NATIONAL AIR POLLUTION CONTROL ADMINISTRATION
DEPARTMENT OF HEALTH EDUCATION AND WELFARE
FINAL REPORT UNDER CONTRACT CPA 22-69-138
JUNE 1970
STUDY OF CONTINUOUS FLOW
COMBUSTION SYSTEMS FOR EXTERNAL
COMBUSTION VEHICLE POWERPLANTS
FOR
THE DIVISION OF MOTOR VEHICLES
RESEARCH AND DEVELOPMENT
C. V. BURKLAND
W. B. LEE
G. BAHN
R. CARLSON
-------
VAX NUM. C*II'OINI4
TABLE OF CONTENTS
SECTION TITLE PAGE
I. INTRODUCTION 1
II. SUMMARY AND CONCLUSIONS if
III. RECOMMENDATIONS 8
IV. ANALYSIS 9
A. Low Emission Combustion 9
1. Introduction 9
2. Fuel Preparation 10
3. Burning as a Chemical Process 12
B. Kinetic Results 12
V. EXPERIMENTAL EQUIPMENT 15
A. Combustion Apparatus 15
1. Air Supply 15
2. Fuel Supply 16
3- Ignitor l8
k. Combustor 19
5. Heat Exchanger 19
B. Fuel Injectors 20
C. Instrumentation 21
1. Instrument List 22
2. Flow Measurement 23
a. Air Supply 23
b. Fuel Supply 2*f
c. Cooling Water 2^
3. Temperature 24
D. Gas Analysis Instrumentation 26
E. Component Checkout 28
F. Operating Procedures 3>O
VI. TEST RESULTS '&
A. Description of Test Conditions >.-
-------
NUM. CtllfOIMIt
Table of Contents (Continued)
Section Title Pace
B. Gas Emission 32
1. Prevaporized, premixed Fuel ij+
2. SUE Burner Configuration 35
3- Pressure atomized Liquid Fuel 35
k. Methane Fuel 36
C. Particulate Emissions 36
1. Particle Collection and Analysis 35
2. Sampling System 37
3. Data Collection and Counting Technique 38
k. Data Evaluation 4l
D. Heat Exchanger Interface /^6
VII. DESIGN CRITERIA 50
A. Experimental Results 50
1. Steady State Emission Criteria 50
a. Temperature 50
b. Residence Time 50
c. Injector Characteristics 51
d. Turbulent Mixing 51
e. Air Addition 51
f. Quenching 52
g. Burning Characteristics 52
2. Other Emission Criteria 52
a. Fuel Modulation 53
b. Startup and Shutdown 53
3- Summary of Design Criteria 5^
VIII. REFERENCES . 57
APPENDICES
A. Combustion Generated Air Pollutants
B. Fuel Selection
C. Test Fuels Specifications
D. Data Summary
E. Data Analysis
-------
LIST OF ILLUSTRATIONS
FIGURE TITLE PAGE
1 Continuous Flow Combustion System for 59
External Combustion Engine
2 Values From the Steady-State Solution for 60
Hydrocarbon Combustion
3 Analytical Model 6^
k Instantaneous Temperature of Recirculating £^
Hydrocarbon/Air Flames
5 Performance of Primary Combustion Zone 5^
6 Afterburner Calculations, Diluent Air 5/1
at 2700 R
7 Afterburning Calculations, Diluent Air 55
at 2*f56°R
8 Combustion Test Rig Schematic 55
9 Air Supply Schematic 67
10 Fuel Supply Schematic 68
11 Combustor and Igniter Schematic gg
12 Burner Assembly - External Combustor 70
13 Combustor Assembly 7^
l^t Combustor Control Area 72
15 Fuel Manifold (SUE Burner Configuration) 7?
16 Gas Analysis Instrumentation Schematic 7^
17 Ideal Gas Temperature Rise 75
18 Typical Pressure Data . 76
19 Typical Wall Temperature Data 77
20 Effect of Overall Equivalence Ratio 78
Vaporized Premixed Injection
21 Variation of Emissions with Heat Release 79
Vaporized Premixed Injection, 0p = 0.8
111
-------
MUM,
List of Illustrations (Continued)
Figure Title Page
22 Variation of Emissions with Heat Release 80
Vaporized Premixed Injection, 0p = 1.0
23 Variation of Emissions with Heat Release 8l
Vaporized Premixed Injection, 0p = 1.2
2k Effect of Overall Equivalence Ratio 82
SUE Burner Configuration
25 Effect of Primary Equivalence Ratio 83
SUE Burner Configuration
26 Variation of Emissions with Heat Release 84
SUE Burner Configuration, 0p = 1.0
27 Variation of Emissions with Heat Release 85
SUE Burner Configuration, 0p = 1.2
28 Effect of Primary Equivalence Ratio 86
Liquid Injection
29 Effect of Overall Equivalence Ratio 87
Liquid Injection
30 Variation of Emissions with Heat Release 88
Liquid Injection, Kerosene
31 Variation of Emissions with Heat Release 89
Liquid Injection, TMH
32 Effect of Overall Equivalence Ratio 90
Methane Fuel
33 Variation of Emissions with Heat Release c,]
Methane Fuel
J)k Particle Collection Schematic <,yf-
35 Particle Collection System 93
36 Particulate Sample - External Combustor y/f
Run 11, Condition 3
37 Particulate Sample - External Combustor .^
Run 14, Condition 8
38 Particulate Sample - External Combustor . <£
Run 14, Condition 8
IV
-------
an
V«M Hurt.
List of Illustrations (Continued)
Figure Title Page
39 Particulate Sample - External Combustor 97
Run Ik, Condition 8
^0 Particulate Sample - External Combustor 98
Run 19, Condition 1
*fl Particulate Sample - External Combustor 99
Run 19» Condition 1
*f2 Particulate Sample - Internal Combustion Engine 100
*O Particulate Sample - Internal Combustion Engine 101
^ Typical Gas Temperature Data 102
^5 Measured Fuel Injector Flow Characteristics 103
B-l Distillation Curves - Hydrocarbon Fuels g_3
E-l Summary of Heat Exchanger Performance E-15
E-2 Equivalence Ratio Based on Carbon Dioxide E-l6
E-3 Conversion Factor to Account for Water Vapor E-17
in Exhaust Gases
E-k Conversion of Gas Emission Data to Unit Mass E-l3
of Fuel: TMH
E-5 Conversion of Gas Emission Data to Unit Mass E-19
of Fuel: Kerosene
E-6 Conversion of Gas Emission Data to Unit Mass E-20
of Fuel: Methane
-------
//ar
VAN MUM. CAU'OtMM
LIST OF TABLES
TABLE TITLE PAGE
I Test Variables 4
II Mass Emission Comparison 5
III Test Summary 53
IV Summary of Particulate Data if2
V Typical Particulate Data 45
VI Heat Exchanger Characteristics I+Q
VII Bankine Engine Boiler Characteristics 49
VIII Recommended Design Criteria for Low 55
Emission Burners
VI
-------
VAN NUTS. C«llfOtNU
I. INTRODUCTION
This report presents the results of a study of continuous flow
combustion systems for external combustion vehicle powerplants conducted by
The Marquardt Company for the Division of Motor Vehicles, Research and
Development, National Air Pollution Control Administration. Analytical
and experimental investigations were carried out in the period June 18, 1969
to June 18, 1970. The particle sampling and analysis reported herein were
performed by personnel of the UCLA School of Engineering under the direction
of Professor A. F. Bush.
The Motor Vehicles Division is conducting a program to develop external
combustion, Rankine Cycle engine technology. The external combustion engine
is intended for eventual use in a passenger-type automobile that would have a
Ifevel of air pollutant emission from the exhaust that is significantly lower
than present internal combustion engines. The external combustion Rankine
engine requires a combustion system that provides a modulated source of thermal
energy.
In a continuous flow combustion system (shown schematically in Figure
l), a fuel such as kerosene is burned at ambient pressure with air supplied
from a blower. The high temperature combustion products are directed through a
vapor generator (boiler), giving up heat, and are then exhausted to the atmos-
phere. In the external combustion engine cycle the vapor from the boiler is
expanded in a reciprocating engine or turbine to produce shaft power, condensed
and returned to the boiler as a liquid.
The purpose of the study was to develop design criteria for the
combustor of such a continuous flow combustion system that had very low emission
of air pollutants. The air pollutants of interest were carbon monoxide, unburn-
ed hydrocarbons, oxides of nitrogen and particulate matter. In addition to
low emissions, it was desired that the combustion system have a high heat
release per unit volume and be capable of being modulated over a range ci' heat
releases corresponding to engine power settings from idle to full throttle.
The specific objectives of- the study were to analytically -jnd
experimentally examine the external combustor variables that were considered
to have a major effect on emissions of air pollutants. These included funl
type, fuel preparation prior to injection, fuel-air ratios, fuel-air mixing,
combustor staging and residence time of the gases in the combustor.
*
The nomenclature used commonly throughout this report is listed
below. Other symbols and abbreviations are defined in the text as they
occur. ' .
-------
//ilarquardt .,
^ lamitmnim
CMI'OtMIA
Properties
0~ . , .. Fuel-Air Ratio
- Equivalence ratio = y. . , . .=:.. . .
Stoichiometric Fuel-Air Ratio
0p - Primary equivalence ratio-occurring in
initial or primary burning zone
0O - Overall equivalence ratio-occurring after
secondary air addition and afterburning
CH - Total hydrocarbons in parts per million
of gas volume
CO - Carbon monoxide in parts per million
of gas volume
CO - Carbon dioxide in volume percent
NO - Oxides of nitrogen in parts per million
of gas volume
NO - Nitric oxide in parts per million of
gas volume (constituting the majority
of the nitrogen oxide compounds in the
combustion gas)
P - Pressure
4P - Pressure loss
T - Temperature
t - Time
Fuels
CH, - Methane
JET A - Aviation turbine fuel, kerosene
TMH - 2,2,5-trimethylhexane
-------
V4M MUM. CAII'OINM
Burner Configuration
IL - Pressure atomizing fuel injection
(high fuel pressure)
IL - Pressure atomizing fuel injection
(low fuel pressure)
IV - Vaporizing, premixed fuel injection
IV - SUE burner configuration using multiple slot nozzles
-------
vtH HUtt.
II. SUMMARY AND CONCLUSIONS
Chemical kinetic studies were employed to better understand how and at
what rate air pollutants are formed in an external combustion process. With
this background, an experimental combustion test rig employing a recirculating
step, staged burner was built. Tests were conducted with various liquid and
vaporized liquid fuel injectors using aviation turbine fuel (Jet A) and
2,2,5-trimethylhexane. A wide range of fuel-air ratios were examined by
individually controlling primary and secondary air flow rates. Fuel flows were
varied from a maximum corresponding to a heat release of 500,000 BTU/hr to
1/30 of this value. Two runs were also made ueing gaseous methane fuel.
The range of test variables is presented in Table I below.
TABLE I
TEST VARIABLES
FUELS: Kerosene, 2,2,5-trimethylhexane, methane
INJECTOR CONFIGURATIONS: Pressure atomizing, vaporizing,
vaporizing premixed
FUEL FLOWS: 0.15 to 4.5 gallons per hour
AIR FLOWS: J>.k to 103 standard cubic feet per
minute
FUEL EQUIVALENCE RATIO:
PRIMARY: 0p = 0.53 to 1-59
OVERALL: 0 = 0.40 to 0.84
NUMBER OF TEST CONDITIONS: IkO
CUMULATIVE COMBUSTION TIME: 22 hours
The tests demonstrated that gaseous and particulate emissions less thar. those
established as the 1980 Federal Research goals can be achieved simultaneously
in a high heat release, low pressure drop, burner configuration. The emission
data measured at steady state conditions is compared to current and future
emission goals for automobiles below.
-------
//flarquardt
-^ laMHHimm
VAN NUTS.
TABLE II
MASS EMISSION COMPARISON
Grams/Gallon of Fuel
Air
Pollutant
CO
CH
NOX
Particulates
1970 (1)
Federal
Standards
230
22
None
None
California (l)
"Low Bnission
Vehicle"
110
5
7.5
None.
1980 (1)
Research
Goals
47-0
3.0
4.0
0.3
Continuous Flow (2)
Combustion System Data
5.6
0.2 (as hexane)
3-1 (as NO)
0.04 (as carbon)
(l) Assuming 10 miles/gallon average fuel consumption
(2) Measured at steady state conditions, averaged over 3^1 turndown
The major conclusions from the tests regarding low emission burner
design criteria are:
Air Addition
It was shown by analysis and experiments that a fuel rich reaction
followed by air introduction and afterburning produce the minimum NO. It
was possible to burn at a higher temperature and avoid a rise in NO formation
by staging the air addition such as to produce a deficiency of oxygen in the
high temperature primary burning zone. Introduction of the secondary air
produced a fairly rapid completion of burning with the CO reaching equilibrium
values.
Temperature
The temperature of the reaction, both in the primary arid :;«r;ondary
burning zones, has a major effect on the formation and destruction of the air
pollutants. Rich burning(primary fuel equivalence ratio, 0p = 1.2J, followed
by air addition to bring the overall fuel equivalence ratio between 0_= 0.6 and
0_= 0.8 gave the optimum results. Gas temperatures in this range oi fuel-air
-------
ratios are also desirable to maximize the thermal efficiency of the vapor
generator and keep the air pumping requirement to a minimum. From the
analysis, a slight preheat of the inlet air appears desirable to minimize
unburned hydrocarbons and more readily oxidize the carbon monoxide.
Residence Time
Residence time requirements are extremely important and it was shown
there is a distinct trade-off in the afterburner between minimizing NO
formation and increasing CO oxidation. The kinetic calculations also indicated
that the primary zone in the experimental test rig could be reduced, as the
reaction was calculated to be complete in considerably less than the full
length. The afterburner section was shown to have two conflicting requirements:
(l) with less time allowed for afterburning, the NO emissions showed a sharp
drop, (2) however, with less time available, the oxidation of the CO to CO
could not be completed. This was confirmed by the experimental results where
NO emissions continued to drop with higher flows through the combustor, whereas
the CO stayed about the same up to a point where the time was insufficient to
complete the oxidation step and the values climbed sharply.
Injector Characteristics
The method of fuel injection and fuel preparation prior to injection
had a significant effect on the steady state emissions. The use of vaporized
fuel produced lower emissions at higher heat release rates. The injection of
liquid kerosene or TMH through pressure atomizing nozzles produced low emissions
up through one-half of design flow. The time required for vaporization of the
fuel droplet was apparently of sufficient magnitude to significantly reduce
the time available for the complete burning of the fuel. Carbureting fuel
injection where the vaporized fuel and air were premixed produced the best
results.
Mixing
The burner used in the tests produced a strong rccircu J .-jtiori /.one;
that aided significantly in gas phase mixing. The inlet velocities and v.-lonitiet
through the primary zone and in the afterburner were sufficiently high to produce-
turbulent flow conditions. The secondary air inlet was sized to produce
vigorous mixing boundaries. The completeness of burning is enhanced by strong
mixing of the fuel and air that avoids pockets of over-rich or over-lean
mixtures. This is especially important where the fuel is not premixed with
the air. In the SUE burner configuration, the injection of the fuel in the
corner of the step at right angles to the high velocity air jet provided
the best possible mixing.
-------
laiRHXHTHW
Quenching
Quenching can occur: (l) locally from cold walls, (2) from heat
extractions from the gases and (3) by introduction of cold air. To prevent
quenching and creation of unburned fuel, the walls of the burner were always
operated hot (500 F and higher). Heat transfer away from the gas was limited
to radiation to a water cooled outer chamber and by convection to the slow
moving secondary air. For a typical run at 100 percent design loads, the
heat transferred away from the burner was less than 10 percent. This produced
a reduction in the actual gas temperature by approximately 150 F. Downstream
air addition was very helpful in completing the combustion reactions, but the
amount and method of introduction are critical with respect to freezing the CO
composition by quenching.
Fuels
Very little differences in emission characteristics were noted between
a commercial grade of kerosene and a pure hydrocarbon, 2,2,5~trimethylhexane
having comparable molecular weights. Combustion tests made with methane and
computer runs using pyrolized fuel fragments strongly indicate that hydrocarbon
fuel blended with lighter fractions than kerosene types could have significantly
lower emissions (including oxides of nitrogen) if used in an external combustion
engine.
Transients
Starting and stopping emission transients were very low with the closely
coupled fuel system and low thermal mass burner used in the experiments. More
significant were the large changes in emissions observed at off-design fuel-air
ratios that occurred in moving from one heat release condition to the n&x.-,.
A control system for the combustion system having rapid response and mair.tair.-
ing narrow fuel-air limits is needed.
-------
V4H NUYl. CtUIOtNIA
III. RECOMMENDATIONS
It is recommended that further development work for low emission,
external combustion systems be directed towards the following goals:
1. Design, fabricate, and test a prototype burner baced on the
experimental design criteria that uses kerosene and hae the full
range of required transient response.
2. Develop an advanced control system for continuous flow combustion
systems having wide fuel modulation limits, narrow fuel-air
limits and rapid transient response.
J>, Establish the significance of low emissions (especially NO^)
associated with lighter fuel fractions and evaluate the
logistics (availability, costs, etc.) of lighter fuels.
-------
IV. ANALYSIS
A. Low Bnission Combustion
1. Introduction
The combustion process under consideration here is the constant-
pressure process operating at steady-state conditions. This process has certain
basic characteristics whether applied to oil burners, jet engines, or the
external combustion system for automotive propulsion. Perhaps the most out-
standing characteristics is the maximum temperature, for this bounds the combustion
environment to which the fuel is subjected and thus largely dictates details
of burning. Another very important characteristic is the sequence of physical
and chemical steps involved in the process of burning. The efficient burning
of a liquid hydrocarbon fuel commences with control of the processes of
injection, atomization, dispersion, and vaporization of the fuel. With the
creation of a combustible mixture in air, a flame may be established in a number
of ways: (l) by continuing auto-ignition; (2) by recirculating ignition viia
hot gases from fuel already burned; (3) by continuous operation of an ignition
source, such as a pilot flame. The second of these is the customary one for
combustion in continuous flow, utilizing one or more flameholding elements. At
very low air flow velocities, such as those induced by natural convection, the
fuel injector itself may effectively serve as the flameholder.
A flame stabilized by a flameholder may operate with a
homogeneous gaseous mixture, conditioned by complete prior vaporization of
liquid fuel droplets, or with a dispersion in the air of suitably small drop-
lets essentially unyaporized, or with combined vapor/liquid fuel conditions as
the combustible mixture reaches the flameholder; this will depend upon such
parameters as temperature, time, and fuel volatility. At temperatures of the
order of flame temperatures, large hydrocarbon molecules crack to smaller
fragments via free radical chain mechanisms before they burn. The oxidation
steps also involve free radical mechanisms as the smaller fragments, whether
themselves stable molecules or free radicals, are converted to HO and CO .
The participation of HO fragments in a multiplicity of shuffling reactions
is extremely important to the ready progress of combustion.
Depending upon the initial conditions of pressure, temperature,
and fuel-air ratio, the final equilibrium products of combustion will reflect
partial completion of reaction versus partial dissociation. In the context
of pollutants, the only equilibrium dissociation products of consequence are
CO and NO. As an intermediate in the combustion process, CO cannot be
avoided as it is ordained by equilibrium considerations. However with ft/.cesfi
air, the CO .can be oxidized to CO and the concentration reduced to v<-.-ry low
values. On the other hand, the creation of NO, as an adverse cirf;um:;c.-irjf:e,
involves kinetics considerations apart from the direct combustion process, and
its final concentration need not necessarily be as equilibrium would dictate.
-------
V4N NUTS. CIll'OINM
Unburned hydrocarbons are primarily attributable to imperfect
fuel distribution, since the equilibrium limits for wide ranges of stoichiometry
are only HO and CO versus H and CO. Particulate carbon arises from too rich
a flame, with fuel cracking and reforming at too high a temperature, for too
long a time, before exposure to oxygen. More complete discussion of specific
combustion-generated air pollutants of interest in this program is presented in
Appendix A.
2. Fuel Preparation
The injection of the fuel and subsequent mixing ie critical to
the process of complete combustion. For good burning, liquid fuels must be
broken up into small droplets (atomized) or changed to the gas phase (vaporized).
Atomization can be accomplished by forcing the fuel through a small orifice under
pressure such as done in conventional oil burners, by air atomizing nozzles, by
ultrasonics, etc. The injector design will dictate the penetration of the fuel
jet into the air stream, and the characteristics of the spray that is created
with respect to both spray pattern and droplet size distribution. The spray
formation from the liquid jet depends upon the interplay between surface, viscous,
and inertial forces. Ultimately a minimum attainable droplet size is imposed
by a practical minimum size of fuel jet, whether this be a "solid" jet, a hollow
cone, or something else in cross section, because droplet size is some direct
function of jet size. In practical terms, the minimum attainable mean droplet
size is of the order of 20 microns, although with strong inertial forces of the
jet and the air stream, somewhat smaller sizes are possible.
After injection of the fuel and formation of a spray, droplet
vaporization and mixing become important. Vaporization prior to entering tho
flame region is not necessarily essential, because small enough dropl^te r:-->n
burn sufficiently rapidly to sustain a flame even if there is no advcJ/ice rfrlf.-a.jt.-
of vapor to feed the flame region. However, the requisite droplet si/.e for /juch
operation is evidently appreciably less than 20 microns, and thun connidcr.Mb, <.
vaporization ahead of the flame region ought to be anticipated arid provi-K-d
for in a practical system. This requirement is similar to that for vo j aU I ix..-jt;ior,
of gasoline in a carburetor in an internal combustion engine.
Relative to vaporization of hydrocarbons, the i.nflijf;ri';<: of
relative velocity between droplets and air stream may not be significant, b'.";;jusfc
small droplets are soon accelerated to the air stream velocity. An influence
of the close proximity of adjacent droplets in a cloud exists, but this
normally relates to a burning cloud with significant radiation and/or to a cloud
of large droplets with significant natural convection currents. Aside from
these factors, either vaporization alone or the vaporization-with burning of a
droplet is described by a simple law, as follows: A= d (D2)/dt.
10
-------
VAN MUM.
The value of the vaporization constant, A, is defined by the
operating conditions, but not by the droplet size. When the environmental
temperature is very high relative to the droplet boiling temperature, the boiling
condition is essentially established and the rate of heat transfer is controlling;
i.e., the value of A is dictated only by heat transfer considerations. When
the environmental temperature is low relative to the droplet boiling temperature,
the diffusivity of the fuel vapor away from the droplet surface is controlling
upon A High speed computer programs designed to solve the droplet vaporization
problem exist. These differ in the interative steps required for solution.
The faster program employs as the independent variable the degree of depression
of the droplet surface temperature below the boiling temperature. The dependent
variables solved for are requisite gas temperature and initial vaporization rate
(zero vapor component in the atmosphere). The other program employs as the
independent variable the system enthalpy. It iterates first to define the
compatible gas temperature and droplet surface temperature and then proceeds to
compute the complete progress of vaporization. The principal output of this
program is the time to achieve complete vaporization, with intermediate ret^s
and temperatures as auxiliary information.
The proper distribution of fuel is important for several
reasons. First of all, with really gross maldistribution it may be impossible
to maintain a flame at a flameholder. Flameholder operating limits of fuel-air
ratio correspond to restrictions of the simple flammability limits of the fuel
under consideration. The restrictions are imposed by factors of limited time
for reaction, heat loss, etc. Clearly, the fuel distribution must be compatible
with the operating limits of a flameholder. Beyond this, too lean a flame
locally may lead to quenching, and thus to unburned hydrocarbons, and too
rich a flame may lead to pyrolysis, and thus to particulate carbon. Therefore,
it is important that no "pockets" either too rich or too lean persist through
the flame region. Finally, nonuniform fuel distribution preserved through
the combustion process likely will result in performance losses.
