November. 1928
I N D U S T R I A L A N D ENGINEERING CHEMISTRY
the switch, which causes the pump motor to start or stop when the contents of the sump reach a predetermined level. Other applications not illustrated include the starting and stopping of the motor of a unit heater; to open or close a steam supply to unit heaters; for the control of the temperature of a drying tumbler, where in addition to stopping the motor which operates the tumbler, the blower to the heater may also be stopped and an alarm sounded to notify the attendant that the tumbler should be emptied. Liquids heat)ed by steam or by hot-water coils may be kept a t an even temperature by such devices and installed therefore to control cookers, stills, washers, and the like. High-pressure booster lines to instantaneous heaters using exhaust steam are also subject to such control, and when the temperature requirements exceed that obtainable with the exhaust steam, installations may be made to cause the high
1155
pressure line to be opened or closed as required. Dry kiln rooms, candy-drying rooms, conditioning chambers, the control of air compressors maintaining pressures on air receiving tanks, maintenance of water level in storage tanks and temperature in commercial refrigeration are among other obvious applications. In the construction of these devices delicate adjustments and light mechanisms have been avoided. Rugged instruments with parts exposed to corroding influences made of materials designed to withstand them, have been produced. The simplicity of operation and the fact that, wherever changes of temperature, pressure, or vacuum occur, devices of this sort may be installed, indicate the possible wide” application in the chemical industry. Accuracy of automatic control is so much to be desired as to require no special emphasis.
Fuel Economy in Burning Portland Cement Clinker‘ Fundamental Data Robert I). Pike PIKE & WEST, 4068 HOLDENST., EMERYVILLE, CALIF.
ROGRESS has been made in fuel economy in burning Portland cement clinker in the direction of recovering heat from the stack gases by use of waste-heat boilers; but in late years there has been little advance in reducing the consumption of fuel in the kiln itself. It can be shown that in the dry process clinker can be burned with from 40 to 50 per cent less fuel than is used in present practice, and that a fuel saving of this magnitude in the burning process itself is far preferable to the recovery of heat from the stack gases of an inefficient kiln. It can be further shown that in properly designed apparatus using the dry process a barrel of clinker can be burned with about 600,000 B. t. u. less than in an equally efficient apparatus using the wet process. When such efficient apparatus has been adopted, the dry process will possess this constant advantage over the wet process, and it is believed that this advantage, when combined with modern dry-process grinding and sampling equipment, will establish a definite economic superiority of the dry process. In attacking the problem of obtaining the maximum practical fuel economy in burning clinker, it is first desirable to visualize an “ideal apparatus” and to calculate the fuel economy for such an apparatus, based upon the latest available fundamental data. An investigation of the ideal apparatus will by comparison serve to emphasize the weak elements of present-day apparatus and to point out the direction for improvement. Such an ideal fuel economy has been sometimes called the theoretical fuel economy and several approximate methods have been suggested for its calculation, none of which appear to be rigorously correct, nor are they based upon the most acceptable fundamental data. The present paper is concerned with the fundamentals of clinker-burning and presents what is believed to be a rigorously correct method for calculating the ideal fuel economy. Later publications will be devoted to methods for calculating the design of efficient practical apparatus. Next to labor the fuel for burning clinker is the largest single item of cost in the manufacture of Portland cement. With reference to actual consumption of fuel per barrel of
clinker in American practice, we quote from certain unpublished sources: There are very few kilns over 175 feet in length in America operating on the dry process. One large company has one kiln 10 feet in diameter by 232 feet in length. They also have a number of kilns 12 feet in diameter and approximately the same length, all using the dry process, but unfortunately operating data such as fuel consumption and output are not available. In one dry process plant where the kilns measure 10 X 164 feet the average B. t. u. required per barrel of clinker is 1,460,000. In another plant operating under similar conditions, except that the kilns measure 10 X 175 feet, 1,240,000 B. t. u.’s are consumed per barrel of clinker. In wet process kilns of 10 X 240 feet using limestone and shale, the average B. t. u. per barrel of clinker burned is 1,560,000. On the other hand, in one of the most modern kiln installations using wet process with 33 per cent water in the slurry and the kilns measuring 10 X 9 X 250 feet the average number of B. t. u.’s required is 1,365,000. In the latter kilns the actual burning surface is exactly 235 feet, the last 15 feet being used as a cooler to convey the thoroughly burned clinker t o the discharge end of the kiln.
The actual costs per barrel, with coal and oil fuels a t different prices, are given in Table I. The barrel of cement weighs 376 pounds, including gypsum, and the barrel of clinker, before adding gypsum, is taken as 365 pounds. If $4.00 per ton is taken as an average of the cost of finished coal of 13,000 B. t. u. per pound a t the kiln in American plants and 150,000,000 barrels as the annual production of cement, the annual fuel bill is something in excess of $30,000,000. A flank attack, as it were, has been made on the cost of fuel in cement-burning by use of waste-heat boilers, in itself an admission of the thermal inefficiency of the rotary kiln. No attempt will be made here to analyze this practice in detail. Suffice it to say that waste-heat boilers have been installed in many cement plants and some records on their design and performance may be found in the literat~re.~#3 Note-There are no direct data available on cost of coal a t cement plants, hut Murray, U. S. Geol. Survey, Professional Paper 123 (1921),makes the following statement: “The cost of coal per short ton ranged from $2.79 2
Received May 16.1928.
8
Baylor, Trans. Am. Inst. Chem. Eng., 10,209 (1917). Anon., “Waste Heat Recovery by the Edge Moor System,” 1920.
