NACA RM E55G28 S. •• • S..

• • s• •• 205
•. S • .S • •
• . .. •
• S • S
•
S •.
...
S • . .. . S. •S••
.. ..• •. S.. S•S • • S•
•. S. S. S S S 555 I.
CHAPTER XIII

COKE DEPOSITION AND SMOKE FORMATION I] iur LBOJE I' ENGINES

By Jerrold D. Wear and Robert R. Hubbard

INTRODUCTION
-X67-87773
- --
In the early development of jet engines, it was occasionally found that exces-
a

sive amounts of coke or other carbonaceous deposits were formed in the combustion
chamber. Sometimes a considerable amount of smoke was noted in the-exhaust gases.
Excessive coke deposits may adversely affect jet-engine performance in several ways.
The formation of excessive amounts of coke on or just downstream of a fuel nozzle
(figs. 116(a) and (b)) changes the fuel-spray pattern and possibly affects combustor
life and performance. Similar effects on performance can result from the deposition
of coke on primary-air entry ports (fig. 116(c)). Sea-level or altitude starting
may be impaired by the deposition of coke on spark-plug electrodes (fig. 116(b)),
deposits either grounding the electrodes completely or causing the spark to occur
at positions other than the intended gap. For some time it was thought that large
deposits of coke in turbojet combustion chambers (fig. 116(a)) might break away and
damage turbine blades; however, experience has indicated that for metal blades this
problem is insignificant. (Cermet turbine blades may be damaged by loose coke de-
posits.) Finally, the deposition of coke may cause high-temperature areas, which
promote liner warping and cracking (fig. 116(d)) from excessive temperature gradients
and variations in thermal-expansion rates. Smoke in the exhaust gases does not gen-
erally impair engine performance but may be undesirable from a tactical or a nuisance
standpoint. Appendix B of reference 1 and references 2 to 4 present data obtained
from full-scale engines operated on test stands and from flight tests that indicate
some effects on performance caused by coke deposits and smoke.

Some information about the mechanism of coke formation is given in reference 5

and chapter IX. The data indicate that (1) high-boiling fuel residuals and partly
polymerized products may be mixed with a large amount of smoke formed in the gas
phase to account for the consistency, structure, and chemical composition of the
soft coke in the dome and (2) the hard deposits on the liner are similar to petro-
leum coke and may result from the liquid-phase thermal cracking of the fuel.

During the early development period of jet engines, it was noted that the ex-
cessive coke deposits and exhaust smoke were generally obtained when fuel-oil-type
fuels were used. Engines using gasoline-type fuels were relatively free from the
deposits and smoke. These results indicated that some type of quality control would
be needed in fuel specifications. Also noted was the effect of engine operating condi-
tions on coke deposition. It is possible that, even with a clean-burning fuel, an ex-
cessive amount of coke could be formed at some operating conditions. In this case,
combustor redesign could possibly reduce the coke to a tolerable level. This chapter
is a summary of the various coke-deposition and exhaust-smoke problems connected- with
the turbojet combustor. Included are (1) the effect of coke deposition on combustor
life or durability and performance; (2) the effect of combustor design, operating con-
ditions, inlet variables, and fuel characteristics on coke deposition; (3) elimination
of coke deposits; (4) the effect of operating conditions and fuel characteristics on
formation of exhaust smoke; and (5) various bench test methods proposed for deter-
mining and controlling fuel quality.

COKE DEPOSITION IN TURBOJET ENGINES

The coke deposits in a turbojet-engine combustor are formed on the liner

walls from 2 to 6 inches downstream of the dome, in the dome, on the fuel
206 •••• •.. • •• •• CNFTI NACA RM E55G28
• • • • •S• S
• :. • •. • •
•S • • • • • • • • •
• S ••. S
.. • • •.• . .••
.: •.: : : : •..

nozzle, and on the spark plug (fig. 116(e)). The deposits on the liner are gen-
erally hard and of a medium gray color, show erosion marks or streaks caused by the
hot gases, and adhere strongly to the liner walls. The dome and nozzle deposits are
sometimes softer than liner deposits and generally darker in color. These deposits
are probably not heated to as high temperatures as are the liner deposits. The
amount of coke deposit obtained in one tubular combustor of a full-scale engine has
been as much as 170 grams (fig. 116(a)); 250 grains has been obtained from full-scale
annular combustors. Although most investigators consider deposit weight as a measure
of the amount of coke, it is possible that volume instead of weight should be
considered.

Effect of Coke on Combustor Life and Performance

As stated previously, large deposits on liner walls can cause severe tempera-
ture gradients that may result in warping and cracking of the liner. Coke deposits
on or just downstream of the fuel nozzle may cause alterations in the fuel-spray
patterns with possible effects on performance.

Life or durability. - Severe liner warping that occurred just downstream of a

large coke formation is shown in figure 116(d). Figure 1 of reference 6 also in-
dicates quite clearly the effect of coke formations on liner warping. Information
presented in reference 6, which was obtained from military flight bases and overhaul
stations, indicates that severe coke deposition can cause liner failure in as little
time as 30 hours. In some engines, five out of eight liners had failed at the 30-
hour check time. Generally, flight stations that reported heavy coking in their en-
gines had much higher rates of liner replacement than did stations that reported
only small amounts of coke build-up. In some cases where the coke deposition was
generally quite low, liner life was as long as 200 hours. Liner life can also be
shortened because of a distortion or deflection of the fuel spray by coke formations
that causes actual burning of part of the liner. This type of failure is also re-
ported in reference 6.

Combustion efficiency and stability. - Loss in fuel heating value due to coke
formation in the combustion chamber is unimportant even under the worst conditions.
As explained in chapter IX, for such conditions, the loss due to coke deposition is
only of the order of 0.004 percent. However, coke formation can affect efficiency
by other means. For example, an increase in efficiency at low fuel-air ratios ac-
comparLied the formation of coke on fuel nozzles in an early type tubular combustor
(fig. 117). The periods of operation shown in this figure were conducted under
various test conditions. This unexpected increase in efficiency was attributed to
a change in the fuel-spray characteristics. Photographs of the fuel sprays in still
air, obtained before and after the coke had been deposited, indicated that the de-
posits caused a large increase in fuel-spray angle and provided improved fuel drop
distribution at low fuel-flow rates. It is noted that cleaning the fuel nozzle after
61 hours of operation reduced combustion efficiency to the original values. The in-
4
sertion of a nozzle shield that prevented the formation of coke on the nozzle tip
resulted in constant combustion efficiency with run time. An opposite effect of
coke deposits on full-scale-engine performance is reported in reference 6. It was
reported from one flight station that if the domes and liners were not cleaned, a
significant loss In engine power and efficiency (increased tail-pipe temperature for
a given engine speed) was observed within 60 hours.

Results of an investigation reported in reference 7 show the effect of coke de-

posits on altitude limits of a single tubular combustor. The altitude limits were
NACA RM E55G28 Fp1'TIAL. . .. .. . ,, • •• • 207
• S •S S. . •S ••S
• • •
S SS • • 5•
• . . . S •s• .
. •. ••5
.
• S. S S
5• 55. •f•5s...
.. S • . . SS
• ••• ••
determined before and after an 80-hour test run, during which the coke deposit pic-
tured in figure 118 was accumulated. The operational limits were reduced somewhat
as a result of the coke deposits, as shown in figure 119. In general, there seems
to be no reason to expect coke deposits to improve either the efficiency or the al-
titude limits of well-designed combustors. Decreases in efficiency and altitude lim-
its would appear more likely, although adequate substantiating data are not
available.

Effect of Combustor Design and Operating Conditions on Coke Deposition

The coke deposits in a combustor are affected by combustor design and operating
conditions or inlet variables. The amount of the coke deposit is determined by the
operating time and by the rates of formation, burning, and erosion. These rates are
in turn affected to different degrees by design and by inlet variables. As previ-
ously mentioned, coke deposits may result from liquid-phase thermal cracking of the
fuel on the liner walls. The softer coke found in the dome is probably high-boiling
fuel residuals mixed with a larger amount of smoke that was formed in the gas phase.
Therefore, a combustor design that permits a large amount of liquid fuel to impinge
on the hot liner walls should accumulate more coke, provided that the temperature of
the liner walls is proper for thermal cracking of the liquid. Preventing the liquid
fuel from getting to the liner walls or using vapor-fuel injection should decrease
the coke deposits.

The direct effects of engine operating conditions or combustor-inlet variables

on combustor deposits are somewhat obscured by the interdependence of these varia-
bles, which include inlet-air pressure, temperature, velocity, and fuel-air ratios.
Investigators employing small-scale or full-scale single tubular combustors have
tried to determine the singular effect of pressure, for example, by holding mass air
flow, temperature, and fuel-air ratio constant. In this case, the inlet-air veloc-
ity varies with pressure. Then to determine whether the pressure or velocity is
affecting the quantity of deposits, tests must be made with constant pressure, tem-
perature, and fuel-air ratio and varying velocity. The change in mass air flow, to
vary velocity, requires a change in fuel flow to maintain constant fuel-air ratio
which in turn requires tests to determine the effect of total fuel flow on deposition.
Because of the large number of tests required, most investigators have not determined
the singular effects of all the combustor-inlet variables. From the available data,
however, some indications can be obtained as to whether or not the effect of the in-
let variables on combustor deposits is in agreement with the deposit formation mech-
anisms previously postulated.

Combustor design. — Several investigators have determined the effect of change

in "air wash" along the liner walls on coke deposition. Data of reference 8 show a 58-
to 77-percent reduction in weight of coke deposits in a small-scale combustor when the
gaps in the liner-wall louvers were increased from 0.030 to 0.050 inch. Information
presented in reference 9 (full-scale tubular combustor) indicated an appreciable in-
crease in coke deposition when one-third of the liner louvers were closed. Unpub-
lished NACA data (full-scale single combustor) show a reduction in coke deposits
from about 30 grams to 1 gram when the combustor was modified so that an annulus of
air was directed downstream surrounding the fuel spray. This annulus of air was ob-
tained by modifying the upstream end of the combustor dome as follows: A 3-inch-
diameter hole was cut in the end of the dome, and a flat plate inches in diameter
was installed at the back of the fuel nozzle so that the plate was centered in the
3-inch opening. The primary air entering the i -inch annulus (1:1T in. from the fuel
 208 09 009 0 00 • •• •• •• • • I'ACA RM E55G28
•S •S • • • S • •S S • •• • •
• • SS • ••-. • . . S S • SS•
•S S• S•• S •S.
•• ..• • . S •• •• • • S.. •• 515 55
nozzle) prevented most of the liquid fuel from impinging on the liner walls. This
method also decreased the amount of liquid fuel recirculating into the dome; smoke
formed in the gas phase could not adhere as readily to the dome surface.

