76-10 JLB
Impact of Gasoline Characteristics on
Fuel Economy and Its Measurement
December 1976
Technology Assessment and Evaluation Branch
Emission Control Technology Division
Office of Mobile Source Air Pollution Control
U.S. Environmental Protection Agency
-------
Background
The expanded role of the EPA in measuring and reporting motor vehicle
fuel economy data, under the terms of the Energy Policy and Conservation
Act, makes it necessary for the EPA to become very well informed on all
technical factors which may influence the accuracy of fuel economy
measurements. Also, the Importance of achieving maximum conservation of
fuel compels the investigation of all factors that may have an impact on
this goal.
Consequently, the purpose of this report is to discuss the impact of the
characteristics of gasoline on the fuel economy of motor vehicles and on
the accuracy of its measurement. It also serves to help explain some of
the differences that may be found between the results of the EPA fuel
economy tests and the values observed by vehicle owners in normal
service.
In principle, other factors such as vehicle weight, vehicle maintenance,
etc., have a greater impact on fuel economy than the characteristics of
gasoline; but a better knowledge of the impact of these fuel characteristics
can contribute to fuel conservation and to the understanding of some of
the variabilities in fuel economy measurements.
This report is based on the analysis of the available literature on the
subject, as well as on the specific information submitted—upon request
from the Emission Control Technology Division of the EPA—by companies
engaged in petroleum refining, fuel additives production, and automobile
manufacturing.
Introduction
The composition of gasoline depends on the origin of the crude petroleum
and on the refining process. Gasoline is principally a mixture of
liquid hydrocarbons having from four to ten carbon atoms, small amounts
of lighter and heavier hydrocarbons, and minute quantities of crude
petroleum impurities such as sulfur and nitrogen. Also, very small
quantities of certain additives are usually added to the gasoline in
order to inhibit such undesirable conditions as knocking, surface ignition,
spark plug fouling, rust, gum formation, and icing of the carburetor.
Apart from the effect of additives, the characteristics of a gasoline
depend on its chemical composition. However, because a gasoline con-
sists of a very large number of different hydrocarbons, its composition
is usually expressed in terms of hydrocarbon types—saturates, olefins,
and aromatics—rather than in terms of individual hydrocarbons.
-------
The characteristics of gasoline which have the greatest impact on fuel
economy are octane rating, density (as a measure of the heating value of
the gasoline), volatility, viscosity, and cleanness. The effect of
changes in these characteristics on fuel economy measurements* are
presented summarily in the following sections. An additional section
refers to the optimum octane ratings of gasoline. Details and comple-
mentary general information regarding the characteristics of gasoline
are given in the Appendix.
The octane number requirements of given populations of motor vehicles
depend basically on the design of their powertrain systems. Thus, such
requirements are outside the specific area of concern of this report and
are not dealt with here.
Except for a few comments, the report does not cover the impact of the
characteristics of gasoline on the emission of air pollutants. This
area of interest will be the subject of a separate or complementary
report.
Impact of Gasoline Density
The fuel economy of motor vehicles is commonly expressed in miles per
gallon (mpg). However, the mpg does not provide an accurate indication
of the fuel economy of the vehicles unless the density of the gasoline
is taken into consideration. This is because the heating value per unit
volume of gasoline increases with its density. Of all the variations in
gasoline characteristics which affect fuel economy, the variations in
density are one of the more significant. For example, data indicates
that in 1974 the density of some commercial gasolines ranged between 5.7
and 6.4 lb/gallon.27 in the extreme case where tests of fuel economy
are made using gasolines falling at the lowest and highest densities of
this range, a difference in fuel economy of as large as 7% would be
indicated if there is no correction made for gasoline density. Or-
dinarily, comparisons are not made involving the extreme cases of gasoline
density and, consequently, the variations in the fuel economy of motor
* To separate the effects of other variables, the impact of the char-
acteristics of gasoline on fuel economy should be determined on the basis
of the same test procedure (that is, for the same vehicle soak con-
ditions, driving cycle, atmospheric conditions, etc.). Therefore, the
report will distinguish between fuel economy differences measured using
only the EPA test procedure and the differences observed when results
from the EPA tests are compared with customer's measurements.
-------
vehicles due to variations in gasoline density are much less than this
figure. However, the corresponding differences in fuel economy can
still be significant, and, therefore, should not be overlooked.*
Thus, for accurate evaluations or comparisons of fuel economy, the mpg
should be measured using gasoline of a fixed density. Alternatively,
the mpg should be corrected for density, to refer all the results to
those of a fixed density of gasoline. In other words, the reported
values of fuel economy should be:
mpg = mpg x Correction Factor.**
Impact of Differences in Volatility
As is indicated in Section IV of the Appendix, there is very little data
regarding the effect of gasoline volatility on fuel economy. Also,
there is the general belief that the impact of this variable on fuel
economy is minor29> for the soak conditions, ambient temperature, and
driving cycle of the EPA test procedure. However, for other operating
conditions the volatility of the gasoline may have a substantial impact
on fuel economy. For instance, volatility greatly affects the engine
starting and engine warm-up period, and these operating modes may have a
very significant impact on the fuel economy of short trips, particularly
at low ambient temperatures. In such an instance, the shorter the trip
and the lower the ambient temperature and the volatility of the gasoline,
the lower would be the fuel economy compared with the fuel economy
measured during the EPA tests.
Impact of Differences in Octane Rating***
For a given motor vehicle, the impact of the octane ratings of different
gasolines would be as follows. If the engine does not knock with either
the fuel used in the EPA tests or the commercial fuel, the difference in
octane ratings of the two fuels should not result in any variations in
fuel economy. If the engine does not knock with the EPA test fuel but
knocks with another fuel, the eventual difference in fuel economy due to
* For the emission tests for certification of automobiles, Indolene HO
III is used as fuel. In 1974 the density of this gasoline ranged bet-
ween 6.10 and 6.22 lb/gallon.27 The maximum variation in fuel economy
which is possible for this change in density is approximately 1%.
** Section III in the Appendix indicates means for determining the cor-
rection factor.
*** This section refers only to the case of given vehicles, with fixed
engine compression ratios and adjustments. The general relationship
among octane rating, compression ratio, and engine efficiency is dis-
cussed in Section II of the Appendix.
-------
the difference in octane ratings would depend on the level of knock. If
the knock is sufficiently severe that spark timing must be retarded,
fuel economy would suffer. Otherwise, the variations in fuel economy
using the EPA test procedure would not be significant enough to be
measurable since knock would occur only during a small fraction of the
EPA driving cycle.
For a driving cycle and operating conditions different from those of the
EPA test procedure, the impact of differences in octane rating on fuel
economy would still be rather small, in general, but could become signi-
ficant. If the engine knocks very lightly its fuel economy could be
slightly better because of the small improvement in engine efficiency
associated with faster combustion. On the other hand, if the level of
knock in any cylinder is beyond its optimum level, the fuel economy
could decrease because of the greater heat losses associated with knock.
The magnitude of the fuel economy difference would depend on the intensity
of knock and on the percentage of time operating under knock. The fuel
economy variation could be very substantial for operation under severe
knock.
Impact of Variations in Other Gasoline Parameters
There is no information to quantify what changes in fuel economy can
result from changes in fuel metering caused by variations in the vo-
latility, viscosity and surface tension of the gasoline. Within the
constraints of the Federal Test Procedure the impact of these variations
are believed to be minor; However, substantial departures from the
conditions of the EPA test procedure—particularly regarding ambient
temperature, because of its effects on viscosity and volatility—could
possibly result in significant variations between the EPA and customer
fuel economies.
As has been indicated above, changes in gasoline density have an impact
on fuel economy, but this is not due to fuel metering effects. Although
variations in density change the carburetor air-fuel ratio, such varia-
tions have a negligible or small effect on the air-fuel equivalence
ratio of the mixture.
The use of unclean gasoline leads to deposits which may alter engine
parameters and adjustments. This can result in significant differences
between the fuel economies measured before and after the alterations
caused by unclean gasoline.
-------
Optimum Octane Rating of Gasoline
With each succeeding new car model year the use of unleaded fuel is
rapidly increasing. This is because the catalyst equipped new cars can
only operate on unleaded fuel due to the deleterious effect of lead
compounds on catalyst efficiency. The only other currently available
antiknock additives can provide only a 2 octane boost or less. On the
other hand, automobile manufacturers are striving to maximize engine
compression ratios for optimum fuel economy within the constraints of
the emission regulations, which retains the need to keep octane levels
up. Thus in order to supply enough unleaded fuel at an octane level
appropriate to satisfy the new vehicles with maximum conservation of
petroleum, new refining techniques must be employed. Refining penalties
for such an action would include the cost of new refining facilities.
Another factor which the petroleum companies must deal with is the
widespread of octane requirements that their product must satisfy.
Historically they have marketed two grades of leaded fuel with approxi-
mately two thirds of the customers using regular grade and the rest
requiring premium fuel. One grade of unleaded fuel has more recently
been added for use by the newer cars. However, because of the many
factors which affect octane requirement, the vehicles designed for the
available 91 RON unleaded fuel actually have requirements varying by
about 10 octane numbers. The time is nearing when companies must decide
whether to switch to two grades of unleaded fuel and retain only a
single grade of leaded fuel. By going to two grades of unleaded fuel,
one lower and the other higher than the present 91 RON, they could still
satisfy the majority of their customers with minimal effect on their
average marketed octane level.
In summary, each petroleum company is faced with a number of difficult
decisions affecting day to day operations and long term planning regarding
the optimum octane number of their products. The goal is to provide no
more octane than is necessary to obtain the most transportation per
barrel of crude.
Conclusions Regarding the Impact of Gasoline Characteristics on Fuel
Economy Measurements
1. For the EPA fuel economy tests, it appears that the effect of
changes in fuel characteristics is minor. However, the available in-
formation permits an accurate assessment only of the impact due to
changes in fuel density. The maximum variation in fuel economy due to
changes in density of the EPA certification fuel was approximately 1%
for the density range of Indolene HO III marketed in 1974.
-------
2. For fuel economy tests carried out with the EPA procedure but using
commercial gasolines, it is also believed that the impact of changes in
the gasoline is usually small. A possible exception is the impact of
changes in gasoline density. For extreme values of density of com-
mercial gasolines a difference in fuel economy of as large as 7%* could
be indicated if there is no correction made for gasoline density.
Ordinarily, comparisons are not made involving the extreme cases of
gasoline density and, consequently, the variations in the fuel economy
of motor vehicles due to variations in gasoline density are much less
than this figure. However, the corresponding differences in fuel economy
can still be significant, and, therefore, should not be overlooked.
3. The variations in fuel economy due to changes in fuel density can
be corrected by referring all the results to those of a fixed density of
gasoline. The present EPA test procedure assumes a constant density and
no such correction is made. Our records on Indolene fuel used for EPA
certification indicate that the density has been held nearly constant
for several years even though commercial fuels have varied on the average.
In the event the density of the certification fuel should vary signi-
ficantly, a correction factor can be applied as discussed in Section III
of the Appendix.
4. Independently of the gasoline used, fuel economy measurements
depend very significantly on the atmospheric conditions (particularly
temperature), vehicle soak times, and driving cycle (particularly trip
length, speeds, and rates of speed changes). In addition, the impact of
certain characteristics of gasoline on fuel economy (mainly volatility
and viscosity) depends on the very same variables. However, there are
no data to quantify (for any given vehicle) the complex relationship
among fuel economy, gasoline characteristics, atmospheric conditions,
and driving cycle. Therefore, what can be said is that the more the
driving cycle, soaking periods, atmospheric conditions, and characteristics
of the gasoline for a vehicle in actual service differ from those of the
EPA test procedure, the greater the differences that can be expected
between the corresponding fuel economies.
5. For a better assessment of the relative impact of the different
characteristics of gasoline on fuel economy, data are needed on the
specific effects of the volatility, viscosity, and surface tension of
gasoline. Likewise, data are needed to evaluate the impact on fuel
economy of burning, in the engines, the evaporative emissions collected
by their control equipment.