1156
I,VDUSTRIAL A X D EXGINEERING C H E M I S T R Y
Vol. 20, No. 11
would therefore require 26 pounds of coal of 13,000 B. t. u., which is the saving effected by waste-heat boilers when operating under ideal conditions of balanced operation with 100 per cent load factor. Table I-Cost of Fuel Only But under actual operating conditions the waste-heat (1,400,000 B. t. u. required per barrel: coal of 13,000 B. t. U. per pound: oil of 19,000 B. t. u. per gallon) ~. pound; we~ghtof 7.9 pounds per . boiler cannot be expected to do so well as this. To get the COST O F FINISHED COST PER CORRESPONDINQ COST actual economic saving there are items of excess labor and COALAT KILN, BARREL OF OF OIL AT KILN, PER TON (2000 LBS.) CLINKER PER BARREL (42 GAL.) excess investment, as compared with ordinary power plants, to be taken into account, as well as the departure from 100 per cent load-factor conditions in practice. However, omitting these details, we will not be far wrong in stating that the saving effected by waste-heat boilers is about 20 pounds of coal per barrel of cement, corresponding to 4 cents per barrel when finished coal is worth $4.00 per ton a t the I n dry-process kilns about 360 pounds of steam a t 200 kiln. pounds pressure and 100”F. superheat can be developed from It seems almost self-evident that, if a fuel-fired apparatus feed water at 200” F. per barrel of clinker. If this steam is is inefficient, the first thing to do is to study that apparatus delivered to a turbine plant, to see if it cannot be improved so as to use less fuel and cost , which consumes 17.5 pounds of steam per kilowatt-hour a t less to operate. It may be the switch board, the power that the answer will be in yield will be about 20 kilothe negative, and if so, then The present-day rotary kiln for burning Portland w a t t - h o u r s per barrel of and then only, waste-heat cement clinker by the dry process has a thermal efficlinker. About 18kilowat-tboilers should be given sericiency of 25 per cent, but efficiencies twice as high as hours are required per barrel ous consideration upon their this should be obtainable. It is more logical to strive in d r y - p r o c e s s p l a n t s . own merits. I n c a r r y i n g for the ultimate possible thermal efficiency than These are the facts which out such a study it is first to accept present low efficiencies and to recover heat underlie the guarantees of necessary to investigate the from the stack gases by use of waste-heat boilers. An waste-heat boiler manufacfundamentals of the probefficient kiln using the dry process would consume turers to furnish all of the lem with a view to establishabout 600,000 B. t. u. per barrel of clinker less than the steam required for generating the limiting or ideal fuel most efficient kiln using the wet process. This fact, ing power in dry-process ceeconomy. combined with recent developments in dry grinding m e n t p l a n t s , if used for The power-plant engineer and sampling equipment should give preference in fugrinding when produced. Of who is accustomed to code ture developments to the dry process. Before proposing course, waste heat cannot tests of boilers, engines, etc., rational methods for increasing the thermal efficiency be stored up like coal in a from which highly accurate of rotary kilns, it is necessary to investigate the fundabin, or oil in a tank, and if h e a t b a l a n c e s a r e conmentals underlying the process for burning clinker. t h e c e m e n t manufacturer structed and carefully conThis is done by first stating and analyzing the heat wants to put clinker into sidered statements of effibalance of actual rotaries and then by developing storage and grind it later, he ciency are made, will be disthrough analytical methods the heat balance of the will have to pay for the necappointed if he expects any ideal apparatus. The comparison of the actual with essary power, either by fircomparable data on the rothe ideal points out the path for progress in increasing ing fuel under the boilers or tary kiln. A few articles in the thermal efficiency of the process. by firing the boilers with exthe literature have come to cess fuel in the kiln. n o t i ~ e .but ~ ~ none ~ , ~ of these To effect full economy the are complete or up to date, waste-heat steam must be used as produced and the clinker and several of them contain statements which are now known ground as made; but, granting this ideal condition, the real to be highly inaccurate. Accurate heat balances of modern saving in fuel effected by the waste-heat boiler is that fuel rotary kilns using both dry and wet processes are much to which would otherwise be burned in the boiler of a modern be desired. power plant to produce the same amount of power. The writer has made many observations of the rotary kiln, Large modern steam-power plants operating a t most eco- and on one occasion ran an organized heat-balance test of nomic load deliver a kilowatt-hour for 1.332 pounds of coal of 125-foot kilns burning dead-burned magnesite for the North13,500 B. t. u. per pound,4 or for 18,000 B. t. u. The oper- west Magnesite Company, a t Chewelah, Wash. It had been ation of the Lakeside Station6 records an important advance planned to publish the results of this test, but the project in fuel economy in power generation. At most economical for publication was abandoned because no means existed for loading this station produces 1 kilowatt-hour for 15,800 weighing the coal fed to the separate kilns, and the calcuB. t. u.’s and, according to the author, any pulverized-fuel lation of the quantity supplied, made indirectly from measurestation with usual equipment should give this economy. ment and analysis of gases passing up the stack, was so However, the results quoted by Weymouth6 very closely different from the general plant average, in spite of every parallel conditions existing in a cement plant in respect precaution for accuracy, that their value was doubtful. to the load factor and the size of power unit. Although the Nevertheless, some methods were developed and conclusions plants mentioned by Weymouth are oil-fired, the results are reached which were of value, and which are applicable and directly comparable with coal-fired plants because the boiler valid to the closely similar operation of burning Portland efficienciesshould be much the same. The results indicate the cement clinker. delivery of 1 kilowatt-hour for about 1.4 pounds of coal or 7 Eckel, “Cements, Limes and Plaster,” p. 437, John Wiley & Sons, 18,500 B. t. u. per kilowatt-hour. Eighteen kilowatt-hours 1922. a t the Wilkes-Barre load center to $8.75 at the Lowell load center. The average cost for the zone was $5.35.” In view of this, the writer’s estimate of $4.00 per ton is probably an understatement.
4 8
6
Hopping, Mech. Eng., 43, 790 (1921). Drewry, I b i d . , 47, 811 (1925). Ibzd , 41, 523 (1919).
8 Meade, “Portland Cement,” p. 175, Chemical Publishing Co.,1911; Soper, Trans. Am. SOC.Mech. Eng., 31,1563 (1910); Richards, J . Am. Chcm. S O L ,26, 81 (1904); Dorman, Tonind.-Zlg., 68, 405 (1914).