The fuel distribution is continuous-flow combustion system is
partly dependent upon the initial spray pattern created at injection. However,
it is largely dependent upon the natural, turbulent flow of the air stream as
a statistical mixing process. The application of fuel distribution correlations
for such flow has been exploited to enhance both flame stabilization and
combustion efficiency. In conjunction with judicious location of fuol injection
elements, assurance can be provided of not only near-stoichiometric burning of
the fuel (for optimum combustion) but also nearly uniform composition of the
exhaust products.
11
-------
VAN NUM. CAIIFOINIA
3- Burning as a Chemical Process
Analysis of the chemical kinetics of hydrocarbon combustion
has been performed and a representative result from homogeneous mixture cal-
culations for the external combustion program is shown in Figure 2. Note first
the temperature history for the case considered. The fuel feed is pyrolyzed
to fragments in approximately 10~^ second, and this mildly endothermic process
causes a slight temperature decrease over this period. Oxidative recovery to
the assigned starting temperature occurs at 2-/2 x 10" 5 second; i.e., .just the
recovery takes I-Yz times as long as the pyrolysis. Appreciable oxidation and
heat release take very much longer, and the histories of concentrations of HO,
CO , and 0 verify this on the figure, changing noticeably only at the longer
times.
It is important that the situation represented by the figure
is a homogeneous one. Under these conditions, where competitive secondary
degradation of the fuel to particulate carbon was allowed for, this did not
eventuate, and the pyrolysis fragments proceeded simply to burn. On the other
hand, if a substantial period of time at high temperature were afforded to the
fuel in the absence of oxygen, secondary pyrolysis steps to produce particulate
carbon could be expected, and such has indeed been calculated for conditions
without oxygen present. There thus is a requirement for rapid vaporization of
any fuel droplets and rapid admixture of the vapors out into the surrounding
air. The ease or rapidity of access to the requisite air for combustion is
dependent upon the droplet size because the effective boundary layer thickness
of fuel vapor around the droplet is proportional to the size. Thus, it is
important in forestalling secondary pyrolysis to work with small droplet sizes
(if not vaporized fuel).
The reaction of primary pyrolysis fragments, such as CH , CH, ,
C^H , etc. to the ultimate products, CO and HO, takes place one discrete
chemical step at a time, and these integrate by multiple parallel paths to a
developed temperature-time profile. Possible limits to completion of reaction
are: (l) equilibrium limits of CO concentration, as noted earlier; (2)
relatively slow scavenging of last traces of intermediates, cr.pecin 1 1 y ot r\<:-->r-
stoichiometric conditions where the 0 concentration also fa I. "le t,o o tro'j
-------
. CAU'OINIA
'IXIKHMATKW '
mixture and of the recirculating products were then taken as inputs for
separate kinetics calculations, ard with these it was established that,
regardless of the proportions (but considering that the primary zone was not
overloaded so as to affect recirculation adversely), the residence time in the
primary burning zone was more than adequate to prepare the fuel very well for
the secondary reaction zone. Indeed, after only 1 millisecond the preparation
appeared to be quite favorable, and one of the mixtures produced after 1
millisecond was employed in a subsequent kinetic calculation of afterburning.
The performance of the primary zone was assessed in a number
of ways. Figure 4 shows temperature-time profiles for three different
proportions of fresh mixture and recirculating products. In each case reaction,
as measured by temperature, is indicated to be substantially complete in 1
millisecond. Figure 5 shows further results for one of the three mixtures, ,
giving concentrations of 0 , CO, C H and total unreacted hydrocarbons in 10
moles/mole of total gas (ppm). The disappearance of 0 is closely tied to
temperature increase, and is essentially complete after 1/2 milliseconds. The
concentration of CO at first decreases while oxygen is plentiful, and then
increases again as oxygen deficiency is felt and the reaction H + CO *>H 0 + CO
proceeds. The summation of all mole fractions of unburned hydrocarbons
initially builds up from an assigned value of 8230 ppm for CnH p fuel to a
maximum about four times as great; this is the result of pyrolysis, and very
grossly might be represented by conversion of CoH p to 2 C H, + C H, + C_H .
With further time allowed, two ensuing processes are exemplified: (l) further
pyrolysis leading to a peak somewhat later in C H concentration, and (2)
further oxidation leading to lower total hydrocarbons. The relative thermal
stability of acetylene is demonstrated by its indicated persistence. These
various concentration profiles indicate that a residence time of 1-3 milliseconds,
after recirculation mixing, would be appropriate for a homogeneous primary zone
with vaporized fuel.
In addition to the simple evidence of these figures, the rate of
disappearance of C H was compared for the primary with the secondary combustion
zone (otherwise discussed below). Thus it was ascertained that the fractional
rate of disappearance of C H at a given concentration level, was very much
slower in the primary. This comparison tended to substantiate the decision
already made, to initiate the afterburning kinetics calculations with the mixture
produced after 1 millisecond in the primary zone. All in all, from v reaction
standpoint the primary combustor thus could be shortened markedly. However,
in actually shortening it, careful attention would need to be paid to the
prevailing aerodynamics, since the record of reaction time needs to be based
upon the effective completion of recirculation and allowance for pyro.iysis
and burning to commence at this point.
For the first of the two secondary /.one calculations, the initial
temperature was taken as 2700 R to represent the condition oT more gradual
addition of secondary air to the effluent from the primary burning zone. Abrupt
13
-------
VAM MUM. ClUI'OINM
and total mixing of the secondary air would have reduced the temperature- to ;>
somewhat lower level at the outset, but this would not have been representative
of the true kinetic situation at the outset of afterburning. On the other hand,
the assigned temperature of 2?00 R could be achieved overall within the burning
cycle by regenerative heating of the secondary air. Figure 6 presents results
of the afterburning calculation, and the point of the discussion immediately
preceding is an explanation of a final indicated temperature on the figure
somewhat higher than the equilibrium flame temperature for 0_ = 0.65 if calculat-
ed without any regenerative heating of the secondary airflow.
Shown on Figure 6 are the profiles of temperature, CO concentration,
NO concentration, and concentration of unburned hydrocarbons. Burning is
essentially complete after about 20 milliseconds, the CO concentration attain-
ing near identity with the equilibrium value (for the prevailing temperature)
at this time. At the temperature level achieved, where the concentration of
atomic oxygen has stabilized, the production of NO proceeds in a steady fashion
(with false appearance of acceleration on a semilog plot) because of the reaction
chain represented by 0 + N = N + NO
* \ £-
i
N + 0- = 0 + NO
J I
Another secondary calculation was performed, starting this time
with a temperature of 2^56 R so as to correspond with no regenerative heating
of the secondary air. Results are presented on Figure 7- Unburned hydrocarbons
endured as traces a little longer than previously, but again the test of
effective completeness of combustion was the decrease of CO concentration nearly
to an equilibrium value, and again about 20 milliseconds was indicated to be
sufficient for this, with a CO level of about 200 ppm resulting. (The equilibrium
value for the point in question, obtained by continuing the kinetics calculation
until equilibrium was finally achieved, was 119 ppm.) By ending the reaction
at about 3300 R, rather than at about 3500 R according to Figure 6, the rate of
production of NO was decreased by a factor of about ten, and the resulting
(homogeneous kinetics) value of NO was negligible. This indicates that NO
observed in practice can be attributed to temperature nonuniformities, keeping
in mind that NO, once formed, is quite stable.
Figure 7 implies a desirability to shorten the afterburning time
consistent with attainment of nearly the equilibrium concentration oi.' CO, which
itself is a function of fuel equivalence ratio (and regenerative heating,
if any). This would lead to the most compact hardware and to minimal production
of NO. It is in the area of predicting suitable design modifications which
will inherently foreshorten allowable reaction times that future kinetics
calculations can be helpful, especially as off-design operation is taken into
account.
-------
NUrS.
V. EXPERIMENTAL EQUIPMENT
A. Combustion Apparatus
The combustion apparatus consisted of five (5) component sections.
These were: (l) air supply, (2) fuel supply, (3) igniter, (*f) combustor and
(5) heat exchanger. A block diagram layout of the arrangement is shown in
Figure 8. Each of these sections is described below:
1. Air Supply
A schematic of the air supply system is shown in Figure 9-
Facility air, at a supply pressure of approximately 95 psig, was used as air
source to the combustor. Choked venturi were used to measure air flows. This
permitted an accurately measured flow of air to be delivered independent of
minor pressure fluctuations in the combustor.
The facility air was regulated to a suitable pressure (usually
80 psig) by an air actuated loader. The air then entered a swirl tank where
entrained water droplets were separated by inertial action. This was followed
by a bank of two large drying tubes in parallel, to complete the water removal
operation. It should be noted that it was neither necessary nor desirable to
obtain complete drying of the air, but rather to remove entrained water droplets
and to insure that the relative humidity was sufficiently low to avoid condensation
in the venturi inlet. It was also of interest at times to know the approximate
humidity of the air entering the combustor. Accordingly, a bleed line was
tapped off the air flow downstream of the main air venturi. Flow through this
line passed over two exposed tip chromel/alumel thermocouples. A moistened wick
was placed around the second of these. The two, therefore, provided a wet and
dry bulb temperature measurement of the supply air. A rotameter in this bleed
line was used to verify that the air velocity over the wet bulb was approximately
1.25 lbm/ft^ sec, which provides an unbiased estimate of wet bulb temperature.
Standard humidity charts were then used to determine the percentage humidity of
the air supplied to the combustor. This bleed flow was activated only during
the period of wet bulb measurement.
Tests of the air supply humidity control system at rnaximurn
air flow of 0.138 Ib/sec were performed. These tests demonstrated a reading of
50 percent humidity after 2 hours 20 minutes flow of saturated air through the
same bed of desiccant.
The air flow was divided, as desired, between primary air
(venturi A) and secondary air (venturi B) by means of valves V, and V .
Readings of P°, and P ,, determined the corresponding mass flow rate.-.'l Valve
Vp, was provided for use in certain low flow rate applications.
15
-------
V«M NU»1.
Pressure adjustment to that of the combustor occurred in the
diverging section of each venturi. The pressure of the primary air stream
(referenced to ambient) was read by manometer in inches of mercury upstream
of the combustor. The pressure of the ;primary air was also referenced against
the secondary air, and against the combustor exit pressure, in two other
manometers. The latter two manometers were read in inches of water.
2. Fuel Supply
Three hydrocarbon fuels were tested in this program as
designated below:
Fuel Grade Purity Source
2,2,5-Trimethylhexane Tech. 95 mole % Phillips Petroleum Co.
Kerosene (Jet A Aeroshell ASTM D1555 Shell Development Co.
Turbine Fuel) Turbine
Fuel 6^0
Methane Tech. 97 mole % Air Products Co.
The 2,2,5trimethylhexane was used as a "base" fuel in that
its properties (particularly the boiling point) are clearly defined, and it is
representative of the molecular class of fuels of prime interest to this program.
Aeroshell Turbine Fuel 6^0 is representative of kerosenes,
and was used in pressurized atomizing nozzle tests.
Methane was used as a comparative fuel in so.ne tests. Specifi-
cations for each of these fuels are given in Appendix C. All these fuels
were tested as received from the supplier.
For the liquid fuels, a nitrogen pressurized fuel syr;tern was
used. This system had the capability of delivering the hydrocarbon fuel at
design flow rate at all conditions ranging from ambient liquid to vapor with
50°F superheat. The system could deliver the fuel at pressures of ambit-nt to
1000 psig.
A schematic of the fuel supply system is shown in Figure 10.
The fuel reservoir was a stainless steel tank of 15 gallon capacity, rated at
1000 psig. The hydrocarbon fuel within this tank was pressurized from a cylinder
of commercial grade nitrogen.
16
-------
The fuel system was designed to deliver fuel in accordance
with two design modes. One calls for the introduction of the fuel into the
combustor in the form of a superheated vapor at low pressure. The other involves
injecting the fuel, as ambient temperature liquid, through a high pressure
atomizing nozzle into the combustor. In accordance with the latter mode, all
components were rated at 1000 psi at ambient temperature. For the former mode,
a 1.6 KW heater, plus a 785 Watt auxiliary heater, was capable of delivering
2,2,5-*rimethylhexane at 50 F superheat to the combustor at design flow rate.
The system could also deliver the fuel at various intermediate states between
the above two cases.
Fuel flow rate was determined by a turbine meter, whose
output was read from a digital readout device. The output of the turbine meter
was calibrated directly in terms of fuel flow rate by collecting a measured
mass of fuel for a measured time, over the flow range of interest.
For tests with methane, the gas cylinder was plumbed in down-
stream of the heating element. A precision rotameter was used to measure methane
flow rate, corrected for the existing gas pressure at the rotameter.
A key feature of the fuel supply system was the three-way
solenoid valve. In the absence of electrical power, the valve was in the bypass
mode in which no fuel entered the combustor, but was rather diverted into the
collection vessel. A water-cooled coil of copper tubing was provided in the
bypass line. This served to condense the superheated fuel vapor (when operating
in this mode) before routing it to the collection vessel.
The solenoid valve was energized by a circuit which included
a photo-conductive cell. This latter device is described in further detail
subsequently, but in essence it de-energized the solenoid valve in the absence
of flame in the combustor. The solenoid valve could not be energized to the
burner mode unless an override button was depressed on the ignition switch box.
It would not remain energized unless combustion was occurring. Thus, in the event
of loss of flame, fuel flow was automatically diverted from the burner into
the bypass mode.
It was necessary that this valve function properly under all
operating conditions including the use of hot hydrocarbon vapors, in excess of
350 F. An existing low temperature valve was reworked. This rework involved:
(a) changing the seat material to Viton A, (b) enlarging the bypass and discharge
ports, and (c) providing external cooling of the solenoid coil while not
significantly reducing the temperature of the hot hydrocarbon vapor passing
through the valve. This valve had a low closure pressure rating (~60 psi).
With some of the liquid pressure spray nozzles, however, operating pressures
of kOO psi or higher were used. For these tests, the modified valve was
replaced with a conventional, small-port solenoid rated for these pressures.
With the use of liquids, the smaller port size was not a problem. The seat
material of this valve was also replaced by Viton A.
17
-------
Him. C*H'O«NM
3- Igniter
Initial igniter component testing was performed with catalytic
igniters. Although promising, these tests indicated that a basic problem of
igniter durability remained. Accordingly, attention was directed to the use of
a high voltage spark igniter.
The high voltage electrical components from a commercial oil
burner were modified for use. This was done by building a housing shown
schematically in Figure 11, which supported the high voltage (10,000 volt)
transformer, and also served as a plenum into which the primary air supply to
the burner was directed. Specifically adapted high voltage electrodes were
prepared. These electrodes passed through the inlet section of the two-inch
diameter .connecting tubing, six inches long, between the primary air plenum
and the combustor. The fuel inlet line was brought in at the upstream end of
this connecting line. Depending on the mode of operation, either a vapor
injector (0.10-inch diameter converging nozzle) or a liquid pressure spray
nozzle was installed. These different nozzles were installed and removed through
the downstream opening of the igniter tube, without having to disturb the high
voltage electrodes. The latter extended the full length of the tube. For use
in the vapor injection mode, the nozzle throat was located approximately k
inches upstream of the arc, permitting mixing of the fuel to occur prior to
reaching the spark. For the liquid injection mode, the spray nozzle was
positioned so that the top of its spray pattern just intercepted the spark.
These conditions were verified experimentally during the preliminary component
tests. Once the desired injector-igniter geometry had been established,
injector-support fittings were prepared which permitted different spray nozzles
to be inserted, without having to readjust the electrodes. The downstream face
of the connector tube was flange-mounted to the combustor.
The photoconductive cell was mounted directly under the trans-
former and was positioned so that it would "see" axially down the length of the
combustor. The cell was a cadmium sulfide device, Honeywell C55^A. It was
designed for use with a commercial, oil fired burner. It was found necessary
to shunt the photo-conductive cell with a 4 K.Q resistor in order to obtain
satisfactory operation of the device over all the combustion conditions encounter-
ed during the program. The output of the photoconductive cell controlled the
position of the 3-way fuel solenoid valve, as described in the preceding section.
Electric current to the high voltage primary coil was controlled
by a separate 110 V switch. A primary current of 2.2 amps was required to
maintain the high voltage arc; an ac ammeter in the primary line was used to
verify, during a combustion test, that the arc was operating properly.
18
-------
NUYJ. CAll'OINM
k. Combustor
A detail of the combustor is shown in Figure 12 and a photo-
graph in Figure 13- The combustor inlet was bolted to the igniter section exit
flange. The combustor consisted of an outer water cooled jacket surrounding
the replaceable inner burner can.
The burner chamber was 5 inches in outside diameter and 3&
inches long with a double wall for water cooling. Secondary air was introduced
as desired into the chamber through the tube on the upstream end. A small
baffle was used downstream of the tube inlet to help distribute the
secondary air around the periphery of the chamber. The replaceable burner cans
consisted of a variable length hot wall section, an afterburner section for the
introduction of secondary air, and a volume downstream that could be operated
with either a hot or a cold wall.
The burner can used in all the tests of this report was of
321 stainless. The combustor wall was 0.028-inch, the wall thickness of the
conical afterburner section was 0.032.
The combustor has eight taps for measuring burner wall temper-
ature; four taps were in use on each run.
Temperature of the hot gases at the combustor exit was
indicated by a Pt/Pt-10% Rh thermocouple (encased in a Pt-6$ Rh sheath), which
passed through the front flange of the transition section. The latter section
reduced the flow area to the 2-inch diameter of the heat exchanger. The
transition section and part of the heat exchanger are shown in Figure 13- This
transition section was cooled by means of an outer water jacket. It was formed
into a relatively close l80 bend, which was AN coupled to the first heat
exchanger section. The combustor inner wall thermocouples can also be seen
in the photograph of Figure 13.
5- Heat Exchanger
The function of the heat exchanger in the experiment was to
cool the gases leaving the combustor to ^00-500 F or less, simulating actual
exhaust conditions. A sampling line to the gas analysis instruments was provided
at this point. Additional gas taps at intermediate positions in the heat
exchanger were provided for possible future use. A single tube heat exchanger,
emphasizing low cost and facility of instrumentation, was selected.
A generalized analysis was performed relating the basic factors
(L, D. t, and AP) for conditions of developing a given temperature drop for a
fixed mass flow rate. Here L = heat exchanger length, D = heat exchanger tube
I.D., t = residence time and AP = pressure drop. The residence time increases
19
-------
approximately as D " and length as D * , while A? changes as D " . Based
on a series of heat transfer calculations, it was decided to construct the
exchanger of 2-inch O.D. stainless steel tubing, 0.016-inch wall. The required
length was 33 ft, with a calculated AP of about 1.5 psi, and a total residence
time of about 130 milliseconds. The gas temperature was reduced to below 2500 F
for all conditions (essentially freezing the reaction) in less than 5 milli-
seconds. These design calculations were based on the maximum design flows.
A comparison of the design calculations with subsequent experimental data is
given in Appendix E.
The different heat exchanger sections are shown in the photo-
graphs of Figures 13 and 14. The first 13 ft of heat exchanger length was cooled
by /2-inch copper tubing, close coiled, through which cooling water flowed. The
last 20 ft of exchanger length was placed in a water cooled trough. There were
four separate cooling water flows to the heat exchanger and combustor. The
individual water flow rates and exit temperatures were measured, permitting a
determination of the amount of heat transferred and thus establishing a heat
balance. Heat balance data are given in Appendix E. Gas temperature measure-
ments were taken at four positions along the heat exchanger length.
B. Fuel Nozzles
Three different configurations of nozzles were tested in this
program. The first was a converging nozzle for vapor injection. The throat
diameter was 0.10 inches. This nozzle was mounted along the centerline of the
ignitor tube as shown in Figure 11. There was a *t inch mixing length downstream
of the nozzle throat, prior to the combustor inlet.
Appendix E presents calculations describing flow conditions in the
nozzle and ignitor section, based on experimental measurements during a
combustion test.
A second nozzle, also for vapor fuels, was a multiple- slot nozzle.
In this arrangement, fuel vapor entered a ring-shaped manifold surrounding the
ignitor tube. Leaving this manifold were six % inch O.D. tubes, spaced
symmetrically around the exterior of and parallel to the ignitor tube. These
tubes passed through the ignitor tube exit flange via Swagelok bulkhead fittings.
Each tube projected about 1/8 inch beyond the inner flange face. A detail
of this nozzle is shown in Figure 15- The end of each tube was scaled, with the
fuel exiting through a slot nozzle. The slots were arranged so that the fuel
vapor of each tube was directed toward the centerline of the combustor. The gap
in the ring manifold was required to permit mounting on the existing ignitor tube
without rework of previously attached fittings. This nozzle configuration
installed in the step burner constituted the Marquardt patented Sudden Expansion
or SUE Burner.
20
-------
VAN NUM. CIH'OIMM
The third type of nozzle was a pressure atomizing spray nozzle,
for liquid fuels. The nozzles were procured from Spraying System Co. Two
nozzles of this type were tested, bearing the designations "K LND SS 1", and
"K LND SS 1.5". The difference between the two lies in the pressure-flow rate
relationship. The '1.51 nozzle required about 380 psi to deliver full flow rate
of 2,2,5-trimethylhexane, while the 'I1 nozzle required 550 psi to deliver the
same fuel at 88$ fuel flow. The orifice diameter was 0.020 inches for both
nozzles. The pressure-flow rate characteristics of each nozzle were measured
using 2,2,5-trimethylhexane, over the full range of flow rates to be tested,
or up to a maximum of 550 psi, depending on the nozzle. The spray angle of
each nozzle is listed by the manufacturer as 70-72 for fully developed
spray. Each nozzle projects a hollow cone spray pattern.
These different injectors are designated as follows in the remainder
of this report:
Vapor nozzles:
0.10"converging nozzle IV]_
6-tube slot nozzle ly
Liquid nozzles:
J4 LND SS 1 (higher pressure) IT
%. LND SS 1.5 (lower pressure) IL
All three internal nozzles were installed and removed from the downstream
opening of the ignitor tube. It was not necessary to remove the latter- or
the high voltage electrodes during these operations.
C. Instrumentation
This section profides a listing of the instrumentation used
(exclusive of gas analysis instruments), followed by a discussion of details
of the test instrumentation.
21
-------
VAH NUM.
1. Instrumentation Lisb
Measurement
Flow. Rates
j Primary air to combustor
1 Secondary air to combustor
Fuel flow to combustor
Pressures
Primary air venturi total
Primary air venturi static
Secondary air venturi total
Secondary air venturi static
Primary air
Differential pressure,
primary vs. secondary air
Differential pressure, pri.
air vs. combustor exit gas
Fuel tank
Fuel heat exchanger exit
Fuel nozzle inlet
Type of Instrument
Venturi
turbine meter with
digital readout
Cox precision rota-
meter (for methane)
Matheson rotometer
603
Bourdon gage
Bourdon gage
Boundon gage
Boudron gage
Manometer
Manometer
Manometer
Bourdon gage
Bourdon gage
Bourdon gage
Bourdon gage
Bourdon gage
Range or type of Ptes/l Out
Venturis of following throat
diameters were used
0.07V
0.115
0.1695
0.170
0.235
0.250
0.296
0.330
5-27 Ib/hr
-k
3.5 x 10 - 0.060 Ib/sec
(std air)
1/10 and 1/30 scale tests
0-100 psig
0-100
0-100
0-200
0-12 in. I-fc gage
0-90 cm H00
C..