6. The available information does not indicate a need for varying the
specifications^ of the fuel for the EPA fuel economy tests.
* 27
For the density range of some gasolines in 1974.
-------
References
1. "The High-Speed Internal - Combustion Engine," H. R. Ricardo,. Blackie &
Son Ltd., 4th ed., 1960.
2. "Thermodynamic Charts for Combustion Processes," part one, text,
B.C. Hottel et al., John Wiley and Sons, Inc., 1949.
3. "The Influence of Compression Ratio and Dissociation on Ideal Otto
Cycle Engine Thermal Efficiency," M. H. Edson, SAE TP-7, 1964.
4. A. E. Felt, Ethyl Corporation, presentation at the SAE Octane Level
Symposium, Houston, Texas, June 4, 1975.
5. "Internal Combustion Engines," E. F. Obert, International Textbook
Company, 3rd. ed., 1968.
6. ASTM Standards, pt. 7, D909-60T p. 1329 and D1656-62T, p. 1316, 1961.
7. "Knock Characteristics of Hydrocarbon Mixtures," E. J. Y. Scott,
Proc. API, 38:111, pp. 90-111, 1958.
8. "Influence of Leaded and Unleaded Fuels on ORI in 1971 Model Cars,
Phase I: 1970-71 CRC Road Rating Program," CRC Report 451, September
1972.
9. "Octane Number Requirement Survey 1973," CRC Report 467, May 1974.
10. "Octane Requirement Increase in 1973 Model Cars, Phase II: 1973
CRC Road Rating Program," CRC Report 476, February 1975.
11. "High Mileage Supplement to the 1973 CRC Octane Number Requirement
Survey," CRC Report 480, August 1975.
12. "Octane Number Requirement Survey 1974," CRC Report 479, August 1975.
.13. "Some Observations of Factors Affecting ORI," F.E. Alquist et al.,
SAE paper 750932.
14. "Some Factors which affect Octane Requirement Increase," J. D. Benson,
GM Research Lab., SAE paper 750933.
15. Courtney, R. L., Chevron Research Company, Discussion of SAE paper 750933,
October 16, 1975.
16. "The Use of Combustion Deposit Analysis for Studying Lubricant-
Induced ORI," H. E. Bachaman et al., SAE paper 750938.
-------
17. "Octane Requirements of 1975 Model Year Automobiles Fueled with
Unleaded Gasoline," Report No. 75-28, JLB, TAEB, ECTD, EPA,
August 1975.
18. "Optimum Octane Number for Unleaded Gasoline," T. 0. Wagner and
L. W. Russum, Amoco Oil Company, SAE paper 730552, May 1973.
19. "Antiknock Compounds - Research, Development, and Refinery
Application," Gibson, H. J. et al., Ethyl Corporation, Fifth
World Petroleum Congress, New York, May 1959.
20. "An Evaluation of Manganese as an Antiknock in Unleaded Gasoline,"
J. E. Faggan et al., Ethyl Corporation, SAE paper 750925, October 1975.
21. "Toxicologic Evaluations of Fuel Additive - Methylcyclopentadienyl
Manganese Tricarbonyl (MMT)," W. Moore et al., EPA, SAE paper 750927.
22. "The Effect of Emission Standards and Gasoline Quality on Fuel Con-
sumption," E. N. Cantwell et al., DuPont, SAE paper 750671, June 1975.
23. Presentation on "Octane Optimization," Exxon Corporation, Washington, D.C.,
May 7, 1975.
24. "Value of High Octane Number Unleaded Gasolines in the U.S.," E. S. Corner
et al., Esso Research and Engineering Company, presentation to the ACS,
Los Angeles, March 28-April 2, 1971.
25. Personal communication from M. Nager, Shell Oil Company, to
J. L. Bascunana, EPA, July 16, 1975.
26. "1975 SAE Handbook," Society of Automotive Engineers, Inc.,
New York, New York, 1975.
27. Personal communication from H. S. Seelig, Amoco Oil Company, to
J. L. Bascunana, EPA, July 22, 1975.
28. American Society for Testing and Materials, 1974 Annual Book of
ASTM Standards.
29. "Automobile Fuel Economy," Motor Vehicle Manufacturers Association,
Detroit, September 21, 1973.
30. Personal communication from J. F. Kendrick, Gulf Oil Company -
U.S., to J. L. Bascunana, EPA, July 25, 1975.
-------
31. "Emissions from Combustion Engines and Their Control," D. J. Patterson
and N. A. Henein, Ann Arbor Science Publishers Inc., 1972.
32. "The Internal Combustion Engine," C. F. Taylor and E. S. Taylor,
International Textbook Commary, 2nd ed. 1961.
33. "Effect of Fuel Density on Carburetion," R. J. Wahreubrock,
Ethyl Corporation, internal memorandum ERM-H13, April 28, 1953.
34. "Influence of Fuel Properties on Metering in Carburetors,"
J. A. Bolt et al., SAE paper 710207.
35. "Symposium on Fuel Injection," September 9, 1955, SAE SP-140.
36. "Symposium on Petrol Injection," Proceedings of the Automobile
Division, The Isntitution of Mechanical Engineers, 1957-58, No. 6.
37. "Exhaust Emission Control by the Ford Programmed Combustion Process-
PROCO," A. Simko et al., SAE paper 720052.
38. "Texaco's Stratified Charge Engine—Multifuel, Efficient, Clean, and
Practical," M. Alperstain et al., SAE paper 740563.
39. "The General Motors Fuel Injection System," J. Dolza, E. Kehoe, and
Z. Arkus-Duntov, SAE preprint, #16, Detroit, January 14-18, 1957.
40. "Electronically Controlled Gasoline Injection System," Robert Bosch
technical bulletin BB 700/1A, January 15, 1968.
41. "Lean Burn Engine Concepts—Emissions and Economy," J. E. A. John,
SAE paper 750930.
42. "An Oil Company's Reaction to Fuel Injection," J. E. Taylor, Gulf
Research and Development Co., SAE SP-140, Symposium on Fuel Injection,
September 9, 1955.
43. "Detonation -and Internal Coolants", E.F. Obert, SAE Quarterly
Transactions, January 1948, Vol. 2, No. 1.
44. "ORI of Today's Vehicles," B. D. Keller et at., SAE paper 760195.
45. "What additives do for gasoline", P. Polss, Hydrocarbon Processing,
February 1973.
46. "Antiknock Antagonists", H. K. Livingston, Industrial and Engineering
Chemistry, Vol. 43, March 1951.
-------
10
47. "Mechanism of Sulfur-Alkyllead Antagonism", R. L. Mieville et at.,
I&EC Product Research and Development, Vol. 6, No. 4, December 1967.
48. "A New Look at High-Compression Engines", Caris et al., SAE Trans-
actions, 1959.
49. "Determination of True Engine Friction," R. E. Gish et al., SAE
Transactions, 1958.
50. "New look at auto-fuel economy vs. refining," E. C. Brown et al.,
The Oil and Gas Journal, Sept. 29, 1975.
51. "Motor Gasolines, Winter 1975-76", ERDA, No. BERC/PPS-76/3 June 1976.
52. "Technique for Determination of Octane Number Requirements of
Passenger Cars", CRC Designation E-15-76, Coordinating Research
Council, Incorporated.
53. "1973 CRC Fuel Rating Program. Part I: Road Octane Performance
in 1973 Model Cars", CRC Report No. 477, Coordinating Research
Council, Inc., Feb. 1975.
54. "ASTM Specifications for Petroleum Products", 1976.
55. "Evaluation of Expressions for Fuel Volatility", CRC report No. 403,
Coordinating Research Council, Inc., November 1967.
56. "The Bosch Continuous Injection System - A Mechanically Operating
System for Continuous Gasoline Injection", Robert Bosch GMBH, 1973.
57. Federal Register, Vol. 38, 1255, Jan. 10, 1973.
58. Federal Register, Vol. 38, 33741, Dec. 6, 1973.
59. Federal Register, Vol. 41, 42675, September 28, 1976.
60. Federal Register, Vol. 41, 38682, September 10, 1976.
-------
APPENDIX
-------
APPENDIX
TABLE OF CONTENTS
Section
I. Insensitivity of the Thermal Efficiency of the
Spark Ignition Engine with regard to the Hydro-
carbon Composition of the Gasoline 1-1
II. Fundamentals concerning the Octane Rating
of Gasoline II-l
II.A. Relationship Between Engine Efficiency and
Compression Ratio II-l
II.B. Knock, the Variables which Affect It, and
Its Effects on Engine Performance II-3
II.C. Knock Rating of Gasolines H-6
II.D. The Octane Requirement Increase (ORI). .... II-7
II.E. Octane Number Requirements versus
Compression Ratio II-8
II.F, Anti-knock Additives 11-10
II.G. Optimum Octane Rating of Gasoline 11-11
III. Density III-l
IV. Volatility IV-1
IV.A. Common Ways of Identifying the Volatility
of Gasoline IV-1
IV.B. Volatility versus Cold Starting, Hot
Starting, and Vapor Lock IV-3
IV.C. Carburetor Icing IV-3
IV.D. Impact of Volatility on Warm-up and Normal
Operation of the Engine IV-A
IV.E. Effects Relative to the Tail-end of the
Distillation Curve IV-4
IV.F. Optimum Volatility, and Impact of Volatility
on Fuel Economy IV-5
IV.G. An Estimate of the Impact of Gasoline
Volatility on Fuel Economy IV-8
-------
ii
Section Page
V. Relationship Between Fuel Characteristics,
Fuel Metering, and Fuel Economy V-l
V.A. Impact of Fuel-Air Mixture Strength on
Engine Efficiency V-l
V.B. Impact of Changes in Fuel Density on Fuel
Metering V-3
V.C. Impact of Variations in Fuel Viscosity on
Fuel Metering V-6
V.D. Impact of Variations in Fuel Volatility and
Surface Tension on Fuel Metering V-8
VI. Gasoline Cleanness and Additives VI-1
VII. Gasoline Injection versus Carburetion VII-1
VILA. Classification of Gasoline Injection System. . VII-1
VII.B. Pros and Cons of Gasoline Injection. ..... VII-2
-------
1-1
I. Insensitivity of the Thermal Efficiency of the Spark Ignition
Engine with regard to the Hydrocarbon Composition of the Gasoline
The thermal efficiency of an engine is defined as the ratio of its
work output to the heating value of the fuel consumed to yield this
output:
Work output
Heating value of the fuel consumed
Except for variations in gasoline composition which might affect
its knocking resistance — an aspect which is discussed in the following
sections of the report — the composition of the gasoline does not
significantly affect the thermal efficiency of the spark ignition engine.
That is, as long as: a) the engine has the same compression ratio, and
b) the gasoline is vaporized and delivered Jo the cylinders at the same
temperature and fuel-air equivalence ratio, the thermal efficiency of
the spark ignition engine is not affected perceptibly by the composition
of the gasoline. In fact, it has been found that under these conditions
the operation of the conventional spark ignition engine with any pure
hydrocarbon fuel results in the same thermal efficiency to within less
than 1 percent. Ricardo was one of the first researchers to find
experimentally this fundamental characteristic. Analytical results
from other investigators corroborate this finding. '
This characteristic of gasoline can be explained basically as
follows. The composition of hydrocarbon fuels affects the thermal
efficiency of the engine mainly through its impact on the specific
heats of CCL and H-0, and on the dissociation of these gases. But it
turns out that the influence of these factors tends to balance in such a
way that neither carbon nor hydrogen is preferable as the main consti-
tuent of the fuel on the combined grounds of dissociation and change of
specific heats. Therefore, the nature and chemical composition of the
fuel, so long as it consists only of carbon and hydrocarbon, has no
significance on the efficiency of an engine.
*
The fuel-air equivalence ratio 0 is defined as:
Mass rate of fuel Fuel-air ratio
0 =
Stoichiometric mass rate of fuel Stoichiometric fuel-air ratio
-------
II-l
II. Fundamentals Concerning the Octane Rating of Gasoline
II.A. Relationship between Engine Efficiency and Compression
Ratio
The efficiency of the spark ignition engine is related fundamentallj-
to its compression ratio. The simplest thermodynamic model of this
engine shows this. That is, the so called "air-cycle" efficiency of the
spark ignition engine is:
= 1 - (l/r)*"1 (II.A.I)
where r is the compression ratio of the engine, and k is the ratio c /c
of the specific heats of the air at constant pressure and constant
volume.