Injection of vaporized fuel (considered as part of combustor design) has an ef-

fect on coke deposits, as shown by data presented in reference 10. Less liner de-
posit was obtained in a small-scale tubular fuel-vaporizing combustor than in a
similar design of a fuel-atomizing combustor operated at comparable conditions.
However, if the fuel is not completely vaporized, the liner deposit may be increased
over the values obtained with the atomizing combustor. Also, deposits may form in
the vaporizing tubes and cause a decrease in fuel flow that will probably result in
tube failure.

Inlet-air pressure. - Increasing the pressure increases the smoke-forming tend-

ency of hydrocarbon flames (ref. 11); thus, combustor deposits should increase with
increase in inlet pressure. This is generally substantiated by information presented
in references 12 and 13 (small-scale tubular combustors; fuel-atomizing and
-prevaporizing, respectively) and reference 14 (full-scale single tubular combustor).
For example, data in figure 120 show a continued increase in total deposit weight
with increase in pressure. In this case, velocity, temperature, and fuel-air ratio
were relatively constant while fuel flow varied. Deposit weight per unit of fuel,
which tends to minimize the variable-fuel-flow effect, also increased with pressure,
but at a decreasing rate. This leveling off of combustor deposits at higher pres-
sures is similar to the pressure effect on the rate of smoke formation shown in ref-
erence 11. This may be caused, in part, by an increased rate of erosion with in-
crease in air density.

Inlet-air temperature. - An increase in the combustor inlet-air temperature

should decrease combustor deposits because of the increased evaporation rate of the
liquid fuel from the liner walls and the resultant decrease in fuel residence time.
Investigations reported in references 13 and 15 (small-scale tubular combustors;
fuel-vaporizing and -atomizing, respectively) and reference 16 (full-scale single
tubular combustor) were conducted to deterthine the effect of inlet-air temperature
on coke deposition. Results of these investigations are somewhat conflicting re-
garding the temperature effect on combustor deposits. For example, data taken at
a constant air velocity (ref. 13) show that the flame-tube deposit of a high-deposit
fuel increased with increase in air temperature up to 250 0 F, then decreased as tem-
perature was further increased up to 400 0 F. Similar data (ref. 15) are shown in
figure 121(a). The data were obtained in a small-scale tubular fuel-atomizing com -
bustor at constant mass air flow. Increase in the air temperature from 100 0 to 2500
F increased the amount of deposits; however, further increase in temperature to 3250
F caused a decrease in deposits. Data for a full-scale single tubular combustor
(ref. 16) for both constant mass air flow and constant inlet-air velocity indicate
a decrease in deposits as inlet-air temperature was increased from 100 0 to 3000 F.
As the air temperature was increased from 300 0 to 5000 F, the deposit quantity in-
creased. Data obtained in a small-scale tubular fuel-atomizing combustor (ref. 17)
at constant reference air velocity show an increase in deposits as air temperature
was increased from 100 0 to 3500 F. The inlet-air pressure was varied from 20 to 50
pounds per square inch absolute. Varying inlet-air temperature from 200 0 to 8600 F
(ref. 18) caused a decrease in deposits at inlet pressures-from 60 to 173 pounds per
square inch absolute, as shown in figure 121(b). The data were obtained at constant
reference air velocity and fuel-air ratio.

It is apparent that the processes that cause combustor deposits are affected
to different degrees by change in inlet-air temperature. With some particular
inlet-air temperature, combustor geometry, and fuel, one process may predominate;
 .. S.. • • •
• . S • • • • . S
•• • •.• • S.. ••
I.
•S • • 5 • S
-
NACARME55G28 • ••• 00I
• S. • •• •a 09
00 000 00 0 0
• , • ••

with other combinations another process may predominate. In some cases the increase
in temperature will increase the vaporization rate of the fuel, thereby decreasing
the fuel residence time on the walls; in other cases the thermal-cracking rate (with
increase in inlet-air temperature) may become greater, relative to the vaporization
rate, and therefore deposits will increase. Of course, if the inlet-air temperature
is high enough, the liner temperature will be above the minimum temperature at which
coke will burn, and, therefore, combustor deposits will not build up on the liner
(fig. 121(b))..

Inlet-air velocity. - An increase in the inlet-air velocity would be expected

to decrease combustor deposits because (1) theincreased scrubbing action of the air on
the liquid fuel decreases fuel residence time on the walls, (2) less fuel impinges on
the liner walls, and (3) the erosion of deposits is greater; Conversely, the fuel re-
circulation increases with increase in air velocity, and the greater amount of liq-
uid fuel in the dome may increase deposits. The relative importance of these dif-
ferent processes is not known. Data of references 14 and 19 (full-scale single
tubular atomizing combustors) show an increase in deposits as the velocity was in-
creased. Information presented in reference 17 for a small-scale tubular atomizing
combustor indicates a decrease in deposits as velocity was increased. Data shown in
figure 122 indicate that, depending on the fuel-air ratio, there will be either an
increase, no change, or a decrease in combustor deposits with increasing inlet ve-
locity, which is proportional to mass air flow. It is apparent that the effect of
inlet velocity on combustor deposits is similar to the inlet-air-temperature effect.
Depending on the operating conditions and combustor geometry, combustor deposits may
increase or decrease with increase in velocity.

Fuel-air ratio. - Increase in fuel-air ratio of fuel flow should increase com-
bustor deposits because of higher local fuel-air ratios and the increased amount of
residual wall fuel. The increased primary-zone temperature resulting from the in-
creased fuel-air ratio should cause either an increase in deposits through increased
therthal cracking or a decrease in deposits because of increased burning and erosion
of the deposits. Investigations reported in reference 13 (small-scale fuel-
vaporizing tubular combustor) and references 14 and 19 (full-scale single tubular
combustors) indicate a general increase in combustor deposits with increase in fuel-
air ratio. Data presented in figure 123 show an increase in total combustor deposits
with increase in fuel-air ratio, with all other variables constant except, of course,
exhaust-gas temperature. However, deposit weight per unit of fuel is constant with
increase in fuel-air ratio. Data shown In figure 122 also indicate an increase in
combustor deposits with increase in fuel-air ratio.

Run time. - The amount of deposit in a combustor depends on the length of the
run time. However, because of increased burning and erosion of deposits as the de-
posit quantity increases, deposition rate remaining constant, the rate of deposit
build-up decreases, and finally at some value of run time a maximum quantity of de-
posits is obtained. This tendency is shown in figure 124.

Summary of effects of combustor design and operating variables. - In summary,

it is noted that certain changes in combustor design and inlet variables affect com-
bustor deposits in the way expected. For example, increased "air wash" along the
liner walls or injection of vaporized fuel caused a decrease in amount of deposits.
Increases in inlet-air pressure and fuel-air ratio both generally caused increases
in coke deposits. Deposit weight increased with run time until an equilibrium level
was reached; the rate of erosion and-burning becomes approximately equal to the dep-
osition rate. Increases in inlet-air temperature or velocity, or both, were expected
to decrease deposits. However, results were conflicting; in some cases deposits in-
creased and in others decreased as air temperature was increased. If the air tem-
perature was high enough (>850 0 F), little deposit was formed. The inlet-air-velocity
 . .. S .. S • .S . . ...
. . •. . . ..•
• .
• S S I • ..
• • • • • •
21D S. • SI I •
• ••. . NACARME55G28
. S • I I S •
00 000 0 go

effect on deposits was similar to the temperature effect; depending on the operating
conditions and combustor design, the deposits either increased or decreased with in-
crease in velocity.

Effect of Fuel Characteristics on Coke Deposition

A substantial quantity of data has been obtained by numerous investigators de-

scribing the effects of various fuel properties on combustor deposits. Since all
inlet variables or :operating conditions can be maintained relatively constant, the
direct effect of féI type on combustor deposits can be determined.

Volatility. - It is indicated earlier in this chapter that combustor deposits

are smaller with gasoline-type fuels than with higher-boiling fuels. This result
is in agreement with the supposition that more easily vaporized fuels should produce
less combustor deposits because of the decrease in fuel residence time. Other data
that substantiate this theory are given in references 9 and 19 (full-scale single
tubular combustors), reference 20 (small-scale tubular combustor), and reference 21
(full-scale annular combustor, 10 1 -in. d.iam.). For example, information presented
in figure 125a) shows a considerable increase in combustor deposit with increase in
the volumetric average boiling temperature (decrease in volatility) at constant
hydrogen-carbon weight ratio. This result suggests that decreasing the liquid-fuel
residence time on the liner walls results in decreased combustor deposits. Hence,
it would seem that the use of vapor fuel would eliminate deposits. However, as in-
dicated in reference 22, some deposits were obtained on the vapor-fuel nozzle and
in the combustor dome when a vapor fuel (propane) was used. These deposits were
formed during incomplete combustion of the fuel in the complete absence of.a liquid
phase.

Fuel composition. - The effect of hydrocarbon type or composition on the tend-

ency of diffusion flames to produce smoke is presented in chapter IX. The fuel
types in order of decreasing tendency to smoke are naphthalene and substituted
naphthalene compounds, aromatics, alkynes, olefins, and normal paraffins. The aro-
matics are about 12 times as smoky as the olefins and about 24 times as smoky as
the normal paraffins. Naphthalene, with its high smoking tendency would be expected
to cause large amounts of combustor deposit. Data of reference 21 (full-scale annu-
lar combustor, 104-in. diam.) show that 139 grams of coke was obtained with a mixture
of a- and -monomethylnaphthalene, 88 grams with triisopropylbenzene, 34
grams with benzene, and 2 grams with a normal paraffin fuel. Other investigators
have shown that aromatic-type fuels generally cause a large amount of coke deposit.
Some of this data is reported in references 9, 23, and 24 (full-scale single
tubular combustors) and references l, 20, and 25 (small-scale tubular combustors. Ar
example of these data is shown in figure 125(b), where the combustor deposits increase
with increase in aromatic content for constant values of volumetric average evaporated
temperature. Some investigators have surmised that the highest-boiling fraction of
the aromatic portion may be causing much of the combustor deposits. Information
presented in reference 12 (small-scale tubular combustor) shows that combustor de-
posits increase as the percentage of aromatics boiling above 4200 F increases. How-
ever, the quantity of coke deposits was not necessarily the same for different types
of fuels that had the seine percentage of aromatics boiling above 4200 F.