0-12 in. HO
0-600 psig
0-60 psig
0-600 psig
0-100 psig
0-600 psig
22
-------
Temperature
Combustor exit gases
Heat exchanger 1
Heat exchanger 2
Heat exchanger 3
Combustor wall 1
Combustor wall 2
Combustor wall 3
Combustor wall k
Fuel heat exchanger exit
Fuel heater element
Cooling water 1
Cooling water 2
Pt/Pt - 10$ Rh
Pt/Pt - 10$ Rh
Pt/Pt - 10$ Rh
chromel/alumel
Pt/Pt - 10$ Rh
Pt/Pt - 10$ Rh
Pt/Pt - 10$ Rh
Pt/Pt - 10% Rh
chromel/alumel
chromel/alumel
chromel/alumel
chromel/alumel
Potentiometer plus 0-20 mv penchart
Potentiometer plus 0-20 mv print chart
0-20 mv print chart
Potentiometer plus 0-20 mv print chart
Potentiometer plus 0-20 mv print chart
Potentiometer plus 0-20 mv pen chart
Potentiometer plus 0-20 mv print chart
Potentiometer plus 0-20 mv print chart
Potentiometer plus 0-20 mv print chart
Potentiometer plus 0-20 mv print chart
Potentiometer plus 0-20 mv print chart
Potentiometer plus 0-20 mv print chart
Cooling water 3
Cooling water if
Cooling water 5
Air supply, dry bulb
Air supply, wet bulb
chromel/alumel Potentiometer plus 0-20 mv print chart
chromel/alumel
chromel/alumel
chromel/alumel
chromel/alumel
Potentiometer
Potentiometer
Potentiometer
Potentiometer
plus
plus
0-20 mv
0-20 mv
print chart
print chart
2. Flow Measurement
a. Air Supply
Except for one condition at 1/30 scale, all air rates to
the combustor were measured by the use of Venturis operating in the choked
condition, or very close to this condition. Eight different Venturis were
used during the program. Each of these Venturis was calibrated with air, using
a Cox Precision Rotameter as standard. The latter had been calibrated by an
outside laboratory 3 months prior to initiation of this program. The
calibrations were performed under conditions of choked flow through each
venturi. It was found that for 6 of these Venturis the measured flow rate
during calibration agreed within J>% with the theoretical equation for choked i'low.
Accordingly, for these Venturis the theoretical flow equation was used in
subsequent tests. For the other two Venturis, a flow equation based on the
measured flow during the calibration was used in subsequent tests.
-------
an
V4M NUM. C«II'OINM
In each flow measurement the venturi inlet (total) and
throat (static) pressures were measured, and the pressure ratio computed and
logged with the data. This was to verify that the nozzle was in fact operating
in the choked state at each test condition. Under certain conditions, and
particularly with the 0.1695 inch venturi, the nozzle was operating slightly
unchoked (total/static ratio approximately 1.7 - 1.8). Whenever this occurred,
the flow rate was corrected by introducing the flow coefficient for a subsonic
nozzle. This correction factor was never less than about 0.985 and was usually
greater than 0.99* The main effect of this condition was that the instantaneous
flow rate through the nozzle was not independent of pressure fluctuations within
the burner. This factor also turned out to be not significant, as pressure
fluctuations within the burner were usually less than 0.1 inch of Hg, and were
considerably less than this at the desired operating conditions.
b. Fuel Supply
For operation with either 2,2,5-trimethylhexane or kerosene,
fuel flow was measured by a turbine meter (Cox Flowmeter Model LF 6-100) with
digital readout (Anadex Model CF-SC&R). The output of the latter was calibrated
directly in terms of mass flow rate of each fuel by collecting a measured amount
of fuel over a measured time interval at steady state conditions. This
procedure was repeated at several points to determine a calibration curve for
each fluid from % to full scale flow rate.
For fuel flow rates less than this, a Matheson No. 603 Rota-
meter was calibrated (by the same procedure) over a range of flow rates from
1/100 to 1/10 scale.
When methane was used as fuel, flow rate was measured by
the same rotameter used to calibrate the air Venturis. Correction was applied
for the pressure existing at the rotameter at each condition.
c. Cooling Water
A rotameter was installed in each of the four cooling
water lines. These were not calibrated, as high accuracy was not needed
in this case.
3- Temperature
Combustion temperature was indicated by a Pt/Pt - 10% Rh
exposed tip thermocouple (all test thermocouples used on this apparatus had
exposed tips). The tip of this thermocouple was approximately at the
combustor centerline, passing radially in through the inlet flange of the
transition section (which was bolted directly to the combustor outlet). The
sheath of this thermocouple was of Pt - 6% Rh, to withstand the extreme
-------
. CtlllOIHIt
conditions to which it was subjected. The sheath O.D. was 0.020 inches. The
output of this thermocouple was displayed on the dual pen recorder in view of
the operator, and (as with all but one of the test thermocouples) could also
be read out simultaneously on a potentiometer. This provided a rapid and
sensitive indication to the operator of such phenomena as ignition, shutdown,
combustion instability, and transitioning from one condition to another.
It was recognized that a simple exposed tip thermocouple would be prone to
radiation errors at the combustion gas temperatures. Installation of a
thermocouple assembly with radiation shields, would have compromised the
design of the combustor-heat exchanger interface.
The magnitude of the radiation error involved in the combustor
thermocouple is analyzed in Appendix E.
The combustor had 8 locations for wall temperature measure-
ment. Location 1 was 3/2 inches from the face of the combustor inlet flange;
there was 4 inch spacing between all of the taps. The last, location 8, was
k inches upstream of the face of the combustor exit flange. During this program,
4 wall thermocouples (all Pt/Pt ~ 10$ Rh) were in use. These were in locations
1, 2, 5 and 8. Locations 1 and 2 were utilized since they covered the region
of the combustor at which possible hot spots were anticipated. Location 5
corresponds to the region where secondary air was entering the combustor; and
thus would detect excessive temperature which might be associated with the onset
of afterburning. Location 8 was in the "dead air" region of the conical portion
of the afterburner can, where there was little or no net flow of secondary air
past the burner wall.
The output of wall thermocouple No. 2 (location 2) was displayed
on the second pen of the dual pen recorder. Thus the operator could always
detect any undesirable increase in wall temperature promptly, and take corrective
action.
Each of these wall thermocouples was installed connected to an
ohmmeter to verify that the tip was in contact with the wall. This condition
was rechecked at various times throughout the program.
Presentation of typical wall temperature data is given in
Section VI-A. No excessively high wall temperatures were measured at any
time during this program. There was no burnout, and post run inspection
revealed no point of significant damage or deterioration of the wall.
The transition section made a l80 bend, whence it was
connected to the first section of heat exchanger tubing. This was a 5 foot
section. A Pt/Pt - 10$ Rh thermocouple ( H.E. l) was installed at the end of
this section. This was followed by a second straight section of heat exchanger
tubing 8 feet long. A second Pt/Pt - 10$ Rh thermocouple (Tn.E. 2^ was installed
25
-------
VtH NUTS. CAU'OINU
at the end of this section. Prom here the heat exchanger entered a trough
through which cooling water flowed. The tubing made a 10 foot pass, close
l80 bend, then another 10 foot straight section, at which point it left the
trough, ending the heat exchanger section. A chromel/alumel thermocouple
(TH.E. 3) recorded this heat exchanger exit flow.
Five cooling water flows were measured, as follows:
T Exit of the combustor cooling water. This stream was
then routed back into the jacketed transition section.
T Exit of the transition section.
T Exit of first (5 foot) section of Cu coil-wrapped heat
c .wo ,
exchanger.
T , Exit of the second (8 foot) section of Cu coil-wrapped
heat exchanger.
T Exit of the cooling water trough.
c w 2
The location of the remaining thermocouples was covered in
the section on fuel and air supply systems.
Each thermocouple was connected to thermocouple extension wire,
arid routed to the l6-channel print chart recorder, potentiometer, or dual pen
recorder, as appropriate. The print chart recorder and potentiometer were
located in the adjacent room.
D. Gas Analysis Instrumentation
Gas analyses were performed by the four instruments shown below:
CONTINUOUS FLOW GAS ANALYZERS
GAS TYPE RANGES
Carbon Dioxide Infrared 0-15#
Beckman Model IR-315A
Carbon Monoxide Infrared 0-1000 PPM
Beckman Model IR-315A 0-5000 PPM
26
-------
uuardt , <»»
Continuous Flow Gas Analyzers (Continued)
GAS TYPE RANGES
Nitric Oxide Infrared 0-200 PPM
Beckman Model IR-315A 0-2000 PPM
Total Hydrocarbons Flame lonization 0-5 PPM
Beckman Model 109A 0-15,000 PPM
(8 ranges)
These instruments were set up in accordance with the schematic
shown in Figure 16. Water vapor in the exhaust gas was removed in the pre-
filter and in the ice bath. Thus the mole fraction water vapor content in the
gases to the instruments was equal to the vapor pressure of water at 0 C
divided by 1 atmosphere. This small amount of water vapor does not have measur-
able influence on the NDIR analysis for CO and CO , and on the hydrocarbon
analyzer. It could have interfered with the NO analyzer, however. Two drying
agents were, therefore, placed in the line leading to the NO analyzer. The
first was "Aquasorb", a trademark for a P-O -based desiccant. This material
was recommended by the instrument manufacturer. The second drying tube contain-
ed Linde Molecular Sieve Type 3-A, 1/16 inch pellets.
The Aquasorb was removed (and replaced with a glass tube) from
the line in Run 15, Condition 5. and in Run 17, Conditions 2 and 3- The
molecular sieve was left in the system. The meter reading was about 2 units
lower with the Aquasorb removed, corresponding to a decrease in apparent NO
reading of about 3 PPM.
This change is relatively small, and in the direction of a lower-
reading, indicating that the use of the Aquasorb in conjunction with the
molecular sieve did not in any way result in an erroneously low reading of
NO.
Upon receipt of these instruments, an electronics check was per-
formed on each, in accordance with supplier recommendations. Each instrument
was calibrated daily when in use, by means of calibration gases procured from
Matheson Company. The concentrations of these calibration gases are:
CO ^05 ppm, balance N
NO 113 ppm, balance N
CO 10-95 vol. %, balance N
CH ^35 ppm propane (1305 ppm C), balance N
26 ppm n-hexane, balance N
(156 ppm C)
27
-------
ft r
Cr/
VAN Nun c*irfO«Nr*
It should be noted that the hydrocarbon analyzer is calibrated
at Marquardt in terms of ppm carbon, and the data of this report are given
in the same units.
Calibration data of each instrument were logged for each day of
use; this permits prompt detection of any loss in sensitivity (or other possible
instrument malfunction). Appropriate adjustments or replacements can then be
made, to assure that the instrument remains within satisfactory operating
limits.
In accordance with Beckman recommendations, the hydrocarbon analyz-
er was calibrated using the ^35 ppm propane (1305 ppm C), and then a reading
taken with the second gas (26 ppm n-hexane). In each case, the reading obtained
was 160 ppm C, in close agreement with value of 156 ppm, based on the concentra-
tion determined by the supplier. .
After standardization and calibration, the 3-way selector valve
of Figure 16 was turned to the N purge position. Just prior to starting an
ignition, the valve was turned to the sample position, which brought gas from
the probe in the heat exchanger tube through each instrument. The flow rate
through each instrument was checked and adjusted as necessary to equal the
value used during the calibration. Each instrument then analyzed continuously
and automatically for its specific component. The only attention necessary
was to switch from one scale to another, depending on the concentration of the
gas being sampled.
The instruments were left in the "sample" mode during startup,
transition, and shutdown.
Initially the sample probe in the heat exchanger exit consisted
of a static pressure-type tap. After combustion test 8, a total pressure
type of probe (of #-inch tubing) was used. This change did not have an effect
on the gas analysis data as determined by comparing the data measured at the
same combustion conditions with the tap and with the probe.
E. Component Checkout
Component tests were performed as necessary during installation
and fabrication of the combustor assembly subsystems. The purpose of these
tests was to assure that each subsystem was operating properly before uniting
them into the final assembly.
28
-------
5. CAII'OtNM
Air Supply System - Twelve tests were performed for the follow-
ing purposes:
(a) Determine opera.ting limits of facility air supply.
(b) Calibration of Venturis and pressure gages.
(c) Operational checkout of various plumbing fixtures and
control valves.
(d) Quantitative checkout of water removal equipment of the
facility air, and of the humidity measurement installation.
Fuel Supply System - Fifteen tests were conducted on the follow-
ing items:
(a) Operational checkout of fuel heater performance at full
scale flow rate.
(b) Operational checkout of vapor injection.
(c) Calibration of pressure vs. flow rate relationship for
liquid injectors.
(d) Operational checkout of 3-way fuel control valve under
different fuel conditions.
Igniter System - Initial testing was done with noble metal catalytic
igniters, with and without external heating of the igniter element. Nine tests
of this type were performed in which the igniter device was coupled with the
fully operational air and supply systems. These tests all involved actual ignition
of the fuel, and combustion for a brief time, over a wide range of conditions.
These tests produced promising results, but the resulting igniter devices were
not sufficiently rugged for the intended purpose.
A high voltage arc igniter had meanwhile been assembled, arid Lwo
tests under operating conditions verified that this device should, perform
satisfactorily over the full range of test conditions.
Preliminary Combustion Tests - The preceding subsystems were
combined, in their operating configuration, with a breadboard, uncooled test
combustor. This arrangement was used for final checkout of the system prior to
the main combustion tests. Ignitions were obtained over a wide range of con-
ditions, and final modifications were made to the configuration of the high
voltage electrodes. Seven tests of this sort were performed, up to full scale
flow rate. Combustion was limited to a maximum of about 60 seconds in these
tests.
A final checkout combustion test was then performed with entire
apparatus, including heat exchanger, assembled. The main phase of combustion
testing was then initiated.
29
-------
Hiirt.
F. Operating Procedure
A set of test conditions was selected for each run, based on the
independent variables of fuel flow rate, overall and primary equivalence ratios.
The tests were run in terms of varying heat release rates. Full scale heat
release fpr^the present jprogram was sel^cjbed_as_^00,^00_jTy^Lr_. The flow rate
of each fuel was controlled, based on its heat of combustion^ to provide a heat
release rate equal to (nominally) I/A-, I/2~, 3A and 100 percent of the full
scale heat release rate. Some tests were run at 1/10 and__l/30_ heat release
jscale,jq observe operation at these low flow conditions.
The required primary and secondary air flows were calculated, and
a selection made of the proper venturi to use for each condition. The
sequence of operating conditions was always selected to require the minimum
number of changes in the air Venturis.
In the case of a run using superheated vapor, fuel flow v/as
established at the desired operating value. The fuel heater was turned on
in steps to bring the fuel up to the desired superheated conditions in about
20 minutes. During this period, all of the fuel flowed into the bypass mode,
through the condenser and into the collection vessel.
When the fuel was up to condition, both air flows were brought up
to operating values. The gas analysis instrumentation was connected to "sample"
mode, and a final check made of the overall test setup. The high voltage
transformer was then activated, and the primary current observed to see if it
corresponded to normal arc conditions. The k KQ shunt resistor to the photo-
conductive cell was shorted out by a switch, and the photocell override button
was depressed. The fuel solenoid was then switched to the burner mode,
introducing fuel into the line leading to the fuel nozzle contained in the
igniter section. It should be noted that this valve would not divert the fuel
into the burner unless the override button was depressed. Immediately after
activating the solenoid valve, however, the override button could be released
without shutting off fuel flow (provided that the shunt resistor was shorted
out). Depending on the flow conditions, the fuel was often introduced into
the fuel nozzle in short spurts, since the initial process was necessarily,
accompanied by some condensation of the fuel as the insulated plumbing downstream
of the fuel solenoid valve became heated. This was continued until the combustor
thermocouple readout on the pen recorder showed a response. This meant the
system was ready to sustain combustion. The fuel valve was left in the burner
mode, and as soon as it was verified that continuous combustion was occurring,
the short on the photocell shunt resistor was switched off. This brought the
photocell back into the circuit. If, at any subsequent time, loss of flame
occurred, the photocell circuit would, as previously describied, shut off fuel
flow to the burner.
30
-------
ftr
Ql
V4M NUrl. CAU'OiHIA
At this time the operator was able to leave the ignition control
switch box, and examine all instruments for proper functioning. It was then
merely a matter of making minor adjustments to fuel and air flow rates, and
perhaps fuel heater setting, to bring the combustor up to the first condition.
Normally, at this time the igniter arc was shut off. Only for conditions
involving operation in "fringe" combustion conditions would it be left on.
After verifying that conditions were at steady state, readings
were taken of all pressure gages, manometers, and other appropriate control
settings. This was followed immediately by reading and recording the output of
each gas analysis instrument.
Conditions were then changed to the next set of desired values.
If this involved a change in air settings only, combustion occurred continuously
during the transition. If a change in Venturis was involved, or a change in
heat release rate (fuel flow), the combustor was shut down simply by turning
off the fuel solenoid valve switch. When the new Venturis and/or the new fuel
flow had been established, the startup procedure described above was repeated.
Runs with injection of liquid fuels, or with methane, were basically
very similar. The main difference was the elimination of the fuel condition-
ing time required in the superheated vapor tests.
The duration of actual cumulative combustion was generally in the
range of 1 to 2 hours for each test.
-------
V4M NUM. i
VI. TEST RESULTS
A. Description of Test Conditions
The schedule of runs is given in Table III. A total of 19 runs
were made with various fuels, fuel injectors and fuel preparation over a wide
range of equivalence ratios and heat release rates. The complete set of raw
data is compiled in Appendix D.
The majority of the data is plotted versus fuel equivalence ratio
or fraction of fuel flow referenced to a design point of 500,000 BTU/hr. The
fuel equivalence ratio is defined as the actual fuel-to-air ratio divided by
the stoichiometric fue]-to-air ratio for the particular fuel under consideration.
The relationship between fuel equivalence ratio, air-to-fuel ratio, and temper-
ature for a typical kerosene-type, hydrocarbon fuel with a heating value of
18,600 BTU/lbm is shown in Figure 17- Samples of the gas temperatures measured
by thermocouples at the burner exit are shown in this figure. The deviation
from the ideal are due to two effects: (l) a heat loss from the burner can to
the water cooled chamber walls, and (2) thermocouple errors. These are discuss-
ed in Appendix E where a heat balance over the complete system is presented.
A typical temperature drop due to heat loss at full design flow was 150 F.
The burner always operated just slightly above atmospheric pressure
as shown by the plotted data in Figure l8. The burner pressure loss was the
order of 5 inches of water or less, whereas the majority of the pressure drop
was taken through the heat exchanger.
The walls of the inner can were cooled by radiation to the outer
chamber and by forced convection to the secondary air flow. Typical wall
temperatures are presented in Figure 19. Although stoichiometric burning
with gas temperatures higher than 3500 F occurred in the primary combustion
zone, the cooling was more than adequate to keep the temperature well within
the range of conventional materials.
B. Gas Emissions
The gas emission data plotted in the following curves is in two
forms: (l) carbon monoxide (CO), nitric oxide (NO) and hydrocarbons (as
carbon) in parts per million of gas passing through the instruments and (2)
the same parts per million corrected to stoichiometric conditions (0~ = 1.0).
Whenever the second condition is used it is noted on the figure. A correction
that has not been included is that correcting the measurements to account
for the total volume of the exhaust gas. This correction is discussed in
Appendix E and is necessary as the instrument sampling system is such that the
water vapor in the gas is removed prior to analysis. Thus the emission data
will be lower when calculated in terms of the total exhaust products. A
typical correction factor is 0.89 for 0_ = 0.8 for kerosene fuel from the graph
in Appendix E.
-------
TABLE III
TEST SUMMARY
Run
No.
1
2
3
4
5 .
6
7
8
9
10
11
12
13
14
15
16
~\ r-
IS
19
Injector
Configuration
Vaporizing,
Premix, Iy
Vaporizing,
Premix, IVl
Vaporizing,
Premix, Iy-
Atomizing, Ij_
Atomizing, Ij^
Atomizing, 1^,
SUE, Iy
SUE, Iy2
Vaporizing,
Premix, Iy
Vaporizing,
Premix, Iy..
Vaporizing,
Premix, Iy.
SUE, Iy2
SUE, Iy2
SUE, Iyp
~
SUE, Iy2
SUE, Iv?
sui;, iv,.
Vaporizing.
Precix, IVl
Atomizing, Ij.,;
Fuel
TMH
TMH
CHk
TMH
Jet A
Jet A
CH^
TMH
TMH
TMH
TMH
TMH
TMH
TMH
TMH
TMH
TMH
TMH
Jet A
Equivalence Ratio
0p 0o
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
.6-1
.8
.8
.4-0
.6-1
.8-0
.4-0
.4-0
.80
.85
.80
.80
.25
.8
.0
.9
.8
.8
.80-1.20
.0-1
.8-1
0.8-1
.2
.2
.2
1.2-1.6
1.0-1
0
.6-1
.2
.0 '
0.6-0.8
0.6-0.8
0.6-0.8
0.4-0.8
0.6-0.8
0.5-0.6
0.4-0.8
0.4-0.8
0.6-0.8
0.78
0.6-0.8
0.6-0.8
0.6-0.8
0.6-0.8
0.6-0.8
0.6-0.8
0.6
0.6-0.8
0.6-0.8
Heat Release
(Ratio to
Design Load)
1/2
1A,
1/4,
iA,
1/2
1/2,
iA,
iA,
1/2,
1
1/10
1/10
1/2
iA,
1/10
1/30
1/2
T/lj.
1/^4-
.1/2, 3/4
1/2
1/2, 3/4
3A
1/2
1/2
3/4
, 1/30
, 1/30
1/2, 3/4
1/2, 3/4, l.O
1/2, 3/4, 1.0
No. of
Points Purpose .
10
5
4
10
11
6
7
12
6
1
4
*
>.
C
K
<
Effect of 0 |;
|q
Effect of Load
Effect of Fuel
Effect of Injection
Effect of 0
Effect of Load
Effect of Injection
Effect of Fuel
Compare to Run 8
Design Maximum
Minimum Flow
(Particles)
Discontinued-Faulty Ignitor
6
12
6
6
4
14
13
Reduced NO
Effect of Load
(Particles)
Low Flow
Minimum Flow
Rich Primary
Reduced NO
(Particles)
Effect of Fuel
Pressure (Particles)
' *
<
z
I
f\
k
1
»
-------
i. CAHFOINIA
To convert the data to a mass basis, a set of graphs has been
prepared and is also presented in the appendix. Thus the data as given in
Figures 20 to 33 can be simply converted to milligrams of pollutants per pound
of fuel burned by applying the appropriate multiplying factor given in
Figures E-5 to E-7 of Appendix E.
The overall fuel equivalence ratios used in the data plots are
those calculated from the carbon dioxide measurements (see Appendix E).
Discrepancies that occur between these measurements and those calculated
from the fuel and air mass flow measurements are also discussed in Appendix E.
1. Prevaporized, Premixed Fuel
The emission results for the basic case where the fuel is
vaporized and mixed with the air prior to ignition are given in Figures
20 to 23.
The pronounced effect on NO of burning with a minimum or
deficiency of oxygen is shown in Figure 20. Very low levels of CO and CH
are present with all primary equivalence ratios at % fuel flow.
The effect of different simulated load conditions are shown
in Figures 21 to 23- Going to richer primary mixture ratios causes a general
decrease in NO (accented at higher flows) but a more difficult problem to keep
CO emissions down at the higher fuel flows. In the analysis (Section IV-B),
it was shown that little or no NO is formed in the primary zone, but large
amounts of CO are created. At the higher flows, the residence time is apparently
decreased below that required to oxidize the majority of the CO to CO . With
less air addition (0n = 0.8), the flow velocities are less (increasing the time)
and the temperature is higher (increasing the oxidation rate constant) and the
CO rise occurs at a higher heat release than for 0O = 0.6. The reason for the
rise in NO at 3/^ fuel flow is not evident. The data results taken together
with the analysis point to several possible directions for further improvements
in simultaneously reducing CO and NO. These would be:
(a) An increase in the afterburner length with a
corresponding reduction in the primary zone.