Of course, the efficiency of an engine also depends on factors
other than r and k — such as engine design and operating load and
speed — and, therefore, the simple theoretical expression given by
equation (II.A.I) is not appropriate for determining absolute values of
the efficiency. But this simple expression can provide fairly close
estimates of the variations in the relative efficiency with changes in
the compression ratio. This is illustrated in figure II.A.I (taken from
reference 4) which compares the relative fuel economy based on equation
(II.A.I) with those corresponding to actual data from nine different
multi-cylinder engines having displacements from 85 to 413 cubic inches.
The theoretical advantage of having a compression ratio as high as
possible is, in practice, limited by certain constraints. The most
important of these constraints are discussed briefly in what follows.
One limitation on higher compression ratios is due to the fact that
the actual, overall efficiency of the engine, the so called "brake
thermal efficiency," starts to decrease beyond a certain value of the
compression ratio. The equation (II.A.I) shown above corresponds to a
thermodynamic process which does not consider the reduced flame speed
and increased heat loss and engine friction, all of which appear to be
associated with the higher compression ratios. ' The highest value
of compression ratio for optimum brake thermal efficiency depends on
engine design and operating conditions, but in general the optimum
compression ratio of the gasoline engine is probably not higher than
around 16 or 17.
Another limitation occurs because as compression ratios increase
beyond about 10, preignition may appear due to glowing of combustion
chamber deposits. The so-called "rumble" is a low-pitched thud probably
caused by early surface ignition raising the pressure greatly in the
cylinders with consequent deflection of mechanical parts.
-------
II-2
1.32
1.28
1.24
Relative
Fuel
Economy
1.20
1. 16
1. 12
1.08
1.04
1.00 *
9 10
C. R.
11 12
Figure li.A.I. Relative Fuel Economy'vs. Compression Ratio
-------
II-3
The limit on compression ratios due to these effects of brake
thermal efficiency change or to surface ignition can be reached only if
the spark ignition engine is operated with a non-knocking fuel. In
practice, using commercial gasolines the compression ratio of the spark
ignition engine is limited to lower values because of the appearance of
knock.
Also, the control of some exhaust pollutants below certain values
may pose another constraint on the compression ratio of the engine.
Significant investigations have been and are being made regarding the
relationship between engine compression ratio, exhaust emissions, and
engine efficiency; but to assess this eventual limiation on the com-
pression ratio due to emission constraints, there is a need for more
comprehensive data — which should include the effect of exhaust emis-
sion controls.
II.B. Knock, the Variables which Affect It, and Its Effects on
Engine Performance
Combustion in the spark ignition engine depends basically on engine
design and fuel quality. Under normal conditions, the flame initiated
at the spark plug spreads evenly across the combustion space until all
the gasoline has been burnt. The spreading of the flame results in an
increase in temperature and pressure in the "end gas," which is that
part of the fuel-air mixture which the flame has not yet reached. This
increase in temperature and pressure in the end gas causes the. gas
mixture to undergo preflame reactions. If the ignition delay — i.e.,
the period which exists before a reaction becomes explosive — of the
end gas fully occurs before the flame arrives, "autoignition" takes
place. With autoignition the combustion process becomes uncontrolled
and an abrupt rise in pressure may occur. This local pressure rise may
induce vibration of the walls of the combustion chamber or other parts
of the engine, resulting in a knocking sound. Thus, "knock" in the
spark ignition engine is due to sudden autoignition of the gas mixture
near the end of the combustion period.
The most significant variables that control autoignition are the
composition of the fuel and the following factors affecting the combustib
mixture:
Temperature
Density
Ignition delay
Fuel-air ratio
Homogeneity
Thus, because of the effects of these variables, knock in the spark
ignition engine:
-------
II-4
Increases with a lower octane rating of the fuel
Increases with the compression ratio
Increases with engine load
Increases with lower engine speed
Increases with inlet air temperature
Increases with engine coolant temperature
Increases with the supercharging of the inlet mixture
Decreases with spark retard
Decreases with higher inlet air humidity
Decreases with higher turbulence of the gas mixture
Decreases for either rich or lean mixtures
Decreases by stratifying the mixture (so that the
end gas is less reactive)
Depends on the combustion chamber design.
Knock of high enough intensity can cause engine damage. The pressure
waves associated with knock increase the rate of heat transfer to — and
hence the temperature of — the susceptible parts, and this may result
in local melting of the material or in its softening to such an extent
that the high local pressure causes erosion. Most-frequently, however,
knock results in engine failure by causing pre-ignition. That is, knock
of high enough intensity heats some part of the combustion chamber to
the point that the fuel-air mixture is ignited before the ignition spark
occurs. Then, pre-ignition, if not checked, gets progressively worse,
culminating in engine failure.
Under knocking the gases in the combustion chamber vibrate with
heat losses as a result, but at some level of light-knock the gain from
the faster combustion associated with knock can be greater than the heat
losses caused by the vibrating gases. Thus, data indicates that the
maximum output and fuel^economy of an engine are obtained under (Controlled
light-knock conditions.
It should be pointed out that advancing the ignition timing beyond
its optimum setting results in deterioration of engine performance.
In the case of optimum ignition timing under light-knock, advancing
the ignition will not only decrease the power and .the fuel economy but
also increase the knock, and thus lead to engine failure. In the case
of an engine operated under conditions such that its octane number
requirement is below the octane number of the fuel, advancing the
spark beyond the optimum timing will decrease the power and the fuel
economy, and may or may not result in engine knock.
*
Supporting information is found in reference 5, pages 109 and
298. This data was originally presented in reference 43.
-------
II-5
Whereas a certain level of light-knock can be beneficial from the
important viewpoint of fuel economy, there are some questions about the
feasibility of extended use of light-kn.ock operation. The first difficulty
associated with permitting light-knock operation is rating the level of
knock. Currently, the only practical method of measuring knock is by
ear; therefore, the intensity of knock may vary according to the particular
observer. Rating the level of knock is much more difficult than simgly
distinguishing the following cases included in the current CRC E-15
procedure for determining octane requirements:
1) no knock
2) borderline knock
3) above borderline knock.
Also, if some level of light knock is permitted within the normal
ranges of speed, load, and weather, it must be considered that the knock
will be higher under other conditions of operation which are more prone
to induce knock.*** Furthermore, even if the knock would always be
within a certain limit, there is a lack of information about the effect
that this knock could have on engine durability and performance if it
would occur for substantial periods of time. There are differences in
opinion of whether the effects of such a substantial occurrence would be
serious or not.
Therefore, it appears that to allow for the general use of light-
knock operation with confidence that it will not cause serious hardship,
two precautions should be taken. First, it would be necessary to establish
a procedure to properly measure the level of light-knock. Second, it
would be required to determine what level of light-knock would be permissible
without penalizing performace or durability. However, the determination
of a feasible level of light-knock is further complicated by the fact
The expression "light-knock" is used in this report to indicate
the region of knock intensity immediately above the borderline of
knock. Some literature uses expressions such as "audible knock" to
refer to the first portion of the knock region beyond "trace" or
"borderline knock".
** 52
CRC E-15 refers to a direct method used for obtaining maximum
octane requirements of cars under normal service. This method was
developed by the Coordinating Research Council for use by all partici-
pants in its periodic new-car Octane Requirement Surveys. The maximum
octane requirement of a car is given by the octane rating of the lowest
octane fuel which is required to avoid knock of a level higher than
border line during the complete CRC E-15 test. For most ordinary driving
conditions,the octane requirement of a car may be substantially below'its'
maximum octane requirement.
***
The calibration of engines for light-knock would decrease the margin
of safety which may be needed to prevent engine damage from heavier knock
under extreme conditions of operation. It is conceivable, however, that
advanced developments in knock sensors and in ignition timing control
technology could, at some future time, ensure that light-knock operation
is maintained within the limits for engine safety and efficiency.
-------
II-6
that different engines may be able to tolerate different levels of
knock. Furthermore, the knock levels of cars of th.e same make and
model may differ substantially (see section II.G).
II.C. Knock Rating of Gasolines
The knock rating of a gasoline is found by comparing its knock
response with that of a blend of "primary reference fuels" (PRF). These
fuels are normal heptane with an "octane number" (ON) of 0, and isooctane
(2, 2, A trimethyl pentane) with an octane number of 100. A blend
containing x percent (bv^volume) isooctane is defined as an x octane
primary reference fuel.
Several methods of knock rating of gasoline are nsed. In each
of these methods a special standard engine must be run' under pre-
scribed operating conditions (of speed, temperature, etc.). The octane
rating of a gasoline may have different values for different tests. Some
fuels are relatively insensitive to such changes while others are quite
sensitive. The two most common octane rating tests are known as the
Research and Motor methods, and their corresponding ratings are indicated
as Research Octane Number (RON) and Motor Octane Number (MON). The
difference between RON and MON is called the "sensitivity" of the fuel.
Because of this sensitivity, the knocking characteristics of^a^gasoline
cannot be determined specifically by a single octane number.
The knock rating of a gasoline depends on its chemical composition
and on the structure of its molecules. The various types of the hydro-
carbons that compose gasoline behave differently in their preflame
reactions and, therefore, in their tendency to knock. Nevertheless,
there is no precise relationship between chemical structure and anti-
*
Also, in multicylinder engines knock occurs first in the cylinder
where conditions are more prone to induce it. Therefore, it would
not be possible, with current engine technology, to have all cylinders
operating at the level of light-knock for best efficiency.
**
The rating of fuels with ON higher than 100 is made in terms of
isooctane plus tetraethyl,lead (TEL), and defined as the "performance
number" (PN) of the fuel. The relationship between ON and PN is such
that ON (above 100) - 100 + (PN-100)/3.
Due to differences between the standard engine for rating fuel and
the engines installed in the vehicles, as well as to differences in
operation, a thorough evaluation of knock for gasoline and engines is
complex. The relationship involved Is expressed by regressions
of road octane rating versus Research and Motor octane numbers and
variables reflecting the hydrocarbon composition of the gasoline. Details
on these matters can be found in reference 53.
-------
II-7
knock performance in an engine. Members of the same hydrocarbon series
may have very different anti-knock characteristics; for example, normal
heptane and normal pentane, both paraffins, have octane numbers of 0 and
62, respectively. In broad terms, however, it can be stated that aromatic
hydrocarbons (e.g., benzene and toluene), and highly branched iso-
paraffins (e.g., isooctane) and olefins (e.g., diisobutylene) have high
knock resistance. In an intermediate position are iso-paraffins with
little branching and hydrocarbons of the naphthene family (e.g.,
cyclohexane), while normal paraffins (e.g., normal heptane) have low
anti-knock values. Also, in general, and always in the case of straight-
chain paraffins, the knock resistance improves as the number of carbon
atoms in the molecule decreases.
The anti-knock characteristics of a blend of two pure hydrocarbons
is usually between those of the constituents. Generally, the knock
rating of the mixture is somewhat lower than if there was a linear
relationship between the knock ratings and the percentages of the two
compounds. However, some hydrocarbons — such as isooctane and di-
isobutylene — blend synergisticly, resulting in a knock rating higher
than that expected from a linear relationship.
In automotive literature reference is made to "chemical" and
"mechanical" octane numbers. The "chemical" octane numbers are those
that have been covered in this section, and they refer specifically to
the fuel and its additives. On the other hand, by "mechanical" octane
numbers are meant those anti-knock gains associated with engine design
(such as combustion chamber design, ignition control, etc.) and engine-
transmission relationships.
II.D. The Octane Requirement Increase (OKI)
As a new motor vehicle goes through its normal break-in process,
deposits accumulate in the combustion chamber and eventually reach a
stabilized level. The sources of these deposits are the fuel, the
lubricant, and the air which enters the combustion chamber. The compo-
sition and the quantity of the deposits depend on: a) the physical and
chemical natures of the fuel and the lubricant, b) the additives in the
fuel and in the lubricant, c) the operating conditions of the engine
(which are determined by the driving pattern of the vehicle), d) the
weather and climate, e) the location in the combustion chamber, and f)
the design of the powertrain system of the vehicle.