November, 1928
INDUSTRIAL S S D E,VGI NEERING CHEMISTRY
1157
Heat Balance of Dry-Process Rotary Kiln
perature, and because the specific heat of the reactants is probably a little greater than that of the products, the actual Table I1 is presented as a typical heat balance, under heat which is liberated a t furnace temperature is somewhat conditions described below, of a dry-process rotary kiln greater. However, this is still a relatively small amount and 125 feet long, and having a 6-foot bore throughout inside derives its importance in considering the heat-transfer relathe brick. The kiln is not supplied with a rotary cooler, and tions involved because it becomes available a t a high temair for combustion enters atJthe temperature of the kiln room. perature. This is not what is known as a modern rotary kiln, but the (3) SENSIDLEHEATIN STACK CASES-A thermocouple difference in fuel economy is sma!l. inserted within the upper mouth of the kiln shows apparent temperature of 1400" to 1500" F. This reading, however, Table 11-Typical Heat Balance of Dry-Process Rotary Kiln 7 f e e t 6 i n c h e s X 45 f e e t (lower e n d ) X 7 Feet X 80 Feet per Barrel of 365 serves merely as a starting point for obtaining the true temP o u n d s of Clinker. Output 500 Barrels Daily perature, because of the correction for radiation. The walls No. ITEM DEBIT CREDIT of the kiln which are "seen" by the thermocouple are cooled B. 1. u. B. t. u . Per cent 1 Heat in fuel, 10 5 gal. oil per bbl. 1,575,000 by contact with the incoming raw mix and are therefore 2 Heat of Formation of silicates, cooler than the gas which passes out of the kiln. Tne gas 1.53 B. t . u. per Ib. 55,800 3 Sensible heat in stack gases a t transfers heat to the thermocouple, which raises its tem1800' F., m. s. h. = 0 . 2 7 736,000 46.3 4 Latent heat in same 82,900 5.1 perature, but the couple loses heat by radiation to the cooler 5 Heat for calcining lime rock dewalls. The actual reading of the thermocouple is the result hydrating clay, 1080 B. t . u. per Ib. 304,000 24.1 of two opposing influences: (1) the heating effect of the gas, 6 Heat for evauoratinp water of raw mix 3,100 0.2 which tends to maintain the couple a t the true temperature 7 Sensible heat of clinker at 2250' F., m. s. h. = 0 . 2 4 5 of the gas; and ( 2 ) the radiation of heat from the couple 196,000 12.0 S Heat lost with dust 40,900 2.4 to the walls, which tends to make the temperature of the 9 1,osses by radiation, convection and conduction from outside couple less than that of the gas. The actual reading of the shell 157,900~ 9.9 - -- - couple is an equilibrium condition which has o ~ l yan in1,630,800 1,630,800 100 0 direct relation to the true temperature of the gas. The a Corresponds to loss of 1165 B t. u per hour from 2824 square feet of differences between the true gas temperature and the reading shell proper. of the couple, under the conditions existing in the upper The apparatus covered by the heat balance is the entire mouth of the kiln, are so great as to invalidate the accuracy shell of the rotary kiln from a point just inside its upper of any heat balance in which the indicated temperature is mouth and through and including its lower end and the out- taken as the true temperature. The magnitude of the corside surface of the firing hood. The degree of accuracy which rection increases very rapidly with increase of temperature may be reasonably assigned to the several items may be and with temperature difference between the couple arid the gathered from the following discussion: walls. (1) FUELCoNsumTIoN-This has never been checked Kreisinger and Barkley11 have developed a method for closely, but is a fair average value. The average value of measuring the radiation correction by use of sirnullaneous 150,000 B. t. u. has been assigned to the gallon of oil. The readings taken with couples of different size from extremely total value is probably accurate to within 5 per cent. There small to large commercial couples. The radiation correcseems to be no reason why oil fuel should be either more or tion reduces rapidly with very small couples, and with less efficient than a good grade of coal, if we expect a greater an ideal couple of no dimension it would be zero. Tne method amount of latent heat in gases of combustion from oil than is to plot the readings of the couples against the diameter of from coal; which, however, is not a n important item. the couple head and then, by means of the mathematical OF SILICATES--A brief review (2) HEATOF FORMATION equation of the line so obtained, to extrapolate to zero size of of the data has been p u b l i ~ h e d . ~The data of Nacken are couple. The writer, in conjunction with George H. West, apparently the most reliable thus far available and have has applied this method to measuring the temperature of been used in this paper. A net figure is employed, which is gases in the regenerators of glass tank furnaces. I n general, the exothermic heat of formation-namely, 180 B. t. u. per the corrections are very large and under conditions approxipound less the estimated heat of partial fusion, assuming mating those existing in the upper end of the rotary kiln that 15 per cent of the clinker is fusedlO-namely, 27 B. t. u. the reading of the thermocouple would be 400" to 500" F. too per pound. This net heat of formation is 153 B. t. 11. per low. Therefore lSOOo F. has been adopted as the temperature pound of clinker-a much lower figure than that quoted by of the stack gases, which is probably accurate within * 5 per earlier authorities. It might be argued that the latent cent. Incidentally, the application of this correction rationalheat of partial fusion would be returned by the solidification izes the heat balance and permits a reasonable interpretation of the clinker in the lower part of the kiln, but this would of the heat losses through the shell. If the observed temperabe the case only if the clinker solidified into the same crystal ture of the exit gas had been taken as the true temperature, a form in which it existed before melting. That this mould large unaccounted-for balance would have existed on the credit occur during the short time of residence in the kiln below side which has ordinarily been grouped in heat balances with the clinkering zone is doubtful, and it seems more likely that the losses through the shell, making these much larger than the part of the clinker which had been melted would leave the could be accounted for on the basis of actual measurement kiln in the form of glass. From this assumption it follows of the temperature of the outside surface of the shell. that none of the latent heat of fusion will be returned and More recently Haslam and Chappell'* have published a this quantity is therefore to be deducted from the exothermic review of this subject which contains a description of the heat. new method. The results of Figure 111,of their paper, roughly The amount of heat that is assigned to the exothermic substantiate the above estimate of a correction of 400" to 500" reactions by these considerations is very small. The heat, as F. in the upper end of the rotary; as also do the data in given, is referred to both products and reactants a t room tem- Table IX, considering that the mass velocity in the cement kiln is about 0.3 pound per second per square foot. The 9 Concrete. 24, 116 (1924~; Nacken, Zement, 11, 245, 257 (1922); C. A , 17,1700 (1923). 10 P. H. Bates in private communication states t h a t 15 per cent is probably a maximum allowance for fusion.
Bur. Mines, Bull. 146. 1'Haslam and Chappell, Bull. Mass. Inst. Tech., 60, No. 8 1 (May, 11
1925).
INDUSTRIAL AND ENGINEERIiYG CHEMISTRY
1158
high temperature level of the gases being measured, the large size of the sheathed commercial couples which are employed, and the fact that the brick surfaces are being continually swept by cold raw mix present all of the conditions which cause a large discrepancy from true readings. It is believed that the high gas-velocity couple described by Haslam and Chappell presents the best method for obtaining true gas temperature in the upper end of a rotary kiln,' but lacking such determinations, the correction noted above is the best that can be assigned in light of present knowledge. The surplus air is 15 per cent, and the constituents of the gas in pounds corresponding to 1 pound of fuel burned are as follows: Pounds 3 08 2.43 1 04 0.48 12.27
COz from fuel COz from mix Hz0 0 2 N2
Total 19.30 Total same per pound of clinker 4.38 pounds Note-With proper burners i t is with practically theoretical air and monoxide, and the correction t o be than 15 per cent surplus air may pulverized coal and oil fuels.
