912
THE JOURNAL OF I N D LTSTRIAL A N D ENGINEERING CHEiWISTRY
~ 7 0 1 . 13,
50.10
TABLE I-SUMMARY OF PHYSICAL CONSTANTS OB THREE'TYPICAL FUELS AND THEIR MIXTURESWITH AIR
Socony Motor Gasoline
128OC.
114
CsHis
0.743
60
High End-point Gasoline
147' C.
128
CsHzo
0.752
57
Socony Kerosene
220°c.
170
CllH25
0.800
45
The vapor pressure curves for the partly condensed mixtures give an indication as to what can be accomplished as to starting in cold weather with various amounts of choking. Table I compares the three fuels i?vestigated in their more important characteristics. It is hoped to continue this work further by determining the foregoing type of data for various types of blended fuels,-naphthene base fuels, etc. SUMVARY
This article may be briefly summarized as containing: 1-A new, simple, and reliable method for the determination of the temperatures of initial condensation of fuels from air-fuel mixtures. 2-An approximate method of determining the temperatures of partial condensation in air-fuel mixtures. 3-Tapor pressure data on three typical automotive fuels: Socony kerosene, Socony gasoline, and an artificially pre-
6OOC. 140OF. 62OC. 144°F. 168°C. 334OF.
128OC. 262OF. 154°C. 309OF. 219'C. 426OF.
169OC. 336OF. 208OC. 403OF. 249OC. 4S0°F.
211'C. 412'F. 252'C. 485'F. 2S2'C. 539'F.
40°C. 104'F. 63OC. 134'F. 116OC. 242'F.
3OC. 36'F. 9oc. 48'F. 73OC. 163'F.
pared high end-point gasoline. 4-A series of charts showing the condensation temperatures of these fuels under different conditions. These results show that the failure to obtain complete vaporization of present-day gasoline is due more to inefficient methods of securing vaporization than to any inherent limitation of the gasoline itself. 6-A suggested method for approximating roughly the initial condensation temperatures of paraffin hydrocarbon fuels from the distillation curves of these fuels. 6-A table of comparison of the more important characteristic properties of the fuels investigated. ACRNOWLEDGME NT The authors desire to express their appreciation to Dr. W. IC. Lewis for his helpful suggestions in the courw of the work, and to The Yale and T o m e Manufacturing Company for their permission to publish that part of the work which was done a t their expense.
The Total Sensible Heats of Motor Fuels and Their Mixtures with Air's2 By Robert E. Wilson and D. P. Barnard, 4th RESEARCH LABORATORY OF APPLIEDCHEMISTRY, MASSACHUSETTS INSTITUTE OF TECHNOLOGY, CAMBRIDGE, MASS.
OBJECTOF WORK A feature which is of paramount importance to the efficient operation of an internal combustion engine, especially of the high-speed type used in motor cars, is that of securing a uniform mixture of fuel and air in the proper proportions before distribution to the cylinders. The only reasonably practicable way of accomplishing this seems to be in havi g the fuel nearly, or preferably completely, vaporized. L n order to vaporize present commercial gasoline or any less volatile fuels completely, i t is necessary to add some heat from external sources.3> It is possible to secure satisfactory vaporization by any one or more of three methods, which may be classed as follows:
;r"
1-Atomization of cold fuel into heated air. 2-Atomization of cold fuef into cold air followed by heating the resulting mixture in the manifold. The "hot spot" aids vaporization partly by heating the mixture as a whole, but t o a greater extent by throwing out the uncondensed drops onto the heated manifo!d surface and vaporizing the liquid directly. 3-Injection of the fuel as a superheated vapor into the cold air.