The deposits increase the octane requirements of the engine.
Significant reasons for this increase are the higher end-gas temperatures
caused by the insulating effect of the deposits, and the increase in
-------
II-8
the compression ratio caused by the volume of the deposits. The pos-
sibility of catalytic effects of the deposits has also been mentioned as
another cause.
The "octane requirement increase" is defined as:
OKI = (octane requirement of engine with stabilized deposits) -
(octane requirement of clean engine).
Data indicate that while cars operated with leaded fuel reached
stabilized octane requirements before accumulating 5,000 miles, cars
operated with unleaded fuel require more miles to reach stabilized
requirements. Also, the ORI for cars operatecLwith..unleaded gasoline is
higher than the ORI for leaded-fuel cars. » » » » Unleaded gasolines
result in combustion chamber deposits which have a composition with a
higher carbon to metal ratio than the deposits from leaded gasolines.
Since carbon is a good heat insulator and metals are good heat conductors,
the increase in carbon to metal ratio of the deposits may be a basic,,
reason for the higher ORI of the cars fueled with unleaded gasoline.
However, the available information regarding how large the differences
are between the ORIs, and between their stabilization mileages, for cars
operated with unleaded s.nd leaded gasolines, is not conclusive. Further-
more, new ORI data must be obtained periodically to take into account
the effect of changes in gasolines, lubricants, and-theirnadditives, and
., ,.,. ,- , . . , , Ljy ±4, ij , ID , 4H
the effect of changes in powertrain system designs.
The ORI is one of the^important determinants of the octane require-
ments of motor vehicles. A better understanding of the mechanisms
which produce the ORI and of the impact of the variables involved would
lead to its decrease and, therefore, to a more efficient use of gasoline
and engines.
II.E. Octane Number Requirements versus Compression Ratio
Although the octane number requirements of engines depend very
strongly on the compression ratio, there is no unique relationship
between these two variables. This is due to the influence of the other
variables (indicated in Section II.B) which affect engine knock. Thus,
the correlations quoted in the literature vary somewhat, depending on
the particular vehicles, fuels, and conditions of operation.
Figure II.E.I presents the general trend of the octane number (MON)
requirements for automobiles versus compression ratios — for the model
years 1956 through 1972 — for different percentages of automobiles
satisfied with a given octane number. This figure is taken from reference
18. It shows the results obtained by a regression analysis of the data
accumulated by the Coordinating Research Council during its surveys for
octane number requirements.
* As indicated before, the octane number requirements of automobiles
are normally determined using the CRC E-15 method. This method is
currently under review for improving its reliability and accuracy.
-------
II-9
CC
UJ
en
5
3
Z
UJ
O
o
O
H-
O
94
92
90
88
86
84
82
80
i r i i
TRACE KNOCK INTENSITY
% SATISFIED
RANGE TO INCLUDE
— MOST DATA "
J I I
7 89 10
COMPRESSION RATIO
11
Figure II.E.I. Motor Octane Number vs. Compression Ratio-
General Trend for Automobiles
-------
11-10
II.F. Anti-Knock Additives
There are certain compounds soluble in gasoline which may decrease
the tendency of the fuel-air mixture to autoignite. It is believed that
the addition of these compounds results in a sufficient reduction of the
chain reactions which produce the highly energized particles or radicals
of pre-flame reactions. Thereby these additives delay the autoignition
of the end gas and permit the normal flame to pass through it without
combustion knock.
Ideally, anti-knock additives should have the following require-
ments:
a. Non-toxic.
b. Do not contribute to the formation of toxic substances.
c. No deposits left in the engine.
d. Relatively low boiling temperature to ensure good distribution
in multicylinder engines.
e.. Complete solubility.
f. Stable.
g. Low cost per unit increase in octane rating.
The most effective and extensively used anti-knock additives are
certain lead compounds. Specifically, tetraethyl lead (TEL), (CJL) Pb,
is the primary commercial anti-knock additive. The burning of TEL
results in lead compounds that condense readily and therefore must be
removed to avoid increased deposits in the combustion chamber. To
prevent these deposits, scavenging agents, mainly chlorine and bromine
compounds, are blended with the TEL.
Another lead compound frequently used in gasoline is tetramethyl
lead (TML), (CH.,), Pb. In general, this compound is less effective as
an anti-knock agent than TEL. However, in some cases it can be more
effective, primarily because it is more volatile than TEL and, therefore,
can provide better distribution in some multicylinder engines. Likewise,
a mixture of both (TEL-TML) may be more effective for some engines.
The lead-compound particulates which are exhausted into the atmosphere
can result in health hazards. Therefore, the EPA has ruled that effective
October 1, 1979, the maximum allowed average content of lead in motor
gasoline will be 0.5 g/gallon. „ Furthermore, the marketing of unleaded
gasoline has become compulsory for motor vehicles equipped with catalytic
converters for the control of exhaust emissions. This is necessary
because the lead compounds poison the catalysts.
Methylcyclopentadienyl manganese tricarbonyl (MMT), CH- (C,-H,)
Mn (C0)_, is an anti-knock agent which has been used to a limited extent
in gasoline. This anti-knock agent is synergistic with TEL; that is,
when a small amount is added to some leaded fuels the octane rating is
raisedqinore than when the same amount is added to the same fuels without
lead.
-------
11-11
Since MMT does not contain lead it is being used in some of the
unleaded gasolines. It has been reported that a concentration of 0.125
grams of manganese per gallon of gasoline is of greatest interest because
an optimum combination of engine durability characteristics and economic
anti-knock value is obtained at about this concentration. It has also
been reported that the use of MMT in a concentration of 0.125 grams
Mn/gallon would provide, on the average, an increase of about two road
octane numbers. This increase could, compared to an equivalent increase
in road octane numbers obtained through gasoline refining, represent
about a 1% savings in crude oil. Thus, although MMT is not as cost-
competitive as lead anti-knock, it appears that its use can be economically
attractive when compared with achieving anti-knock quality by processing,
and can also offer significant savings in petroleum.
Reference 20 includes the following information regarding the use
of MMT in unleaded gasoline for cars equipped with catalytic converters.
The MMT does not lessen the effectiveness of exhaust catalysts in oxidizing
unburnt hydrocarbons and carbon monoxide. When cars on the road were
operated under extreme service conditions some plugging of monolithic
exhaust catalysts occurred using MMT at the concentration of 0.125 grams
Mn/gallon.*
Toxicologic evaluations of MMT have been undertaken by the National
Environmental Research Center of the EPA. Researchers from that center
have indicated in reference 21 that, "The most important factor that
should be considered in determining the environmental impact of MMT is
whether its usage will cause a significant increase in\the ambient
levels of Mn in highly populated areas with high traffic density.
Further studies are needed to determine the lowest atmospheric Mn
concentrations which will produce clinical Mn intoxication and the
possible chronic effects of long term exposure to automotive exhaust
containing Mn."
II.G. Optimum Octane Rating of Gasoline
It has been explained in the previous sections that: 1) engine
efficiency calls for the highest possible compression ratio, 2) compres-
sion ratios are limited by engine knock, and 3) engine knock is inversely
related to the octane rating of the gasoline. Thus, from the viewpoint
of having the maximum fuel economy for mgjor vehicles, gasoline should
have the highest possible octane rating, and the engines should be
designed with the highest possible compression ratio compatible with
this gasoline.
*Investigations are currently being made by several organizations to
evaluate the impact of MMT, for concentrations up to 0.125 g
Mn/gallon, on emission controls and engine performance.
**The benefit of a gasoline with high octane rating is realized only if
the design of the engine — or its manufacture or condition — calls for
such a high octane rating. Use of gasolines of higher octane rating than
the one required by the engine does not improve its efficiency, it is not
economical, and may have negative effects — such as increased combustion
chamber deposits because of the probability of higher levels of lead or
higher levels of aromatic hydrocarbons in the gasoline.
-------
11-12
However, to optimize the conservation of petroleum it is necessary
to consider not only the fuel economy of the vehicle but also the
efficiency of the refining process of gasoline. This is because
at the octane levels of interest the yield of gasoline per barrel of
crude oil is in general inversely related to the octane number. This
is particularly important/ in the case of unleaded fuels which do not
have additives for raising their octane ratings.
Studies have been made for estimating the optimum octane number
of unleaded gasoline for maximum conservation of petroleum considering
both the efficiency of the refining process of gasoline and the fuel
economy of automobiles. The estimates vary, ranging between 93 RON
and 97 RON (approximately-between-SS-and 87 MON for the prevailing
gasoline sensitivities) ' ' ' ' It has also been estimated
that the saving in crude oil corresponding to the optimum octane rating,
compared with the current minimum rating of 91 RON/83^MON for unleaded
motor gasoline,would be approximately 1 to 3 percent.
It should be indicated that some of these studies for optimum
octane rating have^gonsidered availability of more than one grade of
unleaded gasoline. Such consideration is based on the fact that the
distribution of octane number requirements of motor vehicles is rather
wide. Because of the many factors which affect the octane number
requirements, even identical cars may have large differences in octane
requirements. Typically, the distribution of octane number requirements
of identical cars show (when operated with full boiling range unleaded
fuels) a variation of about 10 research octane units, between the 5
and 95 percentiles. The variability of the octane requirements for
the whole car population of a given model year is even higher. Accordingly,
availability of unleaded gasoline with octane rating higher than 91 RON/
83 MON is needed for optimizing petroleum conservation, but availability
of unleaded gasoline of lower than 91 RON/83 MON would also aid petroleum
saving.
Thus it appears that there is some potential for maximizing the
conservation of petroleum by optimizing the octane ratings of unleaded
gasoline. However, some of the studies on optimum octane number have
pointed out constraints against raising the octane rating of the
unleaded gasoline beyond the current minimum rating of 91 RON/83 MON.
The actual saving would be somewhat smaller since the average octane
rating for unleaded gasoline is higher than the minimum 91 RON/83 MON.
The last survey reported by ERDA indicates a national average of 92,1
RON/83.9 MON in summer 1975, and 92.3 RON/84 MON in winter 1975-76.
Reference 50 has considered up to a three grade system which would
have grades of 80 MON, 83 MON, and 88 MON.
-------
11-13
Specifically, it has been indicated that in order to raise the octane
rating the petroleum industry would have to construct additional pro-
cessing facilities, and that this would take time and require large
capital investments. Supposedly, these constraints justify the current
octane rating. But considering the scarcity of petroleum, the opti-
mization of octane ratings should be pursued. Unless a thorough analysis
demonstrates the impossibility or secondary importance of changing it —
or that the necessary capital investment cannot be made available, or
would be better spent in other projects for increasing the availability
of energy — the current rating of 91 RON/83 MON for unleaded gasoline
does not appear strictly appropriate on a permanent basis. In parti-
cular, a reevaluation of the optimum octane ratings of the gasolines of
the future should consider the nationwide marketing of at least two
different grades of unleaded gasoline, to efficiently match the wide
variation in octane requirements of motor vehicles.
In reevaluating future octane ratings of gasoline, an additional
consideration should be taken into account. As has been indicated in
section II.A, the control of some exhaust pollutants below certain
values may pose a limitation on the compression ratio of the engines.
Since a compression ratio constraint would limit the maximum octane
requirement for gasoline, a reevaluation of the optimum octane ratings
should also research thoroughly whether such a constraint on the com-
pression ratio might occur.
-------
III-l
III. Density
Because the densities of the various hydrocarbon compounds are
different, the density of a gasoline varies according to its composition.
For example, since aromatics are heavier than paraffins and olefins, the
density of gasoline increases with its percentage of aromatics.