possible to burn fuel in cement clinker with less than 1 per cent of carbon applied on account of the use of less be calculated. This applies to both
!b+-.ZZGaqe
Figure 1-Asbestos
Chrome[- Alumel couple
Protector for Applying T h e r m o c o u p l e t o Kiln Surface
It is assumed that no carbon monoxide is present in the exit gases. The mean specific heat of these gases a t 1800" F. is about 0.27, and t.his is used through these calculations as the mean specific heat of the stack gases. Note-Values of mean specific heats of gases a t constant pressure are taken from Kuzell and Wigton, Trans. A m . Inst. Mining Eng., 49, 774 (1914),and are as follows: 0 2 = 0,2111 f 0.0000087t h-2 = 0.2411 0.0000099t Air = 0.235 f 0.0000089 t
+ COz = 0.1975 + 0,00003881 - 0.0000000043 HzO = 0.4701 - 0.0000118 t + 0.0000000127
f2
Specific heat in B. t. u. per Ib. per
a
F.; tin
'F.
t2
In considering the mean specific heat of the gases within the furnace, the mean specific heat of the above gases between 2500" and 1800" F. is about 0.30, and this is the figure used later on for the mean specific heat of the gases within the furnace. (4) LATENTHEAT IN STACKGAsEs-In calculating this quantity, the total heat of the steam a t the given partial pressure is used. Fuel oil of the following analysis is assumed: C, 84.03; H2; 11.56; S, 1.16; Nz, 2.80; 02,0.45 per cent. (5) and (6) HEAT FOR CALCININGLIME ROCK AKD DEHYDRATING CLAY; AND FOR EVAPORATING WATER OF RAW Mrx-Column (1) is taken as the analysis of the clinker.
Vol. 20, No. 11 (1) 20.30 2.75 7.90 65.50 2.07 0.19 0.59 0.35 0.35
SiOz FezOa A1203
CaO
MgO so3
Kz0 NazO Ignition loss
(2) 20.39 2.76 7.92 65.72 2.08 0.19 0.59 0.35
-_-
---
100.00
100.00
But the clinker upon leaving the kiln may be assumed tCJ have no ignition loss. The actual analysis given in columii (1) is calculated to an ignition-loss-free basis in column (2). Assume that all of lime in clinker comes from calcium carbonate, all of magnesia from magnesium carbonate: To produce 0.6572 pound CaO requires 1.174 pounds CaC03 To produce 0 . 0 2 ~ 8pound MgO requires 0.044 pound MgC03
The ignition loss of this raw mix has been found to be 35.6 per cent. Assume that the "clay substance" contains 20 per cent &03. The weight of clay substance is then O 0792/0.20 = 0.396 pound. The total weight of raw mix is 0.396 = 1.614, of which 0.396/1.614 X 100 = 1.174 0.044 24.5 per cent is "clay substance." Let X = loss on ignition of the clay. Then
+
+
+
0.6572 +0.0208 0.396(1-X) 1.614
=
o.644
Solving, X = 0.0884. The ignition loss of the clay is therefore 8.84 per cent. KaviasI3 gives data which can be used t o obtain a fair approximation of the heat employed in dehydrating the clay substance. The average clay which he tested had an ignition loss of 13.3 per cent. We will assume, therefore, that the heats of decomposition taken from this table are to be multiplied by the ratio */13.3 to fit the case of the clay used in this mix. We are, furthermore, interested only in data up to 900" C., because that is about the temperature of calcination and beyond this point the clinkering zone is nominally entered. The mean specific heat of clay, 25' to 900' C., may be taken as 0.237 (red clay, Navias).lJ The total apparent specific heat over the range 25" to 900" C. in heating up is 0.597 calorie per gram. The net heat of decomposition is therefore (0 597
-
0.237) X 875 X 1 8 X
8 8 13.3
=
375
B. t . u. per pound of air-dried clay. To each pound of clinker there is, in the raw mix, 0.396 pound of air-dried clay substance. The heat of decomposition of the clay substance per pound of clinker is therefore 0.396 X 375 = 149 B. t. u . per pound of clinker. It should be noted that practically all of this decomposition takes place in the heating zone in a range of 25" to 900" C. Only one clay of those tested by Navias showed decomposition a t temperatures over 900" C. Cohn14 gives the total heat of dehydration of 100 per cent pure clay substance a t 94 calories per gram as 169 B. t. u. per pound. Defining pure clay substance as the compound Al2Oj.2SiO2.3H2O, this would correspond to about 77 B. t. u. per pound, or less than one-quarter of the value found by Navias, and corresponding to 30 B. t. u. per pound of clinker. If the value given by R i ~ h a r d sfor ' ~ heat of dehydration of the compound AI2O3.2SiO2.3H20 is used, a figure of 41 B. t. u. per pound of clinker is obtained by assuming that all the alumina in the raw mix occurs in the form of this compound. Agaiii Mellor and Holdcroft'e determined, by indirect methods, a value only about one-half as great as that used by Cohn. These authorities all agree that dehydration occurs below 900" C., but are in seemingly hopeless disagreement as to the J . A m . Ceram. SOC.,6, 1268 (1923). I b i d , 7 , 475 (1924). 16 "Metallurgical Calculations," p. 291,McGraw-Hill Book CI., 1818 18 Trans. Ceram. SOC.(England), 10, 94 (1910-11).
18
14
November, 1928
I-YD USTRIAL A V D EaVGIAVEERIA-G CHEMISTRY
amount of heat required. Let us examine in more detail the work of Cohn and Navias. The former employed the indirect method of heat input-time curves in a furnace containing the specimen; the latter used a direct calorimetric determination of the heat absorbed on heating up and the heat liberated on cooling. The value of 94 calories per gram for pure clay given by Cohn is calculated graphically from the heat input-time curve. On the other hand, according to him, 1 gram of Zettlitz Kaolin, on being heated from 20" to 900" C., absorbed 386 calories, showing an apparent mean specific lieat of C.495. Taking the true mean specific heat over this range as 0.237 gives the endothermic heat (0.495 - 0.237) X 880 = 227 calories per gram of pure clay substance. This last figure seems to present a more rational interpretation of his data. The clays tested by Navias hhowed an apparent mean specific heat of 0.597 over the range 25" to 900" C. This shows an endothermic heat (0.597 0.237) X 875 = 315 calories per gram. It is thus seen that, if the data of Cohn and Navias are compared on an equal basis, the discrepancy is not so great. Because of the direct method employed, the data of Navias seem the more credible and will be used in calculations. The heat of decomposition of the raw mix is made u p as follows: Heat of calcination, 0 6672 pound CaO 17 Heat of calcination, 0 0208 pound MgO 18 Heat of dehydration of clay Total hrat for c h i n a t i o n and dehydrating per pound clinker
B. 1. u . 900 26
I49
1075
Each pound of clinker will require (omitting dust loss) 1.552 pounds raw mix, and this will contain about 0.3 per cent of water, which must be evaporated in the preheating zone of the dry-process kiln. Heat for evaporating 0.0046 pound water per pound clinker 5 B. t. u. Total heat required for calcination, dehydration, and evaporation in dry process (assuming no dust loss) 1080 B. t. u per pound clinker
(7) SENSIBLEHEATIN CLINKER-A new method has been devised for determining the temperature of the clinker leaving the kiln. The chute is covered with a plate and a pile of clinker allowed to accumulate until the top of the pile is only a few inches below the lip of the kiln. A hot wire optical pyrometer is then sighted on the peak of this pile, giving the temperature of the clinker which was observed to be 2200" F.; 50" F. is added to allow for emissivity correction.'@ Accurate unpublished determinations of the mean or interval specific heat have been made. The following equation gives the value for mean specific heat as a function of temperature: M . s . h. clinker = 0.1764 4685 X t3 t = temperature in C.