Information concerning the effects of normal paraffins, isoparaffins, and cyclo-

paraffins on combustor deposits indicates that these different fuel types have about
the same coke-forming tendencies and that the amount deposited is appreciably less
 :. : •'. : • •• • ••• • •••
NACA RM E55G28 CEN1TAL : :. '. . : . : :. : 2l
•• ••• •• ••• • • •. •. : : : :.. :.
than that obtained from aromatic-type fuels. As shown in chapter IX, except for
the normal paraffins, olefinic-type hydrocarbons have the lowest values of smoking
tendency. In spite of these low values, larger deposits could possibly form because
of the polymerizing characteristics of olefins. However, data from investigations
reported in references 9, 23, and 26 (full-scale single tubular combustors) and ref-
erence 18 (full-scale annular combustor, 1O -in. diam.) show that the quantity of
deposits obtained with olefinic fuels is similar to that obtained from paraffin
fuels of the same boiling range. One investigator (ref. 27) reported heavy deposits
in a full-scale single tubular combustor from a fuel containing a considerable amount
of diolefins. Unreported investigations conducted in a full-scale single tubular
combustor with a fuel blend consisting of 50 percent JP-3 fuel and 50 percent di-
pentene showed an increase of about 25 percent in combustor deposits over those ob-
tained with JP-3 fuel alone. For the same operating conditions, an aromatic fuel
blend with an average boiling point of about 380 0 F caused an increase in deposits
of about 230 percent over those for JP-3 fuel.

Nonhydrocarbon fuel components. - Sulfur comppunds and gum are minor nonhydro-
carbon components present in petroleum-derived aviation fuels that can have an ef-
fect on combustor deposits. Investigations conducted to determine the effects of
sulfur on combustor deposits indicate that the quantity of sulfur in the fuel can
be increased a considerable amount above the present specification MIL-F-5624C max-
imum of 0.4 percent before an appreciable increase in deposits is obtained. Results
of some of these (full-scale single tubular combustor) investigations are shown in
the following table:

Fuel Additive for Sulfur in fuel, Weight of Refer-

increasing percent by deposit, ence
sulfur weight g
content

Straight- None 0.05 4.1 9

run I .5 3.1
kerosene I 1.0 5.0
$ 3.0 7.2

AN-F-58 None 0.034 3.1 28

Thiophene .39 2.3
Thiophene .91 1.2
Disulfide 1.00 3.6
oil

.ffL-F-5624A. None 0.04 7.0 26

grade JP-3 Disulfide .55 8.8
oil
Disulfide 1.03 8.4
oil
None .04 5.6
Mixed butyl 1.05 6.5
mercatansj

It is noted that the sulfur content of a fuel can be increased to 1 percent or more
before a significant increase in combustor deposits is obtained. Thiophene added
to a fuel in amounts of 1 and 3 percent (0.39 and 0.91 percent sulfur) actually
caused a decrease in deposits.
 • • • • .S . . •• ' • t0JFI.MDL NACA PM E55G28
• • S. I •• S
S
•
••S
•
• S I I S S S
S S •I •• S • ••. •• ••• ..
•5Ss•S

The majority of investigations that have been conducted to determine the effect
of gum, either existent or potential, on combustor deposits show, in general, that
the specification MIL-F-5624C maximum limits of 7 and 14 milligrams per 100 milli-
liters of fuel may be increased several times before an appreciable effect
on combustor deposits is obtained. The following data were obtained from some of
these investigations:

Fuel Combustor Gum, Weight of Reference

mg/loo ml deposit,
g
Existent Potential
H
MtL-F-5624A, Full-scale 7 12 12.6 26
grade JP-3 single 77 445 14.7
tubular 165 560 15.4

MIL-F-5624A, Full-scale 9 12 2.1 26

grade JP-3 single 95 103 2.6
tubular fuel
vaporizing

AN-F-58 Full-scale 6.5 3.5 4.3 28

single 4.3 4.2 3.3
tubular 21.1 25.3 4.1
34.1 55.8 4.8

The existent and potential gum contents were determined by the A.S.T.M. Standards
D381-46 and D525-46, respectively. It can be seen that a large increase in gum
content is possible before any significant increase in combustor deposits is ob-
tained. In the data from reference 26, the gum content-of the fuels was increased
without any appreciable change in other fuel characteristics. Information obtained
from reference 29 showing the effect of existent gum content of fuels on deposits
in small-scale tubular combustors (fuel-atomizing and fuel-vaporizing) is presented
in the following table:

Fuel Existent Deposit weight,

gum, mg/hr
mg/loo m
Atomizing- Vaporizing combustor
combustor
flame tube Flame Outside Inside
or liner tube or vaporizer vaporizer
liner tube tube

A174 (base) 6.1 260 9.0 1.5 16.2

A76 58.5 --- 81.2 7.5 377.5

A84 (base) 5.2 216 16.6 7.0 13.5

A98 59.2 185 49.0 16.5 425.0

The data indicate that an existent gum content of about 8 or 9 times the maximum
specification limit may cause large increases in the deposits in the vaporizing-
combustor flame tube or liner and inside the vaporizer tube. Information presented in
reference 30 shows that the vaporizing tubes plugged after about l hours of operation
with a low-aromatic fuel with a gum content of 100 milligrams per 100 milliliters of
fuel. This plugging caused rough burning and unsatisfactory operation of the com-
bustor. Continued combustion with a plugged vaporizer tube would probably result in
a complete tube failure.
 .. I.. • • •
• • • . . . .. .. . 60P . ... ..
* S • • . • • . .S
NACA RM E55G28 VOFEIIAL'. ' '• ' • 1 • • •' • l3
•'5• SS • •
S. 050 •• ••S S • •• •5 S S • 005 55
55

Elimination of Coke Deposits

The effects on combustor deposits of changes in combustor design and fuel in-
jection, variation in combustor inlet-air variables, and changes in fuel quality
can be qualitatively estimated from deposit formation mechanisms previously pre-
sented. From these considerations, methods for decreasing or eliminating combustor
deposits can be proposed.

Combustor design. - Any design changes in the combustor that decrease or prevent
liquid-fuel impingement on the liner walls decrease combustor deposits. Among the
design changes by which deposit reduction may be effected are provisions for (1) fuel
prevaporization, (2) increased "air wash" to the liner walls, (3) change of fuel-
spray angle, and (4) liner temperatures either low enough to prevent thermal decom-
position of the fuel or high enough to burn off the deposits. With fuel prevapori-
zation, additional problems are encountered, such as plugging of vaporizing tubes,
and these problems may overshadow the advantages of decreased combustor deposits.
Increased "air wash" and changed fuel-spray angle may have adverse effects on combus-
tion efficiency, stability, or altitude blow-out. It may be impracticable
to keep liner temperatures below the fuel-decomposition temperature, and very high
liner temperatures shorten liner life (ref. 18). It is apparent that the final com-
bustor design must be a compromise involving several factors, and combustor coke
deposition is only one of these factors.

Fuel-quality control. - Combustor deposits can be greatly affected by choice of

fuel. As previously shown, deposits generally increase as the state of the fuel is
varied from vapor to liquid and as the type of the fuel is varied from high to low
volatility or from paraffin to aromatic. From the standpoint of reducing combustor
deposits, the vapor fuel would be best. As previously mentioned, the special equip-
ment needed to obtain vaporized fuel may be sensitive to coke deposits. However,
continued research may overcome this problem. High-volatility fuels cannot be used
in aircraft because tank losses due to fuel boiling and slugging become excessive
(ref. 31). For liquid-fuel-atomizing combustors, the paraffin-type fuels cause less
deposits than aromatic fuels. However, the availability of paraffin or nonaromatic
fuels is limited. It can be seen that the final fuel will be a compromise depending
on the amount of importance that is assigned to the various considerations.

Fuel additives. - Another means of reducing combustor ' deposits is by use of

fuel additives. Additives effective in reducing or eliminating coke from commercial
furnaces and Diesel engines may also be effective in turbojet combustors. Data of
reference 32 (full-scale single tubular combustor) show a decrease in deposits when
small amounts of tetraethyl lead were added to the fuel. Deposits decreased about
45 percent with a lead concentration of about 0.002 percent. Further increases in
lead concentration eventually cause an increase in deposits because of increased
amounts of lead oxides. With the need of JP fuels and military aviation gasoline
in a national emergency estimated as about 1,000,000 and 300,000 barrels per day,
respectively, the metallic lead needed per day for JP fuels (0.002 percent by weight)
would be about 2 tons and lead required for aviation gasoline (4.6 ml TEL/gal) would
be 55 tons per day. Other additives that reduced deposits were lead naphthenate,
secondary amylnitrate, and commercial fuel-oil additives (refs. 16, 32, and 33; full-
scale single tubular combustors). In small-scale combustors (refs. 34 to 36),, the
decreases in deposits were negligible with additives such as lead naphthenate, U.O.P.
Inhibiter No. 5, aniylnitrates, ditertiary butyl peroxide, and a commercial additive.
Aviation tetraethyl lead in a concentration of 4 cubic centimeters per gallon (0.12
percent lead by weight) caused a considerable increase in deposits because of the
lead deposits added to the coke. Reference 37 presents data for a large number of
 • . S ••S ••
.. S.. S SI S S• ••
214 • S •S I IS S I S • S I, C0FINI NACA RN E55G28
• • • S. SI S • •S• •• ••S •I
•5I•••

additives tested in small-scale atomizing and vaporizing combustors. The additives

consisted of halogen compounds, antioxidants, organometallics, silicates, carbonyls,
high-molecular-weight organics, and water. Tetraethyl lead (0.01 percent by volume)
and iron pentacarbonyl (0.1 percent by volume) were the only additives that caused
any appreciable decrease in combustor liner deposits, and water was the only addi-
tive that caused a decrease of fuel-vaporizer-tube internal deposits.

The results of various investigations to decrease combustor deposits by special

fuel additives are not conclusive enough at the present time to determine whether
the advantages are sufficient to offset such problems as storage stability, delete-
rious effects on engine fuel systems and hot engine parts, and perhaps availability.
The over-all benefits must be determined by further research.