The fuel economy of motor vehicles is commonly expressed in miles
per gallon (mpg). However, the mpg does not provide an accurate indication
of the fuel economy of the vehicles unless the density of the gasoline
is taken into consideration. This is because the density of the gasoline
affects the mpg of a vehicle. The explanation is that £he heating value
per unit volume of gasoline increases with its density. Thus, higher
mpg values are achieved using heavier gasolines. However, this is not
due to a variation of the efficiency of the vehicles, but rather to the
change in gasoline density. In other words, there are fuel economy
variations which are due to changes in fuel density, and therefore, should
not be attributed to the vehicle, but only to the gasoline. As explained
before in Section I, the composition of the gasoline does not signifi-
cantly affect the thermal efficiency of the engine. If the fuel economy
of motor vehicles was measured in miles per pound of gasoline instead of
miles per gallon, the heavier gasolines would result in lower mileages
per pound of fuel because the heating value per unit of mass of gasoline
decreases when its density increases.
Given the same vehicle, driving cycle, and atmospheric conditions,
the relationship between the m^les per gallon and the densities of
gasoline can be expressed as:
(III.2)
mpg
where HV stands for the heating value of the gasoline per unit mass,
£ is the density of the gasoline, and the subscript o refers to the
values corresponding to any particular gasoline used for reference.
*
A correlation of the heating value of commercial full-boiling range
gasolines with density indicates that the
Low Heating Value |Btu/gal| * 47,400 + 10,960.p (III.l)
25
where O is the density in Ib/gal.
** The derivation of this relationship is given at the end of this section.
Also, since the HV is practically a function ofO only, for the given
conditions the ratio mpgo/mpg can be considered a function ofO only.
If consideration of secondary effects (such as the effect of variations
is distillation temperatures) would be desired, it would then be
necessary to expand the function HV to include the additional variables.
-------
III-2
Thus, for accurate evaluations or comparisons of fuel economy, the
mpg should be measured using gasoline of a fixed density. Alternatively,
the mpg should be corrected for density, to refer all the results to
those of a fixed density of gasoline. In other words, the reported
values of fuel economy should be:
mpg = mpg x Correction Factor.
The correction factor determined analytically according to equations
(III.l) and (III.2) is shown in Figure III.l, as a function, of £ (taking
for reference a gasoline with p =6.14 Ib/gal.) The SAE correction
factor for density, which has been found to correlate well with large
volumes of actual fuel economy data, has also been plotted in the same
figure. The two correction factor curves are practically the same.
The densities of the gasolines in the market vary significantly.
Data indicates that in 1974 the density of some commercial gasolines
ranged between 5.7 and 6.4 Ib/gallon. In the extreme case where tests
of fuel economy are made using gasolines falling at the lowest and
highest densities of this range, a difference in fuel economy of as
large as 7% would be indicated if there is no correction made for gaso-
line density. Ordinarily, comparisons are not made for the extreme
cases of gasoline density and, consequently, the variations in the fuel
economy of motor vehicles due to variations in gasoline density are much
less than this figure. However, the corresponding differences in fuel
economy^jan still be significant, and, therefore, should not be over-
looked .
A 9 f\
"Fuel Economy Measurement Road Test Procedure - SAE J1082".
**
For the emission tests for certification of automobiles, Indolene HO
III is used as fuel. In 1974 the density of this gasoline ranged
between 6.10 and 6.22 lb/gal.27 xhe maximum variation in fuel economy
which is possible for this change in density is approximately 1%.
-------
III-3
Figure lll.l Correction Factor for Gasoline Density
-------
III-4
Analytical Derivation of the Correction Factor for Gasoline Density
Consider an arbitrary — but fixed — driving cycle, and a given
vehicle and atmospheric conditions .
Let:
L = length of the driving cycle, [miles/cycle | .
W = work produced by the engine during a cycle, |Btu/ cycle j .
F = average force during the cycle, Btu/mile|, such that W = F.L.
HV = heating value of the gasoline per unit mass, |Btu/lb|.
HV1 = heating value of the gasoline consumed per cycle, jBtu/ cycle] .
mpg = fuel economy.
v = volume of gasoline consumed per cycle, I gal/cycle) .
p= density of the gasoline, jib/gal j .
W = engine efficiency for the given conditions.
o as the subscript to indicate values corresponding to any
particular gasoline used for reference.
As indicated in section I, the thermal efficiency of the engine
is practically independent of the composition of the gasoline. Therefore,
it can be written:
Now, since
V = F.L
HV' = v. a . HV, and
HVj = v . f> . HV ,
0 o Co o
substituting in equation (III. 3) :
F.L F. L
e
v. a .HV
L = L
v.p-HV v . p . HV
t o To o
(III. 4)
-------
III-5
Furthermore,
L/v = mpg, and L/v = mpg ;
therefore, substituting in equation (ill.4):
mpg _ mpgn ,
or
mpg £j .HV (III.5)
which is the equation (III.2) indicated at the beginning of this Section III.
Accordingly, the reported values of fuel economy should be
mpg = mpg x Correction Factor
P HV
where the Correction Factor = V" ..° . (III.6)
p.HV
If the correlation of the heating value of commercial full-boiling
range gasolines with density
Low Heating Value |Btu/gal| = 47,400 + 10,960. ^>
(from reference 25) is used, we can write
HV|Btu/lb| = 10,960 + 47,400/0. (III.7)
Therefore, substituting equation (III.7) into equation (III.6):
Correction Factor , 10,960.fi, + 47,400 . (III.8)
~ 10,960.P + 47,400
The SAE, in its "Fuel Economy Measurement Road Test Procedure - SAE
J1082"26> provides the following correction factor (using a reference
gasoline with specific gravity of 0.737, i.e., Po = 6.14 Ib/gal):
Correction Factor = 1 + 0.8 (0.737 - Specific Gravity). (III.9)
The correction factors for both equations, (III. 8) and (III.9), are
plotted in Figure III.1.It is seen that both methods provide practically
the same values.
-------
IV-1
IV. Volatility
The volatility of a gasoline, that is, its readiness to become
vaporized, affects the performance of motor vehicles in different ways.
Specifically, volatility has an impact on the various aspects of the
driveability of the vehicle — starting, warm-up, acceleration, power,
vapor lock, and carburetor icing. Volatility also affects the fuel
economy, and the dilution cf the lubricant oil of the engine. These
effects of volatility are discussed in the corresponding subsections
that follow.
IV.A. Common Ways of Identifying the Volatility of Gasoline
The hydrocarbons that compose a gasoline have widely divergent
boiling points. Accordingly, gasoline evaporates over a range of tempera-
tures depending upon the types and amounts of hydrocarbons used in
blending the gasoline. Its boiling range may extend from about 80 F to
about 430°F.
The volatility of gasoline is normally assessed by a standard
distillation test (ASTM D86 ) which indicates the percentage of gasoline
evaporated along its temperature boiling range. The temperatures at
which 10, 50, and 90 percent have evaporated are commonly used to
characterize the volatility of automotive gasolines. A typical gasoline
distillation curve is shown in Figure IV.A.I. In general, each of the
vehicle performance parameters affected by gasoline volatility depends
particularly upon certain portions of the distillation curve. There-
fore, to facilitate discussions related to volatility, the distillation
curve is usually divided into three major regions. These regions are
referred to as the Front-End, Mid-Range, and Tail-End, and they cover
the lowest, middle, and highest boiling temperatures, respectively.
Determining the vapor pressure is another means for assessing the
volatility of a gasoline, particularly its tendency towards vapor locking.
The vapor pressure of gasoline is normally measured by the Reid Method
(ASTM D323 ), at a temperature of 100°F, in a closed apparatus at the
standard but arbitrary condition of 4/1 ratio of vapor to liquid.
However, due to the variable composition of gasoline the Reid vapor
pressure is not a conclusive criterion of the vapor locking tendency.
A more reliable indication of the vapor lock propensity is a set of
values showing the relationship between V/L (the ratio of the equilibrium
volumes of the vapor fuel and the liquid fuel) and temperature, at a --
given pressure (ASTM D2533 ). Other expressions for assessing vapor
lock performance of gasolines can be found in reference 55.
-------
IV-2
so
% EVAPORATED
Figure iv.A.l. Typical Distillation Curve for Gasoline
-------
IV-3
IV.B. Volatility versus. Cold Starting, Hot Starting, and Vapor-Lock
Under cold engine conditions — particularly at low ambient tempera-
tures — normal carburetor operation would evaporate only a small percentage
of the gasoline supplied through the carburetor, and this would result
in a fuel-air mixture too lean for ignition. It is for this reason that
carburetors are equipped with chokes which increase the amount of fuel
going into the engine and result in ignitable fuel-air mixtures for cold
starting. Thus, from a fuel standpoint the problem of cold starting is
mainly one of getting sufficient evaporation of fuel. Accordingly, a low
front-end distillation curve is important for cold starting.
Hot starting presents an opposite requirement. If the front-end of
the distillation curve is made excessively low to provide good cold
starting, then the fuel evaporates out of the carburetor bowl when a hot
engine is turned off. If there are no evaporative control devices this
evaporated fuel goes into the intake manifold forming a mixture which
is too rich to burn and results in hard starting. If the vehicles have
evaporative control devices, excessive volatility of the fuel may pose
problems to these devices with regard to capacity for storing the vapor
and control of its disposal.
Also, gasolines with excessive front-end volatility enhance vapor-
lock. This is a condition in which partial or complete interruption
occurs in the liquid fuel flow because of vaporization of the fuel.
Vapor-lock problems can be reduced by proper attention to the design and
location of fuel-system components — keeping fuel lines away from
hot engine parts, avoiding sharp turns and restrictions in fuel lines,
etc., and maintaining the fuel under pressure. But to prevent these
problems, as well as hot starting problems, the front end volatility should be
maintained as low as permitted by opposite requirements. Cold starting,
conservation of crude oil, and gasoline manufacturing favor a high
front-end volatility.
IV.C. Carburetor Icing
Under certain ambient and operating conditions, ice can be formed
in the carburetor. The ice upsets carburetion and results in poor
vehicle performance. Throttle plate icing originates from water in
the air. Ice in the carburetor and in the fuel lines may also
be formed from water in the gasoline.
The formation of the ice in the carburetor is dependent upon
the cooling effect of fuel evaporation. A fuel which evaporates
more rapidly exerts more cooling effect, and therefore is more
prone to result in carburetor icing. Lowering the volatility of the
gasoline decreases the evaporation rate and aids in preventing carburetor
icing. However, carburetor icing can also be controlled by adding
to the gasoline surface-active agents and freezing point depressants.
Also, the recent practice — due to emission control needs — to heat
the inlet air has greatly diminished carburetor icing tendencies.
-------
IV-4
IV.D. Impact of Volatility on Warm-up and Normal Operation of the Engine
The warm-up period of an engine is related very significantly to
the mid-range portion of the distillation curve. This portion includes
a considerable amount of the total quantity of fuel; therefore, it must
be volatile enough to provide appropriate fuel-air ratios under a variety
of operating conditions if operational flexibility is to be achieved.
For long trips the impact of the warm-up period on the overall fuel
economy should be small or negligible because for these trips the warm-
up period represents only a small fraction of the total driving. However,
the impact of the warm-up period on fuel economy may be very significant
in short trips. Thus, keeping the warm-up period to a minimum by
having a gasoline which is relatively volatile in the mid-range section
permits better driveability and fuel economy,
Furthermore, to obtain optimum acceleration, power, smoothness of
operation, and engine efficiency, it is important that the appropriate
fuel-air ratio be delivered to all the cylinders, both during warm-up
and during normal operation of the engine. Consequently, since high
volatility aids good mixture distribution, keeping the mid-range.Aand
tail-end portions of the distillation curve as low as practical will
favor good performance and vehicle fuel economy.
Because hydrocarbons with higher boiling points are generally
heavier than the more volatile ones, and because hydrocarbons of higher
density have higher energy content per gallon, it can be stated that
gasolines with the lowest possible volatility in the mid-range and tail-
end of their distillation ranges increase the mileage. However, as was
explained in section III, this is not due to an increase in the efficiency
of the vehicle. This impact (of higher density due to lower volatility)
on the miles per gallon does not appear when the fuel economy is corrected
for the density of gasoline.