(b) Increase in regenerative heating oi' the secondary
air to speed up the oxidation reaction.
(c) Operation at an overall equivalence ratio greater
than 0.8.
-------
2. SUE Burner Configuration
In the SUE burner configuration, the vaporized fuel is
injected in a fan from slot nozzles to intersect the incoming air at right
angles. Both liquid and gaseous fuel can be used with these nozzles, however,
the slot size becomes extremely small for liquid fuels for the size of burner
under consideration here. The fuel fan is broken up by the high velocity jet
of air from the small diameter inlet duct and the recirculating zone in the
step stabilizes the flame. The initial fuel-air mixture is generally quite
rich regardless of the overall mixture ratio by reason of the fuel-air
mixing. The air in the central core mixes with the annular burning zone as it
proceeds downstream. Thus the mixing process itself gives a degree of staging.
As shown in Figure 2^t, burning richer in the primary zone
did make a significant difference in NO formation. The effect of primary equiva-
lence ratio is shown more clearly in Figure 25- It appeared that an upper
limit of 0p = 1.2 was optimum with regard to keeping NO, CO and CH low simulta-
neously.
The emission data is plotted versus fuel flow for 0p = 1.0
and 1.2 in Figures 26 and 2?. Data was not obtained at the maximum fuel flow
point due to a facility limitation on the fuel pressure. The same type of
variation in NO as occurred in the premixing fuel injection results is
apparent in the 0p = 1.0 runs. In general, the SUE burner configuration gave
lower CO readings but higher NO readings than the premixod fuel system. It
also appeared from the 0 = 1.2 case that the CO rise occurred at a lower
fuel flow.
3- Pressure Atomized Liquid Fuel
The gas emission data taken with pressure atomizing fue]
nozzles spraying directly into the burner are presented in Figure 28 to
31. As discussed in the preceding section, two atomizing nozzles were employed,
differing by the swirl insert that developed the hollow cone spray and by the
pressure required to deliver rated flow. The "low" pressure nozzle delivered
1/2 the design fuel flow at a pressure of 85 psig, whereas the high pressure
nozzle required 160 psig for the same flow.
As shown in Figure 28, the maximum primary equivalence ratio
useable was significantly smaller than that for the vapor injectors. CO
levels rose sharply above 0p = 1.0 and smoke was visible in the exhaust. It
appears that a 0p = 0.8 - 0.9 was optimum for the atomizing injectors. At
Yz design flow, a 0p = 0.87 run showed quite low levels of emissions for overall
equivalence ratios between 0_ = 0.65 and 0.75 (Figure 29)-
35
-------
VAN NU»S.
The effect of heat release is shown in Figure 30 for a
primary equivalence ratio, ,0p = 0.8 and an overall equivalence ratio of
0Q = 0.6. The emissions of CO and CH rise sharply at 3/4 design flow and also
are higher at the 1/4 flow point - at least for the low pressure nozzle
where the AP is only 25 psig for this fuel flow. The higher pressure nozzle
shows better CO and CH characteristics at 1/2 flow and could be expected to be
better at % flow.
The characteristics of the same nozzle with liquid TMH fuel
is shown in Figure 31- The available kerosene data at corresponding equivalence
ratios are also shown for comparison. The TMH exhibits the same characteristics
as the kerosene and the actual data are very close to the same values. Note
that with the higher pressure nozzle, the CO and CH data at 1A fuel flow remains
low.
4. Methane Fuel
A series of runs were made with methane to make a comparison
with a lighter gaseous fuel. Prior analysis (see Appendix B) had indicated
that lighter fuel fractions could burn in an external combustion process with
a lower yield of NO as well as CO and CH. This was confirmed by the data shown
in Figures 32 and 33- The lowest readings of CO and CH were recorded during
these runs. In addition, NO (corrected to stoichiometric) at the corresponding
0p and 0 at 1/2 scale was reduced from the value of 136 PPM for TMH to 89 PPM
for CH, or a reduction of 35%. Facility limitations prevented data being taken
at full fuel flow.
C. Particulate Bnissions
1. Particle Collection and Analysis
The submicron particles that occur in combustion processes
often escape most common collection methods. Those particles less than 0.4
micron are also invisible. In order to study particles of this size range,
indirect techniques based on light scattering properties have been developed.
Unfortunately, these devices cannot disciminate particle size or distributions
in the gas stream accurately. This is because light scattering depends on
particle structure as well as size. There is also a scattering maxima which
occurs at about 0.2 micron. Therefore, many small particles (less than 0.1
micron) could be masked by a few particles of 0.2 micron.
For these tests the particles in the combustor exhaust gases
were collected by a Thermal Precipitator (1-4*). In this collection device,
a strong thermal force field is created by a hot wire heated by an electric
current. The sample is drawn between the heated wire and a cooler surface
(a Formvar coated screen) separated by .02 inch,by a small vacuum pump. When
*Numbers refer to references in Section VIII.
-------
VAH NUM. i
the small particles approach the wire, they are repelled and deposited on the
cooler surface by the strong thermal gradient (about 10,000 F/in.). The efficiency
of the thermal precipitator has been shown to be 100$ for all particle sizes
when the air downstream of the precipitator is observed by light scattering
instruments. When testing for submicron particles, a special settling column
is placed over the sample intake of the precipitator preventing particles larger
than 5 microns from entering the precipitator.
In order to investigate the size and shape of particulate
material in the exhaust gases, it was collected on small electro-mesh nickel
screens suitable for use in an electron microscope. The screen is .001 inch
thick with 200 openings per inch. Each opening is about 60 microns across.
The screens are coated with a thin film of Formvar (in 0.15$ ethylene dichloride)
which bridges the openings in the electromesh screen. This Formvar film is only
a few Angstroms thick and serves as the collection surface.
After collection, the microscope screens are removed from the
precipitator. To produce a three-dimensional image, they can be "shadow cast"
with vaporized paladium metal. The metal is vaporized in a high vacuum by a
large electric current. By shadow casting at about 15 above the horizontal,
a uniform overlay of metal is placed on the microscope screen, except when a
raised particle causes a "shadow". The height of the raised particle is pro-
portional to the length of the shadow. Shadow casting also shows the shape and
structure of the particle better.
The screens, either with or without shadow casting, are
examined with the electron microscope. The microscope operates by the same
principle as optical microscopes. The object to be examined (screen with deposit)
is placed between an electron source and fluorescent screen. The beam of
electrons is absorbed or scattered by the object and becomes dark spots on the
fluorescent screen. The image of the fluorescent screen can be photographed to
produce prints or slides. Particles appear white, the Formvar film appears
gray, and holes in the film appear black. The resolution of the electron beam
is much higher than a light beam (.001 vs. .2 micron). This allows each particle
to be studied individually if desired. Typical magnifications used in atmospheric
studies have been 1300; 10,000; and 18,000. The micrographs presented in
this report are magnified 5i^00 and lS,000 times.
2. Sampling System
Particulate samples were taken at two points in the £ar; i'low
system: (l) in the exhaust pipe just downstream of the gas sample station,
(2) immediately upstream of the third heat exchanger or water ba.th. The
sampling system is diagrammed in Figure J>k.
-------
ftr
" **
The exhaust air is drawn into a box* (l) containing the
precipitator by means of the vacuum pump (2). Once the air enters the box,
the air is circulated by a blower (3) to prevent stratified air within
the box. A heat source (4) is provided in the box to keep the air above its
dew point. A thermometer (5) is mounted on the precipitator (6) to allow
direct temperature readings. The air is drawn through the Thermal Precipitator
by means of the vacuum pump (2). The flow of air drawn through the precipitator
is monitored and regulated by means of a flow meter (8). An electrical source
supplies current to a wire (9) which creates the strong thermal force field
required for precipitation. A current meter (10) is built into the pre-
cipitator. In order to sample continuously, a slide block mechanism (ll) is
driven by an electric motor (12), allowing the slide block and microscope
screens to advance across under the hot wire at a constant, predetermined speed.
Calibration of the flow meter (8) and the slide advance (ll) are needed to
calculate the particle loading per cubic foot of air.
The operation of the precipitator involves preheating the
precipitator box to prevent condensation of the water in the combustor
exhaust. The blower, heater, and pump are turned on before the sampling
begins to stabilize the environment of the box and sample lines. When a sample
of exhaust is desired, the hot wire and slide advance motor are turned on
simultaneously. A stop watch is used to determine the time required for one
slide to cross under the hot wire.
One microscope slide was used for each combustor condition
evaluated. The micrograph print represents a section of the area contained
within one opening of a grid. To ensure that the precipitator box contained
a representative sample of exhaust at each condition, the pump was operated for
several minutes before a sample was taken.
3. Data Collection and Counting Technique
Exhaust samples are presented as photographs enlarged from
negatives exposed by the electron beam of the electron microscope. From the
micrographs it is possible to calculate the number of particles per unit
volume of exhaust gas. This is done by knowing the volume of gas Ga;np]r;, the
area of the deposit, the rate of travel of the slide, the magnification of the
photograph, the area of the deposit encountered, the number of particle:;
counted and the amount of excess air (J). With the aid of a graticule, the size
of the particles is determined. Statistical information is plotted on log
probability paper to obtain the geometric mean (50$) size and the standard
deviation. The amount of excess air is determined from the CO concentrations.
The equation representing the concentration of particles is: 2
*Numbers in parentheses refer to numbers shown in Figure 35-
38
-------
NursCAU,otNM
# of particles
cubic meter
count
field x
11.0 x 10
12
A x R x T
where:
A.. = area of deposit on slide in precipitator (mm )
= (\j x s x 7} for moving slide
w = width of precipitator slide (mm)
s = speed of travel of precipitator slide (mm/sec)
T = time of precipitator slide travel; sample taking
(sec)
A = area of deposit counted on micrograph print
R
Count
Field
11.0(10)
12
(microns)
rate of gas stream flow (cc/min)
(stoichiometric + excess) air
stoichiometric
# of particles counted in A of micrograph print
constant conversion factor giving dimensional
equivalence including magnification factor
W
Wire
O O o
OO <
000
000
OQ
OOO
Metal plate-
containing
microscope
gridr.
-------
The "percent less than" size distribution plot is a simple
graphical solution of the geometric mean particle size. This is obtained by
plotting the % of particles less than a given size vs. that size. Particle
size of small fairly uniform particles is taken as the largest flat cross
section on the micrograph. For large, irregularly shaped particles, the particle
size was arbitrarily defined as .8 of the largest flat cross section on the
micrograph.
The frequency of occurrence of each particle size range is also
plotted by dividing the number of particles of the particular size by the total
number of particles. The geometric mean size and size of greatest frequency
generally are not the same. The geometric mean size is:
Ln loglod d = particle diameter
Mg = antilog - ...
0 ° p n = particle number
and by definition bisects the symmetrical frequency curve yielding the mean
or 50$ size directly.
The standard geometric deviation defined by:
size 50$ size
ag ~ 50# size = 16% size
indicates the range of the geometric particle sizes. The geometric mean size
and standard geometric deviation define the logarathmic probability curve
(symmetrical normal probability) and determine all average diameters (geometric
arithmetic, volume, surface) giving complete description of a non-uniform
particle sample.
Using the mean particle size, an approximate weight equivalence
can be obtained by assuming that the particles are spherical and of known
density. This is a fair assumption for uniform tightly packed particles.
However, the large chains and grossly non-spherical particles are not accurately
represented by this assumption. Therefore, the mass approximations are consider-
able over estimated for the samples yielding large mean particle sizes and
irregularly shaped chains.
In a given sample the volume of particles suspended equals
the mean particle size (diameter defining volume) times the total number of
particles. The mass then equals the density times the volume of particles. For
2,2,5-trimethylhexane, the fuel used in the combustor, the particles are
assumed to be carbon (2.2 gms/cc). For automobile exhaust, 1G# by number of
the particles .1->1^/ are assumed to be lead (11. 0 gms/cc) with the remainder
carbon. Typical data reduction techniques and sample calculations are given in
Appendix E.
-------
H(J1S. CAIIPOINIA
k. Data Evaluation
Typical photomicrographs and particle size distributions are
shown in Figures 36 thru kl and the data are summarized in Table IV.
a. Particle Size
Particle size is determined from the xl8,000 magnification
prints when particles are small. Particle size ranged from 0.033 to 0.66
micron. The minimum particle size occurred for vaporized fuel at 3A fuel flow
scale, Figures 37, 38 & 39. For the liquid kerosene, the minimum particle size was 0.20.
micron which occurred at 1/2 fuel flow, Figures I+Q & 4l. In general, the particle
sizes correspond to the expected sizes for continuous flow combustion processes.
However, there is a wider range of particle sizes within each sample and from
sample to sample than usually found with other combustion processes. The
minimum particle size occurred at mid range fuel flow conditions for both liquid
and vaporized fuels. Vaporized fuels yielded smaller particles at equivalent
fuel flows. Run 14 at 3A fuel flow and Run 19 at 1/2 fuel flow yielded different
particle sizes at different conditions suggesting that the ratio of primary to
secondary air is important in determining the particle size distribution. The
overall amount of excess air is not related to particle size. Particle size
increased at full fuel flow conditions and at low (^_1/10) flow conditions. At
full flow conditions, inadequate residence time in the flame front prevents
complete oxidation. In Run 11, the. combustion was intermittant causing bimodal
size distribution; i.e., the small particles were produced during burning
while the large chains resulted from the "flame-out" periods, Figure 36.
b. Particle Count
The particle count data is all taken from the x5,*fOO
magnification prints which give a larger visual field and presumably a more
representative sample of the deposit. They are also measured downstream
of the water bath as discussed below in Section e(l). Considerable variation in
particle counts per unit volume are evident. They do, however, agree with the
expected values for combustion gas particulate concentration. The liquid kero-
sene produced slightly fewer particles (52 to 9^ x 10° part/cu.ft.) than vaporized
TMH (83 to 210 x 10° part/cu.ft.) and showed smaller spread. The maximum
particle counts occurred for the J>/k fuel flow condition using vaporized fuel,
Figure 40. The concentration of automobile exhaust is between 90 and 2?0 x 10'
particles per cubic foot.
c. Particle Mass (calculated from observed particle
concentrations)
Mass emissions estimated from the mean particle size,
count and particle density is only a rough approximation of the particulate
mass. For large particles, the density is inaccurate because of the void space
-------
TABLE IV
SUMMAKf OF PAKTICULATE DATA
Sample
2285D
2286C
2287C
2287D
2292C
2292D
2293A
2293B
*229l*C
2295C
2297A
*2298C
2299C
3000C
30010
30020
3003C
300l*A
3005A
3006A
Run
u
11
ll*
lU
J-M-
18
18
18
19
19
19
19
19
19
19
19
19
Inject; Fuel ^
Cond. Fuel Flow H?
rp*r
rn»/i
\ Vic
\ 1/30
7 iv2 3A
' TMH 3A
8 JV-
° TMf
3A
I 3A
9 lv, l
9 TMH 1
ivi
11 TMH 1/1*
i IL-
1 Je1
3
1*
5
10
11
13
15
1/2
; A 1/2
1/2
1/2
1/2
3A
1
1/U
iA
0.8
0.8
1.2
1.2
1.0
1.0
1.2
1.2
1.2
0.8
0.8
0.6
1.0
1.0
1.0
1.0
0.8
0.6
'o*
0.6
0.6
0.6
0.6
0.6
0.6
0.8
0.8
0.6
0.6
0.6
0.6
0.6
0.8
0.6
0.8
0.6
0.6
Count
Part/ft3 Part/M3
g 12
x 10* x 10
9.6
170
210
160
100
170
1*6
83
1*1
1*1
78
7^
9^
60
93
5^
52
65
3.U
5.9
7.5
5.7
3.6
6.0
1.6
2.9
1.5
1.5
2.8
2.6
3.U
2.1
3.3
1.9
1.9
2.3
AI gm
m3
8.9 x io5
1.3 x io6
3.2 x IO2
2.1* x 10
1.5 x IO3
2.5 x IO3
2.0 x 10J!
3.2 x 10'
6.0 x 10?
1.6 x KV
2.6 x 10
1.1 x 10"!
2.6 x 10J?
2.1* x lOJ?
3.9 x lOJ?
6.8 x ioi?
3.3 x 10^
l*.l x 10?
M
gra
gm fuel
1.3
1.9
3.2
2.1*
1.5
2.6
2.6
i*.o
6.0
2.0
3.2
1.2
3.1
3.7
U.3
9-U
U.3
5.2
x 10
x 10
x 10^
x 101
x IO3
x lO"3
x 10p
x 10
x IO2
«2
x 10,
x 10,
x 10,
x 10,
x 10,
x 10,
x 10J
St'd.
Mean Dev.
Size(M) (o-)
.60
.56
.033
.07
.07
.u?
.U5
.15
.21
.20
.33
'.^
.1*6
.66
.53
.36
2.50
2.3U
2.36
2.70
2.70
2.61
1.56
2.18
2.2l*
2.25
1.76
2.15
2.15
2.30
2.07
1.80
2.25
Observe
Smoke
^
No |
No
No
No
No
No
No
No
No
No
No
Inter-
mitt ant
No
Inter-
mitt ant
Heavy
Light
No
*Ahead of Water Bath (sanple point 2)
+0 and 0 based or. flow rates: roimded off
-------
ftr
Ol \
in the particles. For elongated particles, the assumption of spherical form is
incorrect. However, given these assumptions, the minimum mass emissions were
32 mgm/m^ and occurred for 3A scale fuel flow with vaporized TMH. This resulted
from the very low mean particle size (.033yw) which defines the average particle
diameter. The maximum mass emission (1.3 x 1CV mgm/m3) occurred for the large
carbon clumps of Run 11. Typical mass emissions are between 16 and kOO mgm/m3.
At equal conditions, the liquid fuel produced larger mass emissions due to the
larger mean particle size characteristic of the liquid fuel.
d. Residual Ash (calculated from observed particle
concentrations)
The estimate of residual ash is as high as 2% by weight
for the Run 11 conditions. The minimum range is as low as .0002% ash for
conditions of Run l*f. The typical ash content is in the range of .02% to .2%.
e. Special Studies
(l) Effect of Heat Exchanger on Particulate Bnissions
Most samples were taken downstream of the final heat
exchanger (water bath) since this is the final combust-
or effluent. Two samples were taken upstream of the
water bath to investigate whether particulates were
being effected. No effect was found in mean
particle size. However, on Run 18, a sharp drop in
standard geometric deviation resulted downstream of
the water bath indicating a smaller range of particle
sizes. This suggests both agglomeration of small
particles and precipitation of large particles. The
total particulate count, however, was increased
downstream of the water bath in both Run l8 and Run
19- This rather surprising observation is attributed
to the configuration of the sample port at the up-
stream sample point. The end of the sample line is
connected to a small diameter tube which is flush to
the inside wall of the exhaust pipe. The sample line
downstream is larger and pointed into the direction
of gas flow. The combination of small flush-fitted
tubing probably caused the abnormally small particle
counts upstream. The total particle count upstream
is about 1/2 the count downstream for both Runs l8
and 19, and the numerical values for the two runs are
similar.
-------
VAN Huff. CtltfOtNU
(2) TMH vs. Kerosene
The vaporized trimethylhexane generally produced
smaller particles but greater total particle counts.
(3) External Combustor Compared to Internal Combustion
Engines in Automobiles and Other Combustion Sources
Total particle counts generally were lower than from
automobiles while the mean particle size was generally
larger. Automobiles have particle size between .15
and .25 micron (see Table V and Figures k2 & ^3). The
external combustor normally operates in the .2 to .5
micron range, with optimum combustion producing
particles less than .1 micron. Mass emissions of
automobiles are difficult to evaluate because of the
uncertainty of lead particulate concentration in the
exhaust stream. If 10$ lead is assumed, the particle
mass emissions are in the range of 35 to l^fO mgm/nr
compared to 16 to 500 mgm/m^ for the external
combustor's usual performance. The optimum mass
emissions of the combustor is less than 1 mgm/m^.
Bnission standards are usually less than 300 mgm/nK
for most stationary combustion processes (Stern,
1968).(5)
f. Physiologic Effect of Inhaled Particulates
Particles larger than 5 microns are trapped in nasal
passages. One to 5 micron particles are deposited in the mid respiratory
tract. Those less than 1 micron reach the alveoli (Surgeon General, 1962)
Therefore, almost all particulate material from both automobiles and the
external combustor will enter the alveoli. Of those which reach the alveoli,
Dautrebande (1962)^' ' reports that 90$ by count will deposit with a minimum
deposition of 40$ at about 0.1 micron. Deposition increases with decreasing
size from .1 micron to .01 micron. The degree of damage is related to number
of particles rather than particle size. Hence efforts to control particle
emissions by reducing mass will not be effective in reducing lung insult if
particle size is reduced rather than the number of particles.
g. Conclusions
(l) The external combustor produces particulate material
somwhat larger in size and lower in concentration than
found with internal combustion engines.
-------
//ilarquardt .,
/orw/wtw/v
TABLE V
TYPICAL PAKTICULATE DATA
(Thermal Precipitator and Electron Microscope)
CONTINUOUS FLAME
Q
Particle Count x 10 Mean Particle Size
particles/cubic foot microns
Coal
Direct Discharge 40 - 125 .1 - .15
Baghouse 500 - 6000 .04 - .1
Natural Gas 30-100 <.!-.!
Oil 20 - 1200 .1 - .2
Marquardt Combustor 40 - 200 .03 - .66
AUTOMOBILE
1957 Chevy 90 - l8o .14 - .23
1968 Plymoutn 160 - 270 .17 - .23
SMOGGY AIR
(Downtown L.A.) 0.5 - 10 «.03
CLEAR AIR .001 - .10 ».03
-------
ftrf
OI \JUUI\Jt V4M NUM.
(2) The range of sizes within each sample and from sample
to sample is greater for the external combustor than
for other combustion sources except incinerators.
(3) Smaller particles are produced when operated on
prevaporized fuel than on liquid fuel.
(*t) Particle size is not dependent on amount of excess
air but appears related to ratio of secondary air
to primary air.
(5) Particle count is not related to excess air.
h. Recommendations
(l) Dynamic tests of the combustor are needed to evaluate
transient response. Similar particulate data is
available from automobiles for direct comparison.
(2) Design and operation of the combustor to produce
large particle agglomerations would aid particle
emission control. The effect of the final heat
exchanger as a particle collection or agglomeration
agent should be further investigated.
D. Heat Exchanger Interface
The interface between the combustion system and the complete
power plant is the heat exchanger. The combustion process cannot be considered
completely independent of the heat exchanger, as the sensible heat from the
combustion gases must be transferred to the heating fluid both efficiently
and in a manner not to increase emissions. For example, consideration of heat
transfer favors high gas temperatures to minimize surface area, volume, weight
and cost of the exchanger limited by material considerations. Excessively lean
burning would reduce the gas temperature sharply which, in turn, reduces the
heat transfer rate (increasing the required heat exchanger surface area and
size), and also reduces the thermal efficiency for a fixed exhaust temperature.
Large excess air rates also increase the power requirements to drive the blower
and so increase its size. In another sense, the presence of relatively cold
surfaces in the heat exchanger can have effects on the combustion process by
freezing incomplete reactions or by affecting the destruction of unwanted,
intermediate compounds. Burning should be completed prior to the time the
gases enter the heat exchanger.