Methods 1 and 2 are frequently combined in the use of the hot spot together with a preheater. These first two methods suffer from a fundamental defect in that it has been found necessary, in order to vaporize the fuel completely in the short time available, to superheat the air or mixture considerably above its theoretical dew point. This results in decreased volumetric efficiency coupled with inReceived July 27, 1021. Published as Contribution No. 37 from the Research Laboratory of Applied Chemistry, Massachusetts Institute of Technology. a See preceding article by the authors, "Condensation Temperatures of Gasoline- and Kerosene-Air Mixtures." 1 2
creased tendency toward knocking. Method 3, however, starts from the other side of equilibrium and is therefore theoretically the best means of securing the proper mixture, as it is necessary to heat the vaporized fuel only to such a tempgrature that the resulting mixture will be approximately a t the dew point. Indeed, experimental results in this Laboratory and elsewhere have shown that the mixture prepared by this method may be as much as 30 to 35 per cent condensed before the distribution is seriously affected, . because the extremely ?mall particles formed by the rapid condensation of the superheated vapor do not tend to separate readily from the air stream. The particles thus produced are, a t least for the first few moments after condensation begins, far smaller than those obtained by the best methods of a t o m i h g a liquid. The practical ppplication of Method 3 requires fairly accurate data on two subjects-first, the dew points of various fuel-air mixtures; and, second, the specific heats, or rather, the total sensible heats, of fuels a t high temperatures. The data on dew points have been covered in the preceding article. It, is the purpose of this article, therefore, to cover the subject of specific and total sensible heats. DATAI N LITERATURE A survey of the available literature failed to yield any data on specific heats of paraffin hydrocarbons at the high temperatures required. Practically the only work done along these lines indicated that up to 100" C. the specific heat, averaged for a large number of petroleum distillates in the liquid state, could be fairly accurately represented by the equation:' I
Bushong and Knight, THISJOURNAL, 12 (1920), 1197.
T H E JOURNAL OF INDUSTRIAL AND ENGINEERIWG CHEiVISTRY
Oct., 1921
FIG. APPARATUS
Cp = 0 5
FOR
913
DETERMINING THE SENSIBLE HEATSOF VAPORIZED FUELS
+ 0.0008 t
Other data 1,233 indicate that the specific heat a t room temperature of light petroleum distillates is very close to 0 5. Such data as are available indicate, as do the vapor pressure curves (see preceding article), that the specific heats of the liquid and the vapor me substantially equal over the temperature range involved. While the existing data on heats of vaporization of hydrocarbons are quite variable, as is shown by Table I, it is nevertheless possible by combining these data with Trouton’s rule, which is known to hold quite well for hydrocarbons, to derive fairly satisfactory values for gasoline and kerosene. For this work, the best values for the heats of vaporization were chosen as 60 cal. per g. for Socony kerosene (the average boiling point of which corresponds to dodecane) and 70 cal. per g. for Socony gasoline (average boiling point of which corresponds to octane). These figures represent a sufficiently good approximation for present purposes. On account of the paucity of data at the higher temperatures, this Laboratory undertook to make some approximate experimental determinations of total sensible heats of two typical motor fuels u p to about 500’ C. The experimental method used in securing these data was probably not accurat? to better than 3 or 4 per cent, but is sufficiently close to the truth for all practical work, especially in view of the rather indefinite composition of the materials investigated. The publication of the data appears worth while in view of the practical importance of the subject and the complete
lack of any similar data in the literature. EXPERIMEKTAL WORK The experimental work was confined to two fuels: Socony kerosene and Socony motor gasoline, samples of which were very kindly furnished by the Boston branch of the Standard Oil Co. The grarity of the kerosene was 45” E%., corresponding to a specific gravity of 0.800; and that of the gasoline was 60” BB., corresponding to a specific gravity of 0.743. The same fuels were used in the previous vapor pressure work, and further data on them will be found in that article. The total heats of these fuels were determined a t various temperatures in the apparatus shown in Fig. 1. The distillate froni the Rector vaporizer1 was continuously condensed in a water-cooled steel coil. The condenser was tightly lagged with about 2 in. of kieselguhr, and the temperatures of the inflowing and outflowing cooling water were read from the thermometers T4and Ts. The cooling water was caught and weighed for each run. The weight of the cooling water multiplied by the difference in the temperatures T4 and T6 gave the heat given up by the fuel in cooling to the discharge temperature, Tg. The thermometers TI and Tj were checked against each other, and showed no deviation over the range of temperatures involved. The vapor temperatures were measured by means of thermocouples placed a t TI and T2, and the air temperature (indicated by the thermocouple T3) around the vapor pipe was kept slightly above that of the vapor in order to prevent more than minor heat losses. A variation of 75” to 100” C. in the air temperature produced
TABLEI-HEATS Hexane CsHir
79.4
.... ....
81.6
Heptane C7Hl6
Octane CsHis
OF VAPORIZATION OF HYDROCARBONS (Calories per Gram) Nonane Decane Undecane Dodecane CnHzo CioHzz ClIH4 C12HZ8
74.1
71.1
....
60 .O
....
63.54
63.1
62.0
74.0 74.5
71.5
(Aromatic-Free Petrol)
....
....
67.5
....
64.5
.... ....
....
....
(Kerosene) 60.0
61.4
60.0
59.0
SOURCE Maybery & Goldstein Syniewski
Ricardo Trouton’s Rule
I=-
.... 1 2
NAL.