SMOKE FORMATION IN TURBOJET ENGINES

As previously indicated, smoke formation in turbojet combustion is less import-

ant than is coke deposition. There is no measurable loss in performance due to
smoke, and, In fact, operating conditions conducive to smoke formation are the same
conditions that yield generally high combustion efficiencies. For this reason, less
work has been done in the field of smoke research in full-scale engines and in single
combustors than has been done in the study of coking. The reverse has been true . in
bench-scale experimentation (ch. IX).

Bench-scale studies reported in chapter IX have shown that smoke can be formed
only in a fuel-rich environment. The primary zone in turbojet combustors certainly
contains local areas that are fuel-rich under most operating conditions, and it is
likely that smoke is often generated in considerable quantities. The fact that
smoke is not found in much higher concentrations in the turbojet exhaust indicates
that much of this material is consumed in passing through the burner. This theory
is confirmed by electron microscopy studies of turbojet smoke (ch. Ix).

Smoke formation has been determined qualitatively in full-scale engines and

single combustors by visual observation. Smoke has been determined quantitatively
by either filtration and optical measurements or by collection and weighing. In
the filtration method, the exhaust is passed through a filter at a given rate for
a given time, and the darkening of the filter is then determined by either visual
rating or by optical methods. This method was used in references 13, 14, 20, and
38. In references 27, 39, and 40, the smoke was trapped by bubbling the exhaust
gases through water, and the smoke was weighed after filtration and drying. There
has been no attempt to report absolute smoke values by either method; the data only
indicate the effect obtained when a small and unreported fraction of the exhaust
gas is sampled. Comparison between laboratories on an absolute basis is obviously
impossible, and the results reported herein are given only as trends.

Effect of Operating Variables on Smoke Formation

A systematic study of the effect of inlet variables on smoke production was

made in a full-scale tubular combustor using the filter-darkening technique (ref.
38). With a typical JP-3 fuel, the exhaust smoke increased from two- to tenfold as
the inlet pressure increased from 35 to 85 inches of mercury absolute. The con-
trolled variables were fuel type, inlet temperature, fuel-air ratio and inlet-air
velocity. This effect of a marked increase in smoke with increasing pressure is
also reported in reference 14 for pressures up to 350 inches of mercury absolute
and in reference 13 for pressures up to 150 inches of mercury absolute. This ef-
fect of pressure is also fully confirmed in the bench-scale work reported in ref-
erence 11 and by the fact that, in flight, turbojet-powered aircraft leave the
heaviest smoke trails at low altitudes.
 .. •.. . S • •S •S S SSè
• . . S ••• 5.
. . • . . . . S S S • • •
• • .. S S S • S • • • • •.
NACA EM E55G28 C0IIkr S ••• 5 . . . . .

With the other variables constant, it was also shown in the work of reference
38 that smoke production increased from two- to fain' fold as the fuel-air ratio in-
creased from 0.008 to 0.016. Above a fuel-air ratio of 0.016, there was a slight
decrease in smoke up to a fuel-air ratio of 0.022. In the tests of reference 13,
• fuel-air ratios were varied up to 0.015 and in the tests of reference 20 up to
0.020. In both cases, the smoke production increased with increasing fuel-air ratio
up to the highest ratios tested. These values are over-all fuel-air ratios, and the
smoke was obviously formed in much richer mixtures. Bench-scale data plotted in
figure IX-4 show that smoke continually increases with increasing fuel-air ratio.
In general, there is agreement that smoke increases with increasing fuel-air ratio.
The slight decrease at higher ratios reported in reference 38 is accounted for by
the probability that increasing flame length at high fuel-air ratios consumed a
large fraction of the smoke that was initially formed.

Both references 13 and 38 show that smoke formation is substantially independent

of inlet-air temperature over the range between 100 0 and 400 0 F. This result is in
full agreement with the bench-scale work reported in reference 41.

It is difficult to fully isolate the effect of inlet-air velocity as a variable

in reference 38 since other air parameters vary simultaneously. However, it appears
that smoke increases slightly with increasing velocity at low fuel-air ratios and
decreases slightly with increasing velocity at high fuel-air ratios. Reference 13
shows a considerable decrease in smoke with increasing velocity at low fuel-air
ratios, but the data are few and scattered. The effect of inlet-air velocity on
smoke would be expected to be a function of combustor geometry, but little agreement
is likely between investigators regarding this relation.

Smoke tests have been made on a full-scale engine running at near-sea-level

conditions with a JP-3 fuel (ref. 38). Engine speed and exhaust-nozzle area were
the controlled variables and fuel-air ratio and inlet-air pressure, temperature, and
velocity were the dependent-variables. Because the various inlet parameters cannot
be independently varied in full-scale-engine testing, the independent effect of the
parameters on smoke cannot be determined. In the tests of reference 38, smoke in-
creased 3 to 4 times as engine speed increased from 5300 to 8000 rpm. Smoke in-
creased to a lesser extent as nozzle area was decreased by 33 percent. The changes
in engine operating conditions that increased smoke also increased fuel-air ratio,
pressure, and temperature. It seems probable that increases in fuel-air ratio and
pressure were the main factors contributing to increased smoke, a theory confirmed
by the observations from both bench-scale and single-combustor studies.

Effect of Fuel Quality on Smoke Formation

A small amount of work has been done in several laboratories relating smoke
formation to fuel quality. Four fuels of low aromatic content but of varying vola-
tility were tested in a full-scale single combustor (ref. 38). As fuel volatility
increased, the maximum values of smoke density did not change appreciably, but the
fuel-air ratio at which maximum smoke was produced shifted to lower ratios. Data
presented in references 19 and 20 and chapter IX indicate that smoke formation is
less dependent on fuel volatility than on other fuel factors. Chapter IX shows
that smoking tendency does not vary greatly with molecular weight in the range of
molecular weights covered by the usual petroleum-derived fuels. However, the effect
of hydrocarbon type on smoke production is quite pronounced, the aromatic fuels
yielding considerably more smoke than the nonaromatic ones. References 19, 20, and
40 show this effect in combustor studies, and chapter IX shows it for bench-scale
flames.
 .. ..• . .. . .. S •. S S .S
• •
26.
•S
• • . •
S. • S• S S • S• • •SS
S•
•5 S.. •
S S
• . S
: cik
• . S..• • ••

• • •• .. . S •5• •S 555. 55
•
NACA PM

Liner coke is quite dependent on both the hydrocarbon type and the volatility
characteristics of fuels, the latter presumably being important since it relates to
the residence time of liquid fuels on combustor walls. Smoke, on the other hand,
is largely dependent on hydrocarbon type but is little affected by fuel volatility.
Therefore, a poor correlation might be expected between the coking and smoking tend-
encies of fuels, and this poor correlation is, in fact, noted in references 20, 27,
and 40.

EVALUATION OF FUEL DEPOSIT-FORMING CHARACTERISTICS

m
Since the influence of fuel quality on coke deposition is a matter of consider-
able importance affecting the performance and reliability of turbojet engines, it
would be desirable to have laboratory tests that could be used to evaluate and con-
trol the quality of such fuels. In general, the most accurate deposit-forming
evaluation of fuels is made by operating complete engines under service conditions.
This approach is prohibitively expensive, and easier methods have been sought.

A fairly close approach to the prototype testing of turbojet fuels has been
made by taking single combustors from full-scale engines and using these in
connected-duct facilities. A considerable amount of deposit-formation work has
been done in such test units. In order to determine whether results of tests with
single full-scale combustors can be related to full-scale-engine coke deposition,
several full-scale-engine tests were conducted. The full-scale engines were oper-
ated on test stands (refs. 2 and 4). Results of some flight test data are also
available and can be used for comparison (appendix B of ref. 1 and refs. 3 and 6).
The results show, in general, a good relation between deposits in full-scale single
çombustors and full-scale engines. However, even full-scale single-combustor test-
ing is too costly in terms of facilities, pdwer, fuels, and time to permit its use
for the control of fuel quality by manufacturers.

Therefore, considerable effort has gone into the development of either a simple
bench-scale test which will correlate with the performance of fuels in engines or
the, correlation of other easily determinable fuel properties with full-scale-engine
performance. Since little quantitative full-scale engine data are available on the
coke-forming tendencies of turbojet fuels, much of this effort has been toward cor-
relating bench-scale results with the results from single-combustor testing. There
has been no complete agreement concerning the best correlating test. The following
discussion describes the progress that has been made to date.

Correlation of Fuel Properties With Coke Deposit Formation

Certain fuel characteristics and related properties are considered to affect

coke deposits. These include hydrogen-carbon ratio, aromatic content, A.S.T.M.
distillation temperatures, gravity, and aniline point. These properties have been
used singly and in combinations to help indicate the coke-forming tendency of a
uel. One drawback to the use of fuel properties to relate combustor deposits to
the fuel is that these properties do not give any indication of the effect of special
fuel additives on deposits.

Hydrogen-carbon ratio and volatility. - Early investigators attempted to relate

the carbon and hydrogen contents of a fuel and some measure of its volatility to
combustor deposits. Deposits from several fuels obtained in a small-scale combustor
(ref. 20) were empirically related to the fuel as follows:
 .. •.. • ••
• S. •.. .•S .S• SSS •. ••• .• ••.. •S ••SS •S
• S •
I
I • Si •• •• .S S.
NACA EM E55G28 C0'EAL S. ••• S S •S •• S S
S
• .5.

nc/n11 +C1) T
Weight of deposits = log ( (1)

where

C 1 . . . C4 constants

ratio of carbon atoms to hydrogen atoms

T A.S.T.M. 10-percent distillation temperature, OF

The fuels varied from low-boiling or high-volatility cycloparaffins to low-

volatility aromatics, and also included high- and low-volatility fuel blends. In
general, the combustor deposits Increased with an increase in the value of equation
(1), although a considerable scatter was obtained.

The logarithm of the combustor-deposit values obtained from 19 fuels (ref. 21;
full-scale annular combustor, 10_in. diam.) were empirically related to the fuels
by a factor designated as the NACA K factor and shown as

Log of combustor deposits = a + b K (2)

K = (t + 600) (0.7) (3)

where
S
a,b constants

t fuel volumetric average evaporated temperature, OF

H/C hydrogen-carbon weight ratio

The fuels consisted of high- and low-volatility paraffins, olefins, aromatics, and
fuel blends. Examples of the relation of the NACA K factor to deposits of various
fuels obtained in a small-scale tubular combustor, a full-scale single tubular com-
bustor, and a full-scale tubular combustor engine are shown in figure 126. The
deposits from most of the fuels can be estimated by this relation, although some
fuel deposits show wide variations.