1*6
-------
VAN Nurs. c/tnroiMU
The rate at which the gas is cooled could affect the kinetics
of the NO formation and CO destruction as described in the analysis. A typical
temperature time profile in the heat exchanger is given in Figure kk. For this
example, the gases are cooled below 2000 F in less than 15 milliseconds. At
this temperature the gas composition is, for practical purposes, frozen and
will not change in the ensuing passage to the sampling station. For very low
flow rates, the cooling is much less rapid and some reactions may continue
past the burner exit. The conditions existing for the maximum burning rate
are shown in Table VI.
For comparison, the data for a Rankine cycle boiler designed by
The Thermo Electron Corporation is shown in Table VII. In this design, an
additional stay time of 9-^ milliseconds at constant temperature occurs in a
transition section between the burner outlet and the boiler inlet. The
residence time for the maximum design load in the first row of heat exchanger
tubes where the temperature is reduced from 3300 to 1900 F is about 5 milliseconds.
-------
OO
STAGE
TOTALS
TABLE VI
COMBUSTION TEST RIG
HEAT EXCHANGEE CHARACTERISTICS
MAXIMUM BURNING RATE CONDITION
GAS TEMPERATURE, F
INLET OUTLET
RESIDENCE TIME
Milli seconds
2060
12.3
2060
519
76.0
108.7
HEAT REMOVAL RATE
BTU/Hr
180,000
91,000
.OOP
^36,000
-------
STAGE
Transition
Vaporizer
Superheater
Preheater
TOTAL
TABLE VII
RANKINE ENGINE BOILER CHARACTERISTICS
THERMOELECTRON DESIGN
MAXIMUM BURNING RATE CONDITION
GAS TEMPERATURE, F
INLET OUTLET
3330
1900
1190
RESIDENCE TIME
Milliseconds
3330
1900
119C
HEAT REMOVAL RATE
BTU/Hr
383,000
339,000
1,576,000
IS
-------
HU11.
VII. DESIGN CRITERIA
A. Experimental Results
1. Steady State Emission Criteria
The principal basis for the design of the Rankine engine
burner is the reduced emissions of carbon monoxide, hydrocarbons, oxides of
nitrogen, and particulate matter. In very general terms, hydrocarbons, carbon
monoxide, and smoke (unburned carbon) are due to incomplete combustion and
reflect fuel injection, vaporization, and mixing problems, plus insufficient
air and too short a burning time. Oxides of nitrogen (NO) are formed in any
combustion process at a rate usually insufficient to achieve full equilibrium
and are destroyed at a negligible rate. Thus, the kinetics are of major importance
in determining at what levels NO will exist, with the major factors affecting
initial reactions being peak temperatures and durations, available oxygen,
and rates of cooling. The burner design variables that have a major bearing
on the emissions have been shown to be:
Temperature (Equivalence ratio, 0)
Residence Time (Lengh/velocity)
Injector Characteristics (Atomization, vaporization)
Turbulent Mixing . (Burner geometry)
Air Staging (Burner geometry)
Quenching (Cold surfaces, air addition)
Fuel Burning Character- (Fuel)
istics
a. Temperature
The temperature of the reaction, both in the primary and
secondary burning zones, has a major effect on the formation and destruction of
the air pollutants. It appeared that rich burning (0^ = 1.2), followed by air
addition to bring the overall equivalence ratio to somewhere between 0- = 0.6
and 0Q = 0.8 gave the optimum results. Gas temperatures in this range of fuel-
air ratios are also desirable to maximize the thermal efficiency of the vapor
generator and keep the air pumping requirement to a minimum. From the analysis,
a slight preheat of the inlet air appears desirable to minimize unburned hydro-
carbons and more readily oxidize the carbon monoxide.
b. Residence Time
Residence time requirements are extremely important and
it was shown there is a distinct trade-off in the afterburner between minimizing
NO formation and increasing CO oxidation. The kinetic calculations also indicate
50
-------
VAM NUM. CAII'OIMIA
that the primary zone in the experimental test rig could possibly be reduced,
as the reaction was calculated to be complete in less than the full length.
The afterburner section was shown to have two conflicting requirements: (l)
with less time allowed for afterburning, the NO emissions showed a sharp drop,
(2) however, with less time available, the oxidation of the CO to CO could not
be completed. This was confirmed by the experimental results where NO emissions
continued to drop with higher flows through the combustor, whereas the CO
stayed about the same up to a point where the time was insufficient to complete
the oxidation step and the values climbed sharply. From the data, this minimum
time appeared to be of the order of 20 milliseconds.
c. Injector Characteristics
The method of injection and fuel preparation prior to
injection had a significant effect on the steady state emissions. The use of
vaporized fuel produced lower emissions at higher heat release rates. The
injection of liquid kerosene or TMH through pressure atomizing nozzles produced
low emissions up through one-half of design flow. The time required for vaporiza-
tion of the fuel droplet was apparently of sufficient magnitude to significantly
reduce the time available for the complete burning of the fuel. Carbureting
fuel injection where the vaporized fuel and air are premixed produced the best
results.
d. Turbulent Mixing
The burner used in all the tests produced a strong re-
circulation zone in the corner of the step that aided significantly in gas phase
mixing. The inlet velocities and velocities through the primary zone and in
the afterburner were sufficiently high to produce highly turbulent flow
conditions. The secondary air inlet holes were also sized to produce vigorous
mixing boundaries. The completeness of burning is enhanced by strong mixing
of the fuel and air that avoids pockets or over-rich ar over-lean mixtures. This
is expecially important where the fuel is not premixed with the air. In the SUE
burner configuration, the injection of the fuel at right angles to the high
velocity air jet in the corner of the step provided the best possible mixing.
e. Air Addition
It was shown by analysis and experiments that a fuel rich
reaction followed by air introduction and afterburning produce the minimum NO.
It was possible to burn at a higher temperature and avoid a rise in NO formation
by staging the air addition such as to produce a deficiency of oxygen in the
high temperature primary burning zone. Introduction of the secondary air produc-
ed a fairly rapid completion of burning with the CO reaching equilibrium values.
Too much air in the afterburner produced a condition where the velocity wan
increased (residence time decreased) and the reaction temperature brought down
too quickly to fully oxidize the CO.
51
-------
V4M Hurl. C4I»OINI«
f. Quenching
Quenching can occur: (l) locally from cold walls, (2)
from heat extractions from the gases, and (3) by introduction of cold air. To
prevent quenching and creation of unburned fuel, the walls of the burner were
always operated hot (500 F and higher). Heat transfer away from the gas warj
limited to radiation to a water cooled outer chamber and by convection to the
slow moving secondary air. For a typical run at ]00 percent design load, tho
heat transferred away from the burner was approximately 10 percent. Thir;
produced a reduction in the actual gas temperature by approximately 150 F. Ac
mentioned previously, downstream air addition was very helpful in completing
the combustion reactions, but the amount and method of introduction are critical
with respect to freezing the CO composition.
g. Fuel Burning Characteristics
There is a wide range of hydrocarbon fuel candidates which
might be considered for the external combustion system based on combustion
performance. These are discussed in Appendix B. Very little difference in emission
characteristics were noted between a commercial grade of kerosene and a pure
hydrocarbon, 2,2,5-trimethylhexane having comparable molecular weights,in the
tests. Combustion tests made with methane and computer runs using pyrolized
fuel fragments strongly indicate that hydrocarbon fuel blended with lighter
fractions than kerosene types could have significantly lower emissions (including
oxides of nitrogen) if used in an external combustion engine.
2. Other Bnission Criteria
In addition to selecting a combustion process that inherently
produces low emissions, consideration must be given to the following in the
design:
a. Level of emissions during a typical drive cycle
with either on-off or modulating control.
b. Level of emissions during startup and shutdown.
c. Possible degradation of emission levels with time.
d. Possible inspection and maintenance that may be required
by State or Federal Laws.
Only the first two are considered to be within the scope of the jjreisent
program.
-------
v«M NUTS.
a. Fuel Modulation
The results obtained at 10:1 and J>0:I turndown with
premixed vaporized fuel strongly suggest that very low emissions at high fuel
turndown can only be obtained with a secondary, or pilot, burner designed
specifically for the lower flow rates. Two types of problems result in any
burner design where large turndowns are required. The first is the combustion
problem of maintaining the stability over a wide range of flows and the
second one is simply that of introducing small amounts of fuel with good mixing.
For example, a 100:1 turndown for the burner corresponds to a minimum fuel flow
of 0.16 gal/hr. As a comparison, a "small" conventional atomizing nozzle used
in oil burners flows 3A gal/hr. In Figure V? the results from three nozzles
flow tested during the current program are-shown. The pressure atomizing nozzle
has a cutoff point of ^fO-50 psi; below which the atomization pattern is
unsatisfactory. Air atomizing nozzles can operate at iQWer pressures and are
reported to have good characteristics for 10:1 turndown*- '. However, -they
require a source of high pressure air and considerable parasitic power. The
air atomizing nozzle tested in the current program did not appear to have
satisfactory atomization at 10:1 turndown, but was not combustion tested.
Assuming a burner can be made to operate over a large turndown ratioi a
practical problem arises of controlling the fuel-air ratio to maintain good
combustion at all points and during rapid transients.
Controlling the fuel-air ratio in transient operations was
not a part of the current program. However, qualitative observations made during
the testing in changing from one condition to the next indicated large emission
peaks could occur where the fuel and air were independently changed and excursions
in equivalence ratio occurred. This showed the need for a control system that
could follow rapid demands for both fuel and air in transient operation. In
general, the combustor gas exit temperature and wall temperature responded rapidly
to change in fuel flow due to the thin wall, low heat capacity burner can.
b. Startup and Shutdown
The startup and shutdown emissions are chiefly a function
of the particular hardware design. For example, in the current program, no
emission overshoot was observed on starting except in the unusual case where
ignition did not occur immediately. In these cases, the hydrocarbon reading
quickly rose to several thousand PPM. On shutdown, the CO and NO meters decreas-
ed to zero in 2-3 seconds with hydrocarbon instrument showing slightly slower
response. Overshoots in CO and CH (but not NO) often occurred on stopping.
The difference between these results and the data observed by others ^° ~ ^-'
on fuel oil heaters is due to the responses of the fuel and air system and the
thermal lag in the combustion chambers. In the test rig, the fuel flow in
controlled by a bypass and fast acting solenoid vaLve closely coupled to tho J.'uo I
injector. The response time between the valve opening and full fuel flow ic
very short. The air flow is brought up prior to fuel injection. In addition,
53
-------
VIM NUVS. CAlirO*NM
V.WWWWW '
the thin wall burner can will heat up in seconds. In oil burners, the air
and fuel flow responses are linked to the rate at which an electric motor driv-
ing the blower and pump comes up to speed and to the filling time of a fairly
large diameter fuel line. The thermal response in the oil burner combustion
chamber is very slow (5-10 minutes in some cases).
For clean startup and shutdown the following criteria
would apply:
(l) Solenoid shutoff valve
(2) Minimize fuel line volume
(3) Air flow to lead fuel on startup
Preheated wall surfaces
3- Summary of Design Criteria
The recommended criteria and features for a low emission
Rankine burner, based on present knowledge, are summarized in Table VIII.
-------
TABLE VIII
RECOMMENDED DESIGN CRITERIA FOR LOW EMISSION BURNERS
A. PERFORMANCE FEATURES
1. Type of Burner - Staged with primary recirculation zone
2. Fuel - Kerosene (Aviation Turbine Fuel)
3. Maximum Heat Release Rate - 2.0 x 10 BTU/Hr-Ft^
*f. Pressure - Atmospheric (8.3 - 1^.7 psia)
5. Inlet Air Temperature - Ambient (-40°F to 120°F)
6. Burner Pressure Loss - 5 in. HO
7. Fuel Pressure - 50 psia maximum
8. Fuel Equivalence Ratio - 0.6 to 0.8
9. Fuel Injector - Carbureting (premixed fuel and air)
10. Ignition - High voltage spark electrodes
11. Gas Temperatures - 3200 F maximum
B. DESIRED SYSTEM FEATURES
1. Low Emissions - Expected levels of emissions will be below
1980 Federal Research Goals.
2. Compact Size - A heat release set by low emission criteria will
determine burner size for a given maximum power
output.
3. Low Noise - Noise levels will depend on velocities arid muffling
by vapor generator. High heat release burners can
be noiny.
4. Positive Ignition - Continuous spark ignition and/or optics I
flame detector' for positive fuel chutoff.
55
-------
VAH NUM, CAU'OtNM
Table VIII - (Continued)
B. Desired System Features - (Continued)
5- Fast Startup - Startup time for burner will depend on response
of blower and fuel pump.
6. Wide Turndown - >10:1
7- Environmental Adaptability - Burner should not be sensitive
to change in ambient conditions.
8. Low Manufacturing Costs - Cannot be estimated at this time.
9- Minimum Maintenance - Ignition electrodes may require periodic
replacement. Small orifices on fuel
injectors should be avoided to prevent
clogging, etc. Other parts of burner should
be maintenance free.
10. Maximum Accessibility - Depends to a large extent on packaging
of complete power plant.
11. Clean, Free of Odors - Efficient and complete combustion will
prevent any carbon (soot) buildup and
eliminate aldehyde or other odors.
56
-------
VAM Nurs CAU'OINM
i:t/Ht1iHWf>\
VIIJ. REFERENCES
1. Bush, A.F., "Analysis of Airborne Particles by Count and Size Using
Thermal Precipitation and Electron Microscopy," Presented at the Tenth
Conference on Methods in Air Pollution and Industrial Hygiene Studies,
San Francisco, February 19, 1969.
2. Bush, A.F., "Instrumentation for Sampling and Analyzing Combustion
Particulates Using an Electron Microscope," Analysis Instrumentation,
Plenum Press, 1964, pp. 267-278.
3. Bush, A.F., "Strange Airborne Particles, "AMA Archives of Industrial
Health. Vol. 15, pp. 1-2, January, 1957.
4. Froula, H., A.F.Bush, and E.S.C.Bowler, "Use of Thermal Precipitator
and Electron Microscope for Evaluation of Airborne Particles,"
Proceedings of the Third National Air Pollution Symposium, April 18-20,
1955, Pasadena, California.
5. Stern, Arthur C., Air Pollution. Vol. 3, 2nd Ed., Academic Press, 1965.
6. "Motor Vehicles, Air Pollution, and Health," a Report of the Surgeon
General to the U.S. Congress, June, 1962.
7. Dautrebande, Lucien, Microaerosols, Academic Press, New York, 1962.
8. Vickers, P.T., "The Design Features of the GMSE-101 - A Vapor Cycle Power-
plant," SAE Paper 700163, Presented at Automotive Engineering Congress,
Detroit, Michigan, January 12-16, 1970.
9. Howekamp, D.P. and Hooper, M.H., "Effect of Combustion-Improving Devices on
Air Pollution Emissions from Residential Oil-Fired Furnaces," Division of
Process Control Engineering, National Air Pollution Control Administration,
NOFI Workshop, September 17-18, 1968.
10. Wasser, J.H., et al, "Effects of Air Fuel Stoichiometry on Air Pollutant
Emissions from an Oil-Fired Test Furnace," NPCA Journal, Vol. 18, No. 5,
May, 1968.
11. Wasser, .1.11., et al, "Effects of Combustion Cns Residence Time on ALr
Pollutant Emissions from an Oil-Fired Test Furnace," Presented at NaM.onal
Oil Fuel. Institute Workshop, September 17-1.8, 1968.
57
-------
NUM.
12. Longwell, J.P. and Weiss, M.A., "Mixing and Distribution of Liquids in High-
Velocity Air Streams," Industrial and Engineering Chemistry, 45, 667-677,
1953.
13. Bahn, G.S. and Builder, C.H., "The Application of High Speed Motion Pictures
in Fuel Injection Studies," American Rocket Society, pp. 463-57, June, 1957.
14. Stern, Arthur C., "Air Pollution," Vol. II, Analysis, Monitoring, and
Surveying, 2nd Ed., Academic Press, 1965.
15. "The Oxides of Nitrogen in Air Pollution," State of California, Department
of Public Health, Bureau of Air Sanitation, January, 1966.
16. Peters, Max S., "Role, Sources and Control of Nitrogen Oxides in Air
Pollution," Presented at the American Institute of Chemical Engineers,
Southern California Section, 5th Annual Technical Exhibit and Meeting,
Los Angeles, California, April 25, 1968.
17. ICRPG Sixth Thermochemistry Working Group Symposium Held at Huntington Beach,
California, March 25-27, 1968.
18. Smith, D.S., Sawyer, R.F., Starkman, E., "Oxides of Nitrogen From Gas
Turbines," Journal of Air Pollution Control Association, Vol. 18, No. 1,
January, 1968.
19. Grumer, J., et al, "Effect of Recycling Combustor Products on Production of
Oxides of Nitrogen, Carbon Monoxide, and Hydrocarbon by Gas Burner Flames,"
AICHE Preprint 37A, Presented at Symposium on Air Pollution Control Through
Applied Combustion Science, New York, N.Y., November 26-30, 1967.
20. Moffat, R.J., "Gas Temperature Measurement," Paper 52. "Temperature, Its
Measurement and Control in Science and Industry," Vol. Ill, Part 2,
Rheinhold, New York, 1962.
21. McAdams, W.H., "Heat Transmission," 3rd Ed., McGraw-Hill, New York, 1954.
22. Dodge, B.F., "Chemical Engineering Thermodynamics," McGraw-Hill, New York,
1944.
-------
CONTINUOUS FLOW COMBUTION SYSTEM
FOR EXTERNAL COMBUSTION ENGINE
ic:
FUEL
STORAGE
SUPERHEATED VAPOR
TO EC ENGINE
FUEL
PUMP
o
AMBIENT
AIR
JELECTRIC j
POWER |-
' SOURCEj
VAPOR
GENERATOR
1=
EXHAUST
SYSTEM
LOW
EMISSION
EXHAUST
PRODUCTS
LIQUID OR GAS FLOWS
ELECTRICAL POWER & CONTROLS
COMMAND
SIGNALS
V// A OF PRIMARY INTEREST TO THIS
PROGRAM
-------
a/)
VAN NUT]. CAIIFOINU
V783I-20
VALUES FROM THE STEADY-STATE SOLUTION
FOR HYDROCARBON COMBUSTION
CONDITIONS: P = 1 ATM, 0 = 1.5, 0.068 MOLES OF FRESH MIXTURE
DILUTING 1 MOLE OF BURNED MIXTURE FROM 3420°R TO 3240°R,
AND RECOVERY CALCULATED FROM 3240°R TO 3420°R
0.14
0.12
0.10
0.08
MOLE
FRACTION
0.06
0.04
0.02
3480
3440
MOLE FRACTION OF
MOLE FRACTION OF CO
TEMPERATURE RECOVERY
MOLE FRACTION OF 0
MOLE FRACTION OF
FUEL (CgHl8) X 10
10 w 10
TIME- SECONDS
60
-------
ANALYTICAL MODEL
a. MODEL FOR ANALYSIS
s^1
if
VAPORIZED
FUEL
+
AIR
DILUTE WITH COLD AIR
REACT
AFTERBURNING
FOR FIXED TIME
RESIDENCE TIME
INCREASING
Z
Z
C
o
>
s
IGNITION
FUEL/AIR
MIX
b0 EXPERIMENTAL BURNER
SECONDARY AIR
O AFTERBURNING
STAY TIMES -10-100 MILLISECONDS
« STAY TIMES »-
5-50 MILLISECONDS
-------
CllLjUdlUI
- ICOMl'ANY
CALIFORNIA
INSTANTANEOUS TEMPERATURE OF RECIRCULATING
HYDROCARBON/AIR FLAMES IN PRIMARY ZONE
4500
4000
3500
TEMPERATURE
- °R
3000
2500
2000
I 1 1 ' T---- --
BASIS: 1 ATM., 1.24 FUEL EQUIVALENCE RATIO, AIR AT ROOM
TEMPERATURE, 38M°R EQUILIBRIUM FLAME TEMPERATURE
FRESH, 88% RECIRCULATING FLOW
10
-3
10
1
23«/, FRESH, 77* RECIRCULATING FLOW
FRESH, 66f, RECIRCULATING FLOW
"2 10"1 1 10
TIME - MILLISECONDS
100
-------
PERFORMANCE OF PRIMARY COMBUSTION ZONE
1.24 EQUIVALENCE RATIO
FRESH MIXTURE + RECIRCULATING PRODUCTS
REACTING FROM 2700°R
!-
^
PPM
35,000
30,000
25,000
20,000
15,000
10,000
5000
0.1 0.2 0.4 0.6 0.8 1 2
TIME, MILLISECONDS
4000
3800
3600
3400
- 3200
<
>
2
2
-<
^n
n
>
r
^i
o
XI
z
TEMPERA-
TURE- °R
2800
6 8 10
-------
CTx
AFTERBURNING AT = 0.65 OF FUEL/AIR MIXTURE AT 0 = 1.24 REACTED FOR 1 MILLISECOND,
AND THEN DILUTED WITH AIR TO TEMPERATURE OF 2700°R
4000
3800
3600
o
uj 3400
Z3
h-
<
£ 3200
2
LU
I
3000
2800
2600
- 35,000
30,000
140
120
0
0.1 0.2 0.40.60.81
i
Z
C
o
I-
Tl
O
Z
>
2 4 6 8 10
TIME - MILLISECONDS
40 6080100
-------
ON
vn
AFTERBURNING AT 0 = 0.65 OF FUEL/AIR MIXTURE AT 4> - 1.24 REACTED FOR 1 MILLISECOND,
AND THEN DILUTED WITH AIR TO TEMPERATURE OF 2456°R
eg
-$
o
4000
3800
3600
LU3400
^3200
^
LU
3000
2800
2600
- 35,000
30,000
- 25,000
o
Q
- = 20,000
o
o
_ ° 15,000
Q.
Q_
- 10,000
5000
140
120
0
0.1
002 0.4 0.60.8 1 2 4 6 8 10 .20
TIME - MILLISECONDS
0
40 60 80100
-------
COMBUSTION TEST RIG SCHEMATIC
MAXIMUM HEAT RELEASE=500,000 BTU/HR
SECONDARY
AIR
t WATER A I WATER 1
L JL I
COMBUSTOR
TEST
CHAMBER
HEAT
EXCHANGER
CONTINUOUS
GAS
SAMPLING
Qj
AMBIENT
EXHAUST
-------
AIR SUPPLY SYSTEM
PURGE AND COOLING AIR i
(TO
FACILITY
AIR
1
V,
A1
DRAI
n?
5n '
PCJ
SECONDARY VFNTIIR 1 R
B
PT
3
_^
MA
s
J
ID
MOMET
L
ERS
PTO-
^
W
COMBUSTOR
-N
D
L
(
P P
KSL TO
<
r-»-
"PT1
CD
u
o
Ul
5
z
z
C
;. CALIFORNI
>
-------
FUEL SUPPLY SYSTEM
ROTAMETER
FROM
AIR SUPPLY
*
HEATER
COOLING AIR
METHANE
Oc
NZ PURGE
FUEL
RESERVOIR-
785 WATT
(PRE-HEATER)
v 110V
3 ?
i^J
:LOW I/
CONTROL)
FILTER
LH*ID-
TURBINE
METER
MAIN HEATER
1 .6 KW
SOLENOID
COOLING
AIR
VARIAC
il
AGITATION
AIR
."COLLECT ON
ESiSflFiSiSb^- COO LING WATER
PRIMARY AIR-
WA
INSULATION
COMBUSTOR
SECONDARY AIR
WB
-------
COMBUSTOR AND IGNITER SCHEMATIC
o
FUEL
INLET
SOLENOID
VALVE
BYPASS ELECTRODES
HIGH VOLTAGE
TRANSFORMER
OPTICAL
FLAME
AIR DETECTOR
INLET
SLOT
NOZZLES
SAME AS ABOVE
LIQUID
INJECTOR
SECONDARY AIR
I
or
if
II
TO AMBIENT
-24
A- PRELIMINARY CAN COMBUSTOR
WATER COOLED WALL
TO HEAT
EXCHANGER
24 (MAX.)