....
65.3
....
M Wilson and Barnard’ 1 From slopes of vapor pressure curves (see preceding article).
....
60.5
Maybery and Goldstein, A m . Chem. J . , 28 (1902), 28, H.E Wales, “Specific Heats of California Petroleums,’’ THISJOUR-
6 (1014).727. V. Syniewski, 2. angew. Chem., 1898, 621.
8
....
20.8T
1
co.
Kindly loaned for the purpose by the Yale and Towne Manufacturing
T H E JOURNAL OF INDUSTRIAL AND ENGINEERING CHEMISTRY
914
FUEL Kerosene
Run
Gasoline
11 12 13 14 15 16 17 18 19 20 21 22 23 24
Length of Run Min.
12 12 12 12 10 10 10
st-
8 14 14 14 12 12 12 12 8 8 8 8 8
TABLE 11 Heat Loss Correction
Determined
Corrected-
24 24 24 24 20 20 20 16 16 16 28 28 28 24 24 24 24
360 341 341 345 285 290 282 255 249 252 336 328 293 266 251 259 254 218 216 235 246 236 7971 7981
16
Q
-I
300 281 281 285 225 230 222 195 189 192 266 258 223 196 181 189 184 148 146 175 186 176
Temperature c.
16 16 16 16 40 20 Water 40 20 1433 1 From Marks and Davis’ “Steam Tables” h at 434’ C. = 797 cal. 1.8 The differencebetween observed and above value 795 - 757 = 41 cal. loss by radiation,, etc., per g. of fuel in 20 min. ’
Kerosene
Vol. 13, No. 10
- -
no appreciable change in the vapor temperatures. There was a t no time a difference of more than 5” or 6” C. between TI and T2. As indicated in the drawing, the vaporizer was heated by means of a gas blast burner. In making the determinations, the apparatus was allowed to run until substantially constant temperatures had been held a t all points for a t least 10 min. The time was then noted, and the condensing water was caught while 500 g. of fuel passed through the generator, readings being frequently taken to check the temperatures.’ All runs were discarded in- which the water temperatures varied over 1’ C. I n most cases the variation was less than 0.5” C. Several check runs were made a t each temperature, the results of which are indicated in Fig. 2. 8ince the condenser was admittedly not adiabatic, it seemed desirable to determine approximate figures for the heat losses and other correction factors by passing water through the generator and condenser, and comparing the total heat of the steam as thus determined with the very accurate values given in the Steam Tables.2 The results
... ...
r
....
tion. This correction remains constant for all determinations, as the same weight of fuel was passed through in each case. RESULTS The experimental results obtained by the foregoing procedure are recorded in Table 11. The total sensible heats have all been corrected to 0” C. It is very desirable to tie in these high temperature data with the values given in the literature for the specific heats of these hydrocarbons in the liquid state. This cannot be done indirectly, but by subtracting the value of r, the heat of vaporization of the fuel in question, from the observed figurp for the total sensible heat a t high temperatures, values may be obtained which should lie on a smooth curve with the data for the sensible heat of the liquids a t temperatures below the boiling p0int.l The values thus obtained are presented in Column 7 of Table 11, and graphically in Fig. 2 . For the comparative purposes a line is drawn to represent the extrapolation of the recent data of Bushong and Knight2 which were obtained from specific heat measurements on light petroleum distillate (liquid onIy) a t temperatures up to 100’ C. It will be noted that the agreement is not bad, considering the unjustifiable extrapolation of Bushong and Knight’s equation: Specific heat = 0.5 0.0008 t, where f = C. The best representative line for the points determined in this Laboratory lies below the curve of Bushong and Knight and accords best with the equation: Specific heat (of either liquid or vapor) = 0.5 0.0006 t It will be noted that values obtained for gasoline lie, on the average, about 5 cal. below the Q-r points for kerosene. This discrepancy, which is in the opposite direction from what might be expected, is very possibly due to the assumption of too large a value for the heat of vaporization for the gasoline. The value obtained from the slope of the vapor pressure curve of the identical fuel was, as indicated in Table 11, 6 5 . 3 cal. per g., but a value of 70 wa8 assumed in determining Q - r because it accorded better with other data in the literature. Assuming, however, that the line Q - r is correct as drawn
+
O
+
FIG. 2
were found to be 41 cal. per g. too low for 500 g. of water vaporized over a period of 20 min. The heat losses were therefore assumed to be 2 cal. per g. per min., and the results of the other determinations were subjected to this correc1 Although runs were made up t o 460’ C. there was no appreciable cracking in the short time that the gas occupied in passing through the system. This was indicated b y the absence of any appreciable change in the distillation curve of the kerosene, and the fact that there was only a trace of noncondensible gas formed or of carbon deposited. Attempts t o use a copper condenser gave trouble with cracking and the depovition of carhon. 2 Marks and Davis, “Steam Tables and Diagrams.”