Aromatic content and volatility. - Both the total aromatic content of a fuel
and the aromatics boiling above 400° or 420 0 F have been used in attempts to relate
combustor deposits to fuel properties. Some of this type of data is presented in
references 12 (small-scale combustor) and 23 (full-scale single combustor). The
full-scale-combustor data show a regular Increase in deposits with increase in total
aromatic content. Although the sniall-scale-conthustor data indicate an increase in
deposits with aromatics and aromatics boiling above 420 0 F, the increase does not
show any regular trend, and different fuels with the same aromatic content gave dif-
ferent deposit values.

- Other investigators have used aromatic content of a fuel and some measure of
volatility as a means of relating combustor deposits to the fuels. For example, as
shown in reference 23 (full-scale single tubular combustor), the logarithms of the
 •• • • • 5•• S •••
S.-... • ••• • •5
. . S • • • S • S * • .•
• • S• S I
•• • ••S S S • •
. ONF1D IAL NACA RM E55G28
:l •• • • • •• , .. ...•
• • I
*01 ••I •

combustor-deposit values of several fuels are plotted against the percentage of aro-
matics by weight plus one-tenth of either the 80- or 90-percent A.S.T.M. distilla-
tion temperature. The combustor deposits of some fuels can be estimated with reason -
able accuracy by this relation; however, wide variations are obtained with other
fuels.

Volatility, density, and aniline-gravity constant. - Fuel properties such as

specific gravity or density, volatility, and aniline point are easily determined and
therefore are desirable for use in a bench or control test to determine the combus-
tor deposit-forming tendency of fuels. Data of reference 19 (full-scale single
tubular combustor) show how combustor deposits of several fuels vary with molal
average boiling point of the fuel and Universal Oil Products (u.o.P.) characteriza-
tion factor.

U.O.P. factor = (4)

Sp.gr. 600/600 F

where

Tb mean average boiling point, OR

There is a general trend of increased combustor deposits with decrease in U.O.P.

factor, although some fuels vary from this trend. Information presented in refer-
ence 40 for a full-scale single tubular combustor also shows a similar relation of
combustor deposits to the U.O.P. characterization factor.

The U.O.P. characterization factor, the-hydrogen-carbon ratio, and, to some

extent, gravity are all related to the aromaticity of a fuel. Therefore, different
combinations of these various fuel properties should give about the same trend of
deposits with fuel.

The aniline-gravity constant (product of the aniline point in OF and the A.P.I.
gravity) is another fuel property that has been used to rate fuels for their coke-
forming tendencies (ref. 1, full-scale single tubular combustors). With this fuel
property there was wide variation In the data.

The A.P.I. gravity has been used to estimate the coking tendency of fuels.
Data presented in reference 1 (full-scale single tubular combustors) show examples
of a general decrease in deposits with increase in A.P.I. gravity (decrease in sp.
gr .) and other examples where this trend is not so evident. Data of reference 42
show a regular decrease In deposits with increase in A.P.I. gravity.

Laboratory Measurement of Fuel Coke-Forming Characteristics

There does not appear to be a chemical or physical or combination chemical-

physical property of a fuel that will consistently give an accurate estimation of
the coke-forming tendency of a fuel. Therefore, investigations were conducted to
determine whether some type of laboratory combustion test would give a satisfactory
coke-forming rating among fuels.

The investigations included smoke-lamp, flame-plate, and small-pot-burner

methods. Although small-scale combustors could be classed as laboratory equipment,
for the data presented herein, they are discussed along with the full-scale
combustors.
 .. •.. . . . .. .. . •.. . •.. ..
• . . S S • • • • S S S • •
NACA RM E55G28 CO'DETI4 S. •
S..
• • • • • •. • ••
.. ••. S. •S• • • •• • S • • • •
.• . . S ••• S..

Smoke-lamp method. - One simple bench test that has been used by several in-
vestigators is the fuel smoke-point determination by some type of smoke lamp. A
simple wick lamp is used to determine the maximum height of a smoke-free flame of
a particular fuel. The different types of lamp and test procedure are described in
chapter IX. The smoke point is defined as the maximum height of a smoke-free flame
in millimeters h; smoking tendency of a fuel is defined as . Information
presented in references 1 (full-scale single tubular combustors) and 12 (small-scale
tubular combustor) shows a trend of increasing deposits with decrease in the smoke
point; however, there is considerable variation in the data. The average deviation of
some of the deposit data of reference 1 from a curve faired through the data is about
16 percent. The data from reference 12 indicate an average deviation of about 60 per-
cent. An example of some of these data is given in figure 127, which is a plot of
smoking tendency (320/h) and combustor deposits of 16 fuels tested in a full-scale
single tubular combustor. This figure indicates some relation between smoke point
and combustor deposits, although some of the data show wide variations. The smoke-
point test is not affected by small amounts of special fuel additives (ch. IX), but
combustor deposits may be decreased by addition of small amounts of these same
additives.

In an attempt to eliminate variations in the data, investigators have included

the boiling point of a certain fraction of the fuel or a certain function of the
boiling point. Data in reference 43 show the relation of combustor deposits ob-
tained with 16 fuels in a full-scale annular combustor (10 .-in. diam.) to a function
of soiling temperature and smoke point. A somewhat different relation is presented
in reference 44 (small-scale tubular combustor) relating combustor deposits to a
function of smoke point and boiling point. Coke deposits are related to

100 (Volume percent boil-

0.1 (90-Percent distil- ing above 4000 F)
lation temperature, °F) + Smoke point of fuel boil- - smoke point, mm (5)
ing above 4000 F, 'on

The average deviation of data from a faired curve was about 21 percent for the an-
nular combustor (ref. 43) and about 14 percent for the small-scale combustor (ref.
44).

The following relation for Smoke Volatility Index (S ill) (ref. 15) has been used
with some success to determine which fuels will produce above average amounts of
combustor deposits in full-scale engines:

volume percent of fuel boiling under 400 0 F

SVI = smoke point + A.S.T.M. distillation temperature
(6)
2.4

This relation is used in current military procurement fuel specification MIL-F-5624C,

grades JP-3 and JP-4, to limit the carbon-forming tendency of these fuels. The spec-
ification requires that fuels have SVI values greater than 54.. Figure 3(a) of ref-
erence 46 shows the SVI values of 21 fuels plotted against deposits obtained in a
full-scale single combustor. The average deviation from a faired curve is about 27
percent. Data showing the relation of the SVI and liner deposits obtained in a
full-scale single fuel-vaporizing combustor are given in reference 30. These data
indicate'a much larger average deviation.

Flame-plate method. - Another bench test that has been used to determine the
coke-forming tendency of a fuel is known as the flame-plate test. Fuel is delivered
 : NACA RN E55G28
• .•
go •. ...
'.: ..: : : : .. ..

dropwise to the surface of a tared stainless steel plate maintained at a constant

elevated temperature. The vaporized fuel is ignited by a bunsen burner pilot, and,
after a specified amount of fuel is burned, the plate is reweighed to determine the
amount of deposits. The apparatus and procedure are more fully described in refer-
ences 12, 47, and 48. Results presented in reference 48 show good agreement between
the deposits obtained from the flame-plate test and deposits obtained from the same
fuels tested in a full-scale single tubular combustor. The effect of such fuel
properties as gravity, percent aromatics, and percent aromatics boiling above 4200
F on flame-plate-deposit quantity is shown in reference 12. In reference 35, in-
formation is presented showing the effect of special fuel additives on flame-plate
deposits. Use of tetraethyl lead, U.O. P . Inhibitor No. 5, and amylnitrate resulted
in decreased flame-plate deposits of 50 to 90 percent. Small amounts of tetraethyl
lead and secondary amylnitrate also reduced deposits in full-scale single combustors.
However, the effect of amylnitrate and U.O.P. Inhibitor No. 5 on deposits in a small-
scale combustor appeared negligible. One drawback to this bench teat is the consid-
erable amount of time required to complete the test (6- hr).
f
Pot-burner method. - Another bench test apparatus that has been used to a lim-
ited extent is the "pot burner" (ref. 49). Fuel is fed to a small combustion chamber
at a known rate and burns as it enters the chamber. The weight of deposit is deter-
mined by weighing the residue scraped from the combustion chamber after a predeter-
mined run time.

Summary of laboratory tests. - In summary, several laboratory and bench tests

have been used for determining the deposit-forming tendency of fuels. The tests
included methods using the smoke point, the smoke point in combination with a par-
ticular boiling point, a flame plate, and a 'pot burner.' None of the methods listed
consistently predicts the combustor-deposit-forming characteristics of various fuels.
However, a method that can be used as a first approximation is the smoke-point method
with some function of the distillation temperature. This type of test also has the
advantage of requiring simple apparatus and only a little time for the determination.
The Smoke Volatility Index defined by equation (s) appears to give as consistent re-
suits as do any of the methods.

SIGNIFICANCE OF COKE DEPOSITION AND SMOKE-FORMING RESEARCH IN APPLICATION

TO JET-ENGINE FUEL SPECIFICATION AND COMBUSTOR DESIGN

The information presented in chapters IX and XIII indicates that coke in a jet-
engine combustor liner 'may be formed from liquid-fuel cracking on the hot liner
walls. The softer coke in the dome appears to be a combination of high-boiling fuel
fractions and partly polymerized products mixed with a large amount of smoke.

A high-volatility paraffin-type fuel will cause the least deposits. Decrease

in fuel volatility causes an increase in combustor deposits but has no significant
effect on exhaust smoke. Increase in aromatic content of a fuel causes large in-
creases in combustor deposits and also exhaust smoke. Minor fuel components, such
as sulfur and gum, have no effect on combustor deposits until their concentrations
are considerably above the maximum permitted by present fuel specifications.

Combustion-chamber inlet-air variables of pressure and fuel-air ratio cause

increases in combustor deposits and exhaust smoke; however, the effects of combustor
inlet-air temperature and velocity on deposits are not conclusive. Future turbojet
engines will have higher inlet-air pressures and mass air flows; these are both con-
ducive to combustor-deposit formations. The higher mass air flows which require
 • S
V. . Si.
•. .. S •. • ... .. . • .. S • •• S.
. •S IS
NACA RM E55G28 • •
C0DTIL .
•. S.. •• ••• .. •. '• .• •• .•' •S S
•.
• . 05 •• • • . ... ..