IV.E. Effects relative to the Tail-End of the Distillation Curve
High boiling temperatures at the tail-end portion of the distillation
curve have some undesirable characteristics other than being a possible
cause for poor mixture distribution among the cylinders. Fuel that does
not vaporize may pass the piston rings into the crankcase of the engine
where it dilutes the oil and decreases its viscosity. Crankcase dilution
is intensified under the low engine temperatures encountered in cold
weather, particularly for short trip driving patterns. Thus, the degree
of crankcase-oil dilution is inversely related to the tail-end volatility.
*
Also, motor vehicles are designed to achieve optimum performance when
their powertrain systems are at normal operating temperatures. Consequently,
before reaching this normal operating condition, the fuel economies of
motor vehicles are below optimum.
**In particular, certain refineries have problems in manufacturing
gasolines with high mid and tail-end volatility.
f
-------
IV-5
Also, lowering the tail-end of the distillation curve of gasoline
eliminates some of the hydrocarbons with high boiling points which
contribute to combustion chamber deposits and to varnish and sludge
deposition inside the engine. This deposition can cause piston ring
plugging and sticking, and valve sticking, resulting in poor operation
and poor fuel economy. Likewise, less spark plug fouling is usually
experienced with gasolines with high tail-end volatility.
Conversely, there are reasons that preclude raising the tail-end
volatility. For one thing, economic considerations require including
relatively heavy fractions in the gasoline. In addition, removal of the
heavy fractions would also eliminate high anti-knock components which
the gasoline gains by catalytic reforming.
IV.F. Optimum Volatility, and Impact of Volatility on Fuel Economy
The foregoing discussion indicates that there are conflicting require-
ments for gasoline volatility, with some performance features requiring
a more volatile fuel while others demand less volatility. Consequently,
the volatility characteristics of the finished fuels must be a compromise
for best overall performance and fuel economy. Optimum satisfaction can
be approached by distinguishing seasons and geographical areas and
distributing gasolines with different volatilities. In this way the
volatility of the gasoline can be consistent wi£h the altitude and
prevailing temperatures in each of these areas.
Due to the multiplicity of factors which affect the influence of
the volatility of gasoline, it is very difficult to isolate the impact
of volatility on fuel economy. In previous paragraphs it has been
indicated that the volatility of the fuel affects the starting, the
warm-up and the normal operation of the engine. But the impact of
volatility on the fuel economy is coupled to the effect of other variables.
Thus, during starting and warm-up operation, the impact of volatility is
affected by:
a. A tmospheric.^ conditions (particularly temperature)
b. Choke type
c. Carburetor design and adjustment ^
d. Inlet manifold design and heating
e. Engine cooling
f. Driving pattern.
*
Gasoline volatility classes and a schedule of seasonal and geographic
volatility classes are included in the "Standard Specifications for
Automotive Gasoline," ASTM D439.
**
The rate at which an automatic choke begins to open is one of the
most critical factors affecting warm-up performance. Fast-opening
chokes improve fuel economy — and CO and HC emissions — but
deteriorate vehicle driveability. However, an appropriate combination
of manifold heating (which improves fuel distribution and warm-up
operation) with a fast-opening automatic choke should provide the
best fuel economy.
-------
IV-6
Similarly, when the powertrain system of the vehicle has reached
its normal operating temperatures, the impact of volatility on fuel
economy also depends on engine design and on the driving pattern.
The difficulty of isolating the specific impact of gasoline volatility
on fuel economy, and also the general belief that this impact is minor
compared with the impact of other variables, have resulted in a very
limited investigation of this point. Consequently, the availability
of data regarding the impact of volatility on fuel economy is very
scarce.
An indication of the change in fuel economy caused by a change in
gasoline volatility can be estimated from the results given in reference
30. These results were obtained for one 1973 and three 1974 model year
automobiles, having engines of 455, 258, 360, and 400 cubic inches of
displacement, respectively. The fuel economy of these vehicles was
measured for eight different gasolines — using the cold start 1972
Federal Test Procedure — and the results were correlated by a regres-
sion analysis. From the regression equation, it is estimated tha.£ for
the given test procedure and model year cars, an increase of 20 F in
the mid-range of the distillation curve — maintaining the same fuel
density — would result in about 1% deterioriation in the fuel economy
(but the statistical significance of this is unknown).
Another item associated with the volatility of gasoline is the one
referring to "fuel evaporative emissions". These evaporative emissions
from the fuel system of a motor vehicle depend on the front-end and mid-
range volatility of the gasoline. Distinction is made between the
"running", "diurnal breathing" and "hot soak" portions of these fuel
evaporations. They are, respectively, the fuel evaporative emissions
that occur during the operation of the vehicle, those resulting from the
daily range in temperature to which the fuel system is exposed, and
those that occur during the hot soak periods which begin immediately
after the engine is turned off.
In motor vehicles equipped with evaporative emission controls (and
in older cars with internally vented carburetors), the running evaporative
emissions are fed into the intake of the engine and are not a loss.
Also, the evaporative emission controls collect nearly all of the diurnal
breathing and hot soak evaporations, which later are drawn in and burned
by the engine when it is operated. Thus, the diurnal breathing and hot
* As shown in IV.G.
** For comparison, in 1974 the average difference between the distillation
temperatures at the 50% evaporative point of the summer and winter
gasolines was about 10 F.
-------
IV-7
soak evaporations are kept from polluting the atmosphere; however, these
evaporations may still result in some loss, as far as fuel saving is
concerned. This loss may occur because of the difficulty in optimizing
the fuel-air ratio of the intake, mixture under the transient purging of
the evaporative control system.
There is no available data about the magnitude of this possible loss.
A very rough estimate of an upper limit for this eventual loss can be
determined on the assumption that: a) all the evaporative emissions
are burnt, but without changing the .-work released by the engine,
b) the average evaporative emissions are 100 g/day per car , c) the
average car travels 10,000 miles/year, and d) the average car has
a fuel economy of 16 mpg. For these assumptions, it is found that the
evaporative losses could be up to 2% of the gasoline supplied to the
vehicle. However, if the burning of the evaporative emissions results
in a relatively significant addition of released work, the actual
fuel loss should be well below 2% for the average automobile equipped
with an adequate evaporative control system. Of course, any burning
of evaporative emissions which yields additional released work
results in a saving of fuel compared to the case of cars without
evaporative emission controls.
-------
IV-8
IV.G.An Estimate of the Impact of Gasoline Volatility on
Fuel Economy
Reference 30 indicates a correlation of fuel economy with gasoline
volatility and density*. The data was taken for one 1973 and three 1974
model year automobiles (having engines of 455, 258, 360, and 400 cubic
inches of displacement, respectively), using the cold start 1972 Federal
Test Procedure, and with eight different gasolines (having the properties
shown in the attached table). The correlation yielded the following
equation (with a multiple correlation coefficient of 0.952):
mpg = 0.142(1 2 _ 0,007£ (0.093 T5Q + 0.029T9Q) + 0.811 (IV.G.I)
where £ is the density of the gasoline in Ib/gal, and T^Q and TgQ are the
temperatures, in °F, at the 50% and 90% distillation points, respectively.
The regression equation(IV.G.I) can be used to estimate the impact
of changes in volatility on fuel economy. First, to simplify (noticing
that the data shows, on the average, that TQQ z 1.5 TSQ), we eliminate
Tgo from the equation, leaving only one distillation variable. That is,
we write equation(IV.G.I) as
mpg a 0.142p2 - 0.007p (0.093 T50 + 0.029 x 1.5 T50) + 0.811
or, mpg z 0.142^2 - 0.00095pT50 + 0.811. (IV.G.2).
Then, differentiating with respect to T-^:
z - 0.00095 P. (IV.G.3)
50 V
Accordingly, for constants, Ampg « - 0.00095P. AT™.
Therefore, the percentage variation in fuel economy, due to a variation
in distillation temperature,is
Ampg x 100/mpg 3 -0.0950. AT5Q/mpg. (IV.G.4)
For the given data, the average^and fuel economy were 6.38 Ib/gal and
12.51 mpg. Therefore, for the given test procedure and model year cars,
an increase of 20°F (for example) in the mid-range of the distillation
curve — maintaining the same fuel density — would result in
0.095 x 6.38 x 20/12.51 = 1% deterioration in the fuel economy
(but the statistical significance of this is unknown),
*The correlation does not consider the effect of gasoline viscosity. How-
ever, as shown in Section V.C, gasoline viscosity may have a significant
impact on fuel economy.
-------
TABLE IV.G.I
FUEL VOLATILITY CHARACTERISTICS
Distillation, D86°C (°F)
Over
End
7. Evap. At:
10
20
30
40
50
60
70
80
90
Driveability Index
(0.093 TSO + 0.029 T90)
Fuel Density,
kg/dn>3 (Ib/gal)
Fuel No.
1
43 (109)
180 (356)
57 (135)
64 (148)
72 (161)
78 (173)
87 (189)
98 (208)
111 (231)
127 (261)
145 (293)
2
29 (84)
219 (427)
i
45 (113)
56 (132)
67 (152)
78 (172) :
92 (198)
109 (228)
135 (275)
163 (325)
185 (365)
3
39 (102)
208 (407)
63 (145)
71 (160)
77 (170)
... .81 (178)
86 (186)
90 (194)
99 (211)
122 (252)
180 (356)
4
24 (76)
165 (329)
37 (98)
49 C120)
62 (143)
82 (180)
114 (237)
131 (267)
136 (277)
139 (283)
144 (292)
5
37 (99)
180 (356)
63 (146)
80 (176)
94 (202)
106 (223)
114 (237)
122 (251)
128 (262)
134 (273)
143 (289)
6
35 (95)
207 (404)
61 (142)
76 (169)
89 (193)
103 (217)
114 (237)
126 (259)
139 (282)
156 (312)
180-(356)
7
28 (82)
213 (416)
38 (101)
48 (118)
59 (138)
81 (178)
114 (237)
132 (269)
149 (300)
164 (327)
186 (366)
8
35 (95)
209 (409)
62 (143)
79 (174)
93 (200)
107 (225)
120 (248)
133 (271)
148 (298)
168 (334)
189 (372)
<
26.1
29.0
27.6
30.5
30.4
32.4
32.7
33.9
0.73 (6.09) 0.73 (6.06) 0.80(6.64) 0.77(6.40) 0.79 (6.58) 0.77(6.39) 0.75 (6.25) 0.79 (6.62)
-------
V-l
V. Relationship Between Fuel Characteristics, Fuel Metering, and
Fuel Economy
Automotive engines are designed taking into consideration the
characteristics of the gasoline that will be used. Ordinarily, the
engine parameters related to the fuel are optimized during the develop-
ment of the engine for the average characteristics of the gasoline
that will be consumed. Consequently, the engines will have during their
commercial use performances and efficiencies which will be affected
somewhat by the variations of the characteristics of the gasolines used.
As explained below, one of the most important parameters affecting
fuel economy is the strength of the fuel-air mixture supplied to the
cylinders of the engine. Therefore, it is pertinent to investigate what
fuel metering changes may occur in the.carburetor as a consequence of
the changes in the characteristics of gasoline. Accordingly, this
section is devoted specifically to these matters.
V.A. . Impact of Fuel-Air Mixture Strength on Engine Efficiency
Theoretical and experimental work indicate that the efficiency of
the gasoline engine is highly dependent on the strength of the mixture
supplied to the cylinders. The efficiency of the engine increases with
leaner air-fuel ratios up to the point where a too lean mixture would
result in misfiring or in a combustion process which is too slow. A
typical plot of fuel consumption versus mixture strength is shown in
Figure V.A.I.
The leaner mixtures provide higher engine efficiencies because the
excess of air decreases the temperatures of combustion. The lower
temperatures have such an effect because:
a. The ratio of the specific heats of the gases, k = c /c ,
increases for decreasing temperatures. This effect is shown in
the constant volume air cycle efficiency,
k-l*
r A .
b. Dissociation of the gases decreases as the temperature
decreases.
c. Heat losses decrease as the temperatures decrease.
Also, operation with leaner mixtures reduces carburetor throttling
resulting in reduced pumping losses and, therefore, in increased engine
efficiency.
*
This is the same equation shown before in section II.A.