CAN
AFTERBURNER
B-MAIN COMBUSTOR
-------
-------
COMBUSTOR ASSEMBLY
rt
i
-------
COMBUSTOR CONTROL AREA
i
-
I
n
g
-------
FUEL MANIFOLD
SUE BURNER CONFIGURATION
U.S. PATENT NO. 3,074,469
.00?
TYP. SLOT DETAIL
1+.0 DIA.
3.0 DIA,
SEE SLOT
DETAIL
1 A" DIA. TUBE
SLEEVE
-------
GAS ANALYSIS INSTRUMENTATION SCHEMATIC
§
-J
O
^rt
PRE-
'Zx, T
VVI 1_
f
COMBUSTION
GASES
PKUBt
FILTER
1 »
_n
|
^~
COLD TRAP
t
I
±
1
]
RELIEF
''AK^yi ( \ 1
£rr>
DIAPHRAGM
PUMP
t t
NDIR NDI
CO NO
L
**p
^
\*
1
1
?
MOLECULAR
SIEVE
AOUASORB
S
y y
1 t
PURGE H2
I
NDIR
CO,
L
V
T
t
FU
H.C.
L
ri
FUEL MIX
601 N,
40' H2
in
ZERO
AIR
FLOWMETER
y
T
QUICK DISCONNECT
-------
maufirdt
VAN NUYS. CALIFORNIA
IDEAL GAS TEMPERATURE RISE
HEATING VALUE = 18,600 BTU/LBM
VH044 10 A
4000
3000
GAS
TEMPERATURE
°F
2000
1000
0
0.2 0.4 0.6 0.8
FUEL EQUIVALENCE RATIO
1*4.8
1.0
7'j
Ki /;u
-------
dr
TYPICAL PRESSURE DATA
VAPORIZED TMH
V8044-9
6.0
5.0
4.0
BURNER
PRESSURE
DROP, 3.0
INCHES
OF WATER
2.0
1.0
ABSOLUTE PRESSURE
0
BURNER
ABSOLUTE
PRESSURE,
' PSIA
1/4 1/2 3/4
FRACTION OF DESIGN FUEL FLOW
76
-------
dK
VAN Nurj. CAtirOINM
TYPICAL WALL TEMPERATURE DATA
FUEL FLOW = 1/2
V6044-I I
2000
1500
WALL
TEMPERATURE,
°F
1000
500
0
O
A
D
V
_0
0.8k
1 .02
0.76
0.79
1 .20
00 FUEL
0.71+ KEROSENE (LIQUID)
0.70 KEROSENE (LIQUID)
0.76 TMH (VAPORIZED)
0.6i+ TMH (VAPORIZED)
TMH (VAPORIZED)
0
10 15 20 25
BURNER STATION, INCHES
77
Figure Yj
-------
THh:
arquardt
ICOMI'ANY-
VAN NUYS. CALIFORNIA
EFFECT OF OVERALL EQUIVALENCE RATIO
VAPORIZED PREMIXED INJECTION
TMH FUEL
1/2 DESIGN FLOW
V8ISO-24
600
500
400
CL
CL
i
O
CJ
300
200
100
0
120i
SYMBOL
100
CL
O.
5
0
0.4
0.5 0.6 0.7
OVERALL EQUIVALENCE RATIO,
-------
-ICOMI'ANY-
VAN NUVS, CALIFORNIA
VARIATION OF EMISSIONS WITH HEAT RELEASE
VAPORIZED, PREMIXED INJECTION
TMH FUEL
0 = 0.8 - CORRECTED TO STOICHIOMETRIC
P
V6I 30-26
700r
600
500
400
Q_
Q.
O
O
300
200
100
0
Q.
Q.
O
s
O
140
120
100
0
0.2 0.4 0.6 0.8
FRACTION OF DESIGN FUEL FLOW
I I J
O.SxlO6 l.OxlO6 l.SxlO6
HEAT RELEASE RATE, Btu/HR-FT3
-------
an
COMI'ANY-
VAN NUYS, CALIFORNIA
VARIATION OF EMISSIONS WITH HEAT RELEASE
VAPORIZED, PREMIXED INJECTION
TMH FUEL
0 = 1.0 - CORRECTED TO STOICHIOMETRIC
600
500
400
QL
O.
, 300
o
o
200
100
0
~ 120
100
a.
a.
o
o
V8I 50-28
20
0
0
0.2 0.4 0.6 0.8 1.0
FRACTION OF DESIGN FUEL FLOW
I I |
0.5x 106 1.0 x 106 1.5 xlO6
HEAT RELEASE RATE, Btu/FT3 - HR
80
Figure
-------
VAN NUVS, CALIFORNIA
V8l50-1'j
VARIATION OF EMISSION WITH HEAT RELEASE
VAPORIZED, PREMIXED INJECTION
TMH FUEL
= 1.2 - CORRECTED TO STOICHIOMETRIC
600r 120
500
400
300
o
o
200
100
0
100
0.
a.
o
v
o
40
20
0
0
0.2 0.4 0.6 0.8
FRACTION OF DESIGN FUEL FLOW
I | I
O.SxlO6 l.OxlO6 l.SxlO6
HEAT RELEASE RATE, Btu/HR-FT3
81
-------
COMI'ANY-
Nuys.
EFFECT OF OVERALL EQUIVALENCE RATIO
SUE BURNER CONFIGURATION
VAPORIZED TMH FUEL
1/2 DESIGN FLOW
V81SO-I 6
600r
500-
400
300
o
o
200
100
1201
OL
Q.
Q_
o
o
100
20
1
0
0.5 0.6 0.7 0.8
OVERALL EQUIVALENCE RATIO, 6
Figure
-------
an
COMPANY'
VAN Nim. CALIFORNIA
EFFECT OF PRIMARY EQUIVALENCE RATIO
SUE BURNER CONFIGURATION
VAPORIZED TMH FUEL
1/2 DESIGN FLOW
600
500
400
Q.
^ 300
o
o
200
100
0
120
100
Q.
Q.
O
O
CH = 100-750 ,
.8
1.0 1.2 1.4 1.6
PRIMARY EQUIVALENCE RATIO, 6
1.8
Figure r->
83
-------
arqua
COMI'ANY-
VAN Nim, CALIFORNIA
VARIATION OF EMISSIONS WITH HEAT RELEASE
SUE BURNER CONFIGURATION
VAPORIZED TMH FUEL
6 = 1.0 - CORRECTED TO STOICHIOMETRIC
V8130-33
700r-
600
500
Q.
CL
O
O
400
300
200
100
0
140
120
100
a.
o_
i
o
0 Oo2 0.4 0.6 0.8
FRACTION OF DESIGN FUEL FLOW
_ | _ | _ I
O.SxlO6 l.OxlO6 l.SxlO
HEAT RELEASE RATE, Btu/HR-FT3
1.0
-------
wquurdt
ICOMI'ANY-
VAN NUYJ. CALIFORNIA
vgiso-17
600
500
400
0.
O_
300
200
100
0
VARIATION OF EMISSIONS WITH HEAT RELEASE
SUE BURNER CONFIGURATION
VAPORIZED TMH FUEL
6 - 1.2 - CORRECTED TO STOICHIOMETRIC
120,
100-
a.
CL
O
O
on
^ u
0
0
0.2 0.4 0.6 0.8 1.0
FRACTION OF DESIGN FUEL FLOW
I | |
0.5xl06 l.OxlO6 1.5xlOfe
HEAT RELEASE RATE - Btu/HR-FT3
Fipjuro ^7
-------
COMI'ANY
VAN Nim. CALIFORNIA
V81 30-23
O.
Q_
o
O
800
700
600
500
400
300
200
100
OL
EFFECT OF PRIMARY EQUIVALENCE RATIO
LIQUID INJECTION
KEROSENE FUEL
1/2 DESIGN FLOW
LOW PRESSURE NOZZLE
160
140
120
CO = U200-7500
o.
a.
5
100 A
CH = 250-i»00
0.6
PRIMARY EQUIVALENCE RATIO, 6
86
-------
""
wqutirdt
ICOMt'ANY-
VAN NUVS. CALIFORNIA
EFFECT OF OVERALL EQUIVALENCE RATIO
LIQUID INJECTION
KEROSENE FUEL
1/2 DESIGN FLOW
HIGH PRESSURE NOZZLE
= 0.87
600r-
500
400
Q.
Q.
O
O
300
200
100
o
D.
Q_
0
o"
VBI SO-2Z
OVERALL EQUIVALENCE RATIO,
Figure P->
-------
arquardt
/COMPANY
VAN NUYS, CALIFORNIA
VARIATION OF EMISSIONS WITH HEAT RELEASE
LIQUID INJECTION
KEROSENE FUEL
0 =0.8- CORRECTED TO STOICHIOMETRIC
V8I30-I2
1 UU
600
500
400
2
CL
CL
i
O
0 300
200
100
0
J-tU
1 20
100
n RO
LL OVJ
CL
IE
0
o AH
2 o*J
40
20
0
.s
GT
SYMBOL
§
\
N1
^-^
Gf
N.
0 NOZZLE
0.6 LOW PRESSURE
0.6 HIGH PRESSURE
NO
CO
CH
^rj
V
CO x 5C
?.
O
I4CC
/CH
//1'
y
00
i = 3800-i^
= 800-1600
\ = 275-69C
00
0 0.2 0.4 0.6 0.8 1.0
FRACTION OF DESIGN FUEL FLOW
O.SxlO6 l.OxlO6 1.5xl06
HEAT RELEASE RATE, Btu/HR-FT3
88 Fin;i;re ^
-------
arquardt
IC.OMI'AHY-
VAN NUYS, CAL/FORN/A
VARIATION OF EMISSIONS WITH HEAT RELEASE
LIQUID INJECTION, HIGH PRESSURE NOZZLE
TMH FUEL
0 = 0.87 - CORRECTED TO STOICHIOMETRIC
VH1 aO-2
700
600
500
Q.
Q.
O
O
400
300
200
100
Q-
Q.
I
O
140
120
100
80
60
40
20
0
CO
CH
SYMBOL
a==
f~^
FUEL
o
O 0.7 TMH
A 0.8 TMH
O 0.7 KEROSENE
Cr
X
X
£S
o-
&^::
\ = 500-60C
i
*
-£$?**
^ «
-co = 350
i
CO = 3800
I
-3050
-1+800
y*CO - c-?n_i -Jin
21+0-WO
(-X^^CH = 275-690
^6"
\ = 1+3-100
0 0.2 0.4 0.6 0.8 1.0
FRACTION OF DESIGN FUEL FLOW
0.5 x 106 1.0 x 106 1.5 x 106
HEAT RELEASE RATE, Btu/HR-FT3
89
Figure
-------
-I COMPANY-
VAN Ntm, CALIFORNIA
EFFECT OF OVERALL EQUIVALENCE RATIO
METHANE FUEL
SUE BURNER CONFIGURATION
1/4 DESIGN FLOW
vaiso-o
600i-
500
400
Q.
Q.
. 300
O
O
200
100
0
120
100
80
Q.
CL
O
o"
0.4 0.5 0.6 0.7
OVERALL EQUIVALENCE RATIO,
0.8
90
Figure.-
-------
-Quardt
-1COMPANY
VAN NUYS. CALIFORNIA
V&\ 50-15
VARIATION OF EMISSIONS WITH HEAT RELEASE
METHANE FUEL
(6=0.8, 0=0.62-0.70
600r-
500
400
CL
CL
O
O
300
200
100
O
CORRECTED TO STOICHIOMETRIC
120
100
a.
CL
o
o
80
60
40
20
0
SYMBOL
O
or
INJECTOR
PREMIXED
SUE BURNER
-NO
0 0.2 0.4 0.6 0.8
FRACTION OF DESIGN FUEL FLOW
| . l i
O.SxlO6 l.OxlO6 l.SxlO6
HEAT RELEASE RATE, Btu/HR-FT3
91
Figure y^
-------
PARTICLE SAMPLING SCHEMATIC
rv>
c
>x
rr>
WATER BATH HEAT EXCHANGER
SAMPLE POINT 2-
COPPER TUBE
HEAT EXCHANGER
t I I ( I I ( (. I ( ( ( ( C
SAMPLE
POINT 1
MARQUARDT BURNER
BOX AND
PRECIPITATOR
PUMP
FLOWMETER
or
si-
is
<
2
Z
-<
r>
>
r
o
a
z
>
-------
PARTICLE COLLECTION SYSTEM
vO
?
e
'-<
o
HEAT SOURCE (k)
EXHAUST PIPE
(5) THERMOMETER
! I
,(1) PRECIPITATOR BOX
JT
(12) ELECTRIC MOTOR
(2) VACUUM PUMP
(3) BLOWER
(6) THERMAL PRECIPITATOR
(10) CURRENT METER
(9) HOT WIRE
(6) THERMAL PRECIPITATOR
LEGEND
AIR FLOW
AIR TUBING
ELECTRICAL SUPPLY
(11) SLIDE
ADVANCE
or
II
-------
I
an
COMPANY-
VAN NUVS, CALIFORNIA
va iso-4
PARTICULATE SAMPLE EXTERNAL COMBUSTOR
SAMPLE 2287D
RUN #11 CONDITION 3
2,2,5 TRIMETHYLHEXANE (VAPOR)
1/30 FUEL FLOW
MAGNIFICATION = XIS^OO
Figure 36
-------
"An
-quardf
1 COMPANY-
VAN NUYS, CALIFORNIA
VB 150-7
PARTICULATE SAMP1E- EXTERNAL COMBUSTOR
SAMPLE 2293A
2,2,5 TRIMETHYLHEXANE (VAPOR)
RUN # 14 CONDITION 8
3/4 FUEL FLOW
MAGNIFICATION = X5400
95
Figure 37
-------
7Hf>//7
//tlarquardt
1 COMPANY-
VAN NUYS, CALIFORNIA
V8150-6
PARTICULATE SAMPLE- EXTERNAL COMBUSTOR
SAMPLE 2293B
RUN #14 CONDITION 8
2,2,5 TRIMETHYLHEXANE (VAPOR)
3/4 FUEL FLOW
MAGNIFICATION = X18,000
96
Figure 38
-------
'quardt
-/COMPANY-
VAN NUVS. CALIFORNIA
VBIHO-30
PARTICULATE SAMPLE - EXTERNAL COMBUSTOR
SAMPLE 2293
SPECIMEN K-4
3/4 FUEL FLOW
RUN #14 CONDITION 8
501-
FREOJJENCY .
0
h
1 1 1 1 1
1.0
0.8
0.6
0.4
0.2
0.1
0.08
0.06
0.04
0.02
0 .04 .07.10.13.16.19.22.25.31.34
SIZE IN MICRONS
0.01
TOTAL COUNT 41 O 178D
1 j !_, , -.
O DATA FROM X18,000 MAGNIFICATION
MEAN SIZE « O.r07 MICRONS
D DATA FROM X 5,400 MAGNIFICATION
I III I I I I I
NO. PARTICLES = 170 x 109 O
CUBIC FOOT EXHAUST 100 x 109 D
STANDARD GEOMETRIC DEVIATION =7.7
2% 5 10 15 20 30 40 50 60 70 80 85 90 95 98%
PERCENT LESS THAN
r i ~-'-C".
97
-------
-ICOMPANY -
VAN NUYS. CALIFORNIA
veiso-3
PARTICULATE SAMPLE - EXTERNAL COMBUSTOR
SAMPLE 2299C
RUN #19 CONDITION 1
JET A (LIQUID)
1/2 FUEL FLOW
MAGNIFICATION = X5400
96
Figure
-------
'quardt
-I COMPANY-
VAN Ntm. CALIFORNIA
VSI50-29
PARTICULATE SAMPLE - EXTERNAL COMBUSTOR
SAMPLE 2299C
SPECIMEN N-5
1/2 FUEL FLOW
RUN #19 CONDITION 1
50r-
FREQUENCY
0
0 0.2 0.4 0.6 0.8 1.0
SIZE IN MICRONS
MEAN SIZE - 0.20 MICRONS
STANDARD GEOMETRIC DEVIATION =2.25
0.01
2% 5 10 15 20 30 40 50 60 70 80 85 90 95 98%
PERCENT LESS THAN
99
-------
an
COMPANY-
VAN NUYS. CAL/FORN/A
V8I50-5
PARTICULATE SAMPLE - INTERNAL COMBUSTION ENGINE
SAMPLE K-2
1968 PLYMOUTH
DOWNTOWN - LOW SPEED
MAGNIFICATION = X5400
100
Figure
-------
quardt
-I COMPANY-
VAN NUVS. CALIFORNIA
V81SO-3I
PARTICULATE SAMPLE - INTERNAL COMBUSTION;ENGINE
SAMPLE K-2
1968 PLYMOUTH
DOWNTOWN - LOW SPEED
100
50
TOTAL COUNT = 214
NO. PARTICLES
I I I I
MEAN SIZE - 0.18 MICRONS
STANDARD GEOMETRIC DEVIATION
I _ i i i
0.1
30 40
50 60 70 80 85 90
PERCENT LESS THAN
95
98'/.
101
-------
arquardt
{COMPANY-
VAN NUYS, CAL/FORN/A
TYPICAL GAS TEMPERATURE DATA
VALORIZED TMH, 3/4 FUEL FLOW
3000
2500
2000
GAS
TEMPERATURE, ,°F
1500
1000
500
0
40 80 120 160
RESIDENCE TIME, MILLISECONDS
200
102
Figure -
-------
NUM.
V 8082-1 1
MEASURED FUEL INJECTOR FLOW CHARACTERISTICS
FRACTION OF
DESIGN FLOW
:(PA1R=10PSI)
PRESSURE ATOMIZING
100
200 300
PRESSURE - PSI
400 500
103
-------
arqudrdt W4M NUTS
ItxtHHiMTinn
APPENDIX A
COMBUSTION GENERATED AIR POLLUTANTS
-------
//ilarquardt VAH
laiRHMfficm
Murs.
The vari.ous pollutants which tnay arise in combustion of keroscne-typ<-
fue1 fall into three general classes: unburned fuel, particulate carbon, and
dissociation species. The unburned fuel may, at one limiting condition, pass
through the combustor essentially unchanged, and at the other limit it may be
very substantially degraded, just short of conversion into particulate carbon.
Regardless of the degree of degradation, the basic cause and means of alleviation
are concerned with the fuel/air mixing process, as discussed below. Formation
of particulate carbon is attributable to inadequate mixing at the microscopic
level, and this is discussed in turn. Under most conceivable combustion condi-
tions, the most prominent dissociation species which might endure at quench
temperatures are CO and NO (aside from Ho, which is of course of no consequence
in the present context).
Hydrocarbons
The problem of traces of unburned hydrocarbons from a hydrocarbon/air
combustor is one of mixing and not directly of reactivity. If mixing were com-
plete with sufficient air available, one would not expect unburned material.
Thus, if a reasonable flame strength is achieved associated with high temperature,
conditions are extremely conducive to reaction (pyrolysis and combustion) of the
last traces of fuel. This premise of the significance of mixing is important
to understanding of the relative amounts of unburned hydrocarbons which may be
expected from different hydrocarbon/air combustion systems.
12
The fuel distribution correlation devised by Longwell has been
employed with success in combustion design. This correlation approximates an
error function. In Reference 13, Builder and Bahn demonstrated that the disper-
sion of fuel in a turbulent air stream reflects a statistical distribution due
to the vagaries of the turbulent flow, consistent with an error function. If
there is a statistical distribution of fuel over the airflow, one can envision
an overall distribution which is generally quite satisfactory, with local pockets
of adverse distribution scattered throughout the flow. The local fuel/air ratio
in some of these adverse pockets may very likely be so far removed from stoichio-
metric as to preclude flames spreading into them from surrounding, more favorable
regions. This barrier to combustion can be adjacent to the next most adverse
regions, with the least flame-spreading capability. Then one can only hope for
local radiative auto-ignition brought about by exposure to some surrounding high
temperature region without the microscopic turbulence required by flame-spreading.
Analysis has shown that if 0.01.% of the airflow had fuel equ I v;i I encr
ratios ($) as adverse as either 0.1 or 1.0, then about 80% of the airflow would
contain fuel of concentration between 0 = 0.5 and 0 2, which are grossly the
f lammnbi I ity limits of hydrocarbon fuel in cold air. The very lean 107. of the
nlr and the very rich 107. would then react with the fuel present only wi. tli
difficulty. However, the summary conclusion of this analysis in that trace
unburned hydrocarbons are not inherent to so-called "off-stoichiomctri.<." design
A-l
-------
VAN NUVS.
in the overall sense, but are attributable to locally very adverse mixtures, and
as such are a problem which can be resolved by proper injector designs.
Particulate Matter
The problem of particulate emission in the external combustion process
is related first to thr- formation of "smoke" particles, and second, to the
emission of submicron particles formed by nonvolatile compounds in the combustion
process.
Smoke evidently arises by catastrophic pyrolysis of fuel vapors when
these have insufficient air to support combustion as the favored process.
Gaseous fuels or vaporized liquid fuels are the easiest to burn without smoke
formation because good mixing with air down to the microscopic level is most
readily accomplished; burning of liquid fuels of good quality in a smoke-free
process is not difficult to achieve if the fuel particle size can be made small
and strong mixing and turbulence created. However, the composition of the liquid
fuel is important in assuming smoke-free combustion. For example, the presence
of substantial concentrations of unsaturated hydrocarbons such as olefins and
diolefins can produce sooty flames as compared to cleaner burning paraffins and
naphthenes.
The second problem relating to submicron particles is a more difficult
one to attack. The knowledge of the history of a single fuel droplet as it
undergoes fractional batch distillation and liquid-phase reactions in passing
through the combustor from one temperature zone to another is necessary in order
to speculate on how such particles may be formed. It has been reported that
gasoline, diesel, and turbine engines all may emit submicron glassy spheres.
Under an electron microscope these appear as chains and rings. Individual fuel
droplets arc known to burn to an outside hard crust if heated rapidly, then to a
vitreous mass that coalesces above 1000°C to nearly colorless glassy spheres.
Carbon Monoxide
Carbon monoxide is the requisite intermediate in the creation of CC^
as an end product in combustion. It also can itself be an end product to a
degree in the incidence of high-temperature equilibrium dissociation, such as
with fuel-rich operation. The latter situation can be overcome by designing and
operating so as to forestall dissociation, but the former situation is inevitable,
Thus, it is necessary to ensure that CO is indeed oxidized to CC^, as called for
by the proper end point of overall reaction. The oxidation of CO in flame is
primarily achieved by progress of the reaction OH + CO = H + C0«, which is
exothermic by 25 kcal/moie. In the flame1 zone there is ready mobility nmonj; Lhr
various H-0 species, so that while Lhe flame endures, n sufficiency oi Oil c.ni In-
consistent wLth the- rate of creation of CO. However, If the flame is qii<-nrhed
by abrupt admixture of excessive cold air, then recombination reactions will tend
A-2
-------
ft ft
UILl1J\JI\JI VtH HUtt.
'tmPHMTKH
to destroy the 0-H dissociation species (especially OH) in the form of 02, ^,
and H?0, thus leaving the CO inert at low temperature. Therefore, the fuel
injection and mixing process is the deciding factor in minimizing air pollution
by CO.
Nitric Oxide
The various deleterious oxides of nitrogen in air pollution derive from
nitric oxide. To reduce NO concentrations, some mechanism must be used that
either (a) affects the formation of NO; or (b) accelerates the decomposition of
NO. Experience has shown that NO in the exhaust from hydrocarbon-air combustion
processes is affected by :
(1) Peak temperature and duration
(2) Availability of oxygen
(3) Rate of cooling of gases
The peak temperature and duration, availability of oxygen, and the
kinetics during mixing and cooling are all strongly affected by burner design.