- more fuel cause an increase in combustor deposits. Because of these considerations,

it may be more practical to use special fuel additives to hold the deposits down to
a tolerable level. In some investigations very small amounts of tetraethyl lead and
lead naphthanate (0'.002 percent lead) caused a Significant decrease in combustor
• deposits.

• Any combustion-chamber and fuel-injector design that prevents liquid fuel from
impinging on the combustor walls or decreases the residence time of liquid fuel on
the hot combustor liner walls will decrease combustor deposits. For example, in-
creasing "air wash" to liner walls should decrease fuel residence time and, therefore,
combustor deposits. Combustion with a properly designed vapor-fuel-injection system
should be relatively free of liner deposits.

As previously indicated, a high-volatility paraffin-type fuel will give the

least combustor deposits. However, considerations of availability require the
final fuel to be a blend of the various types. In an attempt to determine whether
the deposit-forming tendency of various fuel blends can be related to fuel proper-
ties, investigations were made of properties such as aromatic content, aromatics and
volatility, hydrogen-carbon ratio and volatility, density, and ani1inegravity con-
stant. Of these properties, a function of hydrogen-carbon ratio and the 50-percent-
evap orated temperature seemed to give the best relation of fuel properties to combus-
tor deposits.

Laboratory and bench test methods for determining the coke-forming character-
istics of fuels include the smoke-point method with some measure of volatility, the
• flame-plate method, and the "pot-burner" method. The Smoke Volatility Index
Ismoke point +
volume percent boiling under 400 0 F A.S.T.M. distillation temperature\
2.4
seemed to be the best method because of the accuracy of the results, the simple
equipment needed, and the short time required for the determination.

-• Lewis Flight Propulsion Laboratory

National Advisory Committee for Aeronautics
Cleveland, Ohio, December 7, 1955

REFERENCES

1. Jonash, Edmund R., Wear, Jerrold B., and Hibbard, Robert R.: Relations Between
Fuel Properties and Combustion Carbon Deposition. NACA EM E52B14, 1952.

2. Wear, Jerrold. B.: Combustor Deposits Obtained with MIL-F-5624A Fuels in a Full-
Scale Single Combustor and in Three Full-Scale Engines. NACA SM E54D16, 1954.

3. Anon.: Survey of Service Performance of JP-4 Fuel. Bur. Aero., Dept. Navy.

4. Wear, Jerrold D., and Useller, James W.: Carbon Deposition of Several Special
Turbojet-Engine Fuels. NACA EM E51CO2, 1951.

5. Clark, Thomas P.: Examination of Smoke and Carbon from Turbojet-Engine Combus-
tors. NACA EM E52126, 1952.

6. Anon.: Second Ruler Survey on Service Experience with Jet Fuels. Res. and Dev.
Dept., Phillips Petroleum Co., Oct. 1, 1953. (Contract NOas 52-132-c.)

7. Wear, Jerrold D., and Douglass, Howard W.: Carbon Deposition from AN-F-58 Fuels
in a J33 Single Combustor. NACA SM E9D06, 1949.
I
 .. ... . ... S •• •S • • • •5• ••
• •, JIETtL
2 22.. .. I 55 I S • • •
• NACA RM E55G28
.. S.. I I • SI •• • • •s••• ••• ••

I
8. Rogers, J. D.: Combustion Characteristics of Gas Turbine Fuels. Prog. Rep. No.
34, Calif. Res. Corp., Mar. 1951. (Air Force Contract W-33-038ac-9083, AMC
Proj. MX-587.)

9. Horstman, W. W.: Gas Turbine Fuels. Summary Rep. No. 1345, Wood River. Res
Lab., Shell Oil Co., June 15, 1950. (Air Force Contract W-33-038-ac-9813
(14169).)

10. Schirmer, R. M., Fromm, E. H., and Kittredge, G. D.: Evaluation of Fuel Char-
acteristics in Thermojet Engine Combustion Processes. Frog. Rep. No. 3, Rep.
403-14-52R, Res. Div., Phillips Petroleum Co., Apr. 15, 1952. (Navy Contract O)
N-
NOas 52-132-C.) '—I

11. Schalla Rose L., Clark, Thomas P., and McDonald, Glen E.: Formation and Com-
bustion ' of Smoke in Laminar Flames. NACA Rep. 1186, 1954. (Supersedes NACA
lRM t s E51E15, E52G24, E52122, E52126, E53E05, E53J12, and E54E03.)

12. McLean, C. D.: Combustion Characteristics of Gas Turbine Fuels. Summary Rep.,
Calif. Res. Corp., Dec. 1952. (Air Force Contract AF 18(600)-152.)

13. Fromm, E. H.: Evaluation of Fuel Characteristics in Thermojet Engine Combustion

Processes. Frog. Rep. No. 2, Rep. 403-22-53R, Res. Div., Phillips Petroleum
Co., May 28, 1952. (Navy Contract NOas 52-132-c, Amendment I.)

14. Wear, Jerrold D., and Butze, Helmut F.: Preliminary Investigation of the Per-
formance of a Single Tubular Combustor at Pressures up to 12 Atmospheres. Ic

NACA EM E53K09, 1954.

15. Marshall, E. F.: Progress Report, Jan. 1-Feb. 28, 1951. -Frog. Rep. No. 35,
Automotive Lab., Sun Oil Co., Feb. 19, 1951. (Contract No. W-33-038-ac-
9086.)

16. Moore, R. A., and Donnelly, J. J.: Summary Rep. No. 2, Rep. No. 48-1-DX, Res.
and Dev. Div., Socony-Vacuum Labs., Jan. 9, 1948. (U.S. Gov't. Contract
W-33-038-ac-8527, AMC Proj. MX-587.)

17. McLean, C. D.: Combustion Characteristics of Gas Turbine Fuels, Summary Rep.,
Calif. Res. Corp., Jan. 15, 1954. (Air Force Contract AF 18(600)-152.)

18. Butze, Helmut F., and Wear, Jerrold D.: Performance of a Tubular Turbojet Com-
bustor at High Pressures and Temperatures. NACA EM E55A24, 1955.

19. Moore, B. A., and Giaccone, F. W.: Summary Rep. No. 46.23-DX, Res. and Dev.
Div., Socony-Vacuum Labs., Nov. 8, 1946. (U.S. Govtt. Contract W-33-038-ac-
8527, AMC Proj. MX-587.)

20. Starkman, E. S., and Cattaneo, A. G.: The Influences of Fuel Characteristics
on Performance in a Reduced Scale Gas Turbine Combustion Chamber. Rep. No.
S-9861, Shell Dev. Co., Nov. 7, 1945.

21. Wear, Jerrold D., and Jonash, Edmund R.: Carbon Deposition of 19 Fuels in an
Annular Turbojet Combustor. NACA EM E8K22, 1949.

22. McCafferty, Richard J.: Vapor-Fuel-Distribution Effects on Combustion Perform-

ance of a Single Tubular Combustor. NACA EM E50J03, 1950.

a
 I. ••• • • . ••
• S S • • • • S S
I•
•
S ••o .
S S
S.. ••
I • I
4 ••
NACA RM E55G28 IEPLAt•• I,•• ••I•
CO• • •• • I S S SS

23. Marshall, E. F.: Progress Report No. 19, May 1, 1947-June 30, 1948. Dev. Div.,
Sun Oil Co., July 22, 1948. (Contract No. W-33-038-ac-9086 (13420).)

24. Jonash, Edmund H., Wear, Jerrold D., and Cook, William P.: Carbon-Deposition
Characteristics of MIL-F-5624A Fuels Containing High-Boiling Aromatic Hydro-
carbons. NACA EM E52G11, 1952.

25. Marshall, E. F.: Progress Report, May 1-June 30, 1952. Prog. Rep. No. 43,
Automotive Lab., Sun Oil Co., Aug. 1, 1952. (Contract No. W-33-038-ac-9086.)
H
-.1 26. Wear, Jerrold D., and Cook, William P.: Effect of Fuel Properties on Carbon
(0
Deposition in Atomizing and Prevaporizing Turbojet Combustors. NPCA EM
E52C24, 1952.

27. Rogers, J. D.: J-33 Jet Propulsion Burner Tests on Carbon Deposition and Effects
of Wide-Cut Fuel Properties. Summary Rep. for period July 1, 1948 to June 30,
1949, Calif. Res. Corp., Aug. 29, 1949. (Army Contract W-33-038-ac-9083, AMC
Proj. MX-587.)

28. Moore, R. A.: Summary Report No. 4, Rep. No. 49.17-DX, Res. and Dev. Div.,
Socony-Vacuum Labs., June 26, 1950. (U.S. Gov't. Contract W-33-038-ac-8527,
AMC Proj. MX-587.)

29. Fromm, E. H.: Evaluation of Fuel Characteristics in Thermojet Engine Combustion

Processes. Prog. Rep, No. 5, Rep. 403-16-52R, Res. Div., Phillips Petroleum
El
Co., Aug. 13, 1952. (Navy Contract NOas 52-132-c.)

30. Horstman, W. W., and Jackson, J. L.: Fuel Requirements of Pre-Vaporizing Type
Gas Turbine Combustors - Materials Section. Rep. No. 1451, Final Rep. Mar.
17, 1952-June 17, 1954, Wood River Res. Lab., Shell Oil Co., July 30, 1954.
(USAF Contract No. AF 33(616)-76.)

31. Barnett, Henry C., and Hibbard, R. R.: Fuel Characteristics Pertinent to the
Design of Aircraft Fuel Systems. NACA EM E53A21, 1953.

32. Jonash, Edmund R., Wear, Jerrold D., and Cook, William P.: Effect of Fuel Addi-
tives on Carbon Deposition in a J33 Single Combustor. I - Three Metallic-
Organic Additives. NACA EM E52H21, 1952.

33. Hlavin, Vincent F., and Cook, William P.: Effect of Fuel Additives on Carbon
Deposition in a J33 single Combustor. II - Seven Commercial Organo-Metallic
Additives. NACA EM E541123, 1954.

34. Marshall, H. F.: Progress Report, Jan. 1-Feb. 28, 1953. Prog. Rep. No. 47,
Automotive Lab., Sun Oil Co., Mar. 11, 1953. (Contract No. W-33-038-ac-
9086.)