1 This method of treatment of the results is, of course, rigidly applicable only in cases where the heat of vaporization is Substantially constant over the temperature range in question, and hence the specific heats of the liquid and the vapor are equal at a given temperature. That this is substantially true for the paraffin hydrocarbons in question is indicated both by data in the literature and b y the fact that the vapor pressure data obtained by the writers for these fuels gave slraighf Zincs when log f~ was
plotted against 2 LOC.
cit.
1 -. T
Oct., 1921
T H E JOURNAL OF IhTDUSTRIAL A N D ENGINEERING CHEMISTRY
FIG.3
915
FIG. 4
for both fuels (since the divergence is just about within the limit of error of the method), it is then possible to add to this line the fractional amount of the heat of vaporization which corresponds to the fractional amount of the fuel which is vaporized a t any given temperature. This was the method used in obtaining the curved lines shown in Fig. 2, which represent within a probable error of 3 or 4 per cent the values for the total sensible heats of gasoline and of kerosene as a function of temperature a t a pressure of one atmosphere. Figs. 3 and 4 give the sensible heats of kerosene-air and gasoline-air mixtures, respectively. The heats are given in gram calories for 1 3 , ~and . 16 g. of mixture, and also for 12 and 15 g. of air. The heat content of air was calculated from the equation: - 273) 0.0005(T2- 2732) H (per gram) = 6,5(T 29 where H = heat in cal. per g. T = temperature in degrees Kelvin (273 C.) These curves make it possible to read directly the temperature of the resulting air-fuel mixture if the initial temperatures of the air and the fuel are known. If, for example, the fuel (kerosene) enters the mixing chamber at a temperature of 400" C., and the air a t 50" C., then the total heat of 16g. of a 15 : 1 mixture is equal to 304 175, or 479 cnl. The final temperature of the resulting mixture corresponding to this heat content is 10.1" C. By such calculations it is possible to determine just how hot the air or fuel should be heated before mixing in order to give a dry gas mixture a t equilibrium, or indeed to determine any one of the three temperatures if the other two are known. In this connection, there has been considerable speculation as to what happens with regard to the condensation or evaporation of the fuel during the compression cycle of an Otto cycle engine. The increase in pressure taken by itself would, of course, tend to cause condensation, but the heat liberated during the compression would raise the temperature of the mixture and tend to evaporate any uncondensed fuel. Without accurate data as to the change of vapor pressure with temperature and the sensible heat of a given fuel, it was impossible to give a definite answer to these questions, but, from data presented in this and in the previous paper by the writers, definite conclusions can be drawn as follows: Making the calculations on a 5 : 1 compression ratio for Socony kerosene and assuming that the exponent k in the equation, pvk = const., for the compression is 1.2 (which makes a maximum allowance for a cooling during compression), it can readily be shown that the tendency of the
+
+
O
compression as a whole is to vaporize rather than condense the fuel, and indeed, that it would theoretically be possible to vaporize by compression a mixture of kerosene and air which was initially 75 per cent condensed. Actually, of course, the time would probably be insufficient to vaporize. this amount of material, but it seems certain that if the. mixture is nearly enough completely vaporized to give proper distribution, there will be very few, if any, liquid drops present at the moment of ignition, and that the fuel will,, under no circumstances, condense out during compression. SUMMARY This article describes certain approximate methods used by this Laboratory in determining the total sensible heat content of Socony motor gasoline and kerosene and their mixtures with air a t temperatures u p to 500" C. The resulting data are presented in such form as to make it readily possible to calculate, with sufficient accuracy for all practical purposes, the resultant temperature of an air-fuel mixture, knowing the temperature of the two constituents before mixing. Combined with data in a previous article on the dew points of various fuel-air mixtures, this information is very valuable in determining the proper conditions for securing a completely vaporized mixture of air and fuel in various
+
FIG.5
types of carburetors and heated manifolds. The results also indicate that the net effect of the compression stroke on a motor is to vaporize, rather than condense, the fuel, and hence that the most difficult problem in connection with the vaporization of the fuel is to secure satisfactory distri. bution. ,