Although equilibrium calculations are revealing in certain aspects,
the formation and fixation of nitric oxide in a combustion process is dictated
by chemical kinetic considerations. This has also been pointed out by Peters
and was demonstrated by experiments reported by R. F. Sawyer . The
concentrations of nitric oxid found in experiments with both gas turbine engines
and combustion rigs are appreciably less than frozen equilibrium values from the
combustion zone, but often more than final equilibrium values for the situation
after admixture of dilution air.
The essential reaction for formation of NO is 0 -I- ^ = N + NO, with
an activation energy of 75 kcal/mole, corresponding to the heat of reaction.
The atomic nitrogen thus formed either may itself react to form more NO,
N + 02 - 0 + NO, or it may recombine to ^, N + N + M = N- 4- M, depending upon
relative concentrations. All three obvious reactions for the consumption of N,
i.e., the reverse of the first, plus the second and the third, are essentially
on the frequencies of the respective collisions. There is a substantial activa-
tion energy involved in the destruction of NO by the reverse of the second
reaction, and until the concentration of NO has built up to its equilibrium value,
the combination of adverse activation energy arid adverse reactant concentrations
tends to suppress this reverse reaction of destruction.
In general terms, then, NO is formed at a low rate insufficient to
achieve complete equilibrium concentration and is destroyed at a rate so low as
to bo negligible. In order to reduce Lhi- net amount of NO formed, therefon-,
two approaches may be envisioned: (1) combustor operation under condi t. ions
very adverse to kinetic formation of NO, and (2) utilization of chemical
scavengers to remove the NO formed.
A-3
-------
//llarquardt
lamnuwmN
For real understanding of the kinetically-controlled formation of NO
within a flame, it is essential to base the calculation of NO formation on the
concentration of atomic oxygen and atomic nitrogen in the flame. The concentra-
tion of atomic oxygen is certain to be appreciably different than the final
equilibrium value as calculations have shown. As soon as the flame has done its
work and consumption of fuel has been effected, rapid dilution and cooling of the-
flame by addition of excess air is desired, so as to quench the formation of NO.
In Reference 19, Grumer, Harris, Rowe, and Cook have concluded that NO formation
may be reduced "by starting to cool the combustion gases as soon as possible to
about 2300 F, at which temperature concentrations of nitrogen oxides do not
increase within the residence time in most gas appliances." If one accepts this
as an operating principle, the inclusion of the limited NO reaction kinetics
within the much more extensive pyrolysis and combustion kinetics is obviously
required.
Furthermore, practical experience in combustor research and development
lends assurance that a design can be made to operate well when the component
steps are all specified to inhibit the formation of NO by suitable combinations
of temperature, oxygen, concentration, and time. For example, the tendency
toward NO formation is greatly inhibited in the fuel-rich regime where all
available oxygen is needed for carbon oxidation, as calculations have clearly
demonstrated. Thus, first-stage, fuel-rich combustion to a very combustible
mixture of CO and H? at high temperature, followed by some admixture of air and
rapid second-stage reaction (too rapid for much NO to form), and then be followed
by further cooling to reduce the temperature, is indicated by theoretical
considerations (and confirmed by experimental results obtained in this program).
-------
NUM.
APPENDIX B
FUEL SELECTION
-------
vuardt M,
M NUTS. CAU'OINI*
On Table B-l are shown those fuel characteristics which are essentially
constant (within 25%) for liquid hydrocarbon fuels of conceivable interest.
Table B-2 lists variable characteristics and points up the significance of these
characteristics. Table B-3 shows how each significant caracteristic varies as
one moves from a fuel represented by gasoline toward a more volatile fuel such
as propane or methane. The tendency would be to vary in the other direction in
moving from gasoline toward, for example, diesel oil.
Discussion of Fuel Characteristics
Droplets of fuel burn with diffusion flames, if they have not been
already evaporated upstream of the flame region. Clearly, vapor pressure
indicates the feasibility of achieving complete vaporization prior to combustion.
In the diffusion flame around a burning droplet, cracking of the evolved fuel
vapor, to less suitable products, may occur before it is able to mix with air
and burn to the desired final products of CC>2 and H»0. Were it not for this
cracking, the size of the droplet would be important only as this determines
the burning time and thus the requisite combustor length. As it is, the depth
of the diffusion flame around a droplet, in which adverse pyrolysis can occur,
is proportional to the droplet size. Thus, large droplets are more likely to
result in some deleterious pyrolysis of evolved vapors than are small ones.
This consideration argues for use of lighter hydrocarbon fractions as fuels,
since these atomize more easily to small droplets, as influenced by both viscosity
and surface tension.
A burning droplet has as its surface temperature very nearly the boiling
temperature of the liquid. For a given droplet size, and thus a given depth of
diffusion flame, the high temperature exposure of fuel vapors to the liklihood
of pyrolysis will be somewhat lessened by evolution of the vapor at a lower
temperature, even if the maximum temperature of the diffusion flame is the same
(as it is at about the stoichiometric value for all burning droplets). More
important than this, however, is the temperature exposure of the droplet before
vaporization occurs. Thermal degradation may occur in the liquid phase,
accelerating with temperature as liquids of different boiling points are examined.
These considerations relate principally to the possible employment of fuels less
volatile than gasoline, such as fuel oils.
In tables B-2 and B-3, fuel characteristics dealing with combustion
reactions are summarized as ignition energy, burning rate, and smoking tendency/
coking tendency. In gross terms, ignition energy is greater, or ignition is
more difficult, with smaller hydrocarbon molecules (I.e.-., more volaMle fr.icl inns)
.ind c-onl ra rtw i.se, the tendency Low/ird production of smoke or i-ol«- i i: y.ivaier wiih
larger, less volatile molecules. Volatility a8l.de, MM ll a Nee In tin- cn-alli I
t\ combust i hi e mixture of fuel vapor and air, there cannot be Haiti to be j-ond l.,i;.lr
data on comparative burning rates of various fuels under constant-pressure condi-
B-l
-------
, Clil'OINU
tions. To be sure, there are data on flame speeds through flammable mixtures,
but this represents propagation, not burning to completion, and is not really
germane to defining the details of a continuous-flow system.
Analysis
In a continuous-flow combustion system, there is an anchored flame
source, rather than a continuous sparking action or the equivalent. The design
of the system is such that fresh combustible mixture is continuously fed into
the flame source region, and burning material is continuously bled off to ignite
the bulk of the combustible flow which bypasses the flame source region. By
some mixing pattern, combining burning and fresh materials, flame is spread into
the whole flow. Finite-chemical-kinetics calculations have been performed which
have connotation for selection of fuel type for least emission of objectionable
effluent substances. A large hydrocarbon molecule distributed in air in reason-
able proportions for ensuing combustion evidently must be converted to small
fragments before significant reactions providing heat release can occur. Two
calculations were performed with comparable starting conditions (using very hot
air to dictate prompt auto-ignition) to compare the combustibility of nonane
(CgHoQ) and highly pyrolyzed nonane. In the latter case, combustion was essen-
tially completed in about one-tenth of the time required by the former case.
Considering the available time at high temperature to produce NO, the pyrolyzed
fuel is surely preferable in being quicker to burn, and thus an argument is
presented for the use of a more volatile fuel of smaller molecular size,
corresponding to pyrolyzed fuel. The argument can also be made that if the fuel
is burned more rapidly to CCL and H-O, there is less chance of its being
pyrolyzed to free carbon as an objectionable product.
In summary, there is no firm direction for selection of one "most
appropriate" fuel. The benefits of the system may be such that a less refined
fuel than gasoline may be wholly suitable, or a fuel the equivalent of unleaded
gasoline may be called for, or even the benefits of lighter hydrocarbon fractions
may be utilized later to fully realize minimization of pollutants. The influence
of fuel type is one of degree, not of "go/no go" significance, but it is
important that the system can readily utilize lighter hydrocarbons than gasoline
if this should prove to be logistically favorable. The distillation curves for
some common hydrocarbon fuels are presented in Figure B-l.
B-2
-------
an
iXWHMATIM
VAN MUVI. CAl"O)MU
TABLE B-l
FACTORS ESSENTIALLY CONSTANT REGARDLESS OF FUEL
Heating Value
Density
Stoichiometric Fuel/Air Ratio
Identity and Properties of Theoretical Combustion Products
Storage Stability (Fuel Inert Internally and with Materials)
Toxicity
Heat of Vaporization
TABLE B-2
VARIABLE FACTORS EXAMINED AS TO SIGNIFICANCE
Melting Point
Boiling Point and Vapor Pressure
Viscosity
Surface Tension
Flammability Limits
Ignition Energy
Burning Rate
Smoking Tendency and
Coking Tendency
Not a problem
Evaporative losses;
influence on combustibility
Ease of atomization
Ease of atomization
Not a problem
Ease of ignition
Completion of reaction
relative to formation of (NO)
Liklihood of objectionable
effluent
B-3
-------
S/ilarquardt
loURHIMTim
VAN NUM. CAU'OINU
TABLE B-3
TRENDS OF VARIABLE FACTORS MOVING FROM GASOLINE
TOWARD MORE VOLATILE FRACTIONS
Vapor Pressure Increases
Boiling Point Decreases
Viscosity Decreases
Surface Tension Decreases
Ignition Energy Increases
Burning Rate Increases
Smoking Tendency/Coking Tendency Decreases
B-if
-------
APPENDIX C
TEST FUELS SPECIFICATIONS
-------
nn,
HUH. CAII'OINU
FUEL SPECIFICATIONS
AEROSHELL TURBINE FUEL 640
(JET A OR KEROSENE)
PROPERTIES
Gravity, °API
Color, Saybolt
Flash, Tag c.c., °F
Pour Point, °F
Viscosity, cs at 30°F
Copper Strip at 122°F
Copper Strip at 212°F
Corrosion, Silver Strip
Mercaptan Sulfur, 7. wt
Odor
Smoke Point, mm
Aromatics, 7. vol
Freezing Point, ASTM °F
Water Reaction, Inc. or Dec., ml
Interface Rating
ASTM Distillation, °F - I.B:P.
107. Evaporated
5070 Evaporated
907. Evaporated
957. Evaporated
End Point
Recovery, 7. vol
Residue, 7, vol
Lo s s , 7o vo 1
Aniline Gravity Constant
BTU/lb (Calc)
Gum, Existent, Steam Jet Mgs.
Gum, Potential, Steam Jet, Mgs.
TAN-C
SAN-C
Olcfinc, 7. vol
Naphthalenes, (Diaromatics)
Water Scparometcr Index Mod
Luminometcr Number
Thuimal Stability. ACTM-CFR Coker
Pressure Drop, In. Hg
Prehcater Deposit Rating
39.8 (0.826)
30+
136
B-60
10.24
1
1
0
0.0001
ok
20
14.1
B-58
0.5
1
344
365
419
478
491
504
98
1.0
1.0
5532
18,500
1 .
1
Neutral
Nil
0.7
0.31
98
47.3
.05
0
C-l
-------
//ilarquardt NU
IWHHIHATim
S. CAUFOINU
FUEL SPECIFICATIONS
2.2.5-TRIMETHYLHEXANE - TECHNICAL GRADE
95 MOL PERCENT MINIMUM PURITY
Typ.u-i.il Properties
Frcc'/.inc point, F
Boilinf, point, F
Distillation range - Initial Boiling
point, F
Dry point, F
Spei-ilVn: f.rav.ity of liquid at 60/60F
at PO/'lC
API r.ravil.y id. (QV
lieu.-', i.t.v ol' l..i(|u.ixl ;.it 6011', Ib/^al
Vapor pressure at 70F, poia
at 100F, psia
at 130F, psia
Refractive index at 20C
Flash point, approximate, F
uulor, S.'iybolt
Tin I ha- i.'orit'Tit,, v/c-j/iht porcnnt
A'- i 11 i I.y , il i ;: I, i 1.1 .'i I. i iTi /':: i 'Luc
l|i ,n v< i I .'il. i Ic lii:il,l.(T, (M'.'iln:;/ I Oil m'l
Technical Grade
255-3
255.7
0.711
0.3
0.7
1.399
55
+30
Test Method.-;
ASTM D1015
Modified Cottrell
Modified Cottrell
ASTM D ]',
AfJTM D3rJ
ASTM D.1310
ASTM hi.1,.'-
ASTM Hi: i (
AS.TM D." ;'.
AS.TI'l IH ;', ',
C-2
-------
//ihrquardt N0,, Mll,
li:imiiH\THin ^^^
FUEL SPECIFICATION
METHANE
SPECIFICATIONS
COMPONENT
Methane
Oxygen/Argon
Nitrogen
Carbon Dioxide
Ethane
Propane and Higher
Hydrocarbons
W;itcr
Water (Dew Point)
TECHNICAL GRADE
GUARANTEED
97.0 mol. 70
<0.025 mol. 7,
<0.80 mol. %
< 0.01 mol. %
< 2.50 mol. 7,
<0.60 mol. %
< 0.0126 mol. 7,
<-40°F
PHYSICAL PROPERTIES:
Chemical Formula
Molecular Weight
Specific Volume at 60 F, 1 atm.
Boiling Point at 1 atm.
Freezing Point at saturation pressure
Specific Gravity (Air = 1)
Critical Pressure
Critical Temperature
Critical Volume
Flammable Limits in air, by volume
Lower Explosive Lir.iit
Upper Explosive Limit
Autoignition Temperature
Latent Heat of Vaporization at
boiling point, 1 atm.
Latent Heat of Fusion
Heat of Combustion at 25 C
Gross Heat of Combustion at 60 F,
1 atm.
CH.
4
16.04
23.61 cu. ft/lb
-258.7°F
-296.5°F
0.555
673.1 psia
-115.78°F
0.0991 cu. ft/lb
5.37.
14.0%
1000 F
219.22 Btu/lb
0.05562 Btu/lb
978 Btu/cu.ft.
1011.6 Btu/cu.ft.
C-3
-------
lixatHHtmnN
VAN NUM. CAll'OINM
APPENDIX D
DATA SUMMARY
-------
This section presents the pertinent data for each test. The
tabular presentation is in two recurring data formats. The legend for this
table is as follows:
0p Primary equivalence ratio, calculated from flow rates
i
0 Overall equivalence ratio, calculated from flow rates
WA Primary air flow
Wg Secondary air flow
WT Total air flow
Wr Fuel flow
T Fuel temperature at heat exchange outlet
p Fuel pressure at nozzle inlet
P Primary air gage pressure
L0
(PX^-PT-,.) Differential pressure, primary vs. secondary air
(P_, -Pc ) Differential pressure, primary air vs. combustor exit
IQ bL
TC Uncorrected reading of combustor thermocouple
CO Concentration of carbon monoxide in exhaust
NO Concentration of nitric oxide in exhaust
H0C0 Concentration of unburned hydrocarbons in exhaust,
expressed as ppm carbon
C0~ Volume percent carbon dioxide in exhaust
0 Overall equivalence ratio computed from COy concentration
TH £ i First head exchanger thermocouple, 6^ ft. of heat
exchanger length
TH E 2 Second heat exchanger thermocouple, 15 ft. of heat
exchanger length
T Third (exit) heat exchanger thermocouple, 34 ft. of
" " heat exchanger length
TW| Combustor wall thermocouple 1, Station 1.
T^,, Combustor wall thermocoupl c 2, Stntlon 2
TW.. Combustor wall thermocouple 3, Station r>
'^W^ Combustor wall thermocouple ^> Station ^
D-l
-------
VtH NUM. CMI'OIWM
NOTATIONS FOR FUEL CONDITIONS
Fuel
F 2, 2, 5-trimethylhexane
F_ Kerosene
F» Methane
Fuel Condition Suffix
F-l Superheated vapor
F-2 Ambient gas
F-3 Ambient liquid
Fiie I In joe to t:
I,, 0.10" converging vapor injector
I.. "SUE" vapor injector
V2
I Higher pressure liquid atomizing injector
Ll
I, Lower pressure liquid atomizing injector
D-2
-------
SUMMARY OF DATA
COND.
FUEL
SCALE
(NOM.1
WA
LB'SEC
WB
LB'SEC
WT
LB/SEC
LB'HR
TF
°F
PSIG
IN Hg
"
IN
H-0
IN H2O
RU
/
13.
O.o?
Q.OMO
O.O
IZ.b
367
/.£
s.o
/
Z4-40
z
z
c
o
Tl
O
2
c.71
C.'~l 1
<, O
?.
C .
G, 6
in, tin
0 . 0
7
. 7
-A,
/.
/ .PC
o . 05? £
, r
2.6
1,24-
r, .
A - i (
3 *-
/4 -
/*
3
l±Jl
_'/
Id
a
i*- C
.CK?
VTA?
_^_
1.4
I
C
-------
SUMMARY OF DATA
CONDITION
CO
PPM
NO
PPM
H.C.
PPMC
C02
H.E.
°F
il
itGC
1072
71?
z
S.-T
IC3
c. 7
il
1C 2
ts
z
c
n
3
z
V
1 77T
//
£'74
71
7
c.
6 0
SO
-1 05
4-
/o. ?
17
/( 5~C
<£>, 65"
mo
n
IO33
7
7O-4-4-O
?/-«#
122 O
307
1/7
1 1
;-?/; 7*
ft 71
-9/6
111
3&O
7
In-
-------
SUMMARY OF DATA
f:
COND.
FUEL
SCALE
(MOM.)
WA
LB^SEC
WB
LB'SEC
LB'SEC
LB/HR
°F
pf
PSIG
INHg
IN
H20
(P
IN
H20
3
Yz
c.ol
o
, C?
- 3
/.o
3
TA/'JL
2
Z
C
n
3
a
z
r,. c
Y
'
C .
/,, g
4-0
0,3
4-O
~ /.O
?-
3 70
7
O-VI
^,70
?.0 -?*
7/0
ff
13.6.
780
n
£'7
n.c?
n , i
r.c.z
-3
rot?
,06!
o
-C.GR
3-7
C.'M
O.G^l
o
LL£-
c.s
2fr 4-C
( . O
c .
/ . a
2 c -
2.4-4-O
f. .66
C. fit, X
C. .
a
f.Cb
t-ctcl
Z4-CC
t-c 365
,
-------
SUMMARY OF DATA
CONDITION
CO
PPM
NO
PPM
H.C.
PPMC
C02
E.
°F
-
3
73
<14-3
II
2.
10. C
A/ o r
r-
1
4co '5~O
iao<]
1333
t 4
/fe
II 3-3
4-47
f
7C
/4-..TP
I7X
t 330
110-140
loo~i05
C ,£'7
Zoo?
7
4- I o
looo- not
35
7,-
7
11
/IB/
7CQ
/o
Ifc-IZCC
// .
; Sr/5"
PK./ A,
o ,(
.5
1C - §G
I'Z-IC?
, 550
I P. 4-L
>.*r
-------
SUMMARY OF DATA
COND.
FUEL
SCALE
(NOM.i
LB'SEC
WB
LB/SEC
LB'SEC
LB/HR
PSIG
INHg
IN
H2O
(P - P
IN HjO
\o
/a
C.6I
C
.f.
C.//7
/s
42-
2. '7
II
r.
/, 7
R.L
3
'£-~CT/jl? .
la
C .t
C.OZ.3S
r,
t)J r>
2
o
a
r
.
A 7
1C
G.14-
C 6?
±L.
j&
. 7- 3.
C.llb
za.c
- 3. 3
I.CZ1
c ,/=:
-7
^
. c£ m
- 37
c.,c
3
/34-C
G.'ltS
/, 3
y*
C. 0*4-4
c.o (, io
'
C
.4-
r .
r.c
C-C
Zc
si
c ,
c ^ r.
-------
SUMMARY OF DATA
CONDITION
CO
PPM
NO
PPM
H.C.
PPMC
C02
,
w3
^
1C
£00- ft*.
S.4-
/ 7 4-O
73)
IS.
.f
//.
4)7
, t
761
2
C
n
i
(L
;, 7
'7 3
33%
9T-T42
777
1 b
10
$1
/C.
4-KO
ZGC5
1 C
In
3 1C
4-tiO
4-
i O
WO
b
?*c-c
2±
tzo
'05%
U 7
4=-
4-4.
lion
*-/(>
744-
rlc
V74-
inc.
I.4-
4-47
-32$-
1C
.5
I6I3
£5-
c .'7
f, 3-7
74-
7,6
-7
? 7
3
£7
I 3/4
i 2-73
14-4-d
ZL.
A-
c. c-
T'7-7
m
7C
f ,
II
? y v
res
7
f .
7/, -7
-------
SUMMARY OF DATA
COND.
FUEL
SCALE
WA
LB-'SEC
WB
LB'SEC
LB'SEC
LB/HR
PSIG
IN H8
(PT
'
o fi
IN
(pvV
IN H20
or
4-
CdlC.
,6
. 66
I 9
I
O.
+
a
G.I
.f 7
C .4-7
C .C. TK4-
C ,
-a A /c.-r.
<
>
Z
z
<
n
S
*
y?
o. £6 3 /
r--0? Z '7
/. 2
0. 0
C,
G-SO
3/i
C-cC
r-.c
e^G,
-C. 3
C,f,
/9C
f.fe 4-G
c -co / -5"
r.
-------
SUMMARY OF DATA
CONDITION
CO
PPM
NO
PPM
H.C.
PPMC
C02
,
°F
si
*Qj
la
4-
70
4-1
1 t
7 76,
4-47
r . 6,0
203
/^
- 2/5D
/HC,
Tcf
4~/4-
(oO
\zc
/i
4-4-3
fr 0
It-Co
n
*
-------
SUMMARY OF DATA
COND.
FUEL
SCALE
fNOM.i
WA
LB/SEC
W
B
LB'SEC
WT
LB'SEC
LB/HR
pl
PSIG
IN Hg
(PT
T
0 1
IN
(P
IN H2O
o
t
FL^L
rot. '.
Tc C
C
r, S'C
f .'
n .
=577
t. Z
, 3
2
c.
, C <{ il
-f).
c.o&M
/i-
t.az
x" . o 4-?'o
,~ r
01
A.
, c
r.
, a
-Lb
f
r,.
. 3
-r . -7
/ c'X
c.r.GSl
^(.-5-
r.f s/.C
?>Q
3-c- fc 5.5"
7
i.ZC
c. c S'.X
C /
V?
V
-r/7
LJ*_
_!/
/ L
C .C J..11
^.6 -;'7 f
r
^, 6
_!_
^ f
_u^J_
c . c 115
- <-
f - £
«r" c>
r . *>
-------
SUMMARY OF DATA
CONDITION
CO
PPM
NO
PPM
H.C.
PPMC
C02
H.E.
°F
oT
t t !» /^ t
C 72:
dr /J
C- / 7~ / a '
, 4-
ft / /
>.
2
2
O
O
2
JZ
' A/ /
5"
6" /b
33C
76"
it- "3 *r
3/e
?C
3-C-
/r 9
-7
4-14-
4-
15
X 7
/, S'2
tlQ-C
7
f? 7
336,
4-1 V
^T-
1C
ft
0 .
-7
Z
il,7
c. *
//6&
//-?£
-1£3
f-
n.
3^3
s
f
C -b'
//
3V
r -'5"
-
//o. 7
4-1C
7
//, 7
C
4-
-7$
4-fl
Sc
71
V '-b
JL±LL
2£L
n
V
//f-6
it-*)
_£J_
"3 «--
// 3
7/
c
-------
SUMMARY OF DATA
COND.