35. Bert, J. A.: Combustion Characteristics of Gas Turbine Fuels. Prog. Rep. No.
35, Calif. Res. Corp., May 1951. (AF Contract W-33-038-ac-9083, AMC Proj.
NX-587.)

36. Marshall, H. F.: Progress Report, Sept. 1-Oct. 31, 1952. Prog. Rep. No. 45,
Automotive Lab., Sun Oil Co., Dec. 31, 1952. (Contract No. W-33-038-ac-
9086.)

37. Kittredge, G. D., and Fromm, E. H.: Evaluation of Fuel Characteristics in

Thermojét Engine Combustion Processes, Prog. Rep. No. 6, Rep. No. 403-17-52R,
Res. Div., Phillips Petroleum Co., 1952. (Navy Contract NOas 52-132-C.)

C-
 •• S • S •SS •S
.. ••• • GS•
SS
• •S
S SSS • S S •SS
SS •I S S • • • • S S
•• • •• S S
24•••• ,
S
• • • 55 • •• ..•
• • • •
••• C)EAL
S.. NACA EM E55G28
IS

38. Butze, Helmut F.: Effect of Inlet-Air and Fuel Parameters on Smoking Character-
istics of a Single Tubular Turbojet-Engine Combustor. NACA EM E52A18, 1952.

39. Bert, J. A.: G.E. 1-16 Jet Propulsion Burner Tests. Frog. Rep. No. 12,
Calif. Res. Corp., Apr. 19, 1947. (Army Contract W-33-038-ac-9083, AMC *
Proj. MX-587.)

40. Rogers, J. D.: G.E. J-33 Jet Propulsion Burner Tests. Prog. Rep. No. 23,
Calif. Res. Corp., Apr. 30, 1949. (Air Force Contract No. W-33-038ac-9083,
AMC Proj. MX-587.)

41. Clark, T. P.: Influence of External Variables on Smoking of Benzene Flames.

NACA EM E52G24, 1952.

42. Sharp, J. G.: Fuels for Gas-Turbine Aero-Engines. Aircraft Eng., vol. XXIII,
no. 263, Jan. 1951, pp. 2-8.

43. Busch, Arthur M.: Correlation of Laboratory Smoke Test with Carbon Deposition
in Turbojet Combustors. NACA EM E9K04, 1950.

44. Marshall, E. F.: Progress Report Nov. 1-Dec. 31, 1952. Frog. Rep. No. 46,
Automotive Lab., Sun Oil Co., Feb. 4, 1953. (Contract No. W-33-038-ac-9086.)

45. Eaffy, Allan: Method of Predicting the Carbon Forming Tendencies of Jet Fuels.
Tech. Note WCLP 53-204, Power Plant Lab., Wright Air Dev. Center,
Wright-Patterson Air Force Base, July 9, 1953. (E.0. No. R-531-357A.)

46. Jonash, Edmund R., Butze, Helmut F., and Cook, William P.: Correlation of
Turbojet Combustor Carbon Formation with Smoke-Volatility Index, Smoke Point,
and NACA K Factor. NACA PM E55D28, 1955.

47. Rogers, J. D., and Jones, D. R.: Combustion Characteristics of Gas Turbine
Fuels. Frog. Rep. No. 27, Calif. Res. Corp., Dec. 31, 1949. (Air Force
Contract W-33-038ac-9083, AMC Proj. MX-587.)

48. Rogers, J. D., and Jones D. R.: Combustion Characteristics of Gas Turbine Fuels.
Prog. Rep. No. 28, Calif. Res. Corp., Feb. 28, 1950. (Air Force Contract
W-33-038ac-9083, AMC Proj. MX 587.)

49. Anon.: Deposit Forming Tendencies of Several AN-F-58 Jet Engine Fuels Observed
in Bench Tests and in Full Scale Combustor Tests. Refining Dept., Tech, and
Res. Div., Beacon Labs., The Texas Co., July 6, 1950.

50. Jonash, Edmund R., Barnett, Henry C., and Stricker, Edward G.: Investigation
of Carbon Deposition in an 1-16 Jet-Propulsion Engine at Static Sea-Level
Conditions. NACA EM E6KO1, 1947.

51. Wear, Jerrold D., and Locke, Theodore E.: Effect of Retractable Ignition Plug
on Plug Fouling by Carbon Deposits. NACA PM E50F14, 1950.

52. Dittrich, Ralph T.: Effects of Fuel-Nozzle Carbon Deposition on Combustion

Efficiency of Single Tubular-Type, Reverse-Flow, Turbojet Combustor at Simu-
lated Altitude Conditions. NACA TN 1618, 1948.
 .. S.. S S • I.
• •S S ••• • • S• to
S
NACA RM E55G28 COThFDMI!AL.
S
. . . .
us S
to
. . •• • ••
S. S S S S S S
005 S. 555 5 5 •5 to S S S •SS ••

C-14477

(a) On liner after 2-hour run with Diesel fuel oil (ref. 50).

C-34106

(b) On liner after 25-hour rim with JP-4 fuel.

Figure 116. — Coke deposits in full-scale-engine combustors.
 •S • • • ... ..
... . ... . I.
. . . . • • S •' .• ONF.JDTAL NACA RM E55G28
• S •S S •• S • S
•
•S• • S SSS
• • SS SSS
•5S•• S SS •S •• S S SI.

r h%.-
C-28403
(c) On dome and primary-air entry ports after 30-hour run
with JP-3 fuel.
S

S
 S. •.• • • S •• •s S ••• • • S• ••
• • • S S • SS S • SS
• S •• • • S S•
• • . S S S S. S 55 •S
CODLA• ••• S •S• • . S S • S
NACA RN E55G2e 11

C)
/
H
N)

C.)

I
4"
 . S •S • S •
S. ••• • • SS S •• S •• S S •S • •. • . •
•1 S. • • S• S • • • •
• •• • • • . .•S• I • ( O!F •I2!
• S • • • I • • •• ,•• ,• IAL - NACA RN E55G28
I
c
C-
I-
(4
U) 0 0 0
r4 Or1 '-4 C)
Cl) -P0
I
N H+4r14- H+)
a)+'a) d
4)w
N H 0)0
o '-P i—hf)
r.lLt)
0
N) H
c-I Pir4O r40 c
0
i q-iio 0 Co
0>
0 oc-
o PiOPi
24 Co P4 CD
P4
P4 W
'd
a)
r4 0-P
-P4-) a5
4) w
;
1101 o0
°-)
U)+)V)+)
(0
C'1
rlOU)
o
- on G ' B — q-o $
4r
C']O *
0 .r-4 0r10)LI)
-•4-'
ai a)Q: •
i •HH4-i
0-P 0 Ci)
i r4 CO.4
•1-4
r9
Hc-
I 0)bD 'd
H
00 U) .i-400
r 00
0 4
• -Pc-i Ci)
(Q0a)
ON
C) Nd
0
H 0 ci) 0
o HP4Pi
0 i
LO
+)
i-4 000
P4 •
. OH a)
0
H
c •
C-
HO
Hc-i0 H
co 4-4
. ci) w i
0 i4
0 P4
(0 1-1
Q Ix-i
0' 0 0 0 09
0 OD (0
H
quaoaad Qou9ToTjja uosnqmo
 •. ... . .
• SS SS•
.
SS •
SS
S
05 S S..
SS
• ••• ••
SS S•
• • •S S • SS S • S 55
S 5 •S S
NACA RM E55028 •
• S..
••.
S • S• 55
S 5•
55
SS .29
S 555 S.

0)
N-
'-I
N)

S
 .. ... . S.. • •S •S S • • ••• ..
S••
•S • • • • • . . . . . S
• S S S •• S •
• • •. S •• • •
. . . . •
S. ••• S • • •• •• . I ••. .S S NACA RM E55G28

H
60-" - (0
-

I
55
c-4

d, ____

50

_
cd

-p
cc
'-4
III /
40 ____ ____ ____
Combustor

Clean
— — — Containing coke
-

r-1 / deposits after

Cl)
80-hour test -
45

40
50 60 70 80 90 100
Engine speed, percent normal rated

Figure 119. - Altitude operational limits obtained in clean

single tubular combustor and in combustor containing coke
deposited during 80-hour test. Fuel, minimum-quality
MIL-F-5624A, grade JP-3 (ref. 7).

a-
 .. ,.. . . . S. •S • Si. • ••i ..
• • .• • • • • •
S NACA RN E55G28 • S • IAj
CG•'IiEs • •
•
•••
. • . . •. • •. 2L
S. us •. •... • S.. •S• • S •• ••
• •• 5• .•
S 0
OD
H
D U)
a) OJC
cs
o CJQ)Q)
co -PB
H •s .,-
H -i4-
rj
a)
w-4 H •.s
o •r-1 -i
H
U)
cl)CI
U) 0 C.)
Pia)
EO
.j ..
-1
U) 0) N
.'1 PiH
ow 0 Q4) 0
c'J-___ 0)0
H a)
OCH 0
• 4)
U) r- (0
(0
U) •H
U)
(L) 4-'
o
OPi
H Hr-la)
oajbO
4)
D o 0
-1-b C)
-P
ci H4)
H
Q) -H
0
co -I
o 0 -4
-i-) 4)ia5
w tOOl
o opc-i
0 o
Pild (0
0
4)
() -,Lc)
a)OC')
ç- coo
- 00 o 0 00
• OH.—.
o c'
H
CD
fri
0
C')
(0 0 (0 0
H
I!) 0
C') C') H
q.Isodap wpo
 .. S.. S •SS S •• • •• . . S . S • •SS595••
S• • • • SS
•• • ••• I• •, • . .••• . .
. ON?)TAL NACA EM E55G28
. S I •
.. ... I • I •• •• • • S.. •• ••S •S

I
(J
CD
100

.4.)
CD

-i-'
CS (a
80
-P0
.l
CD 4-)
0

dO
0
60
4.)

a)
0
ci)
14

401—
100
Inlet-air temperature, OF

(a) Small-scale combustor. Inlet pressure, 48 inches of

mercury absolute; total mass air flow, 435 pounds per
hour; fuel-air ratio, 0.0133; JP-1 fuel; run time, 2
hours (ref. 15).