+
+ 139144 X
1159
may consider the raw mix as being composed of 75 per cent lime rock and 25 per cent clay substance. The mean specific heat of the clay over this range has been taken as 0.237. B a c k ~ t r o m 'gives ~ the following formula for the average atomic heat of calcite over the interval TI- Tz: 4 23(T2 log Tz - Ti log TI) - 8,387 Tz - Ti in which T = O C. absolute. The value of C, 289" - 1173" C. = 3 46. Av. C,Ti
-
Tz
=
This corresponds to a mean specific heat over this interval of 0.273. Therefore, mean specific heat raw mix from 60" to 1652" F. = 0.2642. Another mean specific heat which we shall use later is that of the calcined mix over the range 1652" t o 2700" F. For the purpose of calculation the material is supposed to pass over this range without change and then to become suddenly converted into clinker. By extrapolating from Backstrom's data the mean specific heat of CaO over this range is taken as 0.243. 1 ; a v i a ~determinations '~~ indicate an approximate mean specific heat for clay over this range of 0.28. Therefore: Mean specific heat calcined mix range 1652" to 2700" F. = 0.253. (8) HEAT O F PARTIAL CALCINATION O F STACK DUSTAND SENSIBLEHEAT IN S.mE-Lleasurements of dust losses show that 57 pounds of raw mix are lost per barrel. This dust is about one-half calcined, indicating an actual weight of dust leaving the kiln of 46 pounds per barrel. B. 1. u. Sensible heat a t temperature of 1900' F. 46 X 1840 X 0 . 2 5 = 21,200 Heat of decomposition 19,700 Total
40,900
(9) LOSSESTHROUGH THE SHELL-It has been customary to report this item by difference, because of the difficulties in the way of measuring it directly. If the uncorrected temperature reading of the exit gases within the mouth
Figure 2-Average Temperatures of Shell (5 Kilns) Showing Maximum and Minimum
t--125 X 10-9 t Z
The figure for heat content of the clinker may be considered accurate within 5 per cent. Although not appearing in the heat balance, it is important to determine the mean specific heat of the raw mix from atmospheric temperature to the temperature of calcination, which for this purpose can be placed a t 900" C. (16.52' F.) (for more accurate estimate see later). For this purpose we Backstrom, J . Am. Chcm. SOC.,47, 2443 (1925). Richards, 09. cit., p. 29. 19 Poote, Fairchild and Harrison, Bur. Standards 113 (1921). 17 18
Tech Paper 170,
of the kiln were taken as their true temperature, and item (8) were then found by difference, it would be more than twice as great as shown in Table 111. As shown there, the item is based upon direct measurement of outside shell temperatures of kilns of the same size burning dead-burned magnesite a t plant of Northwest Magnesite Company a t Chewelah, Wash. The method of obtaining these data will be described. At the time the measurements were taken five kilns were in operation and all were measured. A kiln which had been running a t least 24 hours was selected and marked with chalk every 5 feet from end to end. Readings of the temperature of the surface were taken on the meridian of these marks, from hot toward cold end on one side, and from cold toward hot end on the other side. About 2 hours was the time required to make a complete test. The tempera-
I N D U S T R I A L A N D ENGINEERING CHEMISTRY
1160
ture of the atmosphere away from the kilns to windward was taken a t the beginning of each test. From the temperatures x calculated by known formulas. observed the heat radiated can L A chromel-alumel couple, of 22-gage wires, was used. The couple itself was inserted through a cylindrical pad of asbestos 11/4 inches in diameter by l l / s inches long, with a guard ring of asbestos paper a t its upper end, forming a cylindrical cup 3 / 8 inch deep, into which the couple projected. To determine the temperature of the surface, the couple wires were grasped immediately back of the asbestos pad, and the pressure of the hand was so distributed that the asbestos guard ring and thermocouple were pressed against the kiln simultaneously. The couple wires were directly connected to a Leeds and iYorthrup potentiometer of the manually operated type. This was carried in a box with a thermometer to give the cold-junction temperature. In making a reading one operator placed the couple against the face of the rotating kiln, holding it in one place as long as his own reach and the rotation of the kiln permitted. As the couple had to be periodically moved to a new position, the galvanometer needle never came to a fixed position, but the reading was taken when it had moved to zero a t least three times for a fixed position of the millivolt scale. The couple arrangement and the general set-up of the apparatus in use are shown in Figure 1. The temperature measurements obtained are shown in Figure 2 . Two sets of readings were taken for each kiln-one on each side. The ten sets of readings were averaged t o give the values used in constructing the curve. At the same time the maximum and minimum temperatures were picked out for each point, and are shown by the small circles. The mean temperature of the kiln surface was 261" F., and the total exposed surface of the kiln was 2770 square feet. The average temperature of the atmosphere was 44" F., and the highest average temperature of the shell was 428" F., 15 feet back from the discharge end.