FUEL
SCALE
(NOM.i
WA
LB'SEC
WB
LB'SEC
LB'SEC
W,
LB'HR
PSIG
IN Hg
IN
HjO
(R
IN HjO
I
z
2
C
0
Tl
0
a
z
C.C 1*7
//C
c
- 9, /
/a
l.c'-l
C .61
3. c£ ///
"
-3
X/C1
(c
l^L
-3
-3
uO
"
r Z. '.
X
c. L-r,
2100
C.ctJ
c ; c sr
00
,c
I
1
-M (<-*>'
-c- ''4 a -r
2. /
fl
/. -7/6
(.. t k /
-------
SUMMARY OF DATA
CONDITION
CO
PPM
NO
PPM
H.C.
PPMC
C02
12
bo'
Ito
PC*!
A
o...
m
/GST?
z
n
i
f-f-7
/,*- 12-5
4-4-
'-1,f
64-7
7
y
p.
c' i t^TL L-"73
sr r^~A
L; S ~7~sa
r >> .
/ 5 7" fn L
.- S ' S
D .4- Ts.
t /''
6V:
fC
:?
,(. g C'
l^L
6' 3 7
/ft A-
7 J - / 7
-------
SUMMARY OF DATA
COND.
FUEL
SCALE
(NOM.i
WA
LB'SEC
W8
LB'SEC
LB/SEC
LB'HR
TF
°F
PSIG
INHg
IN
(PT - P. )
To S
IN
HjO
i
pLti
- /
rc
-5,1
lfi.2.
, c Zi'7
2.4- £ 0
r. .
; ,C'lt> 5"
-/
z
0
f-
I, CC
C , S6
-0.3
I.-ZG
~2. 0
-/
i.cc
r .L
« r 'fj "3
r, . IQS 3
l.'l
r.c
C?
1,00
- .-r. /
I,
c so
-C
- 1*73
/5
Z.4
10
1.00
r. *. ? H -7
-2.4-
u
1-2.2.
r,
,.Gl
. o
1ST
C- (.-/
r .c \%t
n . -si
r.
* .r
zz 6s-
/f
i.o£
_£.
>/
i
^^1
i-^L
_c
r^^-/
t.
"0.3
a
.<;/
r ,t C, i 1
/ . C
C ,
l' ' L
C
^_i
-c . *
r
C .g i> 3<-
C C 1 1
- J. A
L ( 6 1C
.1
|_L ._L I. l^-'
^ '
I C
vc
-------
SUMMARY OF DATA
CONDITION
CO
PPM
NO
PPM
H.C.
PPMC
C02
H.E.
Rou
Ice
34
£
7,
1C Iff
'3cc
a
/c
z
z
^
o
r-
3
317
3-6
IC1O
J?
7S
10. 3
1*10
24-0
4-XC
6
3
7
HZ -7
377
4-4-
7
7.Z
r -.
II
72
14-74
7/^
10
ice -2 co
7/Cf
) i
7/
id.
Sbl
23. 1
4-J
T-/C*
if. 0 -//-
ff'j
4-
7/
1C:
a
r- /
iit.-t.tc
r»'r- /-7oo
r
^C 5'
9-
ft s o-
.S- ice
V7
7
-------
SUMMARY OF DATA
CONO
FUEL
SCALE
(NOM.<
LB'SEC
WB
LB'SEC
WT
LB SEC
LB'HR
PSIG
IN Ho
IN
o. ,- Ar /.>'
CfC II (y
o
C&
/ f
C. c
C-
C '?
/f
r .GO
4-6 i
-r t fc c--
.CIS
r .
-t.i K
(.oZ-
r,
-------
SUMMARY OF DATA
CONDITION
CO
PPM
NO
PPM
B.C.
PPMC
C02
,
°F
7?
ffi/C
-QCO- 3&'?C
c.5~
73 i
z
2
C
0
Tl
0
//.
tf.4
C-7?
the
ia
Ic. 7
7
14-7
tftf
Jc.fi
/4-4-I
T7?/
c b >
4-
71
/c.
"6 '7
/so-- vto
iCC
r
/ 2
6~4-
/7
-------
//ilarquardt »
^-^ /(.WPIIMTION
nun,
APPENDIX E
DATA ANALYSIS
-------
VAN NUTS. CAlirOINIA
Detailed Run Analysis
This section presents an expanded analysis of the data for a typical
test, namely, Run No. 10, Condition 1C. This was a full scale test using
2, 2, 5-trimethylhexane superheated vapor, with the 0.10" converging vapor nozzle.
The overall equivalence ratio based on CO. analysis was 0.785. The input data
of this test are given in Appendix D, together with similar data for each test
condition. Parts V through IX give calculations applying to each run.
I. Combustion Temperature
A. Calculated Flame Temperature
A generalized computational procedure was used throughout this
program to calculate flame temperatures. Based on the stoichiometry of
combustion for a given fuel, the composition of the exhaust gases was determined
as a function of 0, the overall equivalence ratio. The standard state heat of
combustion of the fuel was corrected for: (a) the actual fuel conditions at
the fuel nozzle inlet, and (b) the heat removed from the combustor by the cooling
water. This fixed the net amount of heat (per mole of fuel) available to heat up
the combustion, products. A graphical presentation of the mean specific heat of
each component gas, between 0°C and the upper temperature 'T' was used. The
working equation then became:
(1) AHC = £ Ni I C x T - C
where: AH = net heat of combustion per mole of fuel
T = flame temperature, C as measured at the station of
Ni = moles of component 'i' per mole of fuel
flame temperature,
'T ' thermocouple
Cj- = mean specific heat of component 'i' between 0 C and T
o
T. = temperature of fuel at fuel nozzle inlet, C
= mean specific heat of component 'i' between 0 C and T
Equation (1) was then solved for T, the corrected theoretical flame temperature,
by a combination of computation and graphical techniques.
For the case of 2, 2, 5-trimethylhexane, the re la tionsh i.p brl.wccn
Ni anrl 0 is, per mole of fuel:
E-l
-------
an
VtH NUTI. CAll'OINIA
i
H20
co2
°2
N2
Ni
10
9
14 4 '
52.8
0
1)
With this technique, a curve was constructed relating the corrected theoretical
flame temperature to equivalence ratio. For Run 10, Condition 1C, this calculated
flame temperature was 3080 F.
B. Corrected Thermocouple Temperature
The uncorrected thermocouple temperature was 2825 F. As discussed
in Section V-C, a bare thermocouple, exposed to the environment of the combustor
can be expected to introduce some error, primarily due to the radiation from the
thermocouple junction. The situation is aggravated here in that the junction can
'see1 walls which are at a much lower temperature than the surrounding gas stream
whose temperature it is desired to measure.
An extimate was made of the magnitude of the errors to be expected
for this particular installation, using the methods of Reference 20. The
relationship for radiation correction is:
(2)
= T
AD ,T.4 - T \
R ( i v )
h A
c c
where
Tt -
T. =
J
a =
£
gas total temperature, R
thermocouple junction temperature, R
radiation view factor (assumed equal to unity)
Stefan - Boltzman constant
emissivity of junction
area of radiant transfer from junction
coefficient of convectivu heat transfer, gas to junction
area of convectivt- heat transfer, gas to junction.
E-2
-------
//ilarquardt
laiKHlHUTION
VAM NUYS. CAUFOIMIA
Evaluation of the key factors was as follows:
(a) Reference 20, pg. 563, states that t of clean Pt is 0.18 (no
temperature given) and that it can go up to 0.30 or even 0.55 due to contamination
by exhaust. During this program, however, the bead of a similar combustor thermo-
couple was examined under the microscope after Run 6, in which it had been sub-
jected to conditions which caused a heavy carbon deposit on the combustor walls.
The thermocouple bead appeared unaffected, with a shiny luster similar to that
of a new junction. Based on this observation, therefore, and on the data of
McAdams (Reference 21), pg. 475, ( was taken to be 0.18.
(b) The diameter of each thermocuople wire was 0.003"; a spherical
bead of 0.01" diameter was estimated. For the existing configuration, AR .^ A .
(c) The coefficient of convective heat transfer was estimated from
the correlation of Reference 21, pg. 266, to be 170 Btu/ft hr. °F0
These quantities resulted in a computed radiation error of 210 F. The conduction
error, also estimated by the method of Reference 20, was 8 F. No other errors
were found to be significant. It should be noted that this radiation correction
is relatively slight for a bare thermocouple at this temperature; the primary
reasons for this are the fine wire used in the couple, with the resulting small
bead, and the low emissivity of the noble metal surface of the bead.
The above corrections, applied to the measured junction temperature
of 2825 F, gave a corrected measured flame temperature of approximately 3040 F.
This is only 40 F below the computed flame temperature. This is quite close
agreement, although perhaps somewhat fortuitous, since the accuracy of determin-
ing thermocouple error terms is usually not great.
lie Heat Balance
Basic to the goals of this program is the availability of a high flux
of heat energy for transfer to a working fl.uid. Accordingly, the heat removed
from the exhaust gases in the heat exchanger section was computed. This was done
by the following procedure.
The composition of exhaust gases was determined from the measured CO-
content with the assumption of complete combustion. It was assumed that the
composition of this gas remained unchanged after leaving the combustor. The-
composition of this gas mixture was computed to be, in mole fractions:
E-3
-------
l. CAll'OIMU
H20 0.1109
C02 0.0998
02 0.0439
N2 0.7453
with an average molecular weight of 28.6.
The following specific heat equation was computed for this gas
mixture based on the pure component values of page 371, Reference 22.
(3) C = 6.42 + 2.69 x 10~3T - 5.76 x 10"7 T2
P
where
C is in cal/gmol °K, and T is °K.
Based on the measured total flow rate, the enthalpy change of the gas was calcu-
lated between each pair of heat exchanger thermocouples. The corrected measured
value of T was used. Radiation correction was also made for the first heat
exchanger thermocouple (uncorrected reading - 2040 F), but the correction was
only 18 F. No corrections were necessary for the remaining two thermocouples in
the heat exchanger.
The heat absorbed by the cooling water was also calculated. The
results are shown below:
Heat Removed from Heat Absorbed by
Exhaust Gases Cooling Water
Section Btu/min Btu/min
1st Heat Exchanger 3010 2100
T to T, .
c H.E. 1
2nd Heat Exchanger 1520 1330
TH_E. 1 to TH.E. 2
3rd Heat Exchanger 2750 2150
TH.E. 2 t0 TH.E. 3
Totals 7,280 Btu/min 5,580 Btu/min
436,000 Btu/hr 335,000 Btu/hr
-------
First, it should be noted that the calculated heat removed from the exhaust gases
is in good agreement with the nominal heat release of 500,000 Btu/hr, based on
fuel flow. The calculated figure of 435,000 Btu/hr does not include the heat
dissipated in the double wall combustor itself, but only in the heat exchanger
section.
The calculated heat absorbed by the cooling water is 77% of the heat
removed, as calculated from the thermocouple measurement. This is due primarily
to certain aspects of the heat exchanger layout. In the first heat exchanger
section there was a significant amount of radiative heat transfer to the surround-
ings, because two short sections of heat exchanger fittings were not surrounded
by copper cooling coils.
The third section of heat exchanger is the open water trough. The exit
water temperature from this trough was 156 F, so that a significant amount of
vaporization from the water surface occurred.
III. Heat Exchanger Performance
Heat exchanger performance data for Run 10, Condition 1C are summarized
in Figure E-l, in which the following quantities are plotted vs. heat exchanger
length.
(1) Gas temperature, F, based on T and the three heat exchanger
thermocouples.
(2) Cumulative heat transfer, Btu/hr, based on the heat removed
from the gas, as calculated above.
(3) Cumulative residence time, milliseconds. This was determined
by computing the average gas velocity, j/ , at each thermocouple
station, plotting 1/f vs. L, and graphically integrating.
These data show that the corrected gas temperature dropped to 2500 F within
5 milliseconds, and to 20CO F in about 13 milliseconds. Total residence time
in the heat exchanger was .computed to be 109 milliseconds. The graphs of these
three items agree well with a similar plot of calculated quantities performed
during the heat exchanger design, prior to fabrication. The main difference
is that the convective heat transfer coefficient from the copper-coil wrapped
sections was less than that predicted, with radiative heat transfer being
significant in the first section.
The pressure drop across the heat exchanger length of 34 feet, plus
the additional 4 feet of exhaust line, was 3.6 in. Hg, or 1.8 psi. The design
value, based on a 33-ft. length, was 1.5 psi.
E-5
-------
IV o Velocities in Igniter Tube
The igniter tube was 2" O.D. x 0.035" wall x 6" long. The primary air
flow was 0.129 Ib/sec. Therefore, the average air velocity in the igniter tube
(neglecting fuel flow) was 76 ft/sec, which corresponds to a residence time of
4.4 milliseconds for the 4 inches of length between the fuel nozzle exit and the
combustor inlet.
The measured fuel flow was 26.2 Ib/hr = 0.0073 Ib/sec. Assuming sonic
flow at the fuel nozzle exit, with the measured nozzle inlet pressure of 26 psia,
the calculated fuel flow (for choked nozzle flow) is 0.00735 Ib/second. Thus,
the fuel nozzle was operating in the choked condition, with the fuel vapor leaving
the nozzle at sonic velocity. This was calculated to be 550 ft/sec at existing
conditions, in good agreement with the throat velocity calculated from continuity.
This relatively low sonic velocity is a consequence of the high molecular weight
of 2 , 2 , 5-trimethylhexane.
V0 Equivalence Ratio as Function of Cp2 Content
The method of calculation assumed complete combustion of the fuel to
CO- and H20. The levels of CO, NO, and unburned hydrocarbon encountered during
these tests were, in all but a very few exceptional cases, so low as to not exert
a significant, effect.
The equation for stoichiometric combustion was written for each fuel.
Thus, for 2, 2, 5-trimethylhexane,
(1) C9H2Q + 1402 >9C02 + 10 H20
Based on the definition of equivalence ratio, 0, one then obtains an expression
relating the moles of 0« to N~ (per mole of fuel). It was assumed that kerosene
could be adequately represented by an average molecular formula of the type
CnH« . Hence, the equation for stoichiometric combustion for kerosene became:
(2)
These results are summarized below for each fuel in the form of mole fraction of
C0? . The water vapor in the exhaust was removed by the sample gas equipment
upstream of the NDIR's. Accordingly, the water formed in each combustion
reaction was deleted from the following calculations.
E-6
-------
'quardt
Fuel
co
2, 2, 5-trimethylhexane
66.8
Kerosene
Methane
i
9.52
0
These equations for X vs . 0 are plotted in Figures E-2.
co2
ann lysis of C0« is read directly in volume %(= mole7o) COj
The NDIR
VI
(1)
Comparison of Equivalence Ratio Calculations
Two independent methods were used to determine the overall equivalence
ratio, 0Q, defined as:
actual
ft)
where W
stoich
fuel flow rate to combustor
W = total air flow rate to combustor.
The first method was direct substitution in Eq. (1) of the measured
values of fuel and air rates. The second method was based on the C0? analysis
of the exhaust gas, as discussed previously. This latter measurement permits
the calculation of equivalence ratio independent of any of the flow rate measure-
ments, provided that combustion is essentially complete.
In many cases the two valves agreed closely. In other cases, they did
not. . A detailed analysis was made of this question, and possible reasons examined
for those discrepancies which occurred. The analysis is summarized in this
section.
E-7
-------
Assume that the standard relative error of the measurement of fuel flow
and of the total air flow was 3%, and that this cumulative random error from all
sources is normally distributed. It should be noted that, in this discussion,
standard error refers to
-------
VAN Hurt, cunroiNi*
No correlation was observed, however, for any combination of these
factors. There was no set of conditions found which was unique to those cases
showing differences greater than 57». This would indicate that the observed
differences represented a random distribution of cumulative small perturbations.
The trend of the data was apparently not random in certain respects, however.
Thus, of those cases with differences greater than 57,, the sign of the difference
was positive in every case up until Run 17, then negative in every case for
Runs 17 and 18. In Runs 13 and 14, all 18 successive data points showed close
agreement between the two values of equivalence ratio. Particular scrutiny was
given to all the data of these runs, but again there was no set of conditions
found which could be correlated with this apparent non-random behavior.
Since there was no known reason to question any of these data, the
value of equivalence ratio determined from NDIR C0_ analysis was used to repi
sent that data point, due to its statistically-expected smaller t-rror.
VI I o Conversion of Gas Analysis Data to Include Water Vapor Content of
Exhaust Gases
As previously stated, all of the gas analysis data (as measured) are
on a water-free basis. It is of interest to express these results on a "water
basis," that is, the actual concentrations existing in the hot exhaust gas
prior to water vapor removal.
X. = concentration of component 'i1 as measured (^ x 10 )
Define :
Y. = concentration of component 'i1, water basis (- x 10 )
IV
Ni = moles component 'i1 per mole of fuel '
N_^ = moles exhaust, water-free basis, per mole of fuel
H* = moles exhaust, water basis, per mole of fuel
/
The desired conversion factor, K, is:
Y. HL
(3) K = ^ = -^
i
The expression for l^L. had already been obtained in Part V, above.
The expression for N_ was readily obtained from this by the inclusion of the
stoichiometric moles of water.
E-9
-------
Itmfmtmm
Fuel K = Yi/X
66.8 _
2, 2, 5-trimethylhexane 0
Kerosene
+0.5
0
Methane
9.52 _ 1
9 52
The conversion factors for each fuel are plotted in Figure £-3.
VI1I0 Conversion of Gas Analysis Data to Basis of Mass of Pollutant
per Unit Mass of Fuel
Define Li as the loading of component 'i1 in milligrams (mg) per pound
of fuel. It may then be seen that:
(Xi) (M.W. ) (IL. )
(4) Li - r-r. SL x 454 x 10
M.W.f
where, M.W.. = molecular weight of component 'i1
M,W0 = molecular weight of fuel
and, Xi and Nj; have the same meanings as in Part VII, above.
What is desired is a graphical presentation of a factor which is
multiplied by Xi to give Li. Accordingly, define ft. = .
1. jVl
It then follows that, for each fuel:
E-10
-------
VAN Hurt. CAU'OINIA
A. 2, 2, 5 - trimethylhexane
'NO ir - °-530
(6) * - - 0.495
(7) fi - - - 0.254
6
B. Kerosene
<8) "NO ' - °'485
(10)
- °-792
(13) ^C6 - - 0.406
These equations are plotted in Figures E-4, E-5 , and E-6 . A word is necessary
about the form of the results for unburned hydrocarbon. The raw data (Xi), are
reported in terms of ppm carbon. It was necessary to arbitrarily select a
species of hydrocarbon upon which to base the mass calculations. The species
selected was n-hexane. Thus, the mass data, L, is given in terms of milligrams
n-he.xane per pound of fuel. In making the conversion from these graphs
(L-, = /J-, X ) note that the value of X to use is the raw data from the hydro-
C6 C6 C C
carbon analyses, in ppm carbon.
E-ll
-------
-quardt
V/IM NUTS. CAll'OINI*
IXQ Evaluation of Particulate Emissions
Once the mass emissions per unit volume of exhaust products is calculated,
the residual ash of fuel or mass emissions per gram of fuel can be- determined.
The stoichiometric combustion equation is used to give the moles of fuel from the-
total moles of exhaust. The molecular weight of the fuel and the temperature
(5; 75 F) at thich the sample volume was measured yield the mass emissions per
gm. of fuel.
The formulas for each step and example calculations are shown below:
STEPS IN CALCULATION
Particles
JL
cubic meter exhaust
particles 1^C°U" ie ' w
x s x ? x count
cubic meter A- x R x T A- x R x t.
62 \
82mm x 4mm/min x 10X, x count/field j.0 cc
2 9 *
30cc/min 1mm x A
.'. Pfrticles - 11.0 x 1012x^Hn<- x
cubic meter
2. particles
cubic foot exhaust
3
particles particles Itn
cubic foot cubic meter \- _ -3
j j ^ r t
E-12
-------
3.
cubic meter exhaust
cubic meter exhaust
3
/particles ) / A tT | { mean j / Density \
\ cubic metery ^ 3 &J \dia.J ^ J
in' /V\3!
particles - « gms 10>c
-------
/l%
/A Idroudrdt
f r* JUI\JU\JI\Ji VAN NU'S. CAU'OINIA
EXAMPLE CALCULATION
SAMPLE 2294C
1. particles ,. -, , .,12 count ~f ,, n inl2 100 , __
^r-: ; = 11.0 x 10 x x 9 » 11.0 x 10 x -^7: x 1.35
cubic meter 2 * 900
particles . , ,n
- 1.6 x 10
cubic meter
3
2. particles particles 1m . , -.12 1
cubic foot " cubic meter x 35 3 " X x 35.2
- 46 x 109
3. MSms particles fdia\ » i 2 x 10"6
, *; ^ . , XI I -X 1 * X Av
cubic meter cubic meter V I
1.6 x 1012 x (.47)3 x 1.2 x 10"6
2.0 x 105
cubic meter
4. Xgms
^"
r r.
gm fuel cubic meter
gm feet
- 2.0 x 105 x 14.7 x 10~3
.6 x 103
-------
arquardt
ICOMI'ANY
VAN NUYS. CALIFORNIA
SUMMARY OF HEAT EXCHANGER PERFORMANCE
veiso-s
s
.._._^.
.s
\''
fcL
,
r
^
S
...
r
400
COMBUSTOR
EXIT
10 20 30
HEAT EXCHANGER LENGTH
m
o
CQ
CO
o
lOOo
o
LU
CO
50
o
z
LU
O
CO
0
GAS
SAMPLING
STATION
E-15
-e E-l
-------
ON
MOLE % C02
(H20 - FREE BASIS)
w
rv-
EQUIVALENCE RATIO BASED ON CARBON DIOXIDE
COMPLETE COMBUSTION ASSUMED
2,2,5 TRIMETHYLHEXANE
Oo6 0.7 0.8 0.9
EQUIVALENCE RATIO, <#>
5T
s9
is.
2
z.
z
-<
o
I
S
-------
-quardt
-/COMPANY-
VAN Nim. CALIFORNIA
V8I30-I 1
CONVERSION FACTOR TO ACCOUNT FOR WATER VAPOR
IN EXHAUST GASES
_ PPM IN EXHAUST (WATER VAPOR INCLUDED)
~ PPM MEASURED (WATER-FREE BASES)
1.0
0.9
0.8
OVERALL
EQUIVALENCE
RATIO
0.7
0.6
0.5
0.4i
2,2,5 - TRIMETHYLHEXANE
KEROSENE
0.7 0.8 0.9 1.0 1.1
1.2
E-l?
Figure E-3
-------
CONVERSION OF GAS EMISSION DATA TO UNIT MASS OF FUEL
FUEL: 2,2,5 - TRIMETHYLHEXANE
T
(>
OO
oq
r
OVERALL
EQUIVALENCE
RATIO - 0
EMISSION LOAD, MILLIGRAMS 'i'/LB FUEL
PPM 'i', AS MEASURED
HYDROCARBON
(AS HEXANE)
NITRIC OXIDE
CARBON MONOXIDE
0.6
0.5
0.4
i
sa.
z
r-
-------
CONVERSION OF GAS EMISSION DATA TO UNIT MASS OF FUEL
FUEL: KEROSENE (JET FUEL A, ASTM D1555)
H-
oq
TJ
V1
OVERALL
EQUIVALENCE
RATIO - 0
o
till
= X! 01
= EMISSION LOAD, MILLIGRAMS 'i'/LB FUEL
= PPM 'i', AS MEASURED
HYDROCARBON
(AS HEXANE)
NITRIC OXIDE
CARBON MONOXIDE
2
b-
(-
Tl
-------
CONVERSION OF GAS EMISSION DATA TO UNIT MASS OF FUEL
FUEL: METHANE
07
M
r\j
o
OVERALL
EQUIVALENCE
RATIO-0
= EMISSION LOAD, MILLIGRAMS
0.5
0.4
12
14
16
18
------- |