Figure 121. — affect of inlet-air temperature on coke deposition

in tubular combustors.

b4
 •. . .. . . .. . .0 . .. . . .. . . .. ..
• . • • • • • S S
NACA RN E55G28 C0NFIXNfAL S S ••S •
S
• • S
•• •• •.
• • •
23.
•• ••• S • • S
0 000

- 14
I I.
IInlet-air
II
total
pressure,
lb/sq in. abs
12
0 35.
60
86
0---- 173

10

bo A 4.0 g of coke in exhaust

section (for 2-hr run) -
0 B (o ) - 11.5 g of coke in
ID
exhaust section (for 2-
hr run') -
a

ib

C)

-oqH

C)

\
4
PM

'N
\\
2

-B

0 200 400 600 800 1000

Inlet-air temperature, OF
(b) Full-scale single combustor. Inlet reference velocity, 78
feet per second; combustor temperature rise, 11650 F (ref. 18).

Figure 121. - Concluded. Effect of inlet-air temperature on

coke deposition in tubular combustors.
 . .. . .. • .S
. j . . . S S S • •. S ..
•
• ,
•
•
• •••
•corrrm1IA.L
NPCABME55G28 S
• • • •• .. . . S.. •• ••• ••
S. ••S •

2.0

i.e
------
Fuel-air ratio

\ 0 0.0167
1.6
a .0147

------------------------- ) .0111 --
v .0083
1.4

4-,
S
O
P, 1.2
- -
-N- --
S SI
0
()

- ------------------- -
IH 1.0

.8

-
.6 -----------
.4
--
.2

£50 ZOO 220 240 260 280 300 320 340 360
Combustor-inlet mass air flow, lb/hr

Figure 122. - Effect of combustor-inlet mass air flow and fuel-air ratio on coke deposi-
tion in small-scale tubular combustor. Combustor-inlet pressure, 20 pounds per square
inch; JP-1 fuel; run time, 15 minutes (ref. 20).
 .. ... . S •••
• . . go.
S • • . 0S
•
•S • •••
. .
5
. S
•S
SS
5 S 5
NACARME55G28 coN'ILIA S •SS • . S S I • • •

IT
IC I I
Coke deposit
measured in -
a
0
o Grams
o Grams/1000 lb
15 fuel burned

Cc
a)

.016 .018 .020 .022 .024

Fuel-air ratio

Figure 123. - Effect of fuel-air ratio on carbon

deposition in full-scale single tubular com-
bustor. Inlet-air total pressure, 86.2 pounds
per square inch absolute; combustor reference
velocity, 78 feet per second; inlet-air temper-
ature range, 2390 to 2460 F; JP-4 fuel; run
time, 2 hours (ref. 14).

20

Fuel
16 - Symbols represent
V
-
various check runs
bo
AN-P-32
a
If
a
0
/

IIIIIiiII
12

0
C)

18 VY
0
Fr— RC- 18A -
/
4
ix

-
0 4 8 12 16 20
Duration of test, hr

Figure 124. - Effect of run time and fuel type on coke deposi-
tion in full-scale single combustor. Combustor operated at
cyclic conditions (ref. 28).
 •S S • • ••• ..
.. S.. • •S• • IS
••26 . •• ••
• •
, • • •S • •• •C']iETIA.L
• ••, NACABME55G28 S
• .•I.. • • • ..
•.SS'•• •• ... ..
00
U) -- C
I C)
F, F,
3 C)
'-I
0 C)
C C)
04-
o
'-I
Li) Lo ILL Li)
4'-
C)
01 1,
0
1, C))'
o C) II
-:3
I LI) C) C)
C) ,-I
4-,
C)

'-II)
0
03
4-'
CO
C)
401
o
cli
---—--------——---------------i--
CD 0
C) L3C)
:' ,iD4'
C) L04
C C)
2 • 10
1,0 o C)
1, C)
SA
ca 0)',
010 C) 1,'
C) 4.) C)1')
IC) C)
01, 41)
1,
C) 30
bO 1.
:3.0
-,
4'
41 I .00
C) ,-4 4.' C)
1, C) C
)'_40
—\--------- C00°')' •- 0-'
C') C)
C) 0c 0 'C)C)1,
0)
— —
I))
-0--i
IN
— •0 1,4'
C)
C)
N Lo
01.
44
C
0bOC
' C) 03I
C))' S
r0) 0 4' 1,0
-
—---------t-I
0,1'-'
0 C) 1',
.0 4' C)41,
V.
'000
.-.," ,."
M ".0
01. 0
O) C C) 1.1.
0) -
0 4C)
-4" C)I
0)
\
to
C) .-i
\
1. C '-40
0)
CO
,-40
C)
4' 0.
ID
C\J c-'Lo
0-. 0)'.
>- ON
\\
\ C)
C041)
0 0,-i
'-4
4-' -
4' 11)10))'
0
cu 0 ,-I,-i
CU P. 1-,
C)
'0
C)C)
C) 4'C)
.S1
0
N (3 C) . -
\ 0I
CO '0
0
1.0
II) 00
('4 444'
4' 03
C) C)
1. El.
0
-----4)
0 0 0 0 0 0 0 0
CL) C)) 0 10 CL) CL)
4'
'odap )'31°D
 ••. •.. . S I S. SI
.• • i• . ... ..
NACA RN E55G28 C(1tITIt.L •I •
••• • S S I .5 • 15 I '27
05 SI. SI 5.5 S S IS •• 5 . ... ..
0
0) II)
LI)
H 'a
4' 4'
a) 0
'
a) ..-4
0.
o a, a).
44 It..'
LI) 4,0)
0 a).
0 0) ,-1
4-' C
a) HO)
0 a,
a0 o ..d
o
0. a)0
ILI 4-).0
>0
II) a) ,-40J
0
LI) a)
a) .0)1)
a) a)E
a -4
-\----
' a,
o a) OC
-D 00
.-1 a)
1.
44 LI)
LI) a)N
E -4
N o 0
0 CO
_-1
0 CO
0
N • . .-4
a)
\ 4-' a)
4-'
C
.0 H
4-)
_a)
4-' a) a)LI
' C
OC N 0
C LI.'
- LI,
0.
Go
ala)
4) (1 0) a),
Y 4.) 0) 0.a)
a) 0 0
o W 0
4-'
'a
a)
a,- a) 4.'..
F4 a) 0a,
D
-- •I' '.-)
0 Ha) 43 0
a)
> 0) a,.0
C LI-,0-IId ;4
0) ,.0 pa)
---\ 440.
------\---- - a)
4-' ,-4 0)
LI- Ha)
o
' C CO
a-.0 0 .-'C
0
1): C •-0
a) 0010.
4.' .CL))
.'-
----- 44
0
a) Cr-
.-ILI)
Vo 0
0 '
a)o,90
0O
H O
a) ,-p
- .a-,
o a,
o a) .-I
4.' a)
0)0
• ELI-,
) 0 a)
a a) H.*
'I
'-4 1)0
____Th C) 00
C 440
\- 0
0 'a,-4
-------------4 --i----
0
-00

44.
I 'aa)
01,
II) 4-)
C') a,a)
_-) a)
(D 00.
C. CE
0 Ca)
C0 tIa
I' -I
0 0 0 0 0 0 0 0 0),
0 OD '.0 ('4
H H
2 '-'aodap 93103
 •SS I
S. ••• • • S• S 55 •• S S I
SS I SI •S
S S 5• • I S II•
t3t S.
• •
S SI I S
, • • ,
• I
•II
I S
• OND1I•,EAL NACARME55G28
so goo 0 0 •
0
U)
K)
EO
__IIIII1 :
:llt
CH

—r
10
4- 4 El t o -14
Ol 0
- - - ___ - O 0a no
;4). 0
E'w
oc=1-
c-
C)
c)
o
cla
o 0 0
o
c-i
o i .' 0
+)
O
C)
P4-' a)
c_i
Ha)
3)
a)IK) CD
d+'H C'J
H
OH
U0 a)
H -'
H • 0
D a)1-4 .-4
H .i
, p a
-oc')N0
C) 0 0 0 0 0
0
c'.1
0 a) CD K) C')
H
. Ts odp raj -jj' jo 4ua3aad 'c-odrj
 •
•a.a :0,
• • •• • a . S t o . I.. . • ..;
•
. ... ..
•
• NACA RM E55G28 CO']EIA •. I•..
• .
•5 SI. •••5•• • • .5 .. I • •
. . .. .
• • .
• ., -
. .
CAp
I .o
%. :

60

40

a
4)
•1I
0p4

00
2

1
.8' I I I I I I I I
150 200 250 300 350 400 450
I
NACA K factor

(b) In full-scale single tubular combustor.

Combustor-inlet air total pressure, 53.9 inches
of mercury absolute; inlet-air temperature, 2710
F; fuel flow, 127.0 pounds per hour; fuel-air
ratio, 0.0123; run time, 4 hours (ref. 1).

Figure 126. - Continued. Effect of NACA K factor

on coke deposition of several fuels.

p.
 • . S •
. ... S S• S •S 55 • • S • S •
S. • S S • S S S • S
. • . ..
c •• • •• • . S •
2r ,
S. •••
• •
•
• • •
• • S.
, •• • • ••..
... S.
O1AL NACA EM E55G28

100
80 I-.
60
40 o Full-scale single.
combustor, average
of two or more
20
runs
, I o Full-scale engine,
10 _____ _____ _____ _____ • average per
8 combustor
6
0 4
C.)

150 250 350

NACA K factor

(c) In full-scale single combustor and in full-scale

engine. Single combustor-inlet total pressure, 53.9
inches of mercury absolute; inlet-air temperature,
2710 F; fuel flow, 127.0 pounds per hour; fuel-air
ratio, 0.0123. Full-scale engine cyclic running
schedule, 15 minutes at take-off speed and 5 minutes
at idle speed; total run time, about 50 hours (ref.
2).

Figure 126. - Concluded. Effect of NACA K factor on

coke deposition of several fuels.
 • . .. .. •• ••s , •• •
5 • SS
I'IACARME55G28 CONIEIA1 S S •
••• :.. . . S S S • 5

24

20

16

0
a12

0 5 10 15 20 25 30 35 40 45 50
Smoking tendency, 320/h
Figure 127. - Effect of smoking tendency on coke deposition of
several fuels in full-scale single tubular combustor.
Combustor-inlet total pressure, 53.9 inches mercury absolute;
inlet-air temperature, 2710 F; fuel flow, 127.0 pounds per
hour; fuel-air ratio, 0.0123; run time, 4 hours (ref. 1).