that this temperature averaged about 2700' F. This figure varies with the mix as between different plants and, to some extent, with minor changes in mix from day to day in any given plant; but in all that follows the temperature 2700' F. has been adopted as the temperature of clinkering. H e a t Balance of Wet Process
Consider a wet process using a slurry containing 40 per cent mater. If the kiln of Table I1 be made sufficiently longer so that a t a point 125 feet from the lower end all of the water will have been evaporated from the slurry by the heat in the exit gases from the heating zone of the kiln, then the heat balance of Table I1 will be unchanged in its essential aspects, but the weight of the stack gases mill have been increased and their sensible heat decreased to an extent necessary to evaporate the water of the slurry and to heat the vapor to the temperature of the stack gases. The lowered temperature of the stack gases is calculated as follows: Theoretical weight raw mix per barrel of clinker, pounds Assume a dust loss of 3 per cent giving actual weight dry raw mix per barrel, pounds Pounds of water per barrel clinker t o give a 40 per cent water slurry
I
'
1 l ~l/ : I
I
575
593 395
The following shows the make-up of the gases per barrel of clinker after the evaporation of the water of the slurry: CONSTITUENT WEIGHT Pounds
cot
465 481 40 1016
Ha0 0 2 N2
Langmuir20 quotes some experimental data obtained from Carpenterz1giving the heat loss from a 2-inch steam pipe a t 120" F. above room temperature. These data show that, TEMP EXIT GASES
Vol. 20, No. 11
Total
WEIGHT P e r cenl
VOLUME Per cent
23.3 24.0 2.0 50.7
13,84 36.10 1.69 48.37
__
--
--
2002
100.0
100.00
The partial pressure of the water vapor is therefore 0.361 X 14.7 = 5.31 pounds per square inch. At this pressure the temperature of evaporation is 164.7" F. and the total heat of the steam above 60" F. is 1104 B. t. u. per p o ~ n d . ~ 3There are. therefore, 436,000 B. t. u. required to evaporate the 395 pounds of water per barrel of clinker. To provide this heat for evaporation there are 1610 pounds of gas a t 1800" F. This gas in dropping through a temperature of 900" F. will supply the necessary 436,000 B. t. u. for drying the slurry. Finally, we have to consider the resulting temperature after mixing 1610 pounds of stack gases at 900" F. with 295 pounds of water vapor a t 165' F. This is calculated as follows:
OF:
I 50
Sensible heat content of 395 pounds water vapor a t 165' F. above 60' F. Sensible heat content of 1610 pounds of gas a t 900' F. above 60'
F.
Total sensible heat content of mixed gases above 60" F. Mean specific heat of mixed gases Resulting tempera2ure of mixed gas (gas leaving kiln) with 40 per cent slurry, F.
18,600 351,000 369,600 0.302 670
Similarly the temperature of the gas leaving the kiln with 30 and 50 per cent slurries is, respectively, 962" and 300" F. (In these calculations no account has been taken of the sensible heat of the dust.) I n Figure 3 temperature of exit gases is plotted against per cent of water in the slurry. These figures are based upon the fuel consumption of 1,575,000 B. t. u. per barrel, but are known to be closely in accord with practice. This check with known practice in wet process offers as good a practical proof as could be desired, short of direct determinations under the proper conditions as outlined, that the temperature of the stack gases of the dry-process kiln, as stated in Table 11, is substantially correct. It will be understood that in wet-process kilns, because of the low temperatures of the exit gases and the large amount of water vapor present, the radiation correction is very small so that the uncorrected reading of the pyrometer represents substantially the true temperature of the exit gases. 23
Marks and Davis, "Steam Tables," Longmans, Green and Co., 1920.
November, 1925
In-0USTRIAL AlYD EiliGINEERIIYG CHEJIISTRY
If we took the data of Table 11, using 9.5 gallons of oil instead of 10.5, as the fuel consumption, the above calculation for the 40 per cent slurry would give a temperature of exit gases in the neighborhood of 300” F. This indicates roughly that with 40 per cent slurry the lowest possible fuel consumption is 9.5 gallons of oil or 1,425,000 B. t. u. per barrel, and this could only be realized under very exceptional conditions of heat transfer betm-een the wet slurry and gases in the upper part of the kiln, and the use of induced draught. Comparison with Data of Nacken Sackeng presents a statement of the amount of heat to produce clinker from raw mix, the clinker being considered as at its maximum temperature of formation. These data, together with those calculated from the data employed in this paper, are given in Table 111.
1161
remains constant a t the temperature of dissociation (calcination). Of course, this assumption is not strictly true, but it represents a very close approximation to the facts, because there is a definite temperature a t which active calcination begins; and the endothcrmic nature of the reaction tends to maintain this temperature constant as long as any uncalcined material remains. I n this respect the process of calcination is somewhat analogous to the process of boiling mater. Leaving the calcining zone, the material enters the combustion and clinkering zone. Here the fuel is burned and the clinker is formed. But the forming of clinker is an exothermic reaction, which proceeds, a t least theoretically, of its own accord after a certain “kindling” temperature has been reached.
Table 111-Heat C n i t s R e q u i r e d t o Form Clinker
THIS SACKEN PAPER ITEM Sensible heat in 1.55 lbs. raw mix to temperature of calcination 652 149 Dehydration of clay To heat 1 Ib. calcined material t o temperature of 268 clinkering Total sensible heat absorbed 846 1070 900 926 Heat of calcination Exothermic heat of formation of clinker 180 180 54 27 Heat of fusion 1620 Net heat required 1843
That the present writers employ a larger figure for the total heat requirements than Nacken is explained by (1) higher specific heat values for both raw mix and calcined mix, and (2) higher figure for heat of dehydration of clay.