-------
V-2
.1
-t
I
i
.8
1
0.8
0.7
O.S
0.3
0.6 0.8
t.O '.2 t.V t>G
- £y6vV
-------
V-3
Gasolines have chemical and physical characteristics which affect
the carburetion of the mixture supplied to the cylinders of an engine.
For fuel metering requirements, the chemical properties of interest are
those which determine the stoichiometric ratio of fuel to air. The
physical properties which can directly influence the metering process
are density, viscosity, volatility, and surface tension.
The stoichiometric ratio of a fuel-air mixture is determined by the
H/C atom ratio of the fuel, i.e., by the hydrocarbon composition of the
fuel. Of importance are the relative amounts of the paraffin, olefin,
and aromatic series, as well as the percentages of each hydrocarbon
compound in a given series. Thus, different gasolines may have different
stoichiometric fuel-air ratios. Consequently, since there is no one
unique stoichiometric fuel-air ratio for gasolines, and since engine
efficiency — and exhaust emissions —depend on the strength of the
fuel-|ir charge, the significant factor is the fuel-air equivalence
ratio , <(>, rather than the fuel-air ratio itself.
V,B. Impact of Changes in Fuel Density on Fuel Metering
Theoretical analyses indicate that the fuel-air ratio of the mixtures
supplied by the conventional carburetors increases with fuel density
This is confirmed by experimental results. However, it is important to
note that in general the stoichiometric fuel-air ratio of gasoline also
increases with fuel density. Thus, the net effect on the fuel-air
equivalence ratio changes caused by variations in fuel density is usually
negligible or small. Accordingly, the effect of these density variations
on fuel economy — due^g their impact on fuel metering — should also
be negligible or small
Some data regarding the effect of fuel density on the fuel metering
of carburetors is presented in the following Figures V.B.I and V.B.2.
Figure V.B.I (from reference 33) shows the results of flow-stand
tests for two different carburetors. It is seen that for the given
Carter carburetor, the fuel-air ratio increased approximately from 0.069
to 0.076 (i.e., about 10%) when the specific gravity of the fuel changed
from 0.69 (80 ON primary reference fuel) to 0.86 (toluene). However,
the corresponding fuel-air equivalence ratios were 1.04 and 1.03 respec-
tively. Similarly, for the same change in fuel density with the Holley
carburetor the fuel-air ratio changed approximately from 0.074 to 0.082
(i.e., about 10%), and the fuel-air equivalence ratio changed from 1.13
to 1.10. These results indicate, in fact, that the stoichiometric
change associated with the increase in density overcompensated, slightly,
the associated increase in fuel-air ratio.
The definition of <(> was presented earlier in section L
**
See reference 32, p. 423-428 for example.
***
As indicated before in section HI, the changes in gasoline density
may have a significant impact on fuel economy, but this is not due
to fuel metering effects.
-------
V-4
o
a
1
UJ
e
Fuel
Spec! fie
Gravity, °API
80 O.N. Prirary Reference Fuel
IriEor.i: tive Straishtrun
Michir'an Reformed
Dynafuel
Catc Cracked
Balanced Fuel.
50^ Toluene *
Toluene
'Carburetors
80 O.N. PR?
7.1.9
66.3
63.6
61.3
57.3
CC^I,
5U.3
U9.7
31.6
Viscosity,
Centistokes
0.690
0.7U9'
0.739
0.609
0.601
0.616
0.6L3
0.760
0.633
0.679
.085
.030
.075
.070
1. 191*8 Carter 66hS
Dual throat, downdraft, 1§" throat, metering rod jet. Accelerat-
ing pucip jet plugged duririg test.
2. 19U9 Holley AA-1
Dual throat, downdraft, 1-1/16" throat, fixed .jet.
80 O.N. PRF
50% TOLUENE
80 O.N.
TOLUENE
RANGE OF
^COMMERCIAL
FUELS
4=1.13
OO
t^-
70.5
HOLLEY AA-1 CARBURETOR
450 LB'AIR/HR
FULL THROTTLE
57 °AP| GRAtf 1*5.5
CARTER 664-S CARBURETOR
300 LB AIR/Hfl
FULL THROTTLE
35
.65
.70 .75 .00
FUEL SPECIFI C GRAVITY
.85
.90
Figure v.B.l.
Impact of Fuel Density on Fuel Metering.
Flow Stand Tests
-------
V-5
Viscosity
• -.'. Fuel .. D445, cs, 77° F
1 .65
2 .68
3 .65
O 98
w« / w
.0
,N
v /. OO
Qc
«
u
fc /,02
>S!
to
1
Q^ /-**
UT
*
^
^ /.^ffH
Q - 45 MPH •
L *
0 "~~"~\^^ ©
t^^Nfc,^^
~~^~^\^\
^^^^^^ A
A
^' ~~~~ -^~-_ ' ' "
" ^^~--^'X
X
• • • *,
H
(Vj ^^^~-~^^^
GJ
SPKIFIC ffRAVIT/
.739 |.72 |.703
50 55 60 65 70 75
"API GRAVIT/ =. _ • '
Figure V.B.2. Impact
of Fuel Density on Fuel Meterine.
Chassis Dynamometer Test
-------
V-6
Figure V .B.2 (from reference 30) shows the results for a carburetor
installed in a car. The tests were conducted on a chassis dynamometer
at constant ambient temperature. Three different gasolines having
substantially different densities but similar viscosities and volatilities
were used. The changes in fuel density changed the fuel-air ratios
significantly but, again, on an equivalence ratio basis the corresponding
changes in stoichiometry served to compensate for the changes in fuel-
air ratio. Varying the specific gravity of the gasoline over the range
from 0.70 to 0.76 resulted in a relatively small effect on the fuel-air
equivalence ratios.
V.C. Impact of Variations in Fuel Viscosity
Variations in the viscosity of gasoline affect the metering of the
fuel by changing the Reynolds number of the fuel flow. This change in
R generally results in variations in the coefficients of discharge of
the fuel orifices. However, the fuel-air ratio changes resulting from
variations in fuel viscosity depend on the particular channel geometry
and operating points of the carburetor. Accordingly, different carburetors
can be affected differently by the changes in fuel viscosity.
The data regarding the effects of viscosity change on fuel metering
is scarce and inconclusive. During the carburetor flow-stand tests
quoted above in section V.B., from reference 33, the kinematic viscosity
of the fuel ranged from 0.60 to 0.76 centistokes, but no apparent effect
on carburetion was found. On the other hand, the results from reference
30 indicate a definite effect of the variations in viscosity although
the impact changed with the operating mode of the carburetor. The fuel-
air equivalence ratios decreased with increasing viscosity, which is
consistent with the reasoning that increased viscosity should decrease
fuel flow.
These results from reference 30 were obtained by isolating the
effects of viscosity (from the effects of density and volatility) by
proper fuel design and careful testing — whjch included dynamometer
operation at a constant temperature of 75 F. The results are plotted
in Figure V.C.I . The maximum variation in the fuel-air equivalence
ratios^|or these tests — which covered a range from 0.56 to 0.69 centi-
stokes — was about 3.5%. Thus, these particular results indicate
that viscosity variations may cause significant changes in fuel economy
However, these results cannot be considered typical since only one car .
with only one type of carburetor was used for the tests. Furthermore,
all carburetors do not respond to fuel property changes in the same
manner, even at the same operating point.
Gasoline viscosity undergoes large changes with temperature variance;
therefore, the impact of viscosity variations must be determined under
controlled temperature.
**
For comparison, some measurements of the viscosity of commercial
gasolines have shown differences 'of 0.17 cs at 0 F, and 0.06 cs at
100°F.
In a first approximation, for rich fuel-air mixtures, the percentage
variation in fuel economy can be taken as roughly equal — although
opposite in direction — to the percentage variation in fuel-air
equivalence ratio.
-------
V-7
Fuel
1
2
3
4
Viscosity API
D445. cs, .77" F 'Gravity
.56 61.3
.60 61.3
.64 61.0
.69 60.9
Pistillation, D86.-°F
10%
129
130
129
125
507,
221
221
221
219
90?.
323
319
318
319
Carbon ~
Percent
85.96
85,96
86.06
86.02
0.16
\°(72 R3RD
,400 CIJ> ENGUWE
.60
VISCOSITY, CS AT 77°F
Figure v.c.l.
—JTJ=L--_°f. Fuel Viscosity-on Fuel Metering.
Chassis Dynamometer Tests
-------
V-8
V.D. Impact of Variations in Fuel Volatility and Surface Tension
on Fuel Metering
The volatility of the gasoline may have a noticeable effect on fuel
metering. The higher the volatility, the higher the amount of fuel
vaporized in the fuel passages of the carburetor. This tends to de-
crease the fuel-air ratio delivered by the carburetor. On the other
hand, an increase in fuel volatility increases float bowl evaporation,
and this evaporation vented into the intake system tends to increase the
fuel-air ratio. Therefore, the overall effect of changes in fuel volatility
on fuel metering depends on the carburetor design, carburetor temperature,
and carburetor operating point.
As was indicated in section IV.F above, there is very little data
referring to the impact of gasoline volatility on fuel economy. In
particular, there is no data available regarding the impact on fuel
economy that is due exclusively to the changes in fuel metering caused
by variations in fuel volatility.
The surface tension of the gasoline affects fuel metering because
of its effects on orifice flow. The venturi suction required at the
discharge nozzles depends on the surface tension of the fuel. A primary
metering effect of surface tension is its influence on determining at
what air flow the main system fuel-flow is initiated. Likewise, the
minimum pressure differential necessary to initiate flow in the air
bleeding channels is also dependent on the surface tension of the fuel.
However, there is no available information regarding the impact that the
changes in surface tension may have on fuel metering and fuel economy.
There is a general belief that the impact of fuel surface tension on
fuel economy is minimal, and, apparently, no specific efforts have been
devoted to measure this impact.
-------
VI-1
VI. Gasoline Cleanness and Additives
Cleanness as applied to gasoline refers to foreign material which
gets into the fuel as well as to the original impurities in the fuel,
and to the fuel's tendencies to promote deposit formation in engines.
Cleanness is s.n important property of gasoline because foreign material
and fuel deposits alter engine parameters and adjustments and result in
poor engine performance, lower fuel economy, and higher emission of air
pollutants.
The following paragraphs point out important factors in gasoline
cleanness and indicate those additives which are usually added to gaso-
line for preserving its cleanness or to inhibit undesirable performances.
Foreign particulate matter. Substances such as rust, sand, and
fibers must be excluded from gasoline insofar as possible. These materials
interfere with proper operation of fuel pumps and carburetors and will
plug gasoline filters.
Liquid impurities. Contamination with liquids such as diesel fuels
and water must also be avoided. Diesel fuel lowers the anti-knock
quality of gasoline, and it also increases crankcase dilution due to its
high boiling point. Water promotes deterioration of the fuel system and
can cause filter clogging.
Anti-rust additives. To prevent the rusting that might be caused
by traces of water in the gasoline—and also the corrosion arising from
contact with air—anti-rust agents are added to the gasoline (such as
fatty-acid amines, in amounts ranging from 1 to 15 lb/1,000 bbl).
Sulfur. Sulfur or sulfur compounds in the gasoline are objection-
able for various reasons. In some forms, mainly free sulfur and hydrogen
sulfide, it is corrosive and can attack fuel lines, carburetors, and
injection pumps. In s.II forms, the sulfur will unite with oxygen to
form sulfur dioxide, which in the presence of water may form sulfurous
acid. Also, the anti-knock effect of the lead gl^vl compounds added to
gasolines is reduced by the presence of sulfur. ' Presently, sulfur
contents of less than 0.1 percent, by weight, are specified for most
gasolines.
The pulfur emissions emitted by motor vehicles also have an impact
on air quality. In particular, concern has been voiced regarding the
increase in sulfur emissions (mainly sulfur trioxide, sulfuric acid, and
sulfates) from automobiles equipped with catalyst systems for control-
-------
VI-2
ling carbon monoxide .and unburned hydrocarbons. In the presence of a
catalyst, the sulfur dioxide oxidizes into sulfur trioxide which is a
corrosive gas and which combines vigorously with water to form sulfuric
acid. Work is currently underway to improve the techniques for measuring
sulfur emissions and to determine the actual impact of these emissions
from catalyst-equipped cars on air quality.
