from 4.72 to 5.29. Eliminating the ash and moisture and calculating the weight of coal to a inoisture and ash free basis, does not bring the figures into much closer agreement, the variation being 5.15 to 5.60 cu. ft. The yield is not apparently a function of temperature nor of the weight of charge. The question will be taken up later, but the variation may be considered provisionally as due to some unknown
The candle power of the gas varies from 14.1 to 15.5, and the candle feet from 69.7 to 78.9 without any adequate explanation being apparent. The heating value of the gas is only slightly more constant, varying from 622 to 662 B. T. U. per cubic foot. The nunber of heat units in the gas derived from one pound of coal varies from 3110 to 3329 B. T. U. There is an apparent relation between this value and the retort temperature—a point which will be taken up more fully later. The yield of NH3 per ton coal varies from 4.02 to 4.44 pounds with no evident explanation.
The curves of Plate I indicate a much greater uniformity. The curves of candle power and heat value decrease with reasonable regularity, and the gas yield is regular. The curves are introduced mainly to show these points which are characteristic of all the tests on all the coals, the results on which the various tests agree rather than those on which they differ. Particular attention is drawn to the evenness of the quantity of gas evolved from a retort during a half hour, independent of the weight of the coal contained in it. Attention is also directed to the corresponding set of curves showing rate of gas evolution per half hour, referred to a ton of coal as a basis. This last series of curves shows a systematic difference with weight of charge. They are characteristic of the tests on the other coals which are not as fully set forth. These points are taken up more fully further on
PLATE I. Typical Curves Showing Amount and Quality of Gas evolved
from a Charge of Coal in Successive Half-hour Intervals.
DATA OF TESTS ON COAL AA3 FROM HELLIER, KENTUCKY
Data on Tests:
Coal charged, Pounds.
Yields per pound coal as charged
Gas, cu. ft. corrected.
4.72 14.5 69.7 650. 3126.
5.00 14.1 70.5 622. 3110.
5.29 14.2 75.2 630. 3340.
5.17 15.3 77.6 649. 3360.
5.09 15.5 78.9 662. 3370.
Yields per pound coal, M & A free
Gas, cu. ft. corrected.
NH3 pounds per ton coal, as charged..
Duration of carbonization, hours.