u
U
Figure 4-Diagrammatic
Sketch-Burning R o t a r y Kiln
O p e r a t i o n s in a
The material next enters t h e cooling zone, through which it passes in countercurrent to the air for combustion, which cools the clinker and itself becomes heated and passes into T o calculate the fuel economy of a n ideal apparatus for the clinkering zone in countercurrent to the material. burning clinker, we must first define the ideal conditions After passing through the zones of cornbustion and clinkerwhich exist in such an apparatus. It is not sufficient to say ing, the gases enter the calcining zone and pass through this that the ideal fuel economy is mereb that fuel corresponding in countercurrent to the material. The process of calcinato the difference between the heat required for calcining and ation absorbs much heat and under the ideal conditions occurs dehydrating and the net heat liberated by the formation of at a uniform temperature. Furthermore, under ideal consilicates. This would be (referring to Table 11) item (5) ditions it is permissible to assume that the temperature of 304.000 B. t. u. - item (2) 55,800 B. t. u. = 338,200 B. t. u.; the gases leaving the upper end of the calcining zone may and if clinker were t o be burned with this amount of fuel, be equal to the temperature of the calcining material. it would be equivalent to assuming that both the clinker The gases next enter the heating zone and passing through and the stack gases would be discharged at the temperature this in countercurrent to the material, heat the material of the atmosphere, and that there would be no losses of to the calcining temperature, and perform the other operheat through the shell. The latter assumption is proper to ations stated above. As in the calcining zone, in considering the ideal condition, but the first two cannot be accepted ideal conditions, it is permissible to assume that the gases unless supported by a n analysis of the process as a whole. may leave the heating zone at the temperature of the entering The entire process of burning is illustrated in Figure 4. material. Although this sketch illustrates the usual rotary kiln-rotary From these considerations, the criteria of ideal thermal cooler combination, this is merely for convenience of illus- economy can be stated as follows: tration and is not t o be taken as imposing any mechanical 1-There shall be no losses of heat to the outside atmosphere limitations upon the idealized procedure of burning. The uppermost zone of the apparatus illustrated is that through the walls of the apparatus; i. e., insulation shall be perinto which the raw feed enters and from which the gases of fect. 2-The theoretical amount of air shall be used for burning combustion leave. This is known as the “heating zone,” the fuel and combustion shall be perfect. 3-The efficiency of the cooler shall be a maximum-i. e., and in it the raw feed is heated up from atmospheric temperature to the temperature a t which calcination (dissociation) it shall raise the temperature of’the air for combustion to a equal to the temperature of the clinker, leaving the combegins, the water associated with the clay is driven off, and the point bustion and clinkering zone. A condition might be imagined magnesium carbonate is calcined. It is then assumed that where the clinker, in cooling from the temperature a t which it when once the temperature of dissociation has been reached leaves the clinkering zone to atmosphere, would not possess calcining of calcium carbonate, which is the major heat- enough heat to raise the temperature of the air from atmospheric to the temperature of the clinker, but in the ideal clinker burner absorbing reaction, begins abruptly and continues at the the clinker possesses a surplus of heat and must therefore leave uniform temperature of dissociation until finished, when the cooler at a temperature above atmospheric. 4-There is no primary air used for atomizing or conveying the temperature of the material again rises. The period of calcination of calcium carbonate defines the calcining fuel and all air for combustion passes through the cooler. 5-Calcination of calcium carbonate takes place in a sharply zone which, by our assumption, has sharp upper and lower defined zone and a t a uniform temperature corresponding to the boundaries and in which the temperature of the material temperature of dissociation according to the equation: Ideal Clinker Burner
I N D U S T R I A L AND ENGINEERING C H E X I S T R Y
1162
Vol. 20, No. 11
+
CaCO3 +CaO COZ Taking the value for IY given above, the content of the a t the prevailing mean concentration of carbon dioxide in the gases within the furnace, after adding to them one-half of kiln atmosphere. It is not far from the actual fact to assume that the carbon dioxide liberated by calcination to get an average ,~~ calcination takes place at a substantially uniform t e m p e r a t ~ r e . ~ ~composition, is as follows: 6-Heat transfer between gas and material is perfect, permitting the temperature of the gas to be lowered to that of the COKSTITUENT VOLUME material so far as considerations of heat transfer are concerned Per cent in either the calcining or heating zones. co2 27.7 7-There are no dust losses. 9.4 62.9
HzO
N Z
Based upon these criteria of ideal conditions of burning, we can now write the equations governing the ideal process.
That is, the total heat in the fuel per pound of clinker is equal to the heat of calcination in the calcining zone plus the heat of dehydration and calcination in the heating zone, minus the heat of formation of silicates plus the heat in the discharged clinker plus the sensible heat in the exit gas plus the latent heat in the exit gas.
Or, the sensible heat content of the exit gases from the calcining zone equals the heat imparted to the raw material in being heated up to the temperature of calcination plus the heat in the exit gases from the heating zone. ( T , - t)S, = M/A(T, - t)S, ( T , - t ) S , (3) Or, the heat content of the clinker, on discharging from the clinkering zone, is equal to the heat content of the air for combustion which, under the assumed ideal condition, enters the combustion and clinkering zone a t the temperature of clinkering, plus the heat content of the clinker which discharges from the cooler.
+
I n solving for the ideal fuel economy, the value of T , obtained from (3) is placed in (1) giving the following equation in which W is expressed as a function of T :
100.0
The partial pressure of carbon dioxide is therefore 211 mm. In the kilns of Table I1 the partial pressure of the carbon dioxide, calculated in the same way to give a mean value in the calcining zone, is 122 mm. The corresponding temperatures of dissociation are, for the ideal case ( T ,of the formulas), 815OC. (1499" F.),and for the kilns of Table 11.785" C. (1446O F.). Other values in equation (2') follow: G = 0.552, S, = 0.264, Sa S = 0.24, S, = 0,212
Also
=
0 27
Substituting the known values in equation (2') and solxlng gives : 7'3 = 2060" F. This shows that T3,the temperature of the gases leaving the calcining zone, is greater than T,, the temperature of calcination when T , the temperature of the stack gases, is equal to the temperature of the atmosphere, indicating that the latter assumption is proper to the ideal condition. To get the value of T,, the temperature of the discharged clinker, rearrange equation (3) to give (3') as follows: T, =
( T , - t ) ( S ,- WAS,)
S,
f
t
(3')
T , = 652' F.
Then assume that, under the ideal condition, T tion (4)is changed to equation (4') :
=
t, equa(4')
Equation (4') expresses the ideal fuel economy of the clinker-burning process when the exit gases leaving the heating zone are a t the temperature of the entering material which is assumed to be atmospheric. For the purpose of this calculation, assume that oil fuel having a calorific value of 19,000 B. t. u. per pound is used, then: Hc = 900
= 175 H , = 153 Tc = 2700" F. S, = 0.267
Hh
C = 19,000 B. t. u. Sa = 0.26 A = 13.8 pounds f = 1000 B. t. u. t = 60°F.
Summing up, in the ideal apparatus for burning clinker, 405,000 B. t. u. are required per barrel corresponding to 31.2 pounds of coal, or 2.7 gallons of oil; the exit gases leave a t the temperature of the atmosphere and the clinker is discharged a t 652" F. The heat balance of the ideal apparatus for burning clinker is shown in conventional form in Table IV. Equation (1) is the algebraic equivalent of this table. Balance of Ideal Apparatus for Burning Clinker (Per barrel of 365 pounds) ITEM DEBIT CREDIT B. 1. u. B. 1. u. Per cenf Heat in fuel 405.000 ..~~.. ~ . . ~ ~ Heat in formation of silicates 55;800 393,000 85.30 Heat of calcination and dehydration Latent heat in stack gases 21,000 4.55 46,800 10.15 Sensible heat in clinker Table I\'-Heat
~~
Substituting these values in equation (4') gives: W = 0.0584 pound per pound clinker
As a fkal check it is now necessary to refer to equation (2) to see that T3is equal to, or greater than, T , when T = t . For this purpose equation (2) is rearranged in the following form (2') : The value of T,, the temperature of calcination of calcium carbonate, is taken from the determinations of the temperature of dissociation of calcium carbonate made by Johnsons2( 24
LOC.cif. p. 186, Zemenl, 10, 597 (1921). J . Am. Chem. SOC.,83,938 (1910).
*a Nacken, 28
__-
___
-__
460,800
460,800
100.00
Definition of T h e r m a l Efficiency of B u r n i n g Clinker
We will define thermal efficiency as follows: Ideal fuel consumption Actual fuel consumption
x
100
By this definition the thermal efficiency of the apparatus, the heat balance of which is given in Table 11, is: 405 000 X 100 1,575,UUO
= 25.7 per cent
This establishes the theoretical basis and shows that there are no a priori grounds for stating that a project for reducing the fuel consumption in burning clinker by almost one-half is
INDUSTRIAL AND ICNGINEERIAW CHEMISTRY
November, 1928
unrealieable, or even very difficult. Many forms of heatinterchange apparatus exist in which efficiencies far higher than 50 per cent, as above defined, are realized-notably steam boilers; and it would seem that a heat-exchange process, such as burning clinker in which the incoming cold raw materials absorb heat from the gases of combustion, and in which the hot finished product is in convenient form for transfer of heat to the air for combustion, would lend itself to a highly efficient operation, if gone about in a proper way. Fortunately, we do not have to depend entirely upon theoretical grounds, because Portland cement clinker has been aid is today being produced in stack kilns with far less fiiel than is employed in tlie rotary kiln. Muller2' gives a full account of this type of apparat,us and states that at that time (1919) there were in Germany over eight hundred sliaft kilns for burning Portland cement. He gives the fuel economy at 18 to 20 per cent, coal or coke, of the weight of clinker and the capacity from 100 to 250 barrels daily per kiln. Taking coal as having 13,000 B. t,. u. per pound, this corresponds to about from850,000 to 950,000 B. t. u.per barrel of clinker. HansenzBtells of European development of the shaft kiln. His statements indicate an average fuel economy of betweeti 800,000 and 670,000 U. t. u. per barrel, and that with tlie better grades of fuel 50 pounds are required per barrel. The daily output per kiln is about 300 barrels. I n these kilns the raw materials are ground, mixed with the fuel and with 5 to 10 per cent of water, and briqiictted, before being introduced into the kiln. These briquets must hold up u?thout breaking under the considerable shock incident to their being charged into the top of the kiln, and it is not to be expected that all raw materials would make equally sound briquets, or even briquets having, in all cases, sufficient soundness to permit satisfactory operation. Further, the fuel employed must liave a low content of volatile matt.cr-i. e., of tlie naLure,of anthracite, semi-coke, or coke-because obviously if a bituminous coal is employed much of the valuable volatile matter will be distilled in the upper part of the kiln and lost up the stack. These practical difficulties, together with the small capacity, will prevent the widespread adoption of the improved stack kilns described by Hansen, in spite of their high thermal efficiency. The cost of making briquets is also 5 serious drawback. However, we are indebted to t.he stack kiln for furnishing a fairly large scale demonstration of the possibility of burning clinker with 40 to 50 per cent less fuel than is used in the rotary kiln, a n d by implioation of the correctness of the analytical results above set forth. The same comparison between fuel economy of the rotary kin and the shaft kiln has been noted by t.he writer in the closely similar process of burning dead-burned magne~ite.~9In this case the 125foot rotary kilns, using pulverized coal, employ about 9,800,000 B. t. u. per ton of 2000 pounds, while shaft kilns, fed with lump magnesite and coke, require ahout 4,200,000 B. t. u. per ton, the product being equally well burned in both cases. Symbols i i V = weight of fuel per pound of clinker, pounds C = B. t. u. per pound of fuel A = air required for combustion per pound of fuel, pounds T. = temperature of clinkcring, F. I t a = heat of calcination of CaCOa per pound of clinker, B. t. u. IIh = heat of calcination of XgCOa and dehydration of clay per pound clinker, B. t. u. II. = heat liberated in iormation of silicates. B.t.u. p a pound clinker
-
II "Der Scbachtefen in der Zement 1nduettie;'prerented at 42nd zener& meeting of Verein Deutrcher Portland Cement Pabrikaiiten, 1919. $8 Kork Plodurlr, 33, 33 (August 27. 1S21). r BW. ~ i n eB~ ~. I836. . p. 57 (1825).
1163
T, = tempmature of product-,. e , of clinker discharged from
apparatus, * F.
S, = mean mecific heat of clinker at temperature T , above
atmospheric tempmature
G = weight of carbon dioxide liberated in calcination, pounds
T
=
per pound of clinker temperature of gases a t exit from heating zone, 'F.
fuel due to uncondenscd water from burn& .~ hydrozen of fuel = theoretical temperature of calcination, 'F. = temperature of gascs at exit from calcining zone, F. = mean specificheat of gases between T, and t = mean specific heat 01 f e d brtween atmospheric tempera~
I; Tz Sa
S,
S, = mean speeibc heat b i clinker bctween Trand t Acknowledgment
'The work upon which this and subsequent papers to be published in this journal are based was made possible by the unfailing interest and support of Geo. T. Cameron, president of the Santa Cruz Portland Cement Company, of San Francisco and Davenport, Cnlif.
Unbreakable Explosion Pipet' Frederick W. Isles ST-IVDAKD OIL C o r ~ a r voe X 6 w Jsnsru, BAYONNB, N.
I.
T H E pccompanying photograph shows an all-steel explosion pipet wliicti we made u p not long ago to replace tile customary glass one in a Willianis improved gas analyzer. It consists of a 1-inch heavy malleable tee having a '/Finch spark plug in a i/z X 1 inch bnshing in tlie side outlet; a special I-inch steel plug with a capillary hose connection, screwed into the top of tlie tee; and an extra heavy 1-inch nipple, 7% inches long, screwed into the bottom of tile tee. The last-named tias no external threads on the lower end, but, i t is threaded internfilly to receive a special J/rinch steel p l u g w i t h capillary hose connection. AI1 threads are standard pipe t h r e a d s . T h e joints are made up with shellac. This pipet employs m e r c u r y as t,he aspirating fluid aiId has p i n c h c o c k s at both c u d s . After the explosion t h e residue gases arc transferred to an ordinary measuring buret. We have found tliis all-steel pipet very c o n v e n i e n t , as it w h o l l y obviates the hazards and replacement costs incident to tlie use of a, glass explosion buret. in
8 Received O c i o b ~ i ,1928.