Anti-knock additives. Currently, most commercial gasolines contain
lead alkyl compounds. A general account of anti-knock additives has
been included in Section II.F.
Lead, phosphorus. The introduction of catalyst-equipped automobiles
with the 1975 model year resulted in the promulgation of regulations
which require the availability of a grade of gasoline with not more than
0.05 g of lead per gallon and 0.005 g of phosphorus per gallon. This
fuel is required because catalysts have been shown to be poisoned by .„
both lead and phosphorus. Further, in late 1973, the EPA promulgated
regulations requiring the gradual phase-down of lead levels in the r"e-
maining gasoline grades. These regulations to phase-down lead were
based on the Agency's concern regarding the public health consequences
of lead contained in the exhaust gases from motor vehicles. These
regulations were overturned by the courts, but the U.S. Court of Appeals
has ruled that the EPA has the authority to regulate the use of lead in
gasoline. The U.S. Supreme Court has declined to review the ruling, and
the EPA has resumed enforcement of the regulations. However, to ensure
that refiners have sufficient lead time to modify the manufacturing
process of the gasoline without resulting in any gasoline shortage, the
EPA has amended the schedule of the regulations. The definitive re-
gulations limit the average content of lead in motor gasoline to 0.8
g/gallon, effective January 1, 1978*, and to 0.5 g/gallon, effective
October 1, 1979.
"Gum". Unsaturated hydrocarbons and impurities in the fuel have a
tendency to oxidize and polymerize, resulting in viscous liquids and
solids which are described as "gum." Gum formation appears to be due to
some reaction initiated by the formation of peroxides and catalyzed by
the presence of metals, particularly copper, which may have been picked
up during refining and handling operations. The pure stable hydrocarbons
of the paraffin, naphthene, and aromatic families form little gum, while
cracked gasolines have the highest tendency to form gum. A gasoline
with high gum content will cause operating difficulties, such as sticking
valves and piston rings, gum deposits in the inlet manifold, clogging of
carburetor jets, and lacquering of the valve stems, cylinders, and
pistons.
"Suspension of the 0.8 g/gallon standard is conditioned on a showing by
a refiner that he has prior to that time taken and is continuing to take
sufficient actions in procuring and installing equipment or arranging
process and exchange agreements, or both, to insure compliance with the
0.8 g/gallon standard at the earliest practical date, and 0.5 g/gallon.
no later than October 1, 1979".59
-------
VI-3
Freshly manufactured gasoline normally has an insignificant gum
content but upon aging varying amounts of gum may be formed. In storage
gasoline will increase its gum with increased concentrations of oxygen,
with a rise in temperature, with exposure to sunlight, and also on
contact with metals. Thus, distinction must be made between the actual
or preformed gum existing in the gasoline at a particular time and the
potential gum that may be present at some future time. The actual gum
content is no guarantee of the stability of the gasoline against future
gum formation.
Anti-oxidant and Metal Deactivator additives. To ensure storage
stability, anti-oxidant additives known as "inhibitors" are commonly
added to the gasoline (such as phenols or amine compounds, in amounts
ranging from 1 to 15 lb/1,000 bbl.). Such inhibitors react with the ,_
chain carrying free radicals that propagate the auto-oxidation of gasoline.
Other types of additives used to ensure stability are "metal deactivators"
(amine derivatives, in amounts of up to 1 lb/1,000 bbl) which destroy
the catalytic activity of traces of copper.
Detergents. Deposits in the induction system of the engine may
come from air-borne contaminants, preformed gum in the gasoline, and gum
formed in the induction system. Incomplete combustion products and
crankcase vapors have been found to be major contributors to these
deposits. Positive crankcase ventilation and exhaust gas recirculation
increase their formation. Some deposits form in throttle bodies when
the incoming flow strikes the throttle and impacts against the throttle
walls. These deposits restrict the flow of air past the throttle plate
in the idle position and through the idle air bleed and vacuum advance
ports, causing rough idling and stalling. The fuel-air ratio of the
mixture supplied to the cylinders is altered by the carburetor deposits,
and this alteration may significantly affect fuel economy and emissions.
Many gasolines now contain detergent additives to prevent the formation
of these deposits in the induction system and to remove existing deposits.
Alkyl amines and their derivatives are used in amounts ranging from 5 to
40 lb/1,000 bbl. The so called "deposit control" type of detergents
(polybutene amines) are used in amounts of up to 100 lb/1,000 bbl.
"Deposit Modifier" additives. As compression ratios increase
beyond about 10, engine performance tends to become limited by surface
ignition. The higher compression ratios raise the peak temperature in
the combustion chamber and cause combustion chamber deposits to glow
leading to premature ignition. The presence of lead compounds in the
carbonaceous deposits caused by incomplete scavenging of lead alkyl
anti-knocks increases the tendency toward surface ignition. The "deposit
-------
VI-4
modifiers" were introduced to reduce surface ignition. These additives —
generally phosphate esters — act by modifying the deposits to decrease
their glowing temperature and oxidation rate so that their heat release
rate is smaller. Also, lead phosphates are less electrically conductive
than other lead salts. Hence, by increasing the electric resistance of
the deposits the use of additives containing phosphorus decreases spark
plug fouling and, thereby, misfiring. The use of "deposit modifiers,"
however, has been declining and has now been essentially discontinued in
light of the reduction of compression ratios in recent model year vehicles.
Unleaded fuels, of course, do not contain these additives.
Gasoline composition, additives, and the OKI. As has been indicated
in section II.D, the octane requirement increase (OKI) is one of the
important determinants of the octane requirements of motor vehicles. The
deposits which cause the ORI depend, among other things, on the physical
and chemical natures of the gasoline and the lubricant of the engine,
and on the additives in the gasoline and in the lubricant. However, the
mechanisms which produce the ORI are not well understood. Also, the
available information regarding the magnitude of the ORI is not con-
clusive. Therefore, more data is needed. This must include the effects
on ORI of the composition of the gasolines, the lubricants, and the
additives in both fuels and lubricants.
Anti-icing additives. "Surface-active" additives (such as ammonia
salts) are added to prevent ice from adhering to the throttle of the
carburetor. The surface-active additives work by forming a film on the
metal which discourages adhesion of ice. "Freezing point depressants"
(alcohols or other compounds with a large affinity for water) are used
to lower the freezing point of any water present in the gasoline to such
a degree that no ice formation can occur.
Color dyes. These are added to identify different grades of
gasoline. They do not, however, affect gasoline quality.
-------
VII-1
VII. Gasoline Injection versus Carburetion
The information and discussion that have been presented in this
paper apply specifically to gasoline and automotive engines in the case
where the engines are equipped with carburetors — the most common, by
far. However, the number of motor vehicles using gasoline-injection
systems is increasing.
This section indicates the principal types of gasoline-injection
systems. It also discusses, summarily, the most significant changes in
engine performance resulting from the use of gasoline injection instead
of carburetion, and the possible impact on the characteristics of gasoline.
VILA. Classification of Gasoline Injection Systems
The gasoline-injection systems for motor vehicles can be classified
in three basic categories: ' '
a. Direct Cylinder Injection
b. Port Injection
c. Injection Carburetor.
As its name indicates, in the direct cylinder-injection systems the
gasoline is injected directly into the cylinders during the inlet or
compression strokes in a similar fashion as in the Diesel engine.
Because of its higher complexity and cost this injection system is very
uncommon in motor vehicles. The Mercedes-Benz, Model 300 SL, has been
one of the few automobiles equipped with such a system. In stratified
charge engines, however, direct injection is common; examples are the
Ford PROCO and Texaco TCCS .
In the port-injection system, the gasoline is injected by nozzles
located at the inlet ports and aimed toward the inlet valves. The
injection can be continuous (as in the Bosch continuous injection
system ) or timed (as in the Bosch electronic system ).
The basic difference between the injection carburetor (also called
pressure carburetor) and the conventional carburetor is that the pres-
sure difference across the venturi, while providing the metering function,
is not used directly to cause fuel flow. Instead, the fuel is supplied
under pressure by means of a power-driven pump to one or several nozzles
located somewhere downstream from the throttle. Injection carburetors
were developed for aircraft engines and for engines which must operate
in all positions, but the basic principle can be applied to motor vehicle
engines, with the possibility of injecting the fuel at each intake port.
-------
VII-2
VII.B. Pros and Cons of Gasoline Injection
In general, a well developed fuel-injection engine would have the
following advantages when compared with the engine equipped with the
conventional carburetor system:
1. Easier starting due to improved fuel distribution, and because
atomization of the gasoline for fuel injection is practically independent
of cranking speed.
2. Decreased volatility requirements for the gasoline, since fuel
distribution is independent of vaporization. Also, engine warm-up may
be faster and allow leaner fuel-air mixtures.
3. Some decrease in octane number requirements since heat is not
necessary to assist fuel distribution. Also, some decrease in octane
number requirements because of the equal distribution of gasoline com-
ponents — and related combustion chamber deposits — among all the
cylinders. (Relaxation in octane number requirements can also be ob-
tained when the charge is stratified).
4. Decreased fuel-air ratio variations arising from changes in
position or motion, since fuel-air ratio is not dependent on float
level.
5. Better vehicle driveability. In particular, faster throttle
response since the gasoline is injected into, or close to, the cylinders
and need not flow through the inlet manifold.
6. Increased torque and power because of increase in volumetric
efficiency arising from
a. Large inlet manifolds with small pressure losses.
b. Elimination of carburetor pressure loss (except that a
venturi pressure loss remains if a venturi if used for
metering purposes).
c. Elimination of manifold heating.
7. In general, less emissions because of reduced fuel-air ratio
variations among the cylinders. Also, there is a potential for less HC
emissions on deceleration, by -cutting off fuel injection, and less HC
and CO on acceleration by eliminating the acceleration pump (since
manifold wetting is eliminated or minimized, and fuel response is faster).
8. Lower height of engine — and hood — since position of the
injection unit is not critical.
-------
VII-3
On the other hand, gasoline injection systems have the following
disadvantages:
1. Increased complexity and cost.
2. Difficulty of metering equally in all the individual cylinders
the small amount of fuel required for idle and road load, due to injector
tolerances and repeatability. Furthermore, deposit build-up on the fuel
nozzles as mileage increases may seriously upset the fuel delivery
balance between cylinders, leading to lower fuel economy, poor driveability,
and higher emissions. This, and the increased complexity of the systems,
results in increased maintenance requirements.
3. Engine deposits, spark-plug fouling, and crankcase dilution may
be increased if the injected fuel is not atomized properly.
4. Injection systems may be more sensitive to vapor lock problems.
Therefore, they must be carefully designed for.vapor handling capability.
Engines with good carburetors and manifold systems can provide
better performance than with^fuel injection systems improperly designed
or matched with the engines. Therefore, the need for developmental
work, and the higher complexity, cost, and maintenance have limited the
use of gasoline injection. However, the number of motor vehicles equipped
with gasoline injection systems is increasing, and more general use of
these systems is expected in the future.
From the viewpoint of conservation of fuel, it appears that the
general use,of properly designed gasoline injection engines would be
beneficial. This follows from the greater tolerance of gasoline
injection engines for gasoline volatility — which, would permit a
higher yield of gasoline from crude oil — and from the potential of
these engines for better utilization of fuel octane numbers (especially
with sensitive gasolines). In practice, however, the realization of
this improvement in the conservation of fuel would be extremely difficult.
This is due to the very complex and extensive developmental work that
would be required to equip all motor vehicles with satisfactory gasoline-
injection engines for gasoline of wider volatility limits; and also
because the fuel requirements of the existing carbureted engines still
would have to be satisfied. Nevertheless, the potential for improved
fuel conservation associated with gasoline injection exists, and should
not be overlooked. .
*
Also, additional refinements and unconventional carburetion principles -
such as the use of variable throat and supersonic flow effects introduced
in the Dresser carburetor — could improve carburetors further.
